1 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
3 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost SYMBOL_REF,
6 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
7 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
9 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost TRUNCATE.
11 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
12 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
14 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost FMA,
15 FLOAT_EXTEND, FLOAT_TRUNCATE, ABS, SMAX, and SMIN.
17 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
18 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
20 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost comparison
23 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
24 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
26 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costs for
29 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
30 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
32 * config/aarch64/aarch64.c (aarch64_rtx_arith_op_extract_p): New.
33 (aarch64_rtx_costs): Improve costs for SIGN/ZERO_EXTRACT.
35 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
36 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
38 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costs for
41 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
42 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
44 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost
45 ZERO_EXTEND and SIGN_EXTEND better.
47 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
48 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
50 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve cost for
53 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
54 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
56 * config/aarch64/aarch64.c (aarch64_rtx_costs): Use address
57 costs when costing loads and stores to memory.
59 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
60 Philip Tomsich <philipp.tomsich@theobroma-systems.com>
62 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costing
65 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
67 * config/aarch64/aarch64.c (aarch64_rtx_costs): Set default costs.
69 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
70 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
72 * config/aarch64/aarch64.c (aarch64_strip_shift_or_extend): Rename
74 (aarch64_strip_extend): ...this, don't strip shifts, check RTX is
76 (aarch64_rtx_mult_cost): New.
77 (aarch64_rtx_costs): Use it, refactor as appropriate.
79 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
80 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
82 * config/aarch64/aarch64.c (aarch64_build_constant): Conditionally
83 emit instructions, return number of instructions which would
85 (aarch64_add_constant): Update call to aarch64_build_constant.
86 (aarch64_output_mi_thunk): Likewise.
87 (aarch64_rtx_costs): Estimate cost of a CONST_INT, cost
90 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
92 * config/aarch64/aarch64.c (aarch64_rtx_costs_wrapper): New.
93 (TARGET_RTX_COSTS): Call it.
95 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
97 * config/aarch64/aarch64.c (cortexa57_addrcost_table): New.
98 (cortexa57_vector_cost): Likewise.
99 (cortexa57_tunings): Use them.
101 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
103 * config/aarch64/aarch64-protos.h (scale_addr_mode_cost): New.
104 (cpu_addrcost_table): Use it.
105 * config/aarch64/aarch64.c (generic_addrcost_table): Initialize it.
106 (aarch64_address_cost): Rewrite using aarch64_classify_address,
109 2014-05-16 Richard Biener <rguenther@suse.de>
111 * tree-ssa-sccvn.c: Include tree-cfg.h and domwalk.h.
112 (set_ssa_val_to): Handle unexpected sets to VN_TOP.
113 (visit_phi): Ignore edges marked as not executable.
114 (class cond_dom_walker): New.
115 (cond_dom_walker::before_dom_children): Value-number
116 control statements and mark successor edges as not
117 executable if possible.
118 (run_scc_vn): First walk all control statements in
119 dominator order, marking edges as not executable.
120 * tree-inline.c (copy_edges_for_bb): Be not confused
121 about random edge flags.
123 2014-05-16 Richard Biener <rguenther@suse.de>
125 * tree-ssa-sccvn.c (visit_use): Also constant-fold calls.
127 2014-05-15 Peter Bergner <bergner@vnet.ibm.com>
130 * config/rs6000/htmxlintrin.h (_HTM_TBEGIN_STARTED): New define.
131 (__TM_simple_begin): Use it.
132 (__TM_begin): Likewise.
134 2014-05-15 Martin Jambor <mjambor@suse.cz>
137 * ipa-prop.c (update_indirect_edges_after_inlining): Check
138 type_preserved flag when the indirect edge is polymorphic.
140 2014-05-15 Martin Jambor <mjambor@suse.cz>
142 PR tree-optimization/61090
143 * tree-sra.c (sra_modify_expr): Pass the current gsi to
146 2014-05-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
148 * config/arm/arm.c (arm_option_override): Use the SCHED_PRESSURE_MODEL
149 enum name for PARAM_SCHED_PRESSURE_ALGORITHM.
151 2014-05-15 Jakub Jelinek <jakub@redhat.com>
153 PR tree-optimization/61158
154 * fold-const.c (fold_binary_loc): If X is zero-extended and
155 shiftc >= prec, make sure zerobits is all ones instead of
156 invoking undefined behavior.
158 2014-05-15 Zhenqiang Chen <zhenqiang.chen@linaro.org>
160 * regcprop.h: New file.
161 * regcprop.c (skip_debug_insn_p): New decl.
162 (replace_oldest_value_reg): Check skip_debug_insn_p.
163 (copyprop_hardreg_forward_bb_without_debug_insn): New function.
164 * shrink-wrap.c: Include regcprop.h.
165 (prepare_shrink_wrap): Call
166 copyprop_hardreg_forward_bb_without_debug_insn.
168 2014-05-15 Zhenqiang Chen <zhenqiang.chen@linaro.org>
170 * shrink-wrap.h: Update comment.
171 * shrink-wrap.c: Update comment.
172 (next_block_for_reg): Rename to live_edge_for_reg.
173 (live_edge_for_reg): Allow live_edge->dest has two predecessors.
174 (move_insn_for_shrink_wrap): Split live_edge.
175 (prepre_shrink_wrap): One more parameter for move_insn_for_shrink_wrap.
177 2014-05-14 Eric Botcazou <ebotcazou@adacore.com>
179 * config/sparc/sparc-protos.h (sparc_absnegfloat_split_legitimate):
181 * config/sparc/sparc.c (sparc_absnegfloat_split_legitimate): Likewise.
182 * config/sparc/sparc.md (fptype_ut699): New attribute.
183 (in_branch_delay): Return false if -mfix-ut699 is specified and
184 fptype_ut699 is set to single.
185 (truncdfsf2): Add fptype_ut699 attribute.
186 (fix_truncdfsi2): Likewise.
187 (floatsisf2): Change fptype attribute.
188 (fix_truncsfsi2): Likewise.
189 (negtf2_notv9): Delete.
190 (negtf2_v9): Likewise.
191 (negtf2_hq): New instruction.
192 (negtf2): New instruction and splitter.
193 (negdf2_notv9): Rewrite.
194 (abstf2_notv9): Delete.
195 (abstf2_hq_v9): Likewise.
196 (abstf2_v9): Likewise.
197 (abstf2_hq): New instruction.
198 (abstf2): New instruction and splitter.
199 (absdf2_notv9): Rewrite.
201 2014-05-14 Cary Coutant <ccoutant@google.com>
204 * opts.c (common_handle_option): Don't special-case "-g".
205 (set_debug_level): Default to at least level 2 with "-g".
207 2014-05-14 DJ Delorie <dj@redhat.com>
209 * config/msp430/msp430.c (msp430_builtin): Add
210 MSP430_BUILTIN_DELAY_CYCLES.
211 (msp430_init_builtins): Register void __delay_cycles(long long).
212 (msp430_builtin_decl): Add it.
213 (cg_magic_constant): New.
214 (msp430_expand_delay_cycles): New.
215 (msp430_expand_builtin): Call it.
216 (msp430_print_operand_raw): Change integer printing from "int" to
218 * config/msp430/msp430.md (define_constants): Add delay_cycles tags.
219 (delay_cycles_start): New.
220 (delay_cycles_end): New.
221 (delay_cycles_32): New.
222 (delay_cycles_32x): New.
223 (delay_cycles_16): New.
224 (delay_cycles_16x): New.
225 (delay_cycles_2): New.
226 (delay_cycles_1): New.
227 * doc/extend.texi: Document __delay_cycles().
229 2014-05-14 Sandra Loosemore <sandra@codesourcery.com>
231 * config/nios2/nios2.md (nios2_cbranch): Fix paste-o in
232 length attribute computation.
234 2014-05-14 Richard Sandiford <rdsandiford@googlemail.com>
237 * print-rtl.c (print_rtx): Suppress uids if flag_dump_unnumbered.
239 2014-05-14 Richard Sandiford <r.sandiford@uk.ibm.com>
242 * config/sparc/sparc.md: Fix types of low and high in DI constant
243 splitter. Use gen_int_mode in some other splitters.
245 2014-05-14 Martin Jambor <mjambor@suse.cz>
248 * ipa-prop.c (ipa_modify_formal_parameters): Reset DECL_LANG_SPECIFIC.
250 2014-05-14 James Norris <jnorris@codesourcery.com>
252 * omp-low.c (expand_parallel_call): Remove shadow variable.
253 (expand_omp_taskreg): Likewise.
255 2014-05-14 Ilya Tocar <ilya.tocar@intel.com>
257 * common/config/i386/i386-common.c
258 (OPTION_MASK_ISA_CLFLUSHOPT_SET): Define.
259 (OPTION_MASK_ISA_XSAVES_SET): Ditto.
260 (OPTION_MASK_ISA_XSAVEC_SET): Ditto.
261 (OPTION_MASK_ISA_CLFLUSHOPT_UNSET): Ditto.
262 (OPTION_MASK_ISA_XSAVES_UNSET): Ditto.
263 (OPTION_MASK_ISA_XSAVEC_UNSET): Ditto.
264 (ix86_handle_option): Handle OPT_mxsavec, OPT_mxsaves,
266 * config.gcc (i[34567]86-*-*): Add clflushoptintrin.h,
267 xsavecintrin.h, xsavesintrin.h.
269 * config/i386/clflushoptintrin.h: New.
270 * config/i386/xsavecintrin.h: Ditto.
271 * config/i386/xsavesintrin.h: Ditto.
272 * config/i386/cpuid.h (bit_CLFLUSHOPT): Define.
275 * config/i386/driver-i386.c (host_detect_local_cpu): Handle
276 -mclflushopt, -mxsavec, -mxsaves, -mno-xsaves, -mno-xsavec,
278 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
279 OPTION_MASK_ISA_CLFLUSHOPT, OPTION_MASK_ISA_XSAVEC,
280 OPTION_MASK_ISA_XSAVES.
281 * config/i386/i386.c (ix86_target_string): Handle -mclflushopt,
283 (PTA_CLFLUSHOPT) Define.
286 (ix86_option_override_internal): Handle new options.
287 (ix86_valid_target_attribute_inner_p): Ditto.
288 (ix86_builtins): Add IX86_BUILTIN_XSAVEC, IX86_BUILTIN_XSAVEC64,
289 IX86_BUILTIN_XSAVES, IX86_BUILTIN_XRSTORS, IX86_BUILTIN_XSAVES64,
290 IX86_BUILTIN_XRSTORS64, IX86_BUILTIN_CLFLUSHOPT.
291 (bdesc_special_args): Add __builtin_ia32_xsaves, __builtin_ia32_xrstors,
292 __builtin_ia32_xsavec, __builtin_ia32_xsaves64, __builtin_ia32_xrstors64,
293 __builtin_ia32_xsavec64.
294 (ix86_init_mmx_sse_builtins): Add __builtin_ia32_clflushopt.
295 (ix86_expand_builtin): Handle new builtins.
296 * config/i386/i386.h (TARGET_CLFLUSHOPT) Define.
297 (TARGET_CLFLUSHOPT_P): Ditto.
298 (TARGET_XSAVEC): Ditto.
299 (TARGET_XSAVEC_P): Ditto.
300 (TARGET_XSAVES): Ditto.
301 (TARGET_XSAVES_P): Ditto.
302 * config/i386/i386.md (ANY_XSAVE): Add UNSPECV_XSAVEC, UNSPECV_XSAVES.
303 (ANY_XSAVE64)" Add UNSPECV_XSAVEC64, UNSPECV_XSAVES64.
304 (attr xsave): Add xsavec, xsavec64, xsaves, xsaves64.
306 (ANY_XRSTOR64): Ditto.
308 (xrstor): Change into <xrstor>.
309 (xrstor_rex64): Change into <xrstor>_rex64.
310 (xrstor64): Change into <xrstor>64
312 * config/i386/i386.opt (mclflushopt): New.
315 * config/i386/x86intrin.h: Add clflushoptintrin.h, xsavesintrin.h,
317 * doc/invoke.texi: Document new options.
319 2014-05-14 Andrey Belevantsev <abel@ispras.ru>
321 PR rtl-optimization/60866
322 * sel-sched-ir (sel_init_new_insn): New parameter old_seqno.
323 Default it to -1. Pass it down to init_simplejump_data.
324 (init_simplejump_data): New parameter old_seqno. Pass it down
325 to get_seqno_for_a_jump.
326 (get_seqno_for_a_jump): New parameter old_seqno. Use it for
327 initializing new jump seqno as a last resort. Add comment.
328 (sel_redirect_edge_and_branch): Save old seqno of the conditional
329 jump and pass it down to sel_init_new_insn.
330 (sel_redirect_edge_and_branch_force): Likewise.
332 2014-05-14 Georg-Johann Lay <avr@gjlay.de>
334 * config/avr/avr.h (REG_CLASS_CONTENTS): Use unsigned suffix for
335 shifted values to avoid build warning.
337 2014-05-14 Eric Botcazou <ebotcazou@adacore.com>
339 * cfgcleanup.c (try_forward_edges): Use location_t for locations.
340 * cfgrtl.c (rtl_merge_blocks): Fix comment.
341 (cfg_layout_merge_blocks): Likewise.
342 * except.c (emit_to_new_bb_before): Remove prev_bb local variable.
344 2014-05-14 Andrey Belevantsev <abel@ispras.ru>
346 PR rtl-optimization/60901
347 * config/i386/i386.c (ix86_dependencies_evaluation_hook): Check that
348 bb predecessor belongs to the same scheduling region. Adjust comment.
350 2014-05-13 Peter Bergner <bergner@vnet.ibm.com>
352 * doc/sourcebuild.texi: (dfp_hw): Document.
353 (p8vector_hw): Likewise.
354 (powerpc_eabi_ok): Likewise.
355 (powerpc_elfv2): Likewise.
356 (powerpc_htm_ok): Likewise.
357 (ppc_recip_hw): Likewise.
360 2014-05-13 Cary Coutant <ccoutant@google.com>
362 * opts.c (finish_options): Use -ggnu-pubnames with -gsplit-dwarf.
364 2014-05-13 David Malcolm <dmalcolm@redhat.com>
366 * gengtype-parse.c (require3): Eliminate in favor of...
368 (require_template_declaration): Update to support optional single *
371 * gengtype.c (get_ultimate_base_class): Add a non-const overload.
372 (create_user_defined_type): Handle a single level of explicit
373 pointerness within template arguments.
374 (struct write_types_data): Add field "kind".
375 (filter_type_name): Handle "*" character.
376 (write_user_func_for_structure_ptr): Require a write_types_data
377 rather than just a prefix string, so that we can look up the kind
378 of the wtd and use it as an index into wrote_user_func_for_ptr,
379 ensuring that such functions are written at most once. Support
380 subclasses by invoking the marking function of the ultimate base
382 (write_user_func_for_structure_body): Require a write_types_data
383 rather than just a prefix string, so that we can pass this to
384 write_user_func_for_structure_ptr.
385 (write_func_for_structure): Likewise.
386 (ggc_wtd): Add initializer of new "kind" field.
389 * gengtype.h (enum write_types_kinds): New.
390 (struct type): Add field wrote_user_func_for_ptr to the "s"
393 2014-05-13 Richard Sandiford <r.sandiford@uk.ibm.com>
395 * fold-const.c (optimize_bit_field_compare): Use wi:: operations
396 instead of const_binop.
397 (fold_binary_loc): Likewise.
399 2014-05-13 Richard Sandiford <r.sandiford@uk.ibm.com>
401 * tree-dfa.h (get_addr_base_and_unit_offset_1): Update array index
402 calculation to match get_ref_base_and_extent.
404 2014-05-13 Catherine Moore <clm@codesourcery.com>
405 Sandra Loosemore <sandra@codesourcery.com>
407 * configure.ac: Fix assembly for explicit JALR relocation check.
408 * configure: Regenerate.
410 2014-05-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
412 * config/arm/arm.c (neon_itype): Remove NEON_RESULTPAIR.
413 (arm_init_neon_builtins): Remove handling of NEON_RESULTPAIR.
414 Remove associated type declarations and initialisations.
415 (arm_expand_neon_builtin): Likewise.
416 (neon_emit_pair_result_insn): Delete.
417 * config/arm/arm_neon_builtins (vtrn, vzip, vuzp): Delete.
418 * config/arm/neon.md (neon_vtrn<mode>): Delete.
419 (neon_vzip<mode>): Likewise.
420 (neon_vuzp<mode>): Likewise.
422 2014-05-13 Richard Biener <rguenther@suse.de>
425 * tree-inline.c (remap_gimple_stmt): Clear tail call flag,
426 it needs revisiting whether the call still may be tail-called.
428 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
430 * rtl.def (SYMBOL_REF): Remove middle "0" field.
431 * rtl.h (block_symbol): Reduce number of fields to 2.
432 (rtx_def): Add u2.symbol_ref_flags.
433 (SYMBOL_REF_FLAGS): Use it.
434 (SYMBOL_REF_DATA, SET_SYMBOL_REF_DECL, SYMBOL_REF_DECL)
435 (SET_SYMBOL_REF_CONSTANT, SYMBOL_REF_CONSTANT): Lower index.
436 * gengtype.c (adjust_field_rtx_def): Remove SYMBOL_REF_FLAGS handling.
437 Lower index of SYMBOL_REF_DATA.
438 * print-rtl.c (print_rtx): Lower index for SYMBOL_REF_DATA.
439 Print SYMBOL_REF_FLAGS at the same time.
440 * genattrtab.c (attr_rtx_1): Only initialize 1 "0" SYMBOL_REF field.
442 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
444 * rtl.def (VAR_LOCATION): Remove "i" field.
445 * rtl.h (rtx_def): Add u2.var_location_status.
446 (PAT_VAR_LOCATION_STATUS): Use it.
447 (gen_rtx_VAR_LOCATION): Declare.
448 * gengenrtl.c (excluded_rtx): Add VAR_LOCATION.
449 * emit-rtl.c (gen_rtx_VAR_LOCATION): New function.
450 * var-tracking.c (emit_note_insn_var_location): Remove casts.
452 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
454 * rtl.def (scratch): Fix outdated comment and remove "0" field.
455 * gengtype.c (adjust_field_rtx_def): Update accordingly.
457 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
459 * rtl.def (DEBUG_INSN, INSN, JUMP_INSN, CALL_INSN, JUMP_TABLE_DATA)
460 (BARRIER, CODE_LABEL, NOTE): Remove first "i" field.
461 * rtl.h (rtx_def): Add insn_uid to u2 field.
462 (RTX_FLAG_CHECK8): Delete in favor of...
463 (RTL_INSN_CHAIN_FLAG_CHECK): ...this new macro.
464 (INSN_DELETED_P): Update accordingly.
465 (INSN_UID): Use u2.insn_uid.
466 (INSN_CHAIN_CODE_P): Define.
467 (PREV_INSN, NEXT_INSN, BLOCK_FOR_INSN, PATTERN, INSN_LOCATION)
468 (INSN_CODE, REG_NOTES, CALL_INSN_FUNCTION_USAGE, CODE_LABEL_NUMBER)
469 (NOTE_DATA, NOTE_DELETED_LABEL_NAME, NOTE_BLOCK, NOTE_EH_HANDLER)
470 (NOTE_BASIC_BLOCK, NOTE_VAR_LOCATION, NOTE_CFI, NOTE_LABEL_NUMBER)
471 (NOTE_KIND, LABEL_NAME, LABEL_NUSES, JUMP_LABEL, LABEL_REFS): Lower
473 * print-rtl.c (print_rtx): Print INSN_UIDs before the main loop.
474 Update indices for insn-chain rtxes.
475 * gengtype.c (gen_rtx_next): Adjust test for insn-chain rtxes.
476 (adjust_field_rtx_def): Lower '0' indices for all insn-chain rtxes.
477 * emit-rtl.c (gen_label_rtx): Update gen_rtx_LABEL call.
478 * caller-save.c (init_caller_save): Update gen_rtx_INSN calls.
479 * combine.c (try_combine): Likewise.
480 * ira.c (setup_prohibited_mode_move_regs): Likewise.
482 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
484 * rtl.def (REG): Remove middle field.
485 * rtl.h (rtx_def): Add orignal_regno to u2.
486 (ORIGINAL_REGNO): Use it instead of field 1.
487 (REG_ATTRS): Lower field index accordingly.
488 * gengtype.c (adjust_field_rtx_def): Remove handling of
489 ORIGINAL_REGNO. Move REG_ATTRS index down.
490 * print-rtl.c (print_rtx): Move ORIGINAL_REGNO handling to the
491 code that prints the REGNO.
493 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
495 * print-rtl.c (print_rtx): Guard whole '0' block with ifndef
498 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
500 * rtl.h (rtx_def): Mark u2 as GTY ((skip)).
502 2014-05-13 Bin Cheng <bin.cheng@arm.com>
504 * tree-ssa-loop-ivopts.c (contain_complex_addr_expr): New.
505 (alloc_iv): Lower base expressions containing ADDR_EXPR.
507 2014-05-13 Ian Bolton <ian.bolton@arm.com>
509 * config/aarch64/aarch64-protos.h
510 (aarch64_hard_regno_caller_save_mode): New prototype.
511 * config/aarch64/aarch64.c (aarch64_hard_regno_caller_save_mode):
513 * config/aarch64/aarch64.h (HARD_REGNO_CALLER_SAVE_MODE): New macro.
515 2014-05-13 Christian Bruel <christian.bruel@st.com>
517 * target.def (mode_switching): New hook vector.
518 (mode_emit, mode_needed, mode_after, mode_entry): New hooks.
519 (mode_exit, modepriority_to_mode): Likewise.
520 * mode-switching.c (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Hookify.
521 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
522 * target.h: Include tm.h and hard-reg-set.h.
523 * doc/tm.texi.in (EMIT_MODE_SET, MODE_NEEDED, MODE_AFTER, MODE_ENTRY)
524 (MODE_EXIT, MODE_PRIORITY_TO_MODE): Delete and hookify.
525 * doc/tm.texi Regenerate.
526 * config/sh/sh.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
527 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
528 * config/sh/sh.c (sh_emit_mode_set, sh_mode_priority): Hookify.
529 (sh_mode_needed, sh_mode_after, sh_mode_entry, sh_mode_exit): Likewise.
530 * config/i386/i386.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
531 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
532 * config/i386/i386-protos.h (ix86_mode_needed, ix86_mode_after)
533 (ix86_mode_entrym, ix86_emit_mode_set): Remove external declaration.
534 * config/i386/i386.c (ix86_mode_needed, ix86_mode_after, ix86_mode_exit,
535 (ix86_mode_entry, ix86_mode_priority, ix86_emit_mode_set): Hookify.
536 * config/epiphany/epiphany.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY):
538 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
539 * config/epiphany/epiphany-protos.h (epiphany_mode_needed)
540 (emit_set_fp_mode, epiphany_mode_entry_exit, epiphany_mode_after)
541 (epiphany_mode_priority_to_mode): Remove declaration.
542 * config/epiphany/epiphany.c (emit_set_fp_mode): Hookify.
543 (epiphany_mode_needed, epiphany_mode_priority_to_mode): Likewise.
544 (epiphany_mode_entry, epiphany_mode_exit, epiphany_mode_after):
546 (epiphany_mode_priority_to_mode): Change priority type. Hookify.
547 (epiphany_mode_needed, epiphany_mode_entry_exit): Hookify.
548 (epiphany_mode_after, epiphany_mode_entry, emit_set_fp_mode): Hookify.
550 2014-05-13 Jakub Jelinek <jakub@redhat.com>
553 * config/i386/i386.c (ix86_expand_set_or_movmem): If count_exp
554 is const0_rtx, return immediately. Don't test count == 0 when
557 2014-05-13 Zhenqiang Chen <zhenqiang.chen@linaro.org>
559 * Makefile.in: add shrink-wrap.o.
560 * config/i386/i386.c: include "shrink-wrap.h"
561 * function.c: Likewise.
562 (requires_stack_frame_p, next_block_for_reg,
563 move_insn_for_shrink_wrap, prepare_shrink_wrap,
564 dup_block_and_redirect): Move to shrink-wrap.c
565 (thread_prologue_and_epilogue_insns): Extract three code segments
566 as functions in shrink-wrap.c
567 * function.h: Move #ifdef HAVE_simple_return ... #endif block to
569 * shrink-wrap.c: New file.
570 * shrink-wrap.h: New file.
572 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
574 * doc/extend.texi: Reflect current numbers of pragmas. Remove
575 reference to Solaris.
577 2014-05-12 Mike Stump <mikestump@comcast.net>
580 * genattrtab.c (filename): Add.
581 (convert_set_attr_alternative): Improve error message.
582 (check_defs): Restore read_md_filename for error messages.
583 (gen_insn): Save filename.
585 2014-05-12 Dimitris Papavasiliou <dpapavas@gmail.com>
587 * doc/invoke.texi: Document new switches -Wno-shadow-ivar,
588 -fno-local-ivars and -fivar-visibility.
589 * c-family/c.opt: Make -Wshadow also implicitly enable
592 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
594 * doc/tm.texi: Remove reference to deleted macro.
595 * doc/tm.texi.in: Likewise.
597 2014-05-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
600 * config/avr/avr.c (avr_out_store_psi): Use correct constant
603 2014-05-12 Georg-Johann Lay <avr@gjlay.de>
605 * config/arm/arm.h (License): Add GCC Runtime Library Exception.
606 * config/arm/aout.h (License): Same.
607 * config/arm/bpabi.h (License): Same.
608 * config/arm/elf.h (License): Same.
609 * config/arm/linux-elf.h (License): Same.
610 * config/arm/linux-gas.h (License): Same.
611 * config/arm/netbsd-elf.h (License): Same.
612 * config/arm/uclinux-eabi.h (License): Same.
613 * config/arm/uclinux-elf.h (License): Same.
614 * config/arm/vxworks.h (License): Same.
616 2014-05-11 Jakub Jelinek <jakub@redhat.com>
618 * tree.h (OMP_CLAUSE_LINEAR_STMT): Define.
619 * tree.c (omp_clause_num_ops): Increase OMP_CLAUSE_LINEAR
620 number of operands to 3.
621 (walk_tree_1): Walk all operands of OMP_CLAUSE_LINEAR.
622 * tree-nested.c (convert_nonlocal_omp_clauses,
623 convert_local_omp_clauses): Handle OMP_CLAUSE_DEPEND.
624 * gimplify.c (gimplify_scan_omp_clauses): Handle
625 OMP_CLAUSE_LINEAR_STMT.
626 * omp-low.c (lower_rec_input_clauses): Fix typo.
627 (maybe_add_implicit_barrier_cancel, lower_omp_1): Add
628 cast between Fortran boolean_type_node and C _Bool if
631 2014-05-11 Richard Sandiford <rdsandiford@googlemail.com>
633 PR tree-optimization/61136
634 * wide-int.h (multiple_of_p): Define a version that doesn't return
636 * fold-const.c (extract_muldiv_1): Use wi::multiple_of_p instead of an
637 integer_zerop/const_binop pair.
638 (multiple_of_p): Likewise, converting both operands to widest_int
641 2014-05-09 Teresa Johnson <tejohnson@google.com>
643 * cgraphunit.c (analyze_functions): Use correct dump file.
645 2014-05-09 Florian Weimer <fweimer@redhat.com>
647 * cfgexpand.c (stack_protect_decl_p): New function, extracted from
649 (stack_protect_return_slot_p): New function.
650 (expand_used_vars): Call stack_protect_decl_p and
651 stack_protect_return_slot_p for -fstack-protector-strong.
653 2014-05-09 Kenneth Zadeck <zadeck@naturalbridge.com>
656 * fold-const.c (fold_binary_loc): Changed width of mask.
658 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
660 * config/avr/avr-fixed.md (round<mode>3): Use -1U instead of -1 in
661 unsigned int initializers for regno_in, regno_out.
663 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
666 * config/avr/avr.md (cc): Add new attribute set_vzn.
667 (addqi3, addqq3, adduqq3, subqi3, subqq3, subuqq3, negqi2) [cc]:
668 Set cc insn attribute to set_vzn instead of set_zn for alternatives
669 with INC, DEC or NEG.
670 * config/avr/avr.c (avr_notice_update_cc): Handle SET_VZN.
671 (avr_out_plus_1): ADIW sets cc0 to CC_SET_CZN.
672 INC, DEC and ADD+ADC set cc0 to CC_CLOBBER.
674 2014-05-09 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
677 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
679 * wide-int.cc (UTItype): Define.
680 (UDWtype): Define for appropriate W_TYPE_SIZE.
682 2014-05-09 Richard Biener <rguenther@suse.de>
684 * Makefile.in (GTFILES): Remove tree-ssa-propagate.c.
685 * tree-ssa-propagate.c: Do not include gt-tree-ssa-propagate.h.
686 (interesting_ssa_edges, varying_ssa_edges): Move out of GC space.
687 (add_ssa_edge, process_ssa_edge_worklist, ssa_prop_init,
688 ssa_propagate): Adjust.
690 2014-05-08 Jeff Law <law@redhat.com>
692 PR tree-optimization/61009
693 * tree-ssa-threadedge.c (thread_through_normal_block): Return a
694 tri-state rather than a boolean. When a block is too big to
695 thread through, inform caller via negative return value.
696 (thread_across_edge): If a block was too big for normal threading,
697 then it's too big for a joiner too, so remove temporary equivalences
698 and return immediately.
700 2014-05-08 Manuel López-Ibáñez <manu@gcc.gnu.org>
701 Matthias Klose <doko@ubuntu.com>
704 * optc-gen.awk: Fix option handling for -Wunused-parameter.
706 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
709 * config/i386/i386.c (PTA_HASWELL): Remove PTA_RTM.
711 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
714 * config/alpha/alpha.c: Include gimple-iterator.h.
715 (alpha_gimple_fold_builtin): New function. Move
716 ALPHA_BUILTIN_UMULH folding from ...
717 (alpha_fold_builtin): ... here.
718 (TARGET_GIMPLE_FOLD_BUILTIN): New define.
720 2014-05-08 Wei Mi <wmi@google.com>
723 * config/i386/i386.c (ix86_compute_frame_layout): Update
724 preferred_stack_boundary for call, expanded from tls descriptor.
725 * config/i386/i386.md (*tls_global_dynamic_32_gnu): Update RTX
726 to depend on SP register.
727 (*tls_local_dynamic_base_32_gnu): Ditto.
728 (*tls_local_dynamic_32_once): Ditto.
729 (tls_global_dynamic_64_<mode>): Set
730 ix86_tls_descriptor_calls_expanded_in_cfun.
731 (tls_local_dynamic_base_64_<mode>): Ditto.
732 (tls_global_dynamic_32): Set
733 ix86_tls_descriptor_calls_expanded_in_cfun. Update RTX
734 to depend on SP register.
735 (tls_local_dynamic_base_32): Ditto.
737 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
739 * config/arm/arm_neon.h: Update comment.
740 * config/arm/neon-docgen.ml: Delete.
741 * config/arm/neon-gen.ml: Delete.
742 * doc/arm-neon-intrinsics.texi: Update comment.
744 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
746 * config/arm/arm_neon_builtins.def (vadd, vsub): Only define the v2sf
748 (vand, vorr, veor, vorn, vbic): Remove.
749 * config/arm/neon.md (neon_vadd, neon_vsub, neon_vadd_unspec): Adjust
751 (neon_vsub_unspec): Likewise.
752 (neon_vorr, neon_vand, neon_vbic, neon_veor, neon_vorn): Remove.
754 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
756 * config/arm/arm_neon.h (vadd_s8): GNU C implementation
757 (vadd_s16): Likewise.
758 (vadd_s32): Likewise.
759 (vadd_f32): Likewise.
761 (vadd_u16): Likewise.
762 (vadd_u32): Likewise.
763 (vadd_s64): Likewise.
764 (vadd_u64): Likewise.
765 (vaddq_s8): Likewise.
766 (vaddq_s16): Likewise.
767 (vaddq_s32): Likewise.
768 (vaddq_s64): Likewise.
769 (vaddq_f32): Likewise.
770 (vaddq_u8): Likewise.
771 (vaddq_u16): Likewise.
772 (vaddq_u32): Likewise.
773 (vaddq_u64): Likewise.
775 (vmul_s16): Likewise.
776 (vmul_s32): Likewise.
777 (vmul_f32): Likewise.
779 (vmul_u16): Likewise.
780 (vmul_u32): Likewise.
782 (vmulq_s8): Likewise.
783 (vmulq_s16): Likewise.
784 (vmulq_s32): Likewise.
785 (vmulq_f32): Likewise.
786 (vmulq_u8): Likewise.
787 (vmulq_u16): Likewise.
788 (vmulq_u32): Likewise.
790 (vsub_s16): Likewise.
791 (vsub_s32): Likewise.
792 (vsub_f32): Likewise.
794 (vsub_u16): Likewise.
795 (vsub_u32): Likewise.
796 (vsub_s64): Likewise.
797 (vsub_u64): Likewise.
798 (vsubq_s8): Likewise.
799 (vsubq_s16): Likewise.
800 (vsubq_s32): Likewise.
801 (vsubq_s64): Likewise.
802 (vsubq_f32): Likewise.
803 (vsubq_u8): Likewise.
804 (vsubq_u16): Likewise.
805 (vsubq_u32): Likewise.
806 (vsubq_u64): Likewise.
808 (vand_s16): Likewise.
809 (vand_s32): Likewise.
811 (vand_u16): Likewise.
812 (vand_u32): Likewise.
813 (vand_s64): Likewise.
814 (vand_u64): Likewise.
815 (vandq_s8): Likewise.
816 (vandq_s16): Likewise.
817 (vandq_s32): Likewise.
818 (vandq_s64): Likewise.
819 (vandq_u8): Likewise.
820 (vandq_u16): Likewise.
821 (vandq_u32): Likewise.
822 (vandq_u64): Likewise.
824 (vorr_s16): Likewise.
825 (vorr_s32): Likewise.
827 (vorr_u16): Likewise.
828 (vorr_u32): Likewise.
829 (vorr_s64): Likewise.
830 (vorr_u64): Likewise.
831 (vorrq_s8): Likewise.
832 (vorrq_s16): Likewise.
833 (vorrq_s32): Likewise.
834 (vorrq_s64): Likewise.
835 (vorrq_u8): Likewise.
836 (vorrq_u16): Likewise.
837 (vorrq_u32): Likewise.
838 (vorrq_u64): Likewise.
840 (veor_s16): Likewise.
841 (veor_s32): Likewise.
843 (veor_u16): Likewise.
844 (veor_u32): Likewise.
845 (veor_s64): Likewise.
846 (veor_u64): Likewise.
847 (veorq_s8): Likewise.
848 (veorq_s16): Likewise.
849 (veorq_s32): Likewise.
850 (veorq_s64): Likewise.
851 (veorq_u8): Likewise.
852 (veorq_u16): Likewise.
853 (veorq_u32): Likewise.
854 (veorq_u64): Likewise.
856 (vbic_s16): Likewise.
857 (vbic_s32): Likewise.
859 (vbic_u16): Likewise.
860 (vbic_u32): Likewise.
861 (vbic_s64): Likewise.
862 (vbic_u64): Likewise.
863 (vbicq_s8): Likewise.
864 (vbicq_s16): Likewise.
865 (vbicq_s32): Likewise.
866 (vbicq_s64): Likewise.
867 (vbicq_u8): Likewise.
868 (vbicq_u16): Likewise.
869 (vbicq_u32): Likewise.
870 (vbicq_u64): Likewise.
872 (vorn_s16): Likewise.
873 (vorn_s32): Likewise.
875 (vorn_u16): Likewise.
876 (vorn_u32): Likewise.
877 (vorn_s64): Likewise.
878 (vorn_u64): Likewise.
879 (vornq_s8): Likewise.
880 (vornq_s16): Likewise.
881 (vornq_s32): Likewise.
882 (vornq_s64): Likewise.
883 (vornq_u8): Likewise.
884 (vornq_u16): Likewise.
885 (vornq_u32): Likewise.
886 (vornq_u64): Likewise.
888 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
890 * wide-int.cc (UTItype): Define.
891 (UDWtype): Define for appropriate W_TYPE_SIZE.
893 2014-05-08 Marc Glisse <marc.glisse@inria.fr>
895 PR tree-optimization/59100
896 * tree-ssa-phiopt.c: Include tree-inline.h.
897 (neutral_element_p, absorbing_element_p): New functions.
898 (value_replacement): Handle conditional binary operations with a
899 neutral or absorbing element.
901 2014-05-08 Richard Biener <rguenther@suse.de>
903 * tree-ssa-sccvn.c (vn_get_expr_for): Valueize operands before
904 folding the expression.
905 (valueize_expr): Remove.
906 (visit_reference_op_load): Do not valueize the result of
908 (simplify_binary_expression): Likewise.
909 (simplify_unary_expression): Likewise.
911 2014-05-08 Richard Biener <rguenther@suse.de>
913 * gimplify.c (gimplify_call_expr): Use saved fnptrtype for
914 looking at TYPE_ARG_TYPES.
916 2014-05-08 Richard Biener <rguenther@suse.de>
918 * gimple-fold.c (gimple_fold_stmt_to_constant_1): Remove
919 pointer propagation special-case.
921 2014-05-08 Bin Cheng <bin.cheng@arm.com>
923 * tree-affine.c (tree_to_aff_combination): Handle MEM_REF for
924 core part of address expressions.
926 2014-05-08 Alan Modra <amodra@gmail.com>
929 * config/rs6000/rs6000.c (expand_block_move): Allow 64-bit
930 loads and stores when -mno-strict-align at any alignment.
931 (expand_block_clear): Similarly. Also correct calculation of
934 2014-05-07 Thomas Preud'homme <thomas.preudhomme@arm.com>
937 * tree-complex.c (expand_complex_move): Keep line info when expanding
939 * tree-ssa-uninit.c (warn_uninit): New argument. Ignore assignment
940 of complex expression. Use new argument to display correct location
941 for values coming from phi statement.
942 (warn_uninitialized_vars): Adapt to new signature of warn_uninit.
943 (warn_uninitialized_phi): Pass location of phi argument to
945 * tree-ssa.c (ssa_undefined_value_p): For SSA_NAME initialized by a
946 COMPLEX_EXPR, recurse on each part of the COMPLEX_EXPR.
948 2014-05-07 Segher Boessenkool <segher@kernel.crashing.org>
950 * config/rs6000/predicates.md (indexed_address_mem): New.
951 * config/rs6000/rs6000.md (type): Remove load_ext, load_ext_u,
952 load_ext_ux, load_ux, load_u, store_ux, store_u, fpload_ux, fpload_u,
953 fpstore_ux, fpstore_u.
954 (sign_extend, indexed, update): New.
955 (cell_micro): Adjust.
956 (*zero_extend<mode>di2_internal1, *zero_extendsidi2_lfiwzx,
957 *extendsidi2_lfiwax, *extendsidi2_nocell, *extendsfdf2_fpr,
958 *movsi_internal1, *movsi_internal1_single, *movhi_internal,
959 *movqi_internal, *movcc_internal1, mov<mode>_hardfloat,
960 *mov<mode>_softfloat, *mov<mode>_hardfloat32, *mov<mode>_hardfloat64,
961 *mov<mode>_softfloat64, *movdi_internal32, *movdi_internal64,
962 *mov<mode>_string, *ldmsi8, *ldmsi7, *ldmsi6, *ldmsi5, *ldmsi4,
963 *ldmsi3, *stmsi8, *stmsi7, *stmsi6, *stmsi5, *stmsi4, *stmsi3,
964 *movdi_update1, movdi_<mode>_update, movdi_<mode>_update_stack,
965 *movsi_update1, *movsi_update2, movsi_update, movsi_update_stack,
966 *movhi_update1, *movhi_update2, *movhi_update3, *movhi_update4,
967 *movqi_update1, *movqi_update2, *movqi_update3, *movsf_update1,
968 *movsf_update2, *movsf_update3, *movsf_update4, *movdf_update1,
969 *movdf_update2, load_toc_aix_si, load_toc_aix_di, probe_stack_<mode>,
970 *stmw, *lmw, as well as 10 anonymous patterns): Adjust.
972 * config/rs6000/dfp.md (movsd_store, movsd_load): Adjust.
973 * config/rs6000/vsx.md (*vsx_movti_32bit, *vsx_extract_<mode>_load,
974 *vsx_extract_<mode>_store): Adjust.
975 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_microcoded_insn,
976 is_cracked_insn, insn_must_be_first_in_group,
977 insn_must_be_last_in_group): Adjust.
979 * config/rs6000/40x.md (ppc403-load, ppc403-store, ppc405-float):
981 * config/rs6000/440.md (ppc440-load, ppc440-store, ppc440-fpload,
982 ppc440-fpstore): Adjust.
983 * config/rs6000/476.md (ppc476-load, ppc476-store, ppc476-fpload,
984 ppc476-fpstore): Adjust.
985 * config/rs6000/601.md (ppc601-load, ppc601-store, ppc601-fpload,
986 ppc601-fpstore): Adjust.
987 * config/rs6000/603.md (ppc603-load, ppc603-store, ppc603-fpload):
989 * config/rs6000/6xx.md (ppc604-load, ppc604-store, ppc604-fpload):
991 * config/rs6000/7450.md (ppc7450-load, ppc7450-store, ppc7450-fpload,
992 ppc7450-fpstore): Adjust.
993 * config/rs6000/7xx.md (ppc750-load, ppc750-store): Adjust.
994 * config/rs6000/8540.md (ppc8540_load, ppc8540_store): Adjust.
995 * config/rs6000/a2.md (ppca2-load, ppca2-fp-load, ppca2-fp-store):
997 * config/rs6000/cell.md (cell-load, cell-load-ux, cell-load-ext,
998 cell-fpload, cell-fpload-update, cell-store, cell-store-update,
999 cell-fpstore, cell-fpstore-update): Adjust.
1000 * config/rs6000/e300c2c3.md (ppce300c3_load, ppce300c3_fpload,
1001 ppce300c3_store, ppce300c3_fpstore): Adjust.
1002 * config/rs6000/e500mc.md (e500mc_load, e500mc_fpload, e500mc_store,
1003 e500mc_fpstore): Adjust.
1004 * config/rs6000/e500mc64.md (e500mc64_load, e500mc64_fpload,
1005 e500mc64_store, e500mc64_fpstore): Adjust.
1006 * config/rs6000/e5500.md (e5500_load, e5500_fpload, e5500_store,
1007 e5500_fpstore): Adjust.
1008 * config/rs6000/e6500.md (e6500_load, e6500_fpload, e6500_store,
1009 e6500_fpstore): Adjust.
1010 * config/rs6000/mpc.md (mpccore-load, mpccore-store, mpccore-fpload):
1012 * config/rs6000/power4.md (power4-load, power4-load-ext,
1013 power4-load-ext-update, power4-load-ext-update-indexed,
1014 power4-load-update-indexed, power4-load-update, power4-fpload,
1015 power4-fpload-update, power4-store, power4-store-update,
1016 power4-store-update-indexed, power4-fpstore, power4-fpstore-update):
1018 * config/rs6000/power5.md (power5-load, power5-load-ext,
1019 power5-load-ext-update, power5-load-ext-update-indexed,
1020 power5-load-update-indexed, power5-load-update, power5-fpload,
1021 power5-fpload-update, power5-store, power5-store-update,
1022 power5-store-update-indexed, power5-fpstore, power5-fpstore-update):
1024 * config/rs6000/power6.md (power6-load, power6-load-ext,
1025 power6-load-update, power6-load-update-indexed,
1026 power6-load-ext-update, power6-load-ext-update-indexed, power6-fpload,
1027 power6-fpload-update, power6-store, power6-store-update,
1028 power6-store-update-indexed, power6-fpstore, power6-fpstore-update):
1030 * config/rs6000/power7.md (power7-load, power7-load-ext,
1031 power7-load-update, power7-load-update-indexed,
1032 power7-load-ext-update, power7-load-ext-update-indexed, power7-fpload,
1033 power7-fpload-update, power7-store, power7-store-update,
1034 power7-store-update-indexed, power7-fpstore, power7-fpstore-update):
1036 * config/rs6000/power8.md (power8-load, power8-load-update,
1037 power8-load-ext, power8-load-ext-update, power8-fpload,
1038 power8-fpload-update, power8-store, power8-store-update-indexed,
1039 power8-fpstore, power8-fpstore-update): Adjust.
1040 * config/rs6000/rs64.md (rs64a-load, rs64a-store, rs64a-fpload):
1042 * config/rs6000/titan.md (titan_lsu_load, titan_lsu_fpload,
1043 titan_lsu_store, titan_lsu_fpstore): Adjust.
1044 * config/rs6000/xfpu.md (fp-load, fp-store): Adjust.
1046 2014-05-07 Oleg Endo <olegendo@gcc.gnu.org>
1049 * config/sh/sh-mem.cc (sh_expand_strlen): Use loop when emitting
1050 unrolled byte insns. Emit address increments after move insns.
1052 2014-05-07 David Malcolm <dmalcolm@redhat.com>
1054 * gimple.h (gimple_builtin_call_types_compatible_p): Accept a
1055 const_gimple, rather than a gimple.
1056 (gimple_call_builtin_p): Likewise, for the three variants.
1058 * gimple.c (gimple_builtin_call_types_compatible_p): Likewise.
1059 (gimple_call_builtin_p): Likewise, for the three variants.
1061 2014-05-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
1063 PR tree-optimization/61095
1064 * tree-ssanames.c (get_nonzero_bits): Fix type extension in wi::shwi.
1066 2014-05-07 Richard Biener <rguenther@suse.de>
1068 PR tree-optimization/61034
1069 * tree-ssa-alias.c (call_may_clobber_ref_p_1): Export.
1070 (maybe_skip_until): Use translate to take into account
1071 lattices when trying to do disambiguations.
1072 (get_continuation_for_phi_1): Likewise.
1073 (get_continuation_for_phi): Adjust for added translate arguments.
1074 (walk_non_aliased_vuses): Likewise.
1075 * tree-ssa-alias.h (get_continuation_for_phi): Adjust prototype.
1076 (walk_non_aliased_vuses): Likewise.
1077 (call_may_clobber_ref_p_1): Declare.
1078 * tree-ssa-sccvn.c (vn_reference_lookup_3): Also disambiguate against
1079 calls. Stop early if we are only supposed to disambiguate.
1080 * tree-ssa-pre.c (translate_vuse_through_block): Adjust.
1082 2014-05-07 Joern Rennecke <joern.rennecke@embecosm.com>
1084 * config/epiphany/epiphany.c (epiphany_handle_interrupt_attribute):
1085 Emit an error when the function has arguments.
1087 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
1089 * cfgloop.h (unswitch_loops): Remove.
1090 * doc/passes.texi: Remove references to loop-unswitch.c
1091 * timevar.def (TV_LOOP_UNSWITCH): Remove.
1093 2014-05-07 Evgeny Stupachenko <evstupac@gmail.com>
1095 * tree-vect-data-refs.c (vect_grouped_load_supported): New
1096 check for loads group of length 3.
1097 (vect_permute_load_chain): New permutations for loads group of
1099 * tree-vect-stmts.c (vect_model_load_cost): Change cost
1100 of vec_perm_shuffle for the new permutations.
1102 2014-05-07 Alan Lawrence <alan.lawrence@arm.com>
1104 * config/aarch64/arm_neon.h (vtrn1_f32, vtrn1_p8, vtrn1_p16, vtrn1_s8,
1105 vtrn1_s16, vtrn1_s32, vtrn1_u8, vtrn1_u16, vtrn1_u32, vtrn1q_f32,
1106 vtrn1q_f64, vtrn1q_p8, vtrn1q_p16, vtrn1q_s8, vtrn1q_s16, vtrn1q_s32,
1107 vtrn1q_s64, vtrn1q_u8, vtrn1q_u16, vtrn1q_u32, vtrn1q_u64, vtrn2_f32,
1108 vtrn2_p8, vtrn2_p16, vtrn2_s8, vtrn2_s16, vtrn2_s32, vtrn2_u8,
1109 vtrn2_u16, vtrn2_u32, vtrn2q_f32, vtrn2q_f64, vtrn2q_p8, vtrn2q_p16,
1110 vtrn2q_s8, vtrn2q_s16, vtrn2q_s32, vtrn2q_s64, vtrn2q_u8, vtrn2q_u16,
1111 vtrn2q_u32, vtrn2q_u64): Replace temporary asm with __builtin_shuffle.
1113 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
1115 * loop-unswitch.c: Delete.
1117 2014-05-07 Richard Biener <rguenther@suse.de>
1119 * config.gcc: Always set need_64bit_hwint to yes.
1121 2014-05-07 Chung-Ju Wu <jasonwucj@gmail.com>
1123 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Have it in favor
1124 of using optimize_size.
1126 2014-05-06 Mike Stump <mikestump@comcast.net>
1128 * wide-int.h (wi::int_traits <HOST_WIDE_INT>): Always define.
1130 2014-05-06 Joseph Myers <joseph@codesourcery.com>
1132 * config/i386/sse.md (*mov<mode>_internal)
1133 (*<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>)
1134 (*<sse2_avx_avx512f>_loaddqu<mode><mask_name>)
1135 (<sse>_andnot<mode>3, <code><mode>3, *andnot<mode>3)
1136 (*<code><mode>3, *andnot<mode>3<mask_name>)
1137 (<mask_codefor><code><mode>3<mask_name>): Only consider
1138 TARGET_SSE_PACKED_SINGLE_INSN_OPTIMAL for modes of size 16.
1140 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
1143 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
1145 * lra-constraints.c (valid_address_p): Move earlier in file.
1146 Add a constraint argument to the address_info version.
1147 (satisfies_memory_constraint_p): New function.
1148 (satisfies_address_constraint_p): Likewise.
1149 (process_alt_operands, curr_insn_transform): Use them.
1150 (process_address): Pass the constraint to valid_address_p when
1151 checking address operands.
1153 2014-05-06 Richard Sandiford <r.sandiford@uk.ibm.com>
1155 * lto-cgraph.c (compute_ltrans_boundary): Make node variables local
1156 to their respective blocks. Fix inadvertent use of "node".
1158 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
1160 * emit-rtl.c (init_derived_machine_modes): New functionm, split
1162 (init_emit_once): ...here.
1163 * rtl.h (init_derived_machine_modes): Declare.
1164 * toplev.c (do_compile): Call it even if no_backend.
1166 2014-05-06 Kenneth Zadeck <zadeck@naturalbridge.com>
1167 Mike Stump <mikestump@comcast.net>
1168 Richard Sandiford <rdsandiford@googlemail.com>
1169 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1171 * alias.c (ao_ref_from_mem): Use wide-int interfaces.
1172 (rtx_equal_for_memref_p): Update comment.
1173 (adjust_offset_for_component_ref): Use wide-int interfaces.
1174 * builtins.c (get_object_alignment_2): Likewise.
1175 (c_readstr): Likewise.
1176 (target_char_cast): Add comment.
1177 (determine_block_size): Use wide-int interfaces.
1178 (expand_builtin_signbit): Likewise.
1179 (fold_builtin_int_roundingfn): Likewise.
1180 (fold_builtin_bitop): Likewise.
1181 (fold_builtin_bswap): Likewise.
1182 (fold_builtin_logarithm): Use signop.
1183 (fold_builtin_pow): Likewise.
1184 (fold_builtin_memory_op): Use wide-int interfaces.
1185 (fold_builtin_object_size): Likewise.
1186 * cfgloop.c (alloc_loop): Initialize nb_iterations_upper_bound and
1187 nb_iterations_estimate.
1188 (record_niter_bound): Use wide-int interfaces.
1189 (get_estimated_loop_iterations_int): Likewise.
1190 (get_estimated_loop_iterations): Likewise.
1191 (get_max_loop_iterations): Likewise.
1192 * cfgloop.h: Include wide-int.h.
1193 (struct nb_iter_bound): Change bound to widest_int.
1194 (struct loop): Change nb_iterations_upper_bound and
1195 nb_iterations_estimate to widest_int.
1196 (record_niter_bound): Switch to use widest_int.
1197 (get_estimated_loop_iterations): Likewise.
1198 (get_max_loop_iterations): Likewise.
1199 (gcov_type_to_double_int): Rename to gcov_type_to_wide_int and
1200 update for wide-int.
1201 * cgraph.c (cgraph_add_thunk): Use wide-int interfaces.
1202 * combine.c (try_combine): Likewise.
1203 (subst): Use CONST_SCALAR_INT_P rather than CONST_INT_P.
1204 * config/aarch64/aarch64.c (aapcs_vfp_sub_candidate): Use wide-int
1206 (aarch64_float_const_representable_p): Likewise.
1207 * config/arc/arc.c: Include wide-int.h.
1208 (arc_can_use_doloop_p): Use wide-int interfaces.
1209 * config/arm/arm.c (aapcs_vfp_sub_candidate): Likewise.
1210 (vfp3_const_double_index): Likewise.
1211 * config/avr/avr.c (avr_out_round): Likewise.
1212 (avr_fold_builtin): Likewise.
1213 * config/bfin/bfin.c (bfin_local_alignment): Likewise.
1214 (bfin_can_use_doloop_p): Likewise.
1215 * config/darwin.c (darwin_mergeable_constant_section): Likewise.
1216 (machopic_select_rtx_section): Update to handle CONST_WIDE_INT.
1217 * config/i386/i386.c: Include wide-int.h.
1218 (ix86_data_alignment): Use wide-int interfaces.
1219 (ix86_local_alignment): Likewise.
1220 (ix86_emit_swsqrtsf): Update real_from_integer.
1221 * config/msp430/msp430.c (msp430_attr): Use wide-int interfaces.
1222 * config/nds32/nds32.c (nds32_insert_attributes): Likewise.
1223 * config/rs6000/predicates.md (any_operand): Add const_wide_int.
1224 (zero_constant): Likewise.
1225 (input_operand): Likewise.
1226 (splat_input_operand): Likewise.
1227 (non_logical_cint_operand): Change const_double to const_wide_int.
1228 * config/rs6000/rs6000.c (num_insns_constant): Handle CONST_WIDE_INT.
1229 (easy_altivec_constant): Remove comment.
1230 (paired_expand_vector_init): Use CONSTANT_P.
1231 (rs6000_legitimize_address): Handle CONST_WIDE_INT.
1232 (rs6000_emit_move): Update checks.
1233 (rs6000_aggregate_candidate): Use wide-int interfaces.
1234 (rs6000_expand_ternop_builtin): Likewise.
1235 (rs6000_output_move_128bit): Handle CONST_WIDE_INT.
1236 (rs6000_assemble_integer): Likewise.
1237 (rs6000_hash_constant): Likewise.
1238 (output_toc): Likewise.
1239 (rs6000_rtx_costs): Likewise.
1240 (rs6000_emit_swrsqrt); Update call to real_from_integer.
1241 * config/rs6000/rs6000-c.c: Include wide-int.h.
1242 (altivec_resolve_overloaded_builtin): Use wide-int interfaces.
1243 * config/rs6000/rs6000.h (TARGET_SUPPORTS_WIDE_INT): New.
1244 * config/rs6000/rs6000.md: Use const_scalar_int_operand.
1245 Handle CONST_WIDE_INT.
1246 * config/sol2-c.c (solaris_pragma_align): Change low to unsigned HWI.
1247 Use tree_fits_uhwi_p.
1248 * config/sparc/sparc.c: Include wide-int.h.
1249 (sparc_fold_builtin): Use wide-int interfaces.
1250 * config/vax/vax.c: Include wide-int.h.
1251 (vax_float_literal): Use real_from_integer.
1252 * coretypes.h (struct hwivec_def): New.
1254 (const_hwivec): New.
1255 * cse.c (hash_rtx_cb): Handle CONST_WIDE_INT.
1256 (equiv_constant): Handle CONST_WIDE_INT.
1257 * cselib.c (rtx_equal_for_cselib_1): Use CASE_CONST_UNIQUE.
1258 (cselib_hash_rtx): Handle CONST_WIDE_INT.
1259 * dbxout.c (stabstr_U): Use wide-int interfaces.
1260 (dbxout_type): Update to use cst_fits_shwi_p.
1261 * defaults.h (LOG2_BITS_PER_UNIT): Define.
1262 (TARGET_SUPPORTS_WIDE_INT): Add default.
1263 * dfp.c: Include wide-int.h.
1264 (decimal_real_to_integer2): Use wide-int interfaces and rename to
1265 decimal_real_to_integer.
1266 * dfp.h (decimal_real_to_integer2): Return a wide_int and rename to
1267 decimal_real_to_integer.
1268 * doc/generic.texi (Constant expressions): Update for wide_int.
1269 * doc/rtl.texi (const_double): Likewise.
1270 (const_wide_int, CONST_WIDE_INT, CONST_WIDE_INT_VEC): New.
1271 (CONST_WIDE_INT_NUNITS, CONST_WIDE_INT_ELT): New.
1272 * doc/tm.texi.in (REAL_VALUE_TO_INT): Remove.
1273 (REAL_VALUE_FROM_INT): Remove.
1274 (TARGET_SUPPORTS_WIDE_INT): New.
1275 * doc/tm.texi: Regenerate.
1276 * dojump.c (prefer_and_bit_test): Use wide-int interfaces.
1277 * double-int.h: Include wide-int.h.
1278 (struct wi::int_traits): New.
1279 * dwarf2out.c (get_full_len): New.
1280 (dw_val_equal_p): Add case dw_val_class_wide_int.
1281 (size_of_loc_descr): Likewise.
1282 (output_loc_operands): Likewise.
1283 (insert_double): Remove.
1284 (insert_wide_int): New.
1286 (print_die): Add case dw_val_class_wide_int.
1287 (attr_checksum): Likewise.
1288 (attr_checksum_ordered): Likewise.
1289 (same_dw_val_p): Likewise.
1290 (size_of_die): Likewise.
1291 (value_format): Likewise.
1292 (output_die): Likewise.
1293 (double_int_type_size_in_bits): Rename to offset_int_type_size_in_bits.
1295 (clz_loc_descriptor): Use wide-int interfaces.
1296 (mem_loc_descriptor): Likewise. Handle CONST_WIDE_INT.
1297 (loc_descriptor): Use wide-int interfaces. Handle CONST_WIDE_INT.
1298 (round_up_to_align): Use wide-int interfaces.
1299 (field_byte_offset): Likewise.
1300 (insert_double): Rename to insert_wide_int. Use wide-int interfaces.
1301 (add_const_value_attribute): Handle CONST_WIDE_INT. Update
1302 CONST_DOUBLE handling. Use wide-int interfaces.
1303 (add_bound_info): Use tree_fits_uhwi_p. Use wide-int interfaces.
1304 (gen_enumeration_type_die): Use add_AT_wide.
1305 (hash_loc_operands): Add case dw_val_class_wide_int.
1306 (compare_loc_operands): Likewise.
1307 * dwarf2out.h: Include wide-int.h.
1308 (wide_int_ptr): New.
1309 (enum dw_val_class): Add dw_val_class_wide_int.
1310 (struct dw_val_struct): Add val_wide.
1311 * emit-rtl.c (const_wide_int_htab): New.
1312 (const_wide_int_htab_hash): New.
1313 (const_wide_int_htab_eq): New.
1314 (lookup_const_wide_int): New.
1315 (const_double_htab_hash): Use wide-int interfaces.
1316 (const_double_htab_eq): Likewise.
1317 (rtx_to_double_int): Conditionally compile for wide-int.
1318 (immed_double_int_const): Rename to immed_wide_int_const and
1319 update for wide-int.
1320 (immed_double_const): Conditionally compile for wide-int.
1321 (init_emit_once): Use wide-int interfaces.
1322 * explow.c (plus_constant): Likewise.
1323 * expmed.c (mask_rtx): Move further up file. Use wide-int interfaces.
1324 (lshift_value): Use wide-int interfaces.
1325 (expand_mult): Likewise.
1326 (choose_multiplier): Likewise.
1327 (expand_smod_pow2): Likewise.
1328 (make_tree): Likewise.
1329 * expr.c (convert_modes): Consolidate handling of constants.
1330 Use wide-int interfaces.
1331 (emit_group_load_1): Add note.
1332 (store_expr): Update comment.
1333 (get_inner_reference): Use wide-int interfaces.
1334 (expand_constructor): Update comment.
1335 (expand_expr_real_2): Use wide-int interfaces.
1336 (expand_expr_real_1): Likewise.
1337 (reduce_to_bit_field_precision): Likewise.
1338 (const_vector_from_tree): Likewise.
1339 * final.c: Include wide-int-print.h.
1340 (output_addr_const): Handle CONST_WIDE_INT. Use CONST_DOUBLE_AS_INT_P.
1341 * fixed-value.c: Include wide-int.h.
1342 (fixed_from_string): Use wide-int interfaces.
1343 (fixed_to_decimal): Likewise.
1344 (fixed_convert_from_real): Likewise.
1345 (real_convert_from_fixed): Likewise.
1346 * fold-const.h (mem_ref_offset): Return an offset_int.
1347 (div_if_zero_remainder): Remove code parameter.
1348 * fold-const.c (div_if_zero_remainder): Remove code parameter.
1349 Use wide-int interfaces.
1350 (may_negate_without_overflow_p): Use wide-int interfaces.
1351 (negate_expr_p): Likewise.
1352 (fold_negate_expr): Likewise.
1353 (int_const_binop_1): Likewise.
1354 (const_binop): Likewise.
1355 (fold_convert_const_int_from_int): Likewise.
1356 (fold_convert_const_int_from_real): Likewise.
1357 (fold_convert_const_int_from_fixed): Likewise.
1358 (fold_convert_const_fixed_from_int): Likewise.
1359 (all_ones_mask_p): Take an unsigned size. Use wide-int interfaces.
1360 (sign_bit_p): Use wide-int interfaces.
1361 (make_range_step): Likewise.
1362 (build_range_check): Likewise. Pass an integer of the correct type
1363 instead of using integer_one_node.
1364 (range_predecessor): Pass an integer of the correct type instead
1365 of using integer_one_node.
1366 (range_successor): Likewise.
1367 (merge_ranges): Likewise.
1368 (unextend): Use wide-int interfaces.
1369 (extract_muldiv_1): Likewise.
1370 (fold_div_compare): Likewise.
1371 (fold_single_bit_test): Likewise.
1372 (fold_sign_changed_comparison): Likewise.
1373 (try_move_mult_to_index): Update calls to div_if_zero_remainder.
1374 (fold_plusminus_mult_expr): Use wide-int interfaces.
1375 (native_encode_int): Likewise.
1376 (native_interpret_int): Likewise.
1377 (fold_unary_loc): Likewise.
1378 (pointer_may_wrap_p): Likewise.
1379 (size_low_cst): Likewise.
1380 (mask_with_tz): Likewise.
1381 (fold_binary_loc): Likewise.
1382 (fold_ternary_loc): Likewise.
1383 (multiple_of_p): Likewise.
1384 (tree_call_nonnegative_warnv_p): Update calls to
1385 tree_int_cst_min_precision and real_from_integer.
1386 (fold_negate_const): Use wide-int interfaces.
1387 (fold_abs_const): Likewise.
1388 (fold_relational_const): Use tree_int_cst_lt.
1389 (round_up_loc): Use wide-int interfaces.
1390 * genemit.c (gen_exp): Add CONST_WIDE_INT case.
1391 * gengenrtl.c (excluded_rtx): Add CONST_WIDE_INT case.
1392 * gengtype.c: Remove include of double-int.h.
1393 (do_typedef): Use wide-int interfaces.
1394 (open_base_files): Add wide-int.h.
1395 (main): Add offset_int and widest_int typedefs.
1396 * gengtype-lex.l: Handle "^".
1397 (CXX_KEYWORD): Add "static".
1398 * gengtype-parse.c (require3): New.
1399 (require_template_declaration): Handle constant template arguments
1400 and nested templates.
1401 * gengtype-state.c: Don't include "double-int.h".
1402 * genpreds.c (write_one_predicate_function): Update comment.
1403 (write_tm_constrs_h): Add check for hval and lval use in
1405 * genrecog.c (validate_pattern): Add CONST_WIDE_INT case.
1406 (add_to_sequence): Likewise.
1407 * gensupport.c (struct std_pred_table): Add const_scalar_int_operand
1408 and const_double_operand.
1409 * gimple.c (preprocess_case_label_vec_for_gimple): Use wide-int
1411 * gimple-fold.c (get_base_constructor): Likewise.
1412 (fold_array_ctor_reference): Likewise.
1413 (fold_nonarray_ctor_reference): Likewise.
1414 (fold_const_aggregate_ref_1): Likewise.
1415 (gimple_val_nonnegative_real_p): Likewise.
1416 (gimple_fold_indirect_ref): Likewise.
1417 * gimple-pretty-print.c (dump_ssaname_info): Likewise.
1418 * gimple-ssa-strength-reduction.c: Include wide-int-print.h.
1419 (struct slsr_cand_d): Change index to be widest_int.
1420 (struct incr_info_d): Change incr to be widest_int.
1421 (alloc_cand_and_find_basis): Use wide-int interfaces.
1422 (slsr_process_phi): Likewise.
1423 (backtrace_base_for_ref): Likewise. Return a widest_int.
1424 (restructure_reference): Take a widest_int instead of a double_int.
1425 (slsr_process_ref): Use wide-int interfaces.
1426 (create_mul_ssa_cand): Likewise.
1427 (create_mul_imm_cand): Likewise.
1428 (create_add_ssa_cand): Likewise.
1429 (create_add_imm_cand): Take a widest_int instead of a double_int.
1430 (slsr_process_add): Use wide-int interfaces.
1431 (slsr_process_cast): Likewise.
1432 (slsr_process_copy): Likewise.
1433 (dump_candidate): Likewise.
1434 (dump_incr_vec): Likewise.
1435 (replace_ref): Likewise.
1436 (cand_increment): Likewise. Return a widest_int.
1437 (cand_abs_increment): Likewise.
1438 (replace_mult_candidate): Take a widest_int instead of a double_int.
1439 (replace_unconditional_candidate): Use wide-int interfaces.
1440 (incr_vec_index): Take a widest_int instead of a double_int.
1441 (create_add_on_incoming_edge): Likewise.
1442 (create_phi_basis): Use wide-int interfaces.
1443 (replace_conditional_candidate): Likewise.
1444 (record_increment): Take a widest_int instead of a double_int.
1445 (record_phi_increments): Use wide-int interfaces.
1446 (phi_incr_cost): Take a widest_int instead of a double_int.
1447 (lowest_cost_path): Likewise.
1448 (total_savings): Likewise.
1449 (analyze_increments): Use wide-int interfaces.
1450 (ncd_with_phi): Take a widest_int instead of a double_int.
1451 (ncd_of_cand_and_phis): Likewise.
1452 (nearest_common_dominator_for_cands): Likewise.
1453 (insert_initializers): Use wide-int interfaces.
1454 (all_phi_incrs_profitable): Likewise.
1455 (replace_one_candidate): Likewise.
1456 (replace_profitable_candidates): Likewise.
1457 * godump.c: Include wide-int-print.h.
1458 (go_output_typedef): Use wide-int interfaces.
1459 * graphite-clast-to-gimple.c (gmp_cst_to_tree): Likewise.
1460 * graphite-sese-to-poly.c (tree_int_to_gmp): Likewise.
1461 (build_loop_iteration_domains): Likewise.
1462 * hooks.h: Include wide-int.h rather than double-int.h.
1463 (hook_bool_dint_dint_uint_bool_true): Delete.
1464 (hook_bool_wint_wint_uint_bool_true): Declare.
1465 * hooks.c (hook_bool_dint_dint_uint_bool_true): Removed.
1466 (hook_bool_wint_wint_uint_bool_true): New.
1467 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): Use wide-int
1469 (ubsan_expand_si_overflow_mul_check): Likewise.
1470 * ipa-devirt.c (get_polymorphic_call_info): Likewise.
1471 * ipa-prop.c (compute_complex_assign_jump_func): Likewise.
1472 (get_ancestor_addr_info): Likewise.
1473 (ipa_modify_call_arguments): Likewise.
1474 * loop-doloop.c (doloop_modify): Likewise.
1475 (doloop_optimize): Likewise.
1476 * loop-iv.c (iv_number_of_iterations): Likewise.
1477 * loop-unroll.c (decide_unroll_constant_iterations): Likewise.
1478 (unroll_loop_constant_iterations): Likewise.
1479 (decide_unroll_runtime_iterations): Likewise.
1480 (unroll_loop_runtime_iterations): Likewise.
1481 (decide_peel_simple): Likewise.
1482 (decide_unroll_stupid): Likewise.
1483 * lto-streamer-in.c (streamer_read_wi): Add.
1484 (input_cfg): Use wide-int interfaces.
1485 (lto_input_tree_1): Likewise.
1486 * lto-streamer-out.c (streamer_write_wi): Add.
1487 (hash_tree): Use wide-int interfaces.
1488 (output_cfg): Likewise.
1489 * Makefile.in (OBJS): Add wide-int.o and wide-int-print.o.
1490 (GTFILES): Add wide-int.h and signop.h.
1491 (TAGS): Look for .cc files too.
1492 * omp-low.c (scan_omp_1_op): Use wide-int interfaces.
1493 * optabs.c (expand_subword_shift): Likewise.
1494 (expand_doubleword_shift): Likewise.
1495 (expand_absneg_bit): Likewise.
1496 (expand_copysign_absneg): Likewise.
1497 (expand_copysign_bit): Likewise.
1498 * postreload.c (reload_cse_simplify_set): Likewise.
1499 * predict.c (predict_iv_comparison): Likewise.
1500 * pretty-print.h: Include wide-int-print.h.
1502 * print-rtl.c (print_rtx): Add CONST_WIDE_INT case.
1503 * print-tree.c: Include wide-int-print.h.
1504 (print_node_brief): Use wide-int interfaces.
1505 (print_node): Likewise.
1506 * read-rtl.c (validate_const_wide_int): New.
1507 (read_rtx_code): Add CONST_WIDE_INT case.
1508 * real.c: Include wide-int.h.
1509 (real_to_integer2): Delete.
1510 (real_to_integer): New function, returning a wide_int.
1511 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
1512 (ten_to_ptwo): Update call to real_from_integer.
1513 (real_digit): Likewise.
1514 * real.h: Include signop.h, wide-int.h and insn-modes.h.
1515 (real_to_integer2, REAL_VALUE_FROM_INT, REAL_VALUE_FROM_UNSIGNED_INT)
1516 (REAL_VALUE_TO_INT): Delete.
1517 (real_to_integer): Declare a wide-int form.
1518 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
1519 * recog.c (const_int_operand): Improve comment.
1520 (const_scalar_int_operand): New.
1521 (const_double_operand): Add a separate definition for CONST_WIDE_INT.
1522 * rtlanal.c (commutative_operand_precedence): Handle CONST_WIDE_INT.
1523 (split_double): Likewise.
1524 * rtl.c (DEF_RTL_EXPR): Handle CONST_WIDE_INT.
1525 (rtx_size): Likewise.
1526 (rtx_alloc_stat_v): New.
1527 (rtx_alloc_stat): Now calls rtx_alloc_stat_v.
1528 (cwi_output_hex): New.
1529 (iterative_hash_rtx): Handle CONST_WIDE_INT.
1530 (cwi_check_failed_bounds): New.
1531 * rtl.def (CONST_WIDE_INT): New.
1532 * rtl.h: Include <utility> and wide-int.h.
1533 (struct hwivec_def): New.
1534 (CWI_GET_NUM_ELEM): New.
1535 (CWI_PUT_NUM_ELEM): New.
1536 (struct rtx_def): Add num_elem and hwiv.
1537 (CASE_CONST_SCALAR_INT): Modify for TARGET_SUPPORTS_WIDE_INT.
1538 (CASE_CONST_UNIQUE): Likewise.
1539 (CASE_CONST_ANY): Likewise.
1540 (CONST_SCALAR_INT_P): Likewise.
1541 (CONST_WIDE_INT_P): New.
1543 (HWIVEC_CHECK): New.
1544 (cwi_check_failed_bounds): New.
1546 (HWIVEC_CHECK): New.
1547 (CONST_WIDE_INT_VEC) New.
1548 (CONST_WIDE_INT_NUNITS) New.
1549 (CONST_WIDE_INT_ELT) New.
1550 (rtx_mode_t): New type.
1551 (wi::int_traits <rtx_mode_t>): New.
1553 (wi::min_value): New.
1554 (wi::max_value): New.
1556 (const_wide_int_alloc): New.
1557 (immed_wide_int_const): New.
1558 * sched-vis.c (print_value): Handle CONST_WIDE_INT.
1559 * sel-sched-ir.c (lhs_and_rhs_separable_p): Update comment.
1560 * signop.h: New file.
1561 * simplify-rtx.c (mode_signbit_p): Handle CONST_WIDE_INT.
1562 (simplify_const_unary_operation): Use wide-int interfaces.
1563 (simplify_binary_operation_1): Likewise.
1564 (simplify_const_binary_operation): Likewise.
1565 (simplify_const_relational_operation): Likewise.
1566 (simplify_immed_subreg): Likewise.
1567 * stmt.c (expand_case): Likewise.
1568 * stor-layout.h (set_min_and_max_values_for_integral_type): Take a
1569 signop rather than a bool.
1570 * stor-layout.c (layout_type): Use wide-int interfaces.
1571 (initialize_sizetypes): Update calls to
1572 set_min_and_max_values_for_integral_type.
1573 (set_min_and_max_values_for_integral_type): Take a signop rather
1574 than a bool. Use wide-int interfaces.
1575 (fixup_signed_type): Update accordingly. Remove
1576 HOST_BITS_PER_DOUBLE_INT limit.
1577 (fixup_unsigned_type): Likewise.
1578 * system.h (STATIC_CONSTANT_P): New.
1579 (STATIC_ASSERT): New.
1580 * target.def (can_use_doloop_p): Take widest_ints rather than
1582 * target.h: Include wide-int.h rather than double-int.h.
1583 * targhooks.h (can_use_doloop_if_innermost): Take widest_ints rather
1585 * targhooks.c (default_cxx_get_cookie_size): Use tree_int_cst_lt
1586 rather than INT_CST_LT_UNSIGNED.
1587 (can_use_doloop_if_innermost): Take widest_ints rather than
1589 * tree-affine.c: Include wide-int-print.h.
1590 (double_int_ext_for_comb): Delete.
1591 (wide_int_ext_for_comb): New.
1592 (aff_combination_zero): Use wide-int interfaces.
1593 (aff_combination_const): Take a widest_int instead of a double_int.
1594 (aff_combination_elt): Use wide-int interfaces.
1595 (aff_combination_scale): Take a widest_int instead of a double_int.
1596 (aff_combination_add_elt): Likewise.
1597 (aff_combination_add_cst): Likewise.
1598 (aff_combination_add): Use wide-int interfaces.
1599 (aff_combination_convert): Likewise.
1600 (tree_to_aff_combination): Likewise.
1601 (add_elt_to_tree): Take a widest_int instead of a double_int.
1602 (aff_combination_to_tree): Use wide-int interfaces.
1603 (aff_combination_remove_elt): Likewise.
1604 (aff_combination_add_product): Take a widest_int instead of
1606 (aff_combination_mult): Use wide-int interfaces.
1607 (aff_combination_expand): Likewise.
1608 (double_int_constant_multiple_p): Delete.
1609 (wide_int_constant_multiple_p): New.
1610 (aff_combination_constant_multiple_p): Take a widest_int pointer
1611 instead of a double_int pointer.
1612 (print_aff): Use wide-int interfaces.
1613 (get_inner_reference_aff): Take a widest_int pointer
1614 instead of a double_int pointer.
1615 (aff_comb_cannot_overlap_p): Take widest_ints instead of double_ints.
1616 * tree-affine.h: Include wide-int.h.
1617 (struct aff_comb_elt): Change type of coef to widest_int.
1618 (struct affine_tree_combination): Change type of offset to widest_int.
1619 (double_int_ext_for_comb): Delete.
1620 (wide_int_ext_for_comb): New.
1621 (aff_combination_const): Use widest_int instead of double_int.
1622 (aff_combination_scale): Likewise.
1623 (aff_combination_add_elt): Likewise.
1624 (aff_combination_constant_multiple_p): Likewise.
1625 (get_inner_reference_aff): Likewise.
1626 (aff_comb_cannot_overlap_p): Likewise.
1627 (aff_combination_zero_p): Use wide-int interfaces.
1628 * tree.c: Include tree.h.
1629 (init_ttree): Use make_int_cst.
1630 (tree_code_size): Removed code for INTEGER_CST case.
1631 (tree_size): Add INTEGER_CST case.
1632 (make_node_stat): Update comment.
1633 (get_int_cst_ext_nunits, build_new_int_cst, build_int_cstu): New.
1634 (build_int_cst_type): Use wide-int interfaces.
1635 (double_int_to_tree): Likewise.
1636 (double_int_fits_to_tree_p): Delete.
1637 (force_fit_type_double): Delete.
1638 (force_fit_type): New.
1639 (int_cst_hash_hash): Use wide-int interfaces.
1640 (int_cst_hash_eq): Likewise.
1641 (build_int_cst_wide): Delete.
1642 (wide_int_to_tree): New.
1643 (cache_integer_cst): Use wide-int interfaces.
1644 (build_low_bits_mask): Likewise.
1645 (cst_and_fits_in_hwi): Likewise.
1646 (real_value_from_int_cst): Likewise.
1647 (make_int_cst_stat): New.
1648 (integer_zerop): Use wide_int interfaces.
1649 (integer_onep): Likewise.
1650 (integer_all_onesp): Likewise.
1651 (integer_pow2p): Likewise.
1652 (integer_nonzerop): Likewise.
1653 (tree_log2): Likewise.
1654 (tree_floor_log2): Likewise.
1655 (tree_ctz): Likewise.
1656 (int_size_in_bytes): Likewise.
1657 (mem_ref_offset): Return an offset_int rather than a double_int.
1658 (build_type_attribute_qual_variant): Use wide_int interfaces.
1659 (type_hash_eq): Likewise
1660 (tree_int_cst_equal): Likewise.
1661 (tree_int_cst_lt): Delete.
1662 (tree_int_cst_compare): Likewise.
1663 (tree_fits_shwi_p): Use wide_int interfaces.
1664 (tree_fits_uhwi_p): Likewise.
1665 (tree_int_cst_sign_bit): Likewise.
1666 (tree_int_cst_sgn): Likewise.
1667 (tree_int_cst_min_precision): Take a signop rather than a bool.
1668 (simple_cst_equal): Use wide_int interfaces.
1669 (compare_tree_int): Likewise.
1670 (iterative_hash_expr): Likewise.
1671 (int_fits_type_p): Likewise. Use tree_int_cst_lt rather than
1673 (get_type_static_bounds): Use wide_int interfaces.
1674 (tree_int_cst_elt_check_failed): New.
1675 (build_common_tree_nodes): Reordered to set prec before filling in
1677 (int_cst_value): Check cst_and_fits_in_hwi.
1678 (widest_int_cst_value): Use wide_int interfaces.
1679 (upper_bound_in_type): Likewise.
1680 (lower_bound_in_type): Likewise.
1681 (num_ending_zeros): Likewise.
1682 (drop_tree_overflow): Likewise.
1683 * tree-call-cdce.c (check_pow): Update call to real_from_integer.
1684 (gen_conditions_for_pow_cst_base): Likewise.
1685 * tree-cfg.c: Include wide-int.h and wide-int-print.h.
1686 (group_case_labels_stmt): Use wide-int interfaces.
1687 (verify_gimple_assign_binary): Likewise.
1688 (print_loop): Likewise.
1689 * tree-chrec.c (tree_fold_binomial): Likewise.
1690 * tree-core.h (struct tree_base): Add int_length.
1691 (struct tree_int_cst): Change rep of value.
1692 * tree-data-ref.c (dr_analyze_innermost): Use wide-int interfaces.
1693 (dr_may_alias_p): Likewise.
1694 (max_stmt_executions_tree): Likewise.
1695 * tree.def (INTEGER_CST): Update comment.
1696 * tree-dfa.c (get_ref_base_and_extent): Use wide-int interfaces.
1697 * tree-dfa.h (get_addr_base_and_unit_offset_1): Likewise.
1698 * tree-dump.c: Include wide-int.h and wide-int-print.h.
1699 (dequeue_and_dump): Use wide-int interfaces.
1700 * tree.h: Include wide-int.h.
1701 (NULL_TREE): Moved to earlier loc in file.
1702 (TREE_INT_CST_ELT_CHECK): New.
1703 (tree_int_cst_elt_check_failed): New.
1705 (TREE_INT_CST): Delete.
1706 (TREE_INT_CST_LOW): Use wide-int interfaces.
1707 (TREE_INT_CST_HIGH): Delete.
1708 (TREE_INT_CST_NUNITS): New.
1709 (TREE_INT_CST_EXT_NUNITS): Likewise.
1710 (TREE_INT_CST_OFFSET_NUNITS): Likewise.
1711 (TREE_INT_CST_ELT): Likewise.
1712 (INT_CST_LT): Delete.
1713 (tree_int_cst_elt_check): New (two forms).
1714 (type_code_size): Update comment.
1715 (make_int_cst_stat, make_int_cst): New.
1716 (tree_to_double_int): Delete.
1717 (double_int_fits_to_tree_p): Delete.
1718 (force_fit_type_double): Delete.
1719 (build_int_cstu): Replace with out-of-line function.
1720 (build_int_cst_wide): Delete.
1721 (tree_int_cst_lt): Define inline.
1722 (tree_int_cst_le): New.
1723 (tree_int_cst_compare): Define inline.
1724 (tree_int_cst_min_precision): Take a signop rather than a bool.
1725 (wi::int_traits <const_tree>): New.
1726 (wi::int_traits <tree>): New.
1727 (wi::extended_tree): New.
1728 (wi::int_traits <wi::extended_tree>): New.
1729 (wi::to_widest): New.
1730 (wi::to_offset): New.
1731 (wi::fits_to_tree_p): New.
1732 (wi::min_value): New.
1733 (wi::max_value): New.
1734 * tree-inline.c (remap_gimple_op_r): Use wide-int interfaces.
1735 (copy_tree_body_r): Likewise.
1736 * tree-object-size.c (compute_object_offset): Likewise.
1737 (addr_object_size): Likewise.
1738 * tree-predcom.c: Include wide-int-print.h.
1739 (struct dref_d): Change type of offset to widest_int.
1740 (dump_dref): Call wide-int printer.
1741 (aff_combination_dr_offset): Use wide-int interfaces.
1742 (determine_offset): Take a widest_int pointer rather than a
1744 (split_data_refs_to_components): Use wide-int interfaces.
1745 (suitable_component_p): Likewise.
1746 (order_drefs): Likewise.
1747 (add_ref_to_chain): Likewise.
1748 (valid_initializer_p): Likewise.
1749 (determine_roots_comp): Likewise.
1750 * tree-pretty-print.c: Include wide-int-print.h.
1751 (dump_generic_node): Use wide-int interfaces.
1752 * tree-sra.c (sra_ipa_modify_expr): Likewise.
1753 * tree-ssa-address.c (addr_for_mem_ref): Likewise.
1754 (move_fixed_address_to_symbol): Likewise.
1755 (move_hint_to_base): Likewise.
1756 (move_pointer_to_base): Likewise.
1757 (move_variant_to_index): Likewise.
1758 (most_expensive_mult_to_index): Likewise.
1759 (addr_to_parts): Likewise.
1760 (copy_ref_info): Likewise.
1761 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
1762 (indirect_refs_may_alias_p): Likewise.
1763 (stmt_kills_ref_p_1): Likewise.
1764 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
1765 * tree-ssa-ccp.c: Update comment at top of file. Include
1767 (struct prop_value_d): Change type of mask to widest_int.
1768 (extend_mask): New function.
1769 (dump_lattice_value): Use wide-int interfaces.
1770 (get_default_value): Likewise.
1771 (set_constant_value): Likewise.
1772 (set_value_varying): Likewise.
1773 (valid_lattice_transition): Likewise.
1774 (set_lattice_value): Likewise.
1775 (value_to_double_int): Delete.
1776 (value_to_wide_int): New.
1777 (get_value_from_alignment): Use wide-int interfaces.
1778 (get_value_for_expr): Likewise.
1779 (do_dbg_cnt): Likewise.
1780 (ccp_finalize): Likewise.
1781 (ccp_lattice_meet): Likewise.
1782 (bit_value_unop_1): Use widest_ints rather than double_ints.
1783 (bit_value_binop_1): Likewise.
1784 (bit_value_unop): Use wide-int interfaces.
1785 (bit_value_binop): Likewise.
1786 (bit_value_assume_aligned): Likewise.
1787 (evaluate_stmt): Likewise.
1788 (ccp_fold_stmt): Likewise.
1789 (visit_cond_stmt): Likewise.
1790 (ccp_visit_stmt): Likewise.
1791 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Likewise.
1792 (constant_pointer_difference): Likewise.
1793 (associate_pointerplus): Likewise.
1794 (combine_conversions): Likewise.
1795 * tree-ssa-loop.h: Include wide-int.h.
1796 (struct tree_niter_desc): Change type of max to widest_int.
1797 * tree-ssa-loop-im.c (mem_refs_may_alias_p): Use wide-int interfaces.
1798 * tree-ssa-loop-ivcanon.c (remove_exits_and_undefined_stmts): Likewise.
1799 (remove_redundant_iv_tests): Likewise.
1800 (canonicalize_loop_induction_variables): Likewise.
1801 * tree-ssa-loop-ivopts.c (alloc_iv): Likewise.
1802 (constant_multiple_of): Take a widest_int pointer instead of
1803 a double_int pointer.
1804 (get_computation_aff): Use wide-int interfaces.
1805 (ptr_difference_cost): Likewise.
1806 (difference_cost): Likewise.
1807 (get_loop_invariant_expr_id): Likewise.
1808 (get_computation_cost_at): Likewise.
1809 (iv_elimination_compare_lt): Likewise.
1810 (may_eliminate_iv): Likewise.
1811 * tree-ssa-loop-niter.h (estimated_loop_iterations): Use widest_int
1812 instead of double_int.
1813 (max_loop_iterations): Likewise.
1814 (max_stmt_executions): Likewise.
1815 (estimated_stmt_executions): Likewise.
1816 * tree-ssa-loop-niter.c: Include wide-int-print.h.
1817 (split_to_var_and_offset): Use wide-int interfaces.
1818 (determine_value_range): Likewise.
1819 (bound_difference_of_offsetted_base): Likewise.
1820 (bounds_add): Take a widest_int instead of a double_int.
1821 (number_of_iterations_ne_max): Use wide-int interfaces.
1822 (number_of_iterations_ne): Likewise.
1823 (number_of_iterations_lt_to_ne): Likewise.
1824 (assert_loop_rolls_lt): Likewise.
1825 (number_of_iterations_lt): Likewise.
1826 (number_of_iterations_le): Likewise.
1827 (number_of_iterations_cond): Likewise.
1828 (number_of_iterations_exit): Likewise.
1829 (finite_loop_p): Likewise.
1830 (derive_constant_upper_bound_assign): Likewise.
1831 (derive_constant_upper_bound): Return a widest_int.
1832 (derive_constant_upper_bound_ops): Likewise.
1833 (do_warn_aggressive_loop_optimizations): Use wide-int interfaces.
1834 (record_estimate): Take a widest_int rather than a double_int.
1835 (record_nonwrapping_iv): Use wide-int interfaces.
1836 (double_int_cmp): Delete.
1837 (wide_int_cmp): New.
1838 (bound_index): Take a widest_int rather than a double_int.
1839 (discover_iteration_bound_by_body_walk): Use wide-int interfaces.
1840 (maybe_lower_iteration_bound): Likewise.
1841 (estimate_numbers_of_iterations_loop): Likewise.
1842 (estimated_loop_iterations): Take a widest_int pointer than than
1843 a double_int pointer.
1844 (estimated_loop_iterations_int): Use wide-int interfaces.
1845 (max_loop_iterations): Take a widest_int pointer than than
1846 a double_int pointer.
1847 (max_loop_iterations_int): Use wide-int interfaces.
1848 (max_stmt_executions): Take a widest_int pointer than than
1849 a double_int pointer.
1850 (estimated_stmt_executions): Likewise.
1851 (n_of_executions_at_most): Use wide-int interfaces.
1852 (scev_probably_wraps_p): Likewise.
1853 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Update calls
1855 * tree-scalar-evolution.c (simplify_peeled_chrec): Use wide-int
1857 * tree-ssanames.c (set_range_info): Use wide_int_refs rather than
1858 double_ints. Adjust for trailing_wide_ints <3> representation.
1859 (set_nonzero_bits): Likewise.
1860 (get_range_info): Return wide_ints rather than double_ints.
1861 Adjust for trailing_wide_ints <3> representation.
1862 (get_nonzero_bits): Likewise.
1863 (duplicate_ssa_name_range_info): Adjust for trailing_wide_ints <3>
1865 * tree-ssanames.h (struct range_info_def): Replace min, max and
1866 nonzero_bits with a trailing_wide_ints <3>.
1867 (set_range_info): Use wide_int_refs rather than double_ints.
1868 (set_nonzero_bits): Likewise.
1869 (get_range_info): Return wide_ints rather than double_ints.
1870 (get_nonzero_bits): Likewise.
1871 * tree-ssa-phiopt.c (jump_function_from_stmt): Use wide-int interfaces.
1872 * tree-ssa-pre.c (phi_translate_1): Likewise.
1873 * tree-ssa-reassoc.c (decrement_power): Use calls to real_from_integer.
1874 (acceptable_pow_call): Likewise.
1875 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Use wide-int
1877 (vn_reference_fold_indirect): Likewise.
1878 (vn_reference_maybe_forwprop_address): Likewise.
1879 (valueize_refs_1): Likewise.
1880 * tree-ssa-structalias.c (get_constraint_for_ptr_offset): Likewise.
1881 * tree-ssa-uninit.c (is_value_included_in): Use wide-int interfaces,
1882 tree_int_cst_lt and tree_int_cst_le.
1883 * tree-streamer-in.c (unpack_ts_base_value_fields): Use wide-int
1885 (streamer_alloc_tree): Likewise.
1886 * tree-streamer-out.c (pack_ts_int_cst_value_fields): Likewise.
1887 (streamer_write_tree_header): Likewise.
1888 (streamer_write_integer_cst): Likewise.
1889 * tree-switch-conversion.c (emit_case_bit_tests): Likewise.
1890 (build_constructors): Likewise.
1891 (array_value_type): Likewise.
1892 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list): Likewise.
1893 (vect_check_gather): Likewise.
1894 * tree-vect-generic.c (build_replicated_const): Likewise.
1895 (expand_vector_divmod): Likewise.
1896 * tree-vect-loop.c (vect_transform_loop): Likewise.
1897 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
1898 (vect_do_peeling_for_alignment): Likewise.
1899 * tree-vect-patterns.c (vect_recog_divmod_pattern): Likewise.
1900 * tree-vrp.c: Include wide-int.h.
1901 (operand_less_p): Use wide-int interfaces and tree_int_cst_lt.
1902 (extract_range_from_assert): Use wide-int interfaces.
1903 (vrp_int_const_binop): Likewise.
1904 (zero_nonzero_bits_from_vr): Take wide_int pointers rather than
1905 double_int pointers.
1906 (ranges_from_anti_range): Use wide-int interfaces.
1907 (quad_int_cmp): Delete.
1908 (quad_int_pair_sort): Likewise.
1909 (extract_range_from_binary_expr_1): Use wide-int interfaces.
1910 (extract_range_from_unary_expr_1): Likewise.
1911 (adjust_range_with_scev): Likewise.
1912 (masked_increment): Take and return wide_ints rather than double_ints.
1913 (register_edge_assert_for_2): Use wide-int interfaces.
1914 (check_array_ref): Likewise.
1915 (search_for_addr_array): Likewise.
1916 (maybe_set_nonzero_bits): Likewise.
1917 (union_ranges): Pass an integer of the correct type instead of
1918 using integer_one_node.
1919 (intersect_ranges): Likewise.
1920 (simplify_truth_ops_using_ranges): Likewise.
1921 (simplify_bit_ops_using_ranges): Use wide-int interfaces.
1922 (range_fits_type_p): Likewise.
1923 (simplify_cond_using_ranges): Likewise. Take a signop rather than
1925 (simplify_conversion_using_ranges): Use wide-int interfaces.
1926 (simplify_float_conversion_using_ranges): Likewise.
1927 (vrp_finalize): Likewise.
1928 * value-prof.c (gimple_divmod_fixed_value_transform): Likewise.
1929 (gimple_stringops_transform): Likewise.
1930 * varasm.c (decode_addr_const): Likewise.
1931 (const_hash_1): Likewise.
1932 (const_rtx_hash_1): Likewise
1933 (output_constant): Likewise.
1934 (array_size_for_constructor): Likewise.
1935 (output_constructor_regular_field): Likewise.
1936 (output_constructor_bitfield): Likewise.
1937 * var-tracking.c (loc_cmp): Handle CONST_WIDE_INT.
1938 * mkconfig.sh: Include machmode.h to pick up BITS_PER_UNIT for
1940 * gencheck.c: Define BITS_PER_UNIT.
1943 * wide-int-print.cc: New.
1944 * wide-int-print.h: New.
1946 2014-05-06 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1948 * config/avr/avr.c (avr_can_eliminate): Mark unused argument.
1950 2014-05-06 Richard Biener <rguenther@suse.de>
1952 * tree-pass.h (TODO_verify_ssa, TODO_verify_flow,
1953 TODO_verify_stmts, TODO_verify_rtl_sharing): Remove.
1954 (TODO_verify_all): Adjust.
1955 * asan.c: Remove references to TODO_verify_ssa, TODO_verify_flow,
1956 TODO_verify_stmts and TODO_verify_rtl_sharing.
1957 * bb-reorder.c: Likewise.
1958 * cfgexpand.c: Likewise.
1959 * cprop.c: Likewise.
1961 * function.c: Likewise.
1962 * fwprop.c: Likewise.
1964 * gimple-ssa-isolate-paths.c: Likewise.
1965 * gimple-ssa-strength-reduction.c: Likewise.
1966 * ipa-split.c: Likewise.
1967 * loop-init.c: Likewise.
1968 * loop-unroll.c: Likewise.
1969 * lower-subreg.c: Likewise.
1970 * modulo-sched.c: Likewise.
1971 * postreload-gcse.c: Likewise.
1972 * predict.c: Likewise.
1973 * recog.c: Likewise.
1974 * sched-rgn.c: Likewise.
1975 * store-motion.c: Likewise.
1976 * tracer.c: Likewise.
1977 * trans-mem.c: Likewise.
1978 * tree-call-cdce.c: Likewise.
1979 * tree-cfg.c: Likewise.
1980 * tree-cfgcleanup.c: Likewise.
1981 * tree-complex.c: Likewise.
1982 * tree-eh.c: Likewise.
1983 * tree-emutls.c: Likewise.
1984 * tree-if-conv.c: Likewise.
1985 * tree-into-ssa.c: Likewise.
1986 * tree-loop-distribution.c: Likewise.
1987 * tree-object-size.c: Likewise.
1988 * tree-parloops.c: Likewise.
1989 * tree-pass.h: Likewise.
1990 * tree-sra.c: Likewise.
1991 * tree-ssa-ccp.c: Likewise.
1992 * tree-ssa-copy.c: Likewise.
1993 * tree-ssa-copyrename.c: Likewise.
1994 * tree-ssa-dce.c: Likewise.
1995 * tree-ssa-dom.c: Likewise.
1996 * tree-ssa-dse.c: Likewise.
1997 * tree-ssa-forwprop.c: Likewise.
1998 * tree-ssa-ifcombine.c: Likewise.
1999 * tree-ssa-loop-ch.c: Likewise.
2000 * tree-ssa-loop-ivcanon.c: Likewise.
2001 * tree-ssa-loop.c: Likewise.
2002 * tree-ssa-math-opts.c: Likewise.
2003 * tree-ssa-phiopt.c: Likewise.
2004 * tree-ssa-phiprop.c: Likewise.
2005 * tree-ssa-pre.c: Likewise.
2006 * tree-ssa-reassoc.c: Likewise.
2007 * tree-ssa-sink.c: Likewise.
2008 * tree-ssa-strlen.c: Likewise.
2009 * tree-ssa-tail-merge.c: Likewise.
2010 * tree-ssa-uncprop.c: Likewise.
2011 * tree-switch-conversion.c: Likewise.
2012 * tree-tailcall.c: Likewise.
2013 * tree-vect-generic.c: Likewise.
2014 * tree-vectorizer.c: Likewise.
2015 * tree-vrp.c: Likewise.
2017 * var-tracking.c: Likewise.
2018 * bt-load.c: Likewise.
2019 * cfgcleanup.c: Likewise.
2020 * combine-stack-adj.c: Likewise.
2021 * combine.c: Likewise.
2022 * compare-elim.c: Likewise.
2023 * config/epiphany/resolve-sw-modes.c: Likewise.
2024 * config/i386/i386.c: Likewise.
2025 * config/mips/mips.c: Likewise.
2026 * config/s390/s390.c: Likewise.
2027 * config/sh/sh_treg_combine.cc: Likewise.
2028 * config/sparc/sparc.c: Likewise.
2031 * final.c: Likewise.
2032 * ifcvt.c: Likewise.
2033 * mode-switching.c: Likewise.
2034 * passes.c: Likewise.
2035 * postreload.c: Likewise.
2037 * reg-stack.c: Likewise.
2038 * regcprop.c: Likewise.
2039 * regrename.c: Likewise.
2042 2014-05-06 Richard Biener <rguenther@suse.de>
2045 * bitmap.c (debug_bitmap): Dump to stderr, not stdout.
2046 * tree-ssa-structalias.c (dump_solution_for_var): Likewise.
2048 2014-05-05 Jan Hubicka <hubicka@ucw.cz>
2051 * ipa-devirt.c (get_class_context): Allow POD to change to non-POD.
2053 2014-05-05 Radovan Obradovic <robradovic@mips.com>
2054 Tom de Vries <tom@codesourcery.com>
2056 * target.def (call_fusage_contains_non_callee_clobbers): New
2058 * doc/tm.texi.in (@node Stack and Calling): Add Miscellaneous Register
2060 (@node Miscellaneous Register Hooks): New node.
2061 (@hook TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS): New hook.
2062 * doc/tm.texi: Regenerate.
2064 2014-05-05 Marek Polacek <polacek@redhat.com>
2067 * opts.c (common_handle_option): Call error_at instead of warning_at.
2069 2014-05-05 Richard Biener <rguenther@suse.de>
2071 * passes.c (execute_function_todo): Don't reset TODO_verify_ssa
2072 from last_verified if update_ssa ran. Move TODO_verify_rtl_sharing
2073 under the TODO_verify_il umbrella.
2075 2014-05-05 Richard Biener <rguenther@suse.de>
2077 * passes.c (execute_function_todo): Move TODO_verify_flow under
2078 the TODO_verify_ul umbrella.
2080 2014-05-05 Richard Biener <rguenther@suse.de>
2083 * fold-const.c (fold_binary_loc): Consistently avoid canonicalizing
2084 X & CST away from a CST that is the mask of a mode.
2086 2014-05-05 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2088 * config/picochip/picochip-protos.h (picochip_regno_nregs): Change
2089 int argument to enum machine_mode.
2090 (picochip_class_max_nregs): Ditto.
2091 * config/picochip/picochip.c (picochip_regno_nregs): Ditto.
2092 (picochip_class_max_nregs): Ditto.
2094 2014-05-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2096 * target.def: Add new target hook.
2097 * doc/tm.texi: Regenerate.
2098 * targhooks.h (default_keep_leaf_when_profiled): Add prototype.
2099 * targhooks.c (default_keep_leaf_when_profiled): New function.
2101 * config/s390/s390.c (s390_keep_leaf_when_profiled): New function.
2102 (TARGET_KEEP_LEAF_WHEN_PROFILED): Define.
2104 2014-05-05 Bin Cheng <bin.cheng@arm.com>
2106 PR tree-optimization/60363
2107 * gcc/tree-ssa-threadupdate.c (get_value_locus_in_path): New.
2108 (copy_phi_args): New parameters. Call get_value_locus_in_path.
2109 (update_destination_phis): New parameter.
2110 (create_edge_and_update_destination_phis): Ditto.
2111 (ssa_fix_duplicate_block_edges): Pass new arguments.
2112 (thread_single_edge): Ditto.
2114 2014-05-04 Peter Bergner <bergner@vnet.ibm.com>
2116 * config/rs6000/rs6000.h (RS6000_BTM_HARD_FLOAT): New define.
2117 (RS6000_BTM_COMMON): Add RS6000_BTM_HARD_FLOAT.
2118 (TARGET_EXTRA_BUILTINS): Add TARGET_HARD_FLOAT.
2119 * config/rs6000/rs6000-builtin.def (BU_MISC_1):
2120 Use RS6000_BTM_HARD_FLOAT.
2121 (BU_MISC_2): Likewise.
2122 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Handle
2123 RS6000_BTM_HARD_FLOAT.
2124 (rs6000_option_override_internal): Enforce -mhard-float if -mhard-dfp
2126 (rs6000_invalid_builtin): Add hard floating builtin support.
2127 (rs6000_expand_builtin): Relax the gcc_assert to allow the new
2128 hard float builtins.
2129 (rs6000_builtin_mask_names): Add RS6000_BTM_HARD_FLOAT.
2131 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
2133 * config/sh/sh_optimize_sett_clrt.cc (sh_optimize_sett_clrt::execute):
2134 Add missing function* argument.
2136 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
2138 * lra-constraints.c (valid_address_p): Move earlier in file.
2139 Add a constraint argument to the address_info version.
2140 (satisfies_memory_constraint_p): New function.
2141 (satisfies_address_constraint_p): Likewise.
2142 (process_alt_operands, curr_insn_transform): Use them.
2143 (process_address): Pass the constraint to valid_address_p when
2144 checking address operands.
2146 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
2148 * config/mips/mips.c (mips_isa_rev): New variable.
2149 (mips_set_architecture): Set it.
2150 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Set __mips_isa_rev
2152 (ISA_HAS_MUL3, ISA_HAS_FP_CONDMOVE, ISA_HAS_8CC, ISA_HAS_FP4)
2153 (ISA_HAS_PAIRED_SINGLE, ISA_HAS_MADD_MSUB, ISA_HAS_FP_RECIP_RSQRT)
2154 (ISA_HAS_CLZ_CLO, ISA_HAS_ROR, ISA_HAS_WSBH, ISA_HAS_PREFETCH)
2155 (ISA_HAS_SEB_SEH, ISA_HAS_EXT_INS, ISA_HAS_MXHC1)
2156 (ISA_HAS_HILO_INTERLOCKS, ISA_HAS_SYNCI, MIN_FPRS_PER_FMT): Reexpress
2157 conditions in terms of mips_isa_rev.
2158 (mips_isa_rev): Declare.
2160 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
2162 * config/sh/sh-mem.cc: Use tabs instead of spaces.
2163 (prob_unlikely, prob_likely): Make variables const.
2165 2014-05-03 Denis Chertykov <chertykov@gmail.com>
2167 * config/avr/avr.c (avr_adjust_insn_length): Handle JUMP_TABLE_DATA.
2169 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
2171 * config/sh/sh.h (SH_ASM_SPEC): Handle m1, m2*, m3* and m4* cases.
2173 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
2175 * config/sh/sh.h (ROUND_ADVANCE): Delete macro.
2176 (ROUND_REG, PASS_IN_REG_P): Move and rename macros to ...
2177 * config/sh/sh.c (sh_round_reg, sh_pass_in_reg_p): ... these new
2179 (sh_arg_partial_bytes, sh_function_arg, sh_function_arg_advance,
2180 sh_setup_incoming_varargs): Replace usage of PASS_IN_REG_P with
2182 Replace usage of ROUND_REG with sh_round_reg.
2183 Use CEIL instead of ROUND_ADVANCE.
2185 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
2188 * config/sh/sh.c: Include stdlib headers before everything else.
2190 2014-05-02 Jakub Jelinek <jakub@redhat.com>
2192 * gimplify.c (gimplify_adjust_omp_clauses_1): Handle
2193 GOVD_FIRSTPRIVATE | GOVD_LASTPRIVATE.
2194 (gimplify_adjust_omp_clauses): Simd region is never
2195 directly nested in combined parallel. Instead, for linear
2196 with copyin/copyout, if in combined for simd loop, make decl
2197 firstprivate/lastprivate on OMP_FOR.
2198 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
2199 expand_omp_for_static_chunk): When setting endvar, also set
2200 fd->loop.v to the same value.
2202 2014-05-02 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
2204 * hwint.h (zext_hwi): Fix signed overflow for prec == 63.
2206 2014-05-02 Alan Lawrence <alan.lawrence@arm.com>
2208 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_1): Tidy bit-flip
2211 2014-05-02 Marek Polacek <polacek@redhat.com>
2213 * doc/invoke.texi: Describe -fsanitize=float-divide-by-zero.
2215 2014-05-02 Kito Cheng <kito@0xlab.org>
2217 * defaults.h (HONOR_REG_ALLOC_ORDER): Change HONOR_REG_ALLOC_ORDER
2218 to a C expression marco.
2219 * ira-color.c (HONOR_REG_ALLOC_ORDER) : Ditto.
2220 * config/arm/arm.h (HONOR_REG_ALLOC_ORDER): Ditto.
2221 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Ditto.
2222 * doc/tm.texi (HONOR_REG_ALLOC_ORDER): Update document for
2223 HONOR_REG_ALLOC_ORDER.
2224 * doc/tm.texi.in (HONOR_REG_ALLOC_ORDER): Ditto.
2226 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2228 * config/arc/arc.c (TARGET_LRA_P): Undef before redefine.
2230 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2232 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
2234 2014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
2236 * tree-if-conv.c (is_cond_scalar_reduction): New function.
2237 (convert_scalar_cond_reduction): Likewise.
2238 (predicate_scalar_phi): Add recognition and transformation
2239 of simple conditioanl reduction to be vectorizable.
2241 2014-05-01 Marek Polacek <polacek@redhat.com>
2244 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
2246 2014-04-30 Alan Lawrence <alan.lawrence@arm.com>
2248 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
2249 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
2250 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
2251 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
2252 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
2253 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
2254 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
2255 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
2257 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
2259 * config/arc/arc.opt (mlra): Move comment above option name
2260 to avoid mis-parsing as language options.
2262 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2264 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
2265 * config/sol2.h: ... here.
2266 * config/sol2-10.h: Remove.
2268 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
2269 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
2270 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
2271 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
2272 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
2273 * config/sol2.h: ... here.
2274 (SECTION_NAME_FORMAT): Don't redefine.
2275 (STARTFILE_ARCH32_SPEC): Rename to ...
2276 (STARTFILE_ARCH_SPEC): ... this.
2277 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
2278 * config/sparc/sol2.h: ... here.
2279 (SECTION_NAME_FORMAT): Don't undef.
2280 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
2281 (SUBTARGET_EXTRA_SPECS): Remove.
2282 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
2284 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
2285 (MD_STARTFILE_PREFIX): Remove.
2286 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
2287 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
2288 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
2289 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
2290 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
2291 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
2292 * config/i386/sol2.h: ... here.
2293 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
2294 * config/i386/sol2-bi.h: Remove.
2295 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
2296 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
2298 * config/i386/t-sol2-64: Rename to ...
2299 * config/i386/t-sol2: ... this.
2300 * config/sparc/t-sol2-64: Rename to ...
2301 * config/sparc/t-sol2: ... this.
2303 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
2304 sol2_tm_file_head, sol2_tm_file_tail.
2305 Include ${cpu_type}/sol2.h before sol2.h.
2307 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
2308 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
2309 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
2310 Reflect i386/t-sol2-64 renaming.
2311 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
2312 Reflect sparc/t-sol2-64 renaming.
2314 2014-04-30 Richard Biener <rguenther@suse.de>
2316 * passes.c (execute_function_todo): Move TODO_verify_stmts
2317 and TODO_verify_ssa under the TODO_verify_il umbrella.
2318 * tree-ssa.h (verify_ssa): Adjust prototype.
2319 * tree-ssa.c (verify_ssa): Add parameter to tell whether
2320 we should verify SSA operands.
2321 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
2322 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
2323 whether we should verify whether not throwing stmts have EH info.
2324 * graphite-scop-detection.c (create_sese_edges): Adjust.
2325 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
2326 * tree-eh.c (lower_try_finally_switch): Do not add the
2327 default case label twice.
2329 2014-04-30 Marek Polacek <polacek@redhat.com>
2331 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
2332 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
2333 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
2334 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
2336 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
2338 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
2339 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
2340 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
2341 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
2342 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
2343 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
2344 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
2345 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
2347 2014-04-29 David Malcolm <dmalcolm@redhat.com>
2349 * tree-cfg.c (dump_function_to_file): Dump the return type of
2350 functions, in a line to itself before the function body, mimicking
2351 the layout of a C function.
2353 2014-04-29 Jakub Jelinek <jakub@redhat.com>
2355 PR tree-optimization/60971
2356 * tree-tailcall.c (process_assignment): Reject conversions which
2359 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
2361 * calls.c (initialize_argument_information): Always treat
2362 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
2363 (expand_call): Likewise.
2364 (emit_library_call_calue_1): Likewise.
2365 * expr.c (PUSH_ARGS_REVERSED): Do not define.
2366 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
2369 2014-04-29 Nick Clifton <nickc@redhat.com>
2371 * config/msp430/msp430.md (umulsidi): Fix typo.
2372 (mulhisi3): Enable even inside interrupt handlers.
2373 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
2374 bigger return address pushed in large mode.
2376 2014-04-29 Nick Clifton <nickc@redhat.com>
2378 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
2379 (arc_init_reg_tables): Use a machine_mode enum to iterate over
2381 * config/m32r/m32r.c (init_reg_tables): Likewise.
2382 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
2383 enum to hold the modes.
2385 2014-04-29 Richard Biener <rguenther@suse.de>
2387 * dominance.c (free_dominance_info): Add overload with
2389 (dom_info_state): Likewise.
2390 (dom_info_available_p): Likewise.
2391 * basic-block.h (free_dominance_info, dom_info_state,
2392 dom_info_available_p): Declare overloads.
2393 * passes.c (execute_function_todo): Verify that verifiers
2394 don't change dominator info state. Drop dominator info
2395 for IPA pass invocations.
2396 * cgraph.c (release_function_body): Restore asserts that
2397 dominator information is released.
2399 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
2401 * doc/invoke.texi: Fix typo.
2402 * tree-vrp.c: Fix typos.
2403 * gimple.c (infer_nonnull_range): Reorder operands of an && condition.
2405 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2407 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
2409 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
2411 * config/aarch64/aarch64-builtins.c
2412 (aarch64_types_storestruct_lane_qualifiers): New.
2413 (TYPES_STORESTRUCT_LANE): Likewise.
2414 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
2415 (st3_lane): Likewise.
2416 (st4_lane): Likewise.
2417 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
2418 (vec_store_lanesci_lane<mode>): Likewise.
2419 (vec_store_lanesxi_lane<mode>): Likewise.
2420 (aarch64_st2_lane<VQ:mode>): Likewise.
2421 (aarch64_st3_lane<VQ:mode>): Likewise.
2422 (aarch64_st4_lane<VQ:mode>): Likewise.
2423 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
2424 * config/aarch64/arm_neon.h
2425 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
2426 use new macro arguments.
2427 (__ST3_LANE_FUNC): Likewise.
2428 (__ST4_LANE_FUNC): Likewise.
2429 * config/aarch64/iterators.md (V_TWO_ELEM): New.
2430 (V_THREE_ELEM): Likewise.
2431 (V_FOUR_ELEM): Likewise.
2433 2014-04-28 David Malcolm <dmalcolm@redhat.com>
2435 * doc/gimple.texi: Replace the description of the now-defunct
2436 union gimple_statement_d with a diagram showing the
2437 gimple_statement_base class hierarchy and its relationships to
2438 the GSS_ and GIMPLE_ enums.
2440 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
2442 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
2443 * config/aarch64/aarch64.c
2444 (aarch64_cannot_change_mode_class): Weaken conditions.
2445 (aarch64_modes_tieable_p): New.
2446 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
2448 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
2450 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2451 (loadsync_<mode>): Change mode.
2452 (load_quadpti, store_quadpti): New.
2453 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2454 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2456 2014-04-28 Martin Jambor <mjambor@suse.cz>
2458 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
2459 same alias type as the original statement.
2460 (subreplacement_assignment_data): New type.
2461 (handle_unscalarized_data_in_subtree): New type of parameter,
2462 generate new memory accesses with same alias type as the original
2464 (load_assign_lhs_subreplacements): Likewise.
2465 (sra_modify_constructor_assign): Generate new memory accesses with
2466 same alias type as the original statement.
2468 2014-04-28 Richard Biener <rguenther@suse.de>
2470 * tree-pass.h (TODO_verify_il): Define.
2471 (TODO_verify_all): Complete properly.
2472 * passes.c (execute_function_todo): Move existing loop-closed
2473 SSA verification under TODO_verify_il.
2474 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
2475 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
2476 Fix tree sharing issue.
2478 2014-04-28 Richard Biener <rguenther@suse.de>
2481 * builtins.def (DEF_C11_BUILTIN): Add.
2482 (BUILT_IN_ALIGNED_ALLOC): Likewise.
2483 * coretypes.h (enum function_class): Add function_c11_misc.
2484 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
2485 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
2486 (call_may_clobber_ref_p_1): Likewise.
2487 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
2488 (mark_all_reaching_defs_necessary_1): Likewise.
2489 (propagate_necessity): Likewise.
2490 (eliminate_unnecessary_stmts): Likewise.
2491 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
2493 2014-04-28 Richard Biener <rguenther@suse.de>
2495 * tree-vrp.c (vrp_var_may_overflow): Remove.
2496 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
2497 with overflow immediately bump to one before that value and
2498 let iteration figure out overflow status.
2500 2014-04-28 Richard Biener <rguenther@suse.de>
2502 * configure.ac: Do valgrind header checks unconditionally.
2503 Add --enable-valgrind-annotations.
2504 * system.h: Guard valgrind header inclusion with
2505 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
2506 * alloc-pool.c (pool_alloc, pool_free): Use
2507 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
2508 to guard possibly dead code.
2509 * config.in: Regenerated.
2510 * configure: Likewise.
2512 2014-04-28 Jeff Law <law@redhat.com>
2514 PR tree-optimization/60902
2515 * tree-ssa-threadedge.c
2516 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
2517 over real defs when invalidating outputs from statements that do not
2518 produce useful outputs for threading.
2520 2014-04-28 Richard Biener <rguenther@suse.de>
2522 PR tree-optimization/60979
2523 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
2524 SCOPs that end in a block with a successor with abnormal
2527 2014-04-28 Richard Biener <rguenther@suse.de>
2529 * tree-pass.h (execute_pass_list): Adjust prototype.
2530 * passes.c (pass_manager::execute_early_local_passes): Adjust.
2531 (do_per_function): Change callback signature, push all actual
2532 work to the callbals.
2533 (do_per_function_toporder): Likewise.
2534 (execute_function_dump): Adjust.
2535 (execute_function_todo): Likewise.
2536 (clear_last_verified): Likewise.
2537 (verify_curr_properties): Likewise.
2538 (update_properties_after_pass): Likewise.
2539 (execute_pass_list_1): Split out from ...
2540 (execute_pass_list): ... here. Adjust.
2541 (execute_ipa_pass_list): Likewise.
2542 * cgraphunit.c (cgraph_add_new_function): Adjust.
2543 (analyze_function): Likewise.
2544 (expand_function): Likewise.
2545 * cgraph.c (release_function_body): Free dominance info
2546 here instead of asserting it was magically freed elsewhere.
2548 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
2550 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
2551 * configure: Regenerate.
2552 * config/sparc/sparc.opt (muser-mode): New option.
2553 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
2555 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
2556 * doc/invoke.texi (SPARC options): Document -muser-mode.
2558 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
2560 * cselib.c (find_slot_memmode): Delete.
2561 (cselib_hasher): Change compare_type to a struct.
2562 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
2564 (preserve_constants_and_equivs): Adjust for new compare_type.
2565 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
2566 (wrap_constant): Delete.
2567 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
2569 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
2571 * doc/install.texi (Building with profile feedback): Remove
2574 2014-04-26 Tom de Vries <tom@codesourcery.com>
2576 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
2579 2014-04-25 Cary Coutant <ccoutant@google.com>
2582 * dwarf2out.c (should_move_die_to_comdat): A type definition
2583 can contain a subprogram definition, but don't move it to a
2585 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
2586 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
2588 (clone_tree_hash): Rename to...
2589 (clone_tree_partial): ...this; change callers. Copy
2590 DW_TAG_subprogram DIEs as declarations.
2591 (copy_decls_walk): Don't copy children of a declaration into a
2594 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
2597 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
2600 2014-04-25 Jiong Wang <jiong.wang@arm.com>
2602 * config/arm/predicates.md (call_insn_operand): Add long_call check.
2603 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
2605 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
2608 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2610 * config/arm/arm.c (arm_cortex_a8_tune): Initialise T16-related fields.
2612 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2614 PR tree-optimization/60930
2615 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
2616 creating a multiply candidate by folding two constant
2617 multiplicands when the result overflows.
2619 2014-04-25 Jakub Jelinek <jakub@redhat.com>
2621 PR tree-optimization/60960
2622 * tree-vect-generic.c (expand_vector_operation): Only call
2623 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
2625 2014-04-25 Tom de Vries <tom@codesourcery.com>
2627 * expr.c (clobber_reg_mode): New function.
2628 * expr.h (clobber_reg): New function.
2630 2014-04-25 Tom de Vries <tom@codesourcery.com>
2632 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
2635 2014-04-25 Radovan Obradovic <robradovic@mips.com>
2636 Tom de Vries <tom@codesourcery.com>
2638 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
2640 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
2641 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
2642 new argument to find_all_hard_reg_sets call.
2644 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2646 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
2647 Use HOST_WIDE_INT_C for mask literal.
2648 (aarch_rev16_shleft_mask_imm_p): Likewise.
2650 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
2653 * config/sparc/sparc.md (ashlsi3_extend): Delete.
2655 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
2657 PR preprocessor/56540
2658 * config/i386/i386-c.c (ix86_target_macros): Define
2659 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
2661 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2663 * configure.ac (tga_func): Remove.
2664 (LIB_TLS_SPEC): Remove.
2665 * configure: Regenerate.
2666 * config.in: Regenerate.
2667 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
2669 2014-04-25 Richard Biener <rguenther@suse.de>
2672 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
2673 call stmt use/clobber sets during stmt walk instead of
2674 walking the possibly incomplete set of caller edges.
2676 2014-04-25 Richard Biener <rguenther@suse.de>
2679 * passes.c (apply_ipa_transforms): Inline into only caller ...
2680 (execute_one_pass): ... here. Properly bring in function
2681 bodies for nodes we want to apply IPA transforms to.
2683 2014-04-24 Cong Hou <congh@google.com>
2685 PR tree-optimization/60896
2686 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
2687 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
2688 (vect_mark_pattern_stmts): Set the def type of all statements in
2689 PATTERN_DEF_SEQ as vect_internal_def.
2691 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
2693 * doc/extend.texi (PowerPC Built-in Functions): Document new
2694 powerpc extended divide, bcd, pack/unpack 128-bit, builtin functions.
2695 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
2697 * config/rs6000/predicates.md (const_0_to_3_operand): New
2698 predicate to match 0..3 integer constants.
2700 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
2701 to support adding miscellaneous builtin functions.
2702 (BU_DFP_MISC_2): Likewise.
2703 (BU_P7_MISC_1): Likewise.
2704 (BU_P7_MISC_2): Likewise.
2705 (BU_P8V_MISC_3): Likewise.
2706 (BU_MISC_1): Likewise.
2707 (BU_MISC_2): Likewise.
2708 (DIVWE): Add extended divide builtin functions.
2711 (DIVWEUO): Likewise.
2715 (DIVDEUO): Likewise.
2716 (DXEX): Add decimal floating-point builtin functions.
2719 (DDEDPDQ): Likewise.
2721 (DENBCDQ): Likewise.
2728 (CDTBCD): Add new BCD builtin functions.
2732 (BCDADD_LT): Likewise.
2733 (BCDADD_EQ): Likewise.
2734 (BCDADD_GT): Likewise.
2735 (BCDADD_OV): Likewise.
2737 (BCDSUB_LT): Likewise.
2738 (BCDSUB_EQ): Likewise.
2739 (BCDSUB_GT): Likewise.
2740 (BCDSUB_OV): Likewise.
2741 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
2742 (UNPACK_TD): Likewise.
2743 (PACK_TF): Likewise.
2744 (UNPACK_TF): Likewise.
2745 (UNPACK_TF_0): Likewise.
2746 (UNPACK_TF_1): Likewise.
2747 (PACK_V1TI): Likewise.
2748 (UNPACK_V1TI): Likewise.
2750 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
2751 support for decimal floating point builtin functions.
2752 (rs6000_expand_ternop_builtin): Add checks for the new builtin
2753 functions that take constant arguments.
2754 (rs6000_invalid_builtin): Add decimal floating point builtin support.
2755 (rs6000_init_builtins): Setup long double, _Decimal64, and
2756 _Decimal128 types for new builtin functions.
2757 (builtin_function_type): Set the unsigned flags appropriately for
2758 the new builtin functions.
2759 (rs6000_opt_masks): Add support for decimal floating point builtin
2762 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
2763 floating point builtin functions.
2764 (RS6000_BTM_COMMON): Likewise.
2765 (RS6000_BTI_long_double): Likewise.
2766 (RS6000_BTI_dfloat64): Likewise.
2767 (RS6000_BTI_dfloat128): Likewise.
2768 (long_double_type_internal_node): Likewise.
2769 (dfloat64_type_internal_node): Likewise.
2770 (dfloat128_type_internal_node): Likewise.
2772 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
2773 2.07 bcd arithmetic instructions.
2774 (UNSPEC_BCDSUB): Likewise.
2775 (UNSPEC_BCD_OVERFLOW): Likewise.
2776 (UNSPEC_BCD_ADD_SUB): Likewise.
2777 (bcd_add_sub): Likewise.
2778 (BCD_TEST): Likewise.
2779 (bcd<bcd_add_sub>): Likewise.
2780 (bcd<bcd_add_sub>_test): Likewise.
2781 (bcd<bcd_add_sub>_test2): Likewise.
2782 (bcd<bcd_add_sub>_<code>): Likewise.
2783 (peephole2 for combined bcd ops): Likewise.
2785 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
2786 decimal floating point builtin functions.
2787 (UNSPEC_DENBCD): Likewise.
2788 (UNSPEC_DXEX): Likewise.
2789 (UNSPEC_DIEX): Likewise.
2790 (UNSPEC_DSCLI): Likewise.
2791 (UNSPEC_DSCRI): Likewise.
2792 (D64_D128): Likewise.
2793 (dfp_suffix): Likewise.
2794 (dfp_ddedpd_<mode>): Likewise.
2795 (dfp_denbcd_<mode>): Likewise.
2796 (dfp_dxex_<mode>): Likewise.
2797 (dfp_diex_<mode>): Likewise.
2798 (dfp_dscli_<mode>): Likewise.
2799 (dfp_dscri_<mode>): Likewise.
2801 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
2803 (UNSPEC_CDTBCD): Likewise.
2804 (UNSPEC_CBCDTD): Likewise.
2805 (UNSPEC_DIVE): Add support for new extended divide builtin functions.
2806 (UNSPEC_DIVEO): Likewise.
2807 (UNSPEC_DIVEU): Likewise.
2808 (UNSPEC_DIVEUO): Likewise.
2809 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
2810 pack/unpack 128-bit types.
2811 (UNSPEC_PACK_128BIT): Likewise.
2812 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
2813 (udiv<mode>3): Use idiv_ldiv mode attribute.
2814 (div<mode>3): Likewise.
2815 (addg6s): Add new BCD builtin functions.
2818 (UNSPEC_DIV_EXTEND): Add support for new extended divide instructions.
2819 (div_extend): Likewise.
2820 (div<div_extend>_<mode>"): Likewise.
2821 (FP128_64): Add support for new builtin functions to pack/unpack
2823 (unpack<mode>): Likewise.
2824 (unpacktf_0): Likewise.
2825 (unpacktf_1): Likewise.
2826 (unpack<mode>_dm): Likewise.
2827 (unpack<mode>_nodm): Likewise.
2828 (pack<mode>): Likewise.
2829 (unpackv1ti): Likewise.
2830 (packv1ti): Likewise.
2832 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
2834 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
2837 2014-04-24 Jakub Jelinek <jakub@redhat.com>
2839 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
2840 * gimplify.c (omp_is_private): Change last argument's type to int.
2841 Only diagnose lastprivate if the simd argument is 1, only diagnose
2842 linear if the simd argument is 2.
2843 (gimplify_omp_for): Adjust omp_is_private callers. When adding
2844 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
2845 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
2846 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
2847 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
2848 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
2849 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
2850 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
2851 * tree-nested.c (convert_nonlocal_omp_clauses,
2852 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
2854 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
2857 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
2860 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
2862 * flag-types.h (enum ivar_visibility): Add.
2864 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
2866 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
2867 function * argument.
2869 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
2871 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
2873 2014-04-24 Radovan Obradovic <robradovic@mips.com>
2874 Tom de Vries <tom@codesourcery.com>
2876 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
2877 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
2879 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
2880 * emit-rtl.c (try_split): Same.
2882 2014-04-24 Radovan Obradovic <robradovic@mips.com>
2883 Tom de Vries <tom@codesourcery.com>
2885 * common.opt (fuse-caller-save): New option.
2887 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
2889 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of
2890 elements for big-endian.
2892 2014-04-24 Richard Biener <rguenther@suse.de>
2894 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
2895 during TER and instead use the sepops interface for expanding
2896 non-GIMPLE_SINGLE_RHS.
2898 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2900 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
2901 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
2903 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2905 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
2906 assembler 64-bit option.
2907 * configure: Regenerate.
2909 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2911 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
2912 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
2913 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
2914 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
2915 (TARGET_CRYPTO): Take TARGET_SIMD into account.
2917 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2919 * config/aarch64/aarch64-builtins.c
2920 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
2921 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
2922 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
2923 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
2925 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
2926 (Vrevsuff): New mode attribute.
2928 2014-04-24 Terry Guo <terry.guo@arm.com>
2930 * config/arm/arm.h (machine_function): Define variable
2931 after_arm_reorg here.
2932 * config/arm/arm.c (after_arm_reorg): Remove the definition.
2933 (arm_split_constant): Update the way to access variable
2936 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
2938 2014-04-23 Tom de Vries <tom@codesourcery.com>
2940 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
2942 2014-04-23 David Malcolm <dmalcolm@redhat.com>
2944 * is-a.h: Update comments to reflect the following changes to the
2945 "pointerness" of the API, making the template parameter match the
2946 return type, allowing use of is-a.h with typedefs of pointers.
2947 (is_a_helper::cast): Return a T rather then a pointer to a T, so
2948 that the return type matches the parameter to the is_a_helper.
2950 (dyn_cast): Likewise.
2952 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
2953 pointer from the is-a.h API.
2955 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
2956 (is_a_helper <cgraph_node *>::test): ...this, matching change to
2958 (is_a_helper <varpool_node>::test): Likewise, convert to...
2959 (is_a_helper <varpool_node *>::test): ...this.
2961 (varpool_first_variable): Update for removal of implicit pointer
2962 from the is-a.h API.
2963 (varpool_next_variable): Likewise.
2964 (varpool_first_static_initializer): Likewise.
2965 (varpool_next_static_initializer): Likewise.
2966 (varpool_first_defined_variable): Likewise.
2967 (varpool_next_defined_variable): Likewise.
2968 (cgraph_first_defined_function): Likewise.
2969 (cgraph_next_defined_function): Likewise.
2970 (cgraph_first_function): Likewise.
2971 (cgraph_next_function): Likewise.
2972 (cgraph_first_function_with_gimple_body): Likewise.
2973 (cgraph_next_function_with_gimple_body): Likewise.
2974 (cgraph_alias_target): Likewise.
2975 (varpool_alias_target): Likewise.
2976 (cgraph_function_or_thunk_node): Likewise.
2977 (varpool_variable_node): Likewise.
2978 (symtab_real_symbol_p): Likewise.
2979 * cgraphunit.c (referred_to_p): Likewise.
2980 (analyze_functions): Likewise.
2981 (handle_alias_pairs): Likewise.
2982 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
2983 * gimple-ssa.h (gimple_vuse_op): Likewise.
2984 (gimple_vdef_op): Likewise.
2985 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
2986 * gimple.c (gimple_build_asm_1): Likewise.
2987 (gimple_build_try): Likewise.
2988 (gimple_build_resx): Likewise.
2989 (gimple_build_eh_dispatch): Likewise.
2990 (gimple_build_omp_for): Likewise.
2991 (gimple_omp_for_set_clauses): Likewise.
2993 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
2994 (is_a_helper <gimple_statement_asm *>::test): ...this.
2995 (is_a_helper <gimple_statement_bind>::test): Convert to...
2996 (is_a_helper <gimple_statement_bind *>::test): ...this.
2997 (is_a_helper <gimple_statement_call>::test): Convert to...
2998 (is_a_helper <gimple_statement_call *>::test): ...this.
2999 (is_a_helper <gimple_statement_catch>::test): Convert to...
3000 (is_a_helper <gimple_statement_catch *>::test): ...this.
3001 (is_a_helper <gimple_statement_resx>::test): Convert to...
3002 (is_a_helper <gimple_statement_resx *>::test): ...this.
3003 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
3004 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
3005 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
3006 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
3007 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
3008 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
3009 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
3010 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
3011 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
3012 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
3013 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
3014 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
3015 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
3016 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
3017 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
3018 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
3019 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
3020 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
3021 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
3022 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
3023 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
3024 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
3025 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
3026 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
3027 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
3028 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
3029 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
3030 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
3031 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
3032 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
3033 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
3034 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
3035 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
3036 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
3037 (is_a_helper <gimple_statement_phi>::test): Convert to...
3038 (is_a_helper <gimple_statement_phi *>::test): ...this.
3039 (is_a_helper <gimple_statement_transaction>::test): Convert to...
3040 (is_a_helper <gimple_statement_transaction *>::test): ...this.
3041 (is_a_helper <gimple_statement_try>::test): Convert to...
3042 (is_a_helper <gimple_statement_try *>::test): ...this.
3043 (is_a_helper <gimple_statement_wce>::test): Convert to...
3044 (is_a_helper <gimple_statement_wce *>::test): ...this.
3045 (is_a_helper <const gimple_statement_asm>::test): Convert to...
3046 (is_a_helper <const gimple_statement_asm *>::test): ...this.
3047 (is_a_helper <const gimple_statement_bind>::test): Convert to...
3048 (is_a_helper <const gimple_statement_bind *>::test): ...this.
3049 (is_a_helper <const gimple_statement_call>::test): Convert to...
3050 (is_a_helper <const gimple_statement_call *>::test): ...this.
3051 (is_a_helper <const gimple_statement_catch>::test): Convert to...
3052 (is_a_helper <const gimple_statement_catch *>::test): ...this.
3053 (is_a_helper <const gimple_statement_resx>::test): Convert to...
3054 (is_a_helper <const gimple_statement_resx *>::test): ...this.
3055 (is_a_helper <const gimple_statement_eh_dispatch>::test): Convert to...
3056 (is_a_helper <const gimple_statement_eh_dispatch *>::test): ...this.
3057 (is_a_helper <const gimple_statement_eh_filter>::test): Convert to...
3058 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
3059 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
3061 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
3063 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
3065 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
3067 (is_a_helper <const gimple_statement_omp_return>::test): Convert to...
3068 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
3069 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
3071 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
3072 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
3074 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
3075 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
3076 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
3077 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert to...
3078 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
3079 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
3081 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
3082 (is_a_helper <const gimple_statement_omp_target>::test): Convert to...
3083 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
3084 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
3086 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
3087 (is_a_helper <const gimple_statement_omp_single>::test): Convert to...
3088 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
3089 (is_a_helper <const gimple_statement_omp_teams>::test): Convert to...
3090 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
3091 (is_a_helper <const gimple_statement_omp_task>::test): Convert to...
3092 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
3093 (is_a_helper <const gimple_statement_phi>::test): Convert to...
3094 (is_a_helper <const gimple_statement_phi *>::test): ...this.
3095 (is_a_helper <const gimple_statement_transaction>::test): Convert to...
3096 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
3097 (is_a_helper <const gimple_statement_with_ops>::test): Convert to...
3098 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
3099 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
3100 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
3101 (is_a_helper <const gimple_statement_with_memory_ops>::test): Convert
3103 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
3105 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert to...
3106 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
3108 (gimple_use_ops): Update for removal of implicit pointer from the
3110 (gimple_set_use_ops): Likewise.
3111 (gimple_vuse): Likewise.
3112 (gimple_vdef): Likewise.
3113 (gimple_vuse_ptr): Likewise.
3114 (gimple_vdef_ptr): Likewise.
3115 (gimple_set_vuse): Likewise.
3116 (gimple_set_vdef): Likewise.
3117 (gimple_omp_return_set_lhs): Likewise.
3118 (gimple_omp_return_lhs): Likewise.
3119 (gimple_omp_return_lhs_ptr): Likewise.
3120 (gimple_call_fntype): Likewise.
3121 (gimple_call_set_fntype): Likewise.
3122 (gimple_call_set_internal_fn): Likewise.
3123 (gimple_call_use_set): Likewise.
3124 (gimple_call_clobber_set): Likewise.
3125 (gimple_bind_vars): Likewise.
3126 (gimple_bind_set_vars): Likewise.
3127 (gimple_bind_body_ptr): Likewise.
3128 (gimple_bind_set_body): Likewise.
3129 (gimple_bind_add_stmt): Likewise.
3130 (gimple_bind_block): Likewise.
3131 (gimple_bind_set_block): Likewise.
3132 (gimple_asm_ninputs): Likewise.
3133 (gimple_asm_noutputs): Likewise.
3134 (gimple_asm_nclobbers): Likewise.
3135 (gimple_asm_nlabels): Likewise.
3136 (gimple_asm_input_op): Likewise.
3137 (gimple_asm_input_op_ptr): Likewise.
3138 (gimple_asm_output_op): Likewise.
3139 (gimple_asm_output_op_ptr): Likewise.
3140 (gimple_asm_set_output_op): Likewise.
3141 (gimple_asm_clobber_op): Likewise.
3142 (gimple_asm_set_clobber_op): Likewise.
3143 (gimple_asm_label_op): Likewise.
3144 (gimple_asm_set_label_op): Likewise.
3145 (gimple_asm_string): Likewise.
3146 (gimple_catch_types): Likewise.
3147 (gimple_catch_types_ptr): Likewise.
3148 (gimple_catch_handler_ptr): Likewise.
3149 (gimple_catch_set_types): Likewise.
3150 (gimple_catch_set_handler): Likewise.
3151 (gimple_eh_filter_types): Likewise.
3152 (gimple_eh_filter_types_ptr): Likewise.
3153 (gimple_eh_filter_failure_ptr): Likewise.
3154 (gimple_eh_filter_set_types): Likewise.
3155 (gimple_eh_filter_set_failure): Likewise.
3156 (gimple_eh_must_not_throw_fndecl): Likewise.
3157 (gimple_eh_must_not_throw_set_fndecl): Likewise.
3158 (gimple_eh_else_n_body_ptr): Likewise.
3159 (gimple_eh_else_e_body_ptr): Likewise.
3160 (gimple_eh_else_set_n_body): Likewise.
3161 (gimple_eh_else_set_e_body): Likewise.
3162 (gimple_try_eval_ptr): Likewise.
3163 (gimple_try_cleanup_ptr): Likewise.
3164 (gimple_try_set_eval): Likewise.
3165 (gimple_try_set_cleanup): Likewise.
3166 (gimple_wce_cleanup_ptr): Likewise.
3167 (gimple_wce_set_cleanup): Likewise.
3168 (gimple_phi_capacity): Likewise.
3169 (gimple_phi_num_args): Likewise.
3170 (gimple_phi_result): Likewise.
3171 (gimple_phi_result_ptr): Likewise.
3172 (gimple_phi_set_result): Likewise.
3173 (gimple_phi_arg): Likewise.
3174 (gimple_phi_set_arg): Likewise.
3175 (gimple_resx_region): Likewise.
3176 (gimple_resx_set_region): Likewise.
3177 (gimple_eh_dispatch_region): Likewise.
3178 (gimple_eh_dispatch_set_region): Likewise.
3179 (gimple_omp_critical_name): Likewise.
3180 (gimple_omp_critical_name_ptr): Likewise.
3181 (gimple_omp_critical_set_name): Likewise.
3182 (gimple_omp_for_clauses): Likewise.
3183 (gimple_omp_for_clauses_ptr): Likewise.
3184 (gimple_omp_for_set_clauses): Likewise.
3185 (gimple_omp_for_collapse): Likewise.
3186 (gimple_omp_for_index): Likewise.
3187 (gimple_omp_for_index_ptr): Likewise.
3188 (gimple_omp_for_set_index): Likewise.
3189 (gimple_omp_for_initial): Likewise.
3190 (gimple_omp_for_initial_ptr): Likewise.
3191 (gimple_omp_for_set_initial): Likewise.
3192 (gimple_omp_for_final): Likewise.
3193 (gimple_omp_for_final_ptr): Likewise.
3194 (gimple_omp_for_set_final): Likewise.
3195 (gimple_omp_for_incr): Likewise.
3196 (gimple_omp_for_incr_ptr): Likewise.
3197 (gimple_omp_for_set_incr): Likewise.
3198 (gimple_omp_for_pre_body_ptr): Likewise.
3199 (gimple_omp_for_set_pre_body): Likewise.
3200 (gimple_omp_parallel_clauses): Likewise.
3201 (gimple_omp_parallel_clauses_ptr): Likewise.
3202 (gimple_omp_parallel_set_clauses): Likewise.
3203 (gimple_omp_parallel_child_fn): Likewise.
3204 (gimple_omp_parallel_child_fn_ptr): Likewise.
3205 (gimple_omp_parallel_set_child_fn): Likewise.
3206 (gimple_omp_parallel_data_arg): Likewise.
3207 (gimple_omp_parallel_data_arg_ptr): Likewise.
3208 (gimple_omp_parallel_set_data_arg): Likewise.
3209 (gimple_omp_task_clauses): Likewise.
3210 (gimple_omp_task_clauses_ptr): Likewise.
3211 (gimple_omp_task_set_clauses): Likewise.
3212 (gimple_omp_task_child_fn): Likewise.
3213 (gimple_omp_task_child_fn_ptr): Likewise.
3214 (gimple_omp_task_set_child_fn): Likewise.
3215 (gimple_omp_task_data_arg): Likewise.
3216 (gimple_omp_task_data_arg_ptr): Likewise.
3217 (gimple_omp_task_set_data_arg): Likewise.
3218 (gimple_omp_taskreg_clauses): Likewise.
3219 (gimple_omp_taskreg_clauses_ptr): Likewise.
3220 (gimple_omp_taskreg_set_clauses): Likewise.
3221 (gimple_omp_taskreg_child_fn): Likewise.
3222 (gimple_omp_taskreg_child_fn_ptr): Likewise.
3223 (gimple_omp_taskreg_set_child_fn): Likewise.
3224 (gimple_omp_taskreg_data_arg): Likewise.
3225 (gimple_omp_taskreg_data_arg_ptr): Likewise.
3226 (gimple_omp_taskreg_set_data_arg): Likewise.
3227 (gimple_omp_task_copy_fn): Likewise.
3228 (gimple_omp_task_copy_fn_ptr): Likewise.
3229 (gimple_omp_task_set_copy_fn): Likewise.
3230 (gimple_omp_task_arg_size): Likewise.
3231 (gimple_omp_task_arg_size_ptr): Likewise.
3232 (gimple_omp_task_set_arg_size): Likewise.
3233 (gimple_omp_task_arg_align): Likewise.
3234 (gimple_omp_task_arg_align_ptr): Likewise.
3235 (gimple_omp_task_set_arg_align): Likewise.
3236 (gimple_omp_single_clauses): Likewise.
3237 (gimple_omp_single_clauses_ptr): Likewise.
3238 (gimple_omp_single_set_clauses): Likewise.
3239 (gimple_omp_target_clauses): Likewise.
3240 (gimple_omp_target_clauses_ptr): Likewise.
3241 (gimple_omp_target_set_clauses): Likewise.
3242 (gimple_omp_target_child_fn): Likewise.
3243 (gimple_omp_target_child_fn_ptr): Likewise.
3244 (gimple_omp_target_set_child_fn): Likewise.
3245 (gimple_omp_target_data_arg): Likewise.
3246 (gimple_omp_target_data_arg_ptr): Likewise.
3247 (gimple_omp_target_set_data_arg): Likewise.
3248 (gimple_omp_teams_clauses): Likewise.
3249 (gimple_omp_teams_clauses_ptr): Likewise.
3250 (gimple_omp_teams_set_clauses): Likewise.
3251 (gimple_omp_sections_clauses): Likewise.
3252 (gimple_omp_sections_clauses_ptr): Likewise.
3253 (gimple_omp_sections_set_clauses): Likewise.
3254 (gimple_omp_sections_control): Likewise.
3255 (gimple_omp_sections_control_ptr): Likewise.
3256 (gimple_omp_sections_set_control): Likewise.
3257 (gimple_omp_for_set_cond): Likewise.
3258 (gimple_omp_for_cond): Likewise.
3259 (gimple_omp_atomic_store_set_val): Likewise.
3260 (gimple_omp_atomic_store_val): Likewise.
3261 (gimple_omp_atomic_store_val_ptr): Likewise.
3262 (gimple_omp_atomic_load_set_lhs): Likewise.
3263 (gimple_omp_atomic_load_lhs): Likewise.
3264 (gimple_omp_atomic_load_lhs_ptr): Likewise.
3265 (gimple_omp_atomic_load_set_rhs): Likewise.
3266 (gimple_omp_atomic_load_rhs): Likewise.
3267 (gimple_omp_atomic_load_rhs_ptr): Likewise.
3268 (gimple_omp_continue_control_def): Likewise.
3269 (gimple_omp_continue_control_def_ptr): Likewise.
3270 (gimple_omp_continue_set_control_def): Likewise.
3271 (gimple_omp_continue_control_use): Likewise.
3272 (gimple_omp_continue_control_use_ptr): Likewise.
3273 (gimple_omp_continue_set_control_use): Likewise.
3274 (gimple_transaction_body_ptr): Likewise.
3275 (gimple_transaction_label): Likewise.
3276 (gimple_transaction_label_ptr): Likewise.
3277 (gimple_transaction_set_body): Likewise.
3278 (gimple_transaction_set_label): Likewise.
3280 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
3281 * ipa-inline-analysis.c (inline_write_summary): Likewise.
3282 * ipa-ref.c (ipa_record_reference): Likewise.
3283 * ipa-reference.c (analyze_function): Likewise.
3284 (ipa_reference_write_optimization_summary): Likewise.
3285 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
3286 (address_taken_from_non_vtable_p): Likewise.
3287 (comdat_can_be_unshared_p_1): Likewise.
3288 * lto-cgraph.c (lto_output_ref): Likewise.
3289 (add_references): Likewise.
3290 (compute_ltrans_boundary): Likewise.
3291 (output_symtab): Likewise.
3292 (input_ref): Likewise.
3293 (input_cgraph_1): Likewise.
3294 (output_cgraph_opt_summary): Likewise.
3295 * lto-streamer-out.c (lto_output): Likewise.
3296 (output_symbol_p): Likewise.
3297 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
3298 (lsei_start_function_in_partition): Likewise.
3299 (lsei_next_variable_in_partition): Likewise.
3300 (lsei_start_variable_in_partition): Likewise.
3301 * symtab.c (insert_to_assembler_name_hash): Likewise.
3302 (unlink_from_assembler_name_hash): Likewise.
3303 (symtab_unregister_node): Likewise.
3304 (symtab_remove_node): Likewise.
3305 (dump_symtab_node): Likewise.
3306 (verify_symtab_base): Likewise.
3307 (verify_symtab_node): Likewise.
3308 (symtab_make_decl_local): Likewise.
3309 (symtab_alias_ultimate_target): Likewise.
3310 (symtab_resolve_alias): Likewise.
3311 (symtab_get_symbol_partitioning_class): Likewise.
3312 * tree-phinodes.c (allocate_phi_node): Likewise.
3313 (reserve_phi_args_for_new_edge): Likewise.
3314 (remove_phi_args): Likewise.
3315 * varpool.c (varpool_node_for_asm): Likewise.
3316 (varpool_remove_unreferenced_decls): Likewise.
3318 2014-04-23 Jeff Law <law@redhat.com>
3320 PR tree-optimization/60902
3321 * tree-ssa-threadedge.c
3322 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
3323 invalidate outputs from statements that do not produce useful
3324 outputs for threading.
3326 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
3328 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
3329 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
3330 machine descriptions for Stack Smashing Protector.
3332 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
3334 * aarch64.md (<optab>_rol<mode>3): New pattern.
3335 (<optab>_rolsi3_uxtw): Likewise.
3336 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
3338 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
3340 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
3341 (arm_cortex_a12_tune): Likewise.
3343 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3345 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
3347 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3349 * config/arm/arm.md (arm_rev16si2): New pattern.
3350 (arm_rev16si2_alt): Likewise.
3351 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
3353 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3355 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
3356 (rev16<mode>2_alt): Likewise.
3357 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
3358 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
3359 (aarch_rev16_shleft_mask_imm_p): Likewise.
3360 (aarch_rev16_p_1): Likewise.
3361 (aarch_rev16_p): Likewise.
3362 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
3363 (aarch_rev16_shright_mask_imm_p): Likewise.
3364 (aarch_rev16_shleft_mask_imm_p): Likewise.
3366 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3368 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
3369 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
3371 (cortex_a53_extra_costs): Likewise.
3372 (cortex_a57_extra_costs): Likewise.
3373 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
3374 (cortexa7_extra_costs): Likewise.
3375 (cortexa8_extra_costs): Likewise.
3376 (cortexa12_extra_costs): Likewise.
3377 (cortexa15_extra_costs): Likewise.
3378 (v7m_extra_costs): Likewise.
3379 (arm_new_rtx_costs): Handle BSWAP.
3381 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3383 * config/arm/arm.c (cortexa8_extra_costs): New table.
3384 (arm_cortex_a8_tune): New tuning struct.
3385 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
3387 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3389 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
3391 2014-04-23 Richard Biener <rguenther@suse.de>
3393 * Makefile.in (OBJS): Remove loop-unswitch.o.
3394 * tree-pass.h (make_pass_rtl_unswitch): Remove.
3395 * passes.def (pass_rtl_unswitch): Likewise.
3396 * loop-init.c (gate_rtl_unswitch): Likewise.
3397 (rtl_unswitch): Likewise.
3398 (pass_data_rtl_unswitch): Likewise.
3399 (pass_rtl_unswitch): Likewise.
3400 (make_pass_rtl_unswitch): Likewise.
3401 * rtl.h (reversed_condition): Likewise.
3402 (compare_and_jump_seq): Likewise.
3403 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
3405 * loop-unroll.c (compare_and_jump_seq): Likewise.
3407 2014-04-23 Richard Biener <rguenther@suse.de>
3409 PR tree-optimization/60903
3410 * tree-ssa-loop-im.c (analyze_memory_references): Remove
3411 commented code block.
3412 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
3413 loop flags to newly created BBs and edges.
3415 2014-04-23 Nick Clifton <nickc@redhat.com>
3417 * config/msp430/msp430.c (msp430_handle_option): Move function
3419 (msp430_option_override): Simplify mcu and mcpu option handling.
3420 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
3421 support for -mhwmult command line option.
3422 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
3423 -mhwmult command line option.
3424 (msp430_hwmult_enabled): Delete.
3425 (msp43o_output_labelref): Add support for -mhwmult command line option.
3426 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
3427 (umulsidi3): Likewise.
3428 * config/msp430/msp430.opt (mmcu): Add Report attribute.
3429 (mcpu, mlarge, msmall): Likewise.
3430 (mhwmult): New option.
3431 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
3433 (msp430_is_f5_mcu): Remove prototype.
3434 (msp430_use_f5_series_hwmult): Add prototype.
3435 * config/msp430/msp430-opts.h: New file.
3436 * common/config/msp430: New directory.
3437 * common/config/msp430/msp430-common.c: New file.
3438 * config.gcc (msp430): Remove target_has_targetm_common.
3439 * doc/invoke.texi: Document -mhwmult command line option.
3441 2014-04-23 Nick Clifton <nickc@redhat.com>
3443 * config/i386/cygwin.h (ENDFILE_SPEC): Include
3444 default-manifest.o if it can be found in the search path.
3445 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
3447 2014-04-23 Terry Guo <terry.guo@arm.com>
3449 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
3451 2014-04-23 Richard Biener <rguenther@suse.de>
3454 * tree-inline.c (declare_return_variable): Use mark_addressable.
3456 2014-04-23 Richard Biener <rguenther@suse.de>
3459 * loop-init.c (loop_optimizer_init): Make sure to apply
3460 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
3462 2014-04-22 Jakub Jelinek <jakub@redhat.com>
3465 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
3467 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
3468 if flag_sanitize_undefined_trap_on_error.
3469 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
3470 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
3471 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
3472 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
3473 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
3474 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
3475 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
3476 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
3477 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
3478 * ubsan.c (ubsan_instrument_unreachable): Return
3479 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
3480 (ubsan_expand_null_ifn): Emit __builtin_trap ()
3481 if flag_sanitize_undefined_trap_on_error and
3482 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
3483 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
3484 instrument_bool_enum_load): Emit __builtin_trap () if
3485 flag_sanitize_undefined_trap_on_error and
3486 __builtin_handle_*_abort () if !flag_sanitize_recover.
3487 * doc/invoke.texi (-fsanitize-recover,
3488 -fsanitize-undefined-trap-on-error): Document.
3490 2014-04-22 Christian Bruel <christian.bruel@st.com>
3492 * config/sh/sh.md (mov<mode>): Replace movQIHI.
3493 Force immediates to SImode.
3495 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
3497 * config/nios2/nios2.md (UNSPEC_ROUND): New.
3499 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
3500 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
3501 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
3502 (nios2_fpu_insn): Add entry for round.
3503 (N2FPU_NO_ERRNO_P): Define.
3504 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
3506 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
3508 2014-04-22 Richard Henderson <rth@redhat.com>
3510 * config/aarch64/aarch64 (addti3, subti3): New expanders.
3511 (add<GPI>3_compare0): Remove leading * from name.
3512 (add<GPI>3_carryin): Likewise.
3513 (sub<GPI>3_compare0): Likewise.
3514 (sub<GPI>3_carryin): Likewise.
3515 (<su_optab>mulditi3): New expander.
3516 (multi3): New expander.
3517 (madd<GPI>): Remove leading * from name.
3519 2014-04-22 Martin Jambor <mjambor@suse.cz>
3521 * cgraphclones.c (cgraph_function_versioning): Copy
3522 ipa_transforms_to_apply instead of asserting it is empty.
3524 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
3527 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
3528 on count_exp to get mode.
3530 2014-04-22 Andrew Pinski <apinski@cavium.com>
3532 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
3533 Handle TLS for ILP32.
3534 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
3535 (tlsie_small_<mode>): this and handle PTR.
3536 (tlsie_small_sidi): New pattern.
3537 (tlsle_small): Change to an expand to handle ILP32.
3538 (tlsle_small_<mode>): New pattern.
3539 (tlsdesc_small): Rename to ...
3540 (tlsdesc_small_<mode>): this and handle PTR.
3542 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3544 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
3546 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
3548 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
3549 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
3550 (aarch64_types_signed_poly_qualifiers): Likewise.
3551 (aarch64_types_unsigned_signed_qualifiers): Likewise.
3552 (aarch64_types_poly_signed_qualifiers): Likewise.
3553 (TYPES_REINTERP_SS): Type macro added.
3554 (TYPES_REINTERP_SU): Likewise.
3555 (TYPES_REINTERP_SP): Likewise.
3556 (TYPES_REINTERP_US): Likewise.
3557 (TYPES_REINTERP_PS): Likewise.
3558 (aarch64_fold_builtin): New expression folding added.
3559 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
3560 Declarations removed.
3561 (REINTERP_SS): Declarations added.
3562 (REINTERP_US): Likewise.
3563 (REINTERP_PS): Likewise.
3564 (REINTERP_SU): Likewise.
3565 (REINTERP_SP): Likewise.
3566 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
3567 (vreinterpretq_p8_f64): Likewise.
3568 (vreinterpret_p16_f64): Likewise.
3569 (vreinterpretq_p16_f64): Likewise.
3570 (vreinterpret_f32_f64): Likewise.
3571 (vreinterpretq_f32_f64): Likewise.
3572 (vreinterpret_f64_f32): Likewise.
3573 (vreinterpret_f64_p8): Likewise.
3574 (vreinterpret_f64_p16): Likewise.
3575 (vreinterpret_f64_s8): Likewise.
3576 (vreinterpret_f64_s16): Likewise.
3577 (vreinterpret_f64_s32): Likewise.
3578 (vreinterpret_f64_s64): Likewise.
3579 (vreinterpret_f64_u8): Likewise.
3580 (vreinterpret_f64_u16): Likewise.
3581 (vreinterpret_f64_u32): Likewise.
3582 (vreinterpret_f64_u64): Likewise.
3583 (vreinterpretq_f64_f32): Likewise.
3584 (vreinterpretq_f64_p8): Likewise.
3585 (vreinterpretq_f64_p16): Likewise.
3586 (vreinterpretq_f64_s8): Likewise.
3587 (vreinterpretq_f64_s16): Likewise.
3588 (vreinterpretq_f64_s32): Likewise.
3589 (vreinterpretq_f64_s64): Likewise.
3590 (vreinterpretq_f64_u8): Likewise.
3591 (vreinterpretq_f64_u16): Likewise.
3592 (vreinterpretq_f64_u32): Likewise.
3593 (vreinterpretq_f64_u64): Likewise.
3594 (vreinterpret_s64_f64): Likewise.
3595 (vreinterpretq_s64_f64): Likewise.
3596 (vreinterpret_u64_f64): Likewise.
3597 (vreinterpretq_u64_f64): Likewise.
3598 (vreinterpret_s8_f64): Likewise.
3599 (vreinterpretq_s8_f64): Likewise.
3600 (vreinterpret_s16_f64): Likewise.
3601 (vreinterpretq_s16_f64): Likewise.
3602 (vreinterpret_s32_f64): Likewise.
3603 (vreinterpretq_s32_f64): Likewise.
3604 (vreinterpret_u8_f64): Likewise.
3605 (vreinterpretq_u8_f64): Likewise.
3606 (vreinterpret_u16_f64): Likewise.
3607 (vreinterpretq_u16_f64): Likewise.
3608 (vreinterpret_u32_f64): Likewise.
3609 (vreinterpretq_u32_f64): Likewise.
3611 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
3613 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
3614 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
3615 (vreinterpret_p8_s8): Likewise.
3616 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
3617 (vreinterpret_p8_s16): Likewise.
3618 (vreinterpret_p8_s32): Likewise.
3619 (vreinterpret_p8_s64): Likewise.
3620 (vreinterpret_p8_f32): Likewise.
3621 (vreinterpret_p8_u8): Likewise.
3622 (vreinterpret_p8_u16): Likewise.
3623 (vreinterpret_p8_u32): Likewise.
3624 (vreinterpret_p8_u64): Likewise.
3625 (vreinterpret_p8_p16): Likewise.
3626 (vreinterpretq_p8_s8): Likewise.
3627 (vreinterpretq_p8_s16): Likewise.
3628 (vreinterpretq_p8_s32): Likewise.
3629 (vreinterpretq_p8_s64): Likewise.
3630 (vreinterpretq_p8_f32): Likewise.
3631 (vreinterpretq_p8_u8): Likewise.
3632 (vreinterpretq_p8_u16): Likewise.
3633 (vreinterpretq_p8_u32): Likewise.
3634 (vreinterpretq_p8_u64): Likewise.
3635 (vreinterpretq_p8_p16): Likewise.
3636 (vreinterpret_p16_s8): Likewise.
3637 (vreinterpret_p16_s16): Likewise.
3638 (vreinterpret_p16_s32): Likewise.
3639 (vreinterpret_p16_s64): Likewise.
3640 (vreinterpret_p16_f32): Likewise.
3641 (vreinterpret_p16_u8): Likewise.
3642 (vreinterpret_p16_u16): Likewise.
3643 (vreinterpret_p16_u32): Likewise.
3644 (vreinterpret_p16_u64): Likewise.
3645 (vreinterpret_p16_p8): Likewise.
3646 (vreinterpretq_p16_s8): Likewise.
3647 (vreinterpretq_p16_s16): Likewise.
3648 (vreinterpretq_p16_s32): Likewise.
3649 (vreinterpretq_p16_s64): Likewise.
3650 (vreinterpretq_p16_f32): Likewise.
3651 (vreinterpretq_p16_u8): Likewise.
3652 (vreinterpretq_p16_u16): Likewise.
3653 (vreinterpretq_p16_u32): Likewise.
3654 (vreinterpretq_p16_u64): Likewise.
3655 (vreinterpretq_p16_p8): Likewise.
3656 (vreinterpret_f32_s8): Likewise.
3657 (vreinterpret_f32_s16): Likewise.
3658 (vreinterpret_f32_s32): Likewise.
3659 (vreinterpret_f32_s64): Likewise.
3660 (vreinterpret_f32_u8): Likewise.
3661 (vreinterpret_f32_u16): Likewise.
3662 (vreinterpret_f32_u32): Likewise.
3663 (vreinterpret_f32_u64): Likewise.
3664 (vreinterpret_f32_p8): Likewise.
3665 (vreinterpret_f32_p16): Likewise.
3666 (vreinterpretq_f32_s8): Likewise.
3667 (vreinterpretq_f32_s16): Likewise.
3668 (vreinterpretq_f32_s32): Likewise.
3669 (vreinterpretq_f32_s64): Likewise.
3670 (vreinterpretq_f32_u8): Likewise.
3671 (vreinterpretq_f32_u16): Likewise.
3672 (vreinterpretq_f32_u32): Likewise.
3673 (vreinterpretq_f32_u64): Likewise.
3674 (vreinterpretq_f32_p8): Likewise.
3675 (vreinterpretq_f32_p16): Likewise.
3676 (vreinterpret_s64_s8): Likewise.
3677 (vreinterpret_s64_s16): Likewise.
3678 (vreinterpret_s64_s32): Likewise.
3679 (vreinterpret_s64_f32): Likewise.
3680 (vreinterpret_s64_u8): Likewise.
3681 (vreinterpret_s64_u16): Likewise.
3682 (vreinterpret_s64_u32): Likewise.
3683 (vreinterpret_s64_u64): Likewise.
3684 (vreinterpret_s64_p8): Likewise.
3685 (vreinterpret_s64_p16): Likewise.
3686 (vreinterpretq_s64_s8): Likewise.
3687 (vreinterpretq_s64_s16): Likewise.
3688 (vreinterpretq_s64_s32): Likewise.
3689 (vreinterpretq_s64_f32): Likewise.
3690 (vreinterpretq_s64_u8): Likewise.
3691 (vreinterpretq_s64_u16): Likewise.
3692 (vreinterpretq_s64_u32): Likewise.
3693 (vreinterpretq_s64_u64): Likewise.
3694 (vreinterpretq_s64_p8): Likewise.
3695 (vreinterpretq_s64_p16): Likewise.
3696 (vreinterpret_u64_s8): Likewise.
3697 (vreinterpret_u64_s16): Likewise.
3698 (vreinterpret_u64_s32): Likewise.
3699 (vreinterpret_u64_s64): Likewise.
3700 (vreinterpret_u64_f32): Likewise.
3701 (vreinterpret_u64_u8): Likewise.
3702 (vreinterpret_u64_u16): Likewise.
3703 (vreinterpret_u64_u32): Likewise.
3704 (vreinterpret_u64_p8): Likewise.
3705 (vreinterpret_u64_p16): Likewise.
3706 (vreinterpretq_u64_s8): Likewise.
3707 (vreinterpretq_u64_s16): Likewise.
3708 (vreinterpretq_u64_s32): Likewise.
3709 (vreinterpretq_u64_s64): Likewise.
3710 (vreinterpretq_u64_f32): Likewise.
3711 (vreinterpretq_u64_u8): Likewise.
3712 (vreinterpretq_u64_u16): Likewise.
3713 (vreinterpretq_u64_u32): Likewise.
3714 (vreinterpretq_u64_p8): Likewise.
3715 (vreinterpretq_u64_p16): Likewise.
3716 (vreinterpret_s8_s16): Likewise.
3717 (vreinterpret_s8_s32): Likewise.
3718 (vreinterpret_s8_s64): Likewise.
3719 (vreinterpret_s8_f32): Likewise.
3720 (vreinterpret_s8_u8): Likewise.
3721 (vreinterpret_s8_u16): Likewise.
3722 (vreinterpret_s8_u32): Likewise.
3723 (vreinterpret_s8_u64): Likewise.
3724 (vreinterpret_s8_p8): Likewise.
3725 (vreinterpret_s8_p16): Likewise.
3726 (vreinterpretq_s8_s16): Likewise.
3727 (vreinterpretq_s8_s32): Likewise.
3728 (vreinterpretq_s8_s64): Likewise.
3729 (vreinterpretq_s8_f32): Likewise.
3730 (vreinterpretq_s8_u8): Likewise.
3731 (vreinterpretq_s8_u16): Likewise.
3732 (vreinterpretq_s8_u32): Likewise.
3733 (vreinterpretq_s8_u64): Likewise.
3734 (vreinterpretq_s8_p8): Likewise.
3735 (vreinterpretq_s8_p16): Likewise.
3736 (vreinterpret_s16_s8): Likewise.
3737 (vreinterpret_s16_s32): Likewise.
3738 (vreinterpret_s16_s64): Likewise.
3739 (vreinterpret_s16_f32): Likewise.
3740 (vreinterpret_s16_u8): Likewise.
3741 (vreinterpret_s16_u16): Likewise.
3742 (vreinterpret_s16_u32): Likewise.
3743 (vreinterpret_s16_u64): Likewise.
3744 (vreinterpret_s16_p8): Likewise.
3745 (vreinterpret_s16_p16): Likewise.
3746 (vreinterpretq_s16_s8): Likewise.
3747 (vreinterpretq_s16_s32): Likewise.
3748 (vreinterpretq_s16_s64): Likewise.
3749 (vreinterpretq_s16_f32): Likewise.
3750 (vreinterpretq_s16_u8): Likewise.
3751 (vreinterpretq_s16_u16): Likewise.
3752 (vreinterpretq_s16_u32): Likewise.
3753 (vreinterpretq_s16_u64): Likewise.
3754 (vreinterpretq_s16_p8): Likewise.
3755 (vreinterpretq_s16_p16): Likewise.
3756 (vreinterpret_s32_s8): Likewise.
3757 (vreinterpret_s32_s16): Likewise.
3758 (vreinterpret_s32_s64): Likewise.
3759 (vreinterpret_s32_f32): Likewise.
3760 (vreinterpret_s32_u8): Likewise.
3761 (vreinterpret_s32_u16): Likewise.
3762 (vreinterpret_s32_u32): Likewise.
3763 (vreinterpret_s32_u64): Likewise.
3764 (vreinterpret_s32_p8): Likewise.
3765 (vreinterpret_s32_p16): Likewise.
3766 (vreinterpretq_s32_s8): Likewise.
3767 (vreinterpretq_s32_s16): Likewise.
3768 (vreinterpretq_s32_s64): Likewise.
3769 (vreinterpretq_s32_f32): Likewise.
3770 (vreinterpretq_s32_u8): Likewise.
3771 (vreinterpretq_s32_u16): Likewise.
3772 (vreinterpretq_s32_u32): Likewise.
3773 (vreinterpretq_s32_u64): Likewise.
3774 (vreinterpretq_s32_p8): Likewise.
3775 (vreinterpretq_s32_p16): Likewise.
3776 (vreinterpret_u8_s8): Likewise.
3777 (vreinterpret_u8_s16): Likewise.
3778 (vreinterpret_u8_s32): Likewise.
3779 (vreinterpret_u8_s64): Likewise.
3780 (vreinterpret_u8_f32): Likewise.
3781 (vreinterpret_u8_u16): Likewise.
3782 (vreinterpret_u8_u32): Likewise.
3783 (vreinterpret_u8_u64): Likewise.
3784 (vreinterpret_u8_p8): Likewise.
3785 (vreinterpret_u8_p16): Likewise.
3786 (vreinterpretq_u8_s8): Likewise.
3787 (vreinterpretq_u8_s16): Likewise.
3788 (vreinterpretq_u8_s32): Likewise.
3789 (vreinterpretq_u8_s64): Likewise.
3790 (vreinterpretq_u8_f32): Likewise.
3791 (vreinterpretq_u8_u16): Likewise.
3792 (vreinterpretq_u8_u32): Likewise.
3793 (vreinterpretq_u8_u64): Likewise.
3794 (vreinterpretq_u8_p8): Likewise.
3795 (vreinterpretq_u8_p16): Likewise.
3796 (vreinterpret_u16_s8): Likewise.
3797 (vreinterpret_u16_s16): Likewise.
3798 (vreinterpret_u16_s32): Likewise.
3799 (vreinterpret_u16_s64): Likewise.
3800 (vreinterpret_u16_f32): Likewise.
3801 (vreinterpret_u16_u8): Likewise.
3802 (vreinterpret_u16_u32): Likewise.
3803 (vreinterpret_u16_u64): Likewise.
3804 (vreinterpret_u16_p8): Likewise.
3805 (vreinterpret_u16_p16): Likewise.
3806 (vreinterpretq_u16_s8): Likewise.
3807 (vreinterpretq_u16_s16): Likewise.
3808 (vreinterpretq_u16_s32): Likewise.
3809 (vreinterpretq_u16_s64): Likewise.
3810 (vreinterpretq_u16_f32): Likewise.
3811 (vreinterpretq_u16_u8): Likewise.
3812 (vreinterpretq_u16_u32): Likewise.
3813 (vreinterpretq_u16_u64): Likewise.
3814 (vreinterpretq_u16_p8): Likewise.
3815 (vreinterpretq_u16_p16): Likewise.
3816 (vreinterpret_u32_s8): Likewise.
3817 (vreinterpret_u32_s16): Likewise.
3818 (vreinterpret_u32_s32): Likewise.
3819 (vreinterpret_u32_s64): Likewise.
3820 (vreinterpret_u32_f32): Likewise.
3821 (vreinterpret_u32_u8): Likewise.
3822 (vreinterpret_u32_u16): Likewise.
3823 (vreinterpret_u32_u64): Likewise.
3824 (vreinterpret_u32_p8): Likewise.
3825 (vreinterpret_u32_p16): Likewise.
3826 (vreinterpretq_u32_s8): Likewise.
3827 (vreinterpretq_u32_s16): Likewise.
3828 (vreinterpretq_u32_s32): Likewise.
3829 (vreinterpretq_u32_s64): Likewise.
3830 (vreinterpretq_u32_f32): Likewise.
3831 (vreinterpretq_u32_u8): Likewise.
3832 (vreinterpretq_u32_u16): Likewise.
3833 (vreinterpretq_u32_u64): Likewise.
3834 (vreinterpretq_u32_p8): Likewise.
3835 (vreinterpretq_u32_p16): Likewise.
3837 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
3839 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
3841 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator extended.
3843 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
3844 (vqnegd_s64): Likewise.
3845 (vqabs_s64): Likewise.
3846 (vqabsd_s64): Likewise.
3848 2014-04-22 Richard Henderson <rth@redhat.com>
3850 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
3851 computation to the top of the loop.
3853 2014-04-22 Renlin <renlin.li@arm.com>
3854 Jiong Wang <jiong.wang@arm.com>
3856 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
3857 * config/aarch64/aarch64.c (aarch64_layout_frame)
3858 (aarch64_initial_elimination_offset): Likewise.
3860 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
3862 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
3865 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
3867 * machmode.h (bitwise_mode_for_mode): Declare.
3868 * stor-layout.h (bitwise_type_for_mode): Likewise.
3869 * stor-layout.c (bitwise_mode_for_mode): New function.
3870 (bitwise_type_for_mode): Likewise.
3871 * builtins.c (fold_builtin_memory_op): Use it instead of
3872 int_mode_for_mode and build_nonstandard_integer_type.
3874 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3876 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
3877 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
3878 (*-*-solaris2*): Simplify.
3879 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
3880 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
3881 *-*-solaris2.9* handling.
3883 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
3885 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
3886 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
3888 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
3889 * configure: Regenerate.
3891 * config/i386/sol2-9.h: Remove.
3893 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
3894 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
3895 Remove Solaris 9 references.
3897 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
3899 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
3900 (floatuns<GPI:mode><GPF:mode>2): Remove.
3901 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
3902 and floatuns conversions.
3903 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
3904 and floatuns conversions.
3905 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
3906 (w1,w2): New mode attributes for inequal width conversions.
3908 2014-04-22 Renlin Li <Renlin.Li@arm.com>
3910 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
3911 the output asm format.
3913 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
3915 * config/aarch64/aarch64-simd.md
3916 (aarch64_cm<optab>di): Always split.
3917 (*aarch64_cm<optab>di): New.
3918 (aarch64_cmtstdi): Always split.
3919 (*aarch64_cmtstdi): New.
3921 2014-04-22 Jakub Jelinek <jakub@redhat.com>
3923 PR tree-optimization/60823
3924 * omp-low.c (ipa_simd_modify_function_body): Go through
3925 all SSA_NAMEs and for those refering to vector arguments
3926 which are going to be replaced adjust SSA_NAME_VAR and,
3927 if it is a default definition, change it into a non-default
3928 definition assigned at the beginning of function from new_decl.
3929 (ipa_simd_modify_stmt_ops): Rewritten.
3930 * tree-dfa.c (set_ssa_default_def): When removing default def,
3931 check for NULL loc instead of NULL *loc.
3933 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3935 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
3936 restrictions on core registers for DImode values in Thumb2.
3938 2014-04-22 Ian Bolton <ian.bolton@arm.com>
3940 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
3941 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
3943 2014-04-22 Ian Bolton <ian.bolton@arm.com>
3945 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
3946 (*iordi_notzesidi_di): Likewise.
3947 (*iordi_notsesidi_di): Likewise.
3949 2014-04-22 Ian Bolton <ian.bolton@arm.com>
3951 * config/arm/arm-protos.h (tune_params): New struct members.
3952 * config/arm/arm.c: Initialise tune_params per processor.
3953 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
3954 for speed, based on new tune_params.
3956 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
3958 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro added.
3959 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added macro.
3960 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment corrected.
3961 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
3962 * config/aarch64/arm_neon.h (vrnd_f64): Added.
3963 (vrnda_f64): Likewise.
3964 (vrndi_f64): Likewise.
3965 (vrndm_f64): Likewise.
3966 (vrndn_f64): Likewise.
3967 (vrndp_f64): Likewise.
3968 (vrndx_f64): Likewise.
3970 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3972 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
3973 GET_MODE_SIZE argument is enum machine_mode.
3975 2014-04-22 Jakub Jelinek <jakub@redhat.com>
3978 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
3979 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
3981 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
3984 * asan.c (asan_emit_stack_protection): Force the base to align to
3985 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
3986 appropriate bits if STRICT_ALIGNMENT.
3987 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
3989 (expand_used_vars): Leave a space in the stack frame for alignment
3990 if STRICT_ALIGNMENT.
3992 2014-04-21 David Malcolm <dmalcolm@redhat.com>
3994 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
3996 (gimple_store_p): Likewise.
3997 (gimple_assign_load_p): Likewise.
3998 (gimple_assign_cast_p): Likewise.
3999 (gimple_clobber_p): Likewise.
4001 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
4002 rather than a gimple.
4003 (gimple_assign_cast_p): Likewise.
4005 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
4008 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
4009 If mode is DDmode and TARGET_E500_DOUBLE allow move.
4011 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
4012 more debug information for E500 if -mdebug=reg.
4014 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
4017 * config/i386/i386.c (ix86_expand_builtin)
4018 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
4019 register for target RTX.
4020 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
4022 2014-04-18 Cong Hou <congh@google.com>
4024 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
4025 the widen-mult pattern by handling two operands with different sizes,
4026 and operands whose size is smaller than half of the result type.
4028 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
4030 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
4031 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
4032 (do_estimate_edge_time): Compute it.
4033 * ipa-inline.c (want_inline_small_function_p): Bypass
4034 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
4036 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
4038 * ipa-inline.c (spec_rem): New static variable.
4039 (dump_overall_stats): New function.
4040 (dump_inline_stats): New function.
4042 2014-04-18 Richard Henderson <rth@redhat.com>
4044 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
4045 to GET_MODE_SIZE, not a reg_class_t.
4047 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4049 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
4050 (vsx_xxmrglw_<mode>): Likewise.
4052 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
4055 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
4056 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
4057 (rs6000_init_hard_regno_mode_ok): Likewise.
4059 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
4061 * ipa-inline.c (inline_small_functions): Account only non-cold
4063 * doc/invoke.texi (inline-unit-growth): Update documentation.
4065 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
4067 * config/rs6000/rs6000.md (addti3, subti3): New.
4069 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
4072 * config/i386/i386.c (ix86_expand_clear): Remove outdated
4073 comment. Check optimize_insn_for_size_p instead of
4074 optimize_insn_for_speed_p.
4076 2014-04-17 Martin Jambor <mjambor@suse.cz>
4078 * gimple-iterator.c (gsi_start_edge): New function.
4079 * gimple-iterator.h (gsi_start_edge): Declare.
4080 * tree-sra.c (single_non_eh_succ): New function.
4081 (disqualify_ops_if_throwing_stmt): Renamed to
4082 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
4083 having one non-EH successor BB.
4084 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
4085 generate loads into replacements.
4086 (sra_modify_assign): Likewise and and also use the simple path for
4088 (sra_modify_function_body): Commit statements on edges.
4090 2014-04-17 Richard Biener <rguenther@suse.de>
4093 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
4094 comparison results and add clarifying comment.
4096 2014-04-17 Jakub Jelinek <jakub@redhat.com>
4098 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
4099 (blank_mode): Initialize it.
4100 (emit_mode_size_inline, emit_mode_nunits_inline,
4101 emit_mode_inner_inline): New functions.
4102 (emit_insn_modes_h): Call them and surround their output with
4103 #if GCC_VERSION >= 4001 ... #endif.
4104 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
4105 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
4106 mode_* arrays if the argument is __builtin_constant_p.
4107 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
4108 is enum machine_mode.
4110 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
4112 * passes.c (opt_pass::execute): Adjust.
4113 (pass_manager::execute_pass_mode_switching): Likewise.
4114 (early_local_passes::execute): Likewise.
4115 (execute_one_pass): Pass cfun to the pass's execute method.
4116 * tree-pass.h (opt_pass::execute): Add function * argument.
4117 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
4118 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
4119 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
4120 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
4121 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
4122 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
4123 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
4124 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
4125 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
4126 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
4127 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
4128 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
4129 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
4130 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
4131 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
4132 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
4133 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
4134 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
4135 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
4136 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
4137 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
4138 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
4139 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
4140 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
4141 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
4142 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
4143 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
4144 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
4145 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
4148 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
4150 * passes.c (opt_pass::gate): Take function * argument.
4151 (gate_all_early_local_passes): Merge into
4152 (early_local_passes::gate): this.
4153 (gate_all_early_optimizations): Merge into
4154 (all_early_optimizations::gate): this.
4155 (gate_all_optimizations): Mege into
4156 (all_optimizations::gate): this.
4157 (gate_all_optimizations_g): Merge into
4158 (all_optimizations_g::gate): this.
4159 (gate_rest_of_compilation): Mege into
4160 (rest_of_compilation::gate): this.
4161 (gate_postreload): Merge into
4162 (postreload::gate): this.
4163 (dump_one_pass): Pass cfun to the pass's gate method.
4164 (execute_ipa_summary_passes): Likewise.
4165 (execute_one_pass): Likewise.
4166 (ipa_write_summaries_2): Likewise.
4167 (ipa_write_optimization_summaries_1): Likewise.
4168 (ipa_read_summaries_1): Likewise.
4169 (ipa_read_optimization_summaries_1): Likewise.
4170 (execute_ipa_stmt_fixups): Likewise.
4171 * tree-pass.h (opt_pass::gate): Add function * argument.
4172 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
4173 combine-stack-adj.c, combine.c, compare-elim.c,
4174 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
4175 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
4176 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
4177 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
4178 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
4179 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
4180 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
4181 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
4182 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
4183 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
4184 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
4185 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
4186 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
4187 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
4188 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
4189 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
4190 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
4191 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
4192 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
4193 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
4194 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
4195 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
4196 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
4197 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
4198 var-tracking.c, vtable-verify.c, web.c: Adjust.
4200 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
4202 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
4203 * configure: Regenerate.
4205 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
4207 * passes.c (dump_one_pass): don't check pass->has_gate.
4208 (execute_ipa_summary_passes): Likewise.
4209 (execute_one_pass): Likewise.
4210 (ipa_write_summaries_2): Likewise.
4211 (ipa_write_optimization_summaries_1): Likewise.
4212 (ipa_read_optimization_summaries_1): Likewise.
4213 (execute_ipa_stmt_fixups): Likewise.
4214 * tree-pass.h (pass_data::has_gate): Remove.
4215 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
4216 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
4217 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
4218 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
4219 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
4220 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
4221 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
4222 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
4223 gimple-low.c, gimple-ssa-isolate-paths.c,
4224 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
4225 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
4226 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
4227 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
4228 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
4229 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
4230 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
4231 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
4232 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
4233 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
4234 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
4235 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
4236 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
4237 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
4238 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
4239 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
4240 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
4241 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
4242 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
4243 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
4244 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
4245 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
4248 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
4250 * pass_manager.h (pass_manager::register_dump_files_1): Remove
4252 * passes.c (pass_manager::register_dump_files_1): Merge into
4253 (pass_manager::register_dump_files): this, and remove its handling of
4254 properties since the pass always has the properties anyway.
4255 (pass_manager::pass_manager): Adjust.
4257 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
4259 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
4260 * passes.c (pass_manager::register_dump_files_1): Remove dead code
4261 dealing with properties.
4262 (pass_manager::register_dump_files): Adjust.
4264 2014-03-20 Mark Wielaard <mjw@redhat.com>
4266 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
4267 then represent the bound as normal constant value.
4269 2014-04-17 Jakub Jelinek <jakub@redhat.com>
4272 Forward port from 4.8 branch
4273 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
4275 * config/i386/bmiintrin.h (_blsi_u32): New.
4279 (_blsmsk_u32): Ditto.
4280 (_blsmsk_u64): Ditto.
4281 (_tzcnt_u32): Ditto.
4282 (_tzcnt_u64): Ditto.
4284 2014-04-17 Kito Cheng <kito@0xlab.org>
4286 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
4288 2014-04-17 Richard Biener <rguenther@suse.de>
4291 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
4292 boolean results for comparisons.
4294 2014-04-17 Richard Biener <rguenther@suse.de>
4296 PR tree-optimization/60836
4297 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
4298 initial PHI args to be gimple values.
4300 2014-04-17 Richard Biener <rguenther@suse.de>
4302 PR tree-optimization/60841
4303 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
4304 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
4305 of stmts to SLP build.
4306 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
4307 (vect_analyze_slp): Likewise.
4308 (vect_analyze_slp_instance): Likewise.
4309 (vect_build_slp_tree): Limit overall SLP tree growth.
4310 * tree-vectorizer.h (vect_analyze_data_refs,
4311 vect_analyze_slp): Adjust prototypes.
4313 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
4315 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
4318 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
4320 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
4321 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
4322 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
4323 for TARGET_SLOW_PSHUFB
4325 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
4327 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
4328 * config/i386/i386.c (intel_cost): Ditto.
4330 2014-04-17 Joey Ye <joey.ye@arm.com>
4332 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
4334 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
4336 * opts.c (common_handle_option): Disable -fipa-reference coorectly
4339 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
4341 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
4342 (type_all_derivations_known_p): New predicate.
4343 (type_all_ctors_visible_p): New predicate.
4344 (type_possibly_instantiated_p): New predicate.
4345 (get_odr_type): Compute all_derivations_known.
4346 (dump_odr_type): Dump the flag.
4347 (maybe_record_type): Cleanup.
4348 (record_target_from_binfo): Add bases_to_consider array;
4349 record bases for types w/o instances and skip CXX destructor.
4350 (possible_polymorphic_call_targets_1): Add bases_to_consider
4351 and consider_construction parameters; check if type may have instance.
4352 (get_polymorphic_call_info): Set maybe_in_construction to true
4353 when we know nothing.
4354 (record_targets_from_bases): Skip CXX destructors; they are
4355 never called for types in construction.
4356 (possible_polymorphic_call_targets): Do not record target when
4357 type may not have instance.
4359 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
4362 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
4363 external aliases alive, too.
4365 2014-04-16 Andrew Pinski <apinski@cavium.com>
4367 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
4370 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
4372 * final.c (compute_alignments): Do not apply loop alignment to a block
4373 falling through to the exit.
4375 2014-04-16 Catherine Moore <clm@codesourcery.com>
4377 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
4378 Adjust constraints for microMIPS store patterns.
4380 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
4382 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
4384 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
4386 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
4387 (append_use): Run at -O0.
4388 (append_vdef): Likewise.
4389 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
4390 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
4392 2014-04-16 Jakub Jelinek <jakub@redhat.com>
4394 PR tree-optimization/60844
4395 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
4396 (propagate_op_to_single_use, remove_visited_stmt_chain,
4397 linearize_expr, repropagate_negates, reassociate_bb): Use it
4398 instead of gsi_remove.
4400 2014-04-16 Martin Jambor <mjambor@suse.cz>
4402 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
4403 ipa_transforms_to_apply.
4404 (cgraph_function_versioning): Assert that old_node has empty
4405 ipa_transforms_to_apply.
4406 * trans-mem.c (ipa_tm_create_version): Likewise.
4407 * tree-inline.c (tree_function_versioning): Do not duplicate
4408 ipa_transforms_to_apply.
4410 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4413 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
4415 Pass necessary as flags on 64-bit Solaris/x86.
4416 Use lowercase relocs for x86_64-*-*.
4417 * configure: Regenerate.
4419 2014-04-15 Jan Hubicka <jh@suse.cz>
4421 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
4422 (maybe_record_node, likely_target_p): Use it.
4424 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4427 Revert following patch
4429 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
4432 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
4433 software floating point or no floating point registers, do not
4434 allow any type in the FPRs. Eliminate a test for SPE SIMD types
4435 in GPRs that occurs after we tested for GPRs that would never be
4438 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
4439 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
4440 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
4441 specifically allow DDmode, since that does not use the SPE SIMD
4444 2014-03-21 Mark Wielaard <mjw@redhat.com>
4446 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
4447 as unsigned or int depending on type and value used.
4449 2014-04-15 Richard Biener <rguenther@suse.de>
4451 PR rtl-optimization/56965
4452 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
4453 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
4455 * alias.c (true_dependence_1): Do not call
4456 nonoverlapping_component_refs_p.
4457 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
4458 nonoverlapping_component_refs_p.
4459 (indirect_refs_may_alias_p): Likewise.
4461 2014-04-15 Teresa Johnson <tejohnson@google.com>
4463 * cfg.c (dump_bb_info): Fix flags check.
4464 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
4466 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4468 PR rtl-optimization/60663
4469 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
4472 2014-04-15 Richard Biener <rguenther@suse.de>
4474 * lto-streamer.h (LTO_major_version): Bump to 4.
4476 2014-04-15 Richard Biener <rguenther@suse.de>
4478 * common.opt (lto_partition_model): New enum.
4479 (flto-partition=): Merge separate options with a single with argument,
4480 add -flto-partition=one support.
4481 * flag-types.h (enum lto_partition_model): Declare.
4482 * opts.c (finish_options): Remove duplicate -flto-partition=
4484 * lto-wrapper.c (run_gcc): Adjust.
4486 2014-04-15 Richard Biener <rguenther@suse.de>
4488 * alias.c (ncr_compar): New function.
4489 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
4491 2014-04-15 Richard Biener <rguenther@suse.de>
4493 * alias.c (record_component_aliases): Do not walk BINFOs.
4495 2014-04-15 Richard Biener <rguenther@suse.de>
4497 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
4498 Add struct function argument and adjust.
4499 (find_func_aliases_for_call): Likewise.
4500 (find_func_aliases): Likewise.
4501 (find_func_clobbers): Likewise.
4502 (intra_create_variable_infos): Likewise.
4503 (compute_points_to_sets): Likewise.
4504 (ipa_pta_execute): Adjust. Do not push/pop cfun.
4506 2014-04-15 Richard Biener <rguenther@suse.de>
4508 * tree.c (iterative_hash_expr): Use enum tree_code_class
4509 to store TREE_CODE_CLASS.
4510 (tree_block): Likewise.
4511 (tree_set_block): Likewise.
4512 * tree.h (fold_build_pointer_plus_loc): Use
4513 convert_to_ptrofftype_loc.
4515 2014-04-15 Jakub Jelinek <jakub@redhat.com>
4518 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
4521 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
4523 * cfgloop.h (struct loop): Move force_vectorize down.
4524 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
4525 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
4526 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
4527 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
4528 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
4529 * tree-core.h (enum annot_expr_kind): Add new kind values.
4530 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
4531 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
4533 * tree.def (ANNOTATE_EXPR): Tweak comment.
4535 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
4537 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
4540 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
4542 * tree.h (TYPE_IDENTIFIER): Declare.
4543 * tree.c (subrange_type_for_debug_p): Use it.
4544 * godump.c (go_format_type): Likewise.
4545 * dwarf2out.c (is_cxx_auto, modified_type_die,
4546 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
4547 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
4549 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
4552 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
4554 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
4556 * config/i386/i386.c (examine_argument): Return bool. Return true if
4557 parameter should be passed in memory.
4558 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
4559 (construct_container): Update calls to examine_argument.
4560 (function_arg_advance_64): Ditto.
4561 (return_in_memory_32): Merge with ix86_return_in_memory.
4562 (return_in_memory_64): Ditto.
4563 (return_in_memory_ms_64): Ditto.
4565 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
4567 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
4568 * coverage.c (coverage_compute_profile_id): Handle externally visible
4571 2014-04-14 Martin Jambor <mjambor@suse.cz>
4573 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
4574 DECL_DISREGARD_INLINE_LIMITS functions.
4576 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
4579 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
4581 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
4584 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
4585 optimize_insn_for_speed_p instead of
4586 optimize_function_for_speed_p.
4588 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
4590 * doc/invoke.texi (free): Document AArch64.
4592 2014-04-14 Richard Biener <rguenther@suse.de>
4594 PR tree-optimization/60042
4595 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
4596 (insert_into_preds_of_block): Do not prevent PHI insertion
4597 for REFERENCE exprs here ...
4598 (eliminate_dom_walker::before_dom_children): ... but prevent
4599 their use here under similar conditions when applied to the
4600 IL after PRE optimizations.
4602 2014-04-14 Richard Biener <rguenther@suse.de>
4604 * passes.def: Move early points-to after early SRA.
4606 2014-04-14 Richard Biener <rguenther@suse.de>
4608 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
4609 check for which sign-changes we allow when forwarding
4610 a converted value into a switch.
4612 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
4614 * stor-layout.c (place_field): Finalize non-constant offset for the
4617 2014-04-14 Richard Biener <rguenther@suse.de>
4619 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
4621 (expand_switch_using_bit_tests_p): Likewise.
4622 (process_switch): Compute and pass on speed_p based on the
4624 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
4625 optimize_bb_for_speed_p.
4627 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
4629 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
4630 * function.h (struct function): Rename has_force_vect_loops into
4631 has_force_vectorize_loops.
4632 * lto-streamer-in.c (input_cfg): Adjust for renaming.
4633 (input_struct_function_base): Likewise.
4634 * lto-streamer-out.c (output_cfg): Likewise.
4635 (output_struct_function_base): Likewise.
4636 * omp-low.c (expand_omp_simd): Likewise.
4637 * tree-cfg.c (move_sese_region_to_fn): Likewise.
4638 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
4639 (version_loop_for_if_conversion): Likewise.
4640 (tree_if_conversion): Likewise.
4641 (main_tree_if_conversion): Likewise.
4642 (gate_tree_if_conversion): Likewise.
4643 * tree-inline.c (copy_loops): Likewise.
4644 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
4645 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
4646 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
4647 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
4648 * tree-vectorizer.c (vectorize_loops): Likewise.
4649 * tree-vectorizer.h (unlimited_cost_model): Likewise.
4651 2014-04-14 Richard Biener <rguenther@suse.de>
4654 * lto-streamer-out.c (wrap_refs): New function.
4655 (lto_output): Wrap symbol references in global initializes in
4656 type-preserving MEM_REFs.
4658 2014-04-14 Christian Bruel <christian.bruel@st.com>
4660 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
4662 2014-04-14 Christian Bruel <christian.bruel@st.com>
4664 * config/sh/sh.md (setmemqi): New expand pattern.
4665 * config/sh/sh.h (CLEAR_RATIO): Define.
4666 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
4667 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
4669 2014-04-14 Richard Biener <rguenther@suse.de>
4672 * fold-const.c (negate_expr_p): Don't negate directional rounding
4674 (fold_negate_expr): Likewise.
4676 2014-04-14 Richard Biener <rguenther@suse.de>
4678 PR tree-optimization/59817
4679 PR tree-optimization/60453
4680 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
4681 recursion to catch all CHRECs in the scalar evolution and restrict
4682 the predicate for the remains appropriately.
4684 2014-04-12 Catherine Moore <clm@codesourcery.com>
4686 * config/mips/constraints.md: Add new register constraint "kb".
4687 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
4688 (*movhi_internal): Likewise.
4689 (*movqi_internal): Likewise.
4690 * config/mips/mips.h (M16_STORE_REGS): New register class.
4691 (REG_CLASS_NAMES): Add M16_STORE_REGS.
4692 (REG_CLASS_CONTENTS): Likewise.
4693 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
4695 2014-04-11 Tobias Burnus <burnus@net-b.de>
4698 * doc/invoke.texi (-Wformat-signedness): Document it.
4699 (Wformat=2): Mention that this enables -Wformat-signedness.
4701 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
4703 * common/config/epiphany/epiphany-common.c
4704 (epiphany_option_optimization_table): Enable section anchors by
4705 default at -O1 or higher.
4706 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
4707 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
4708 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
4709 carries no extra cost.
4710 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
4711 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
4712 * config/epiphany/predicates.md (memclob_operand): New predicate.
4713 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
4714 Use memclob_operand predicate and X constraint for operand 3.
4716 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
4718 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
4719 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
4722 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
4724 PR rtl-optimization/60651
4725 * mode-switching.c (optimize_mode_switching): Make sure to emit
4726 sets of a lower numbered entity before sets of a higher numbered
4727 entity to a mode of the same or lower priority.
4728 When creating a seginfo for a basic block that starts with a code
4729 label, move the insertion point past the code label.
4730 (new_seginfo): Document and enforce requirement that
4731 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
4732 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
4733 * doc/tm.texi: Regenerate.
4735 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
4738 * config/arc/arc.c (arc_save_restore): Fix assert typo.
4740 2013-04-11 Jakub Jelinek <jakub@redhat.com>
4742 * BASE-VER: Set to 4.10.0.
4744 2014-04-11 Tobias Burnus <burnus@net-b.de>
4747 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
4748 * doc/gcc.texi (Service): Update description in the @menu
4749 * doc/invoke.texi (Option Summary): Remove misplaced and
4752 2014-04-11 Steve Ellcey <sellcey@mips.com>
4753 Jakub Jelinek <jakub@redhat.com>
4756 * expr.c (convert_move): Use emit_store_flag_force instead of
4757 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
4760 2014-04-11 Richard Biener <rguenther@suse.de>
4763 * varasm.c (assemble_alias): Avoid endless error reporting
4764 recursion by setting TREE_ASM_WRITTEN.
4766 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4768 * config/s390/s390.md: Add a splitter for NOT rtx.
4770 2014-04-11 Jakub Jelinek <jakub@redhat.com>
4772 PR rtl-optimization/60663
4773 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
4775 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
4776 Jakub Jelinek <jakub@redhat.com>
4779 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
4780 flag from decl_node to node.
4782 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4785 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
4786 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
4787 ameliorating the cases where it can be.
4789 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
4792 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
4794 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
4795 (loadsync_<mode>): Change mode.
4796 (load_quadpti, store_quadpti): New.
4797 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
4798 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
4799 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
4801 2014-04-09 Cong Hou <congh@google.com>
4804 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
4807 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4809 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
4810 instead of vnor to exploit possible fusion opportunity in the
4812 (altivec_expand_vec_perm_const_le): Likewise.
4814 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
4816 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
4817 (loadsync_<mode>): Change mode.
4818 (load_quadpti, store_quadpti): New.
4819 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
4820 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
4822 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
4825 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
4826 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
4827 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
4829 2014-04-08 Richard Biener <rguenther@suse.de>
4832 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
4833 a 64bit widest int print double-int similar to on HWI64 hosts.
4835 2014-04-08 Richard Biener <rguenther@suse.de>
4837 PR tree-optimization/60785
4838 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
4839 default defs properly.
4841 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
4843 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
4844 (Weffc++): Likewise.
4846 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
4848 * ipa-devirt.c (maybe_record_node): When node is not recorded,
4849 set completep to false rather than true.
4851 2014-04-07 Douglas B Rupp <rupp@adacore.com>
4854 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
4855 ARM_TARGET2_DWARF_FORMAT.
4857 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
4860 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
4861 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
4864 2014-04-07 Richard Biener <rguenther@suse.de>
4866 PR tree-optimization/60766
4867 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
4868 (may_eliminate_iv): Convert cand_value_at result to desired type.
4870 2014-04-07 Jason Merrill <jason@redhat.com>
4873 * common.opt (-fno-gnu-unique): Add.
4874 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
4876 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4878 * haifa-sched.c: Fix outdated function reference and minor
4879 grammar errors in introductory comment.
4881 2014-04-07 Richard Biener <rguenther@suse.de>
4884 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
4886 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
4888 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
4891 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
4892 size accounting for thunks.
4893 (pa_asm_output_mi_thunk): Use final_start_function() and
4894 final_end_function() to output function start and end directives.
4896 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
4898 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
4899 device specific ISA/ feature information. Remove short_sp and
4900 errata_skip ds. Add avr_device_specific_features enum to have device
4902 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
4903 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
4904 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
4905 updated device specific info.
4906 * config/avr/avr-mcus.def: Merge device specific details to
4907 dev_attribute field.
4908 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
4910 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
4911 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
4912 assembler if RMW isa supported by current device.
4913 * config/avr/genmultilib.awk: Update as device info structure changed.
4914 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
4916 2014-04-04 Cong Hou <congh@google.com>
4918 PR tree-optimization/60656
4919 * tree-vect-stmts.c (supportable_widening_operation):
4920 Fix a bug that elements in a vector with vect_used_by_reduction
4921 property are incorrectly reordered when the operation on it is not
4922 consistant with the one in reduction operation.
4924 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
4926 PR rtl-optimization/60155
4927 * gcse.c (record_set_data): New function.
4928 (single_set_gcse): New function.
4929 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
4930 (hoist_code): Likewise.
4931 (get_pressure_class_and_nregs): Likewise.
4933 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
4935 * explow.c (probe_stack_range): Emit a final optimization blockage.
4937 2014-04-04 Anthony Green <green@moxielogic.com>
4939 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
4942 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
4945 * lto-cgraph.c (input_overwrite_node): Check that partitioning
4946 flags are set only during streaming.
4947 * ipa.c (process_references, walk_polymorphic_call_targets,
4948 symtab_remove_unreachable_nodes): Drop bodies of always inline
4949 after early inlining.
4950 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
4952 2014-04-04 Jakub Jelinek <jakub@redhat.com>
4953 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4956 * dwarf2out.c (const_ok_for_output_1): Reject expressions
4959 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4962 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
4964 (cortex_a53_fdivd): Likewise.
4966 2014-04-04 Martin Jambor <mjambor@suse.cz>
4969 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
4971 * cgraph.c (clone_of_p): Also return true if thunks match.
4972 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
4973 cgraph_function_or_thunk_node and an obsolete comment.
4974 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
4976 (build_function_decl_skip_args): Likewise.
4977 (set_new_clone_decl_and_node_flags): New function.
4978 (duplicate_thunk_for_node): Likewise.
4979 (redirect_edge_duplicating_thunks): Likewise.
4980 (cgraph_clone_node): New parameter args_to_skip, pass it to
4981 redirect_edge_duplicating_thunks which is called instead of
4982 cgraph_redirect_edge_callee.
4983 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
4984 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
4986 2014-04-04 Jeff Law <law@redhat.com>
4989 * config/arm/predicates.md (const_int_I_operand): New predicate.
4990 (const_int_M_operand): Similarly.
4991 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
4993 (insv_t2, extv_reg, extzv_t2): Likewise.
4994 (load_multiple_with_writeback): Similarly for const_int_I_operand.
4995 (pop_multiple_with_writeback_and_return): Likewise.
4996 (vfp_pop_multiple_with_writeback): Likewise
4998 2014-04-04 Richard Biener <rguenther@suse.de>
5001 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
5002 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
5004 * gimplify.h (gimple_add_tmp_var_fn): Declare.
5005 * gimplify.c (gimple_add_tmp_var_fn): New function.
5006 * gimple-expr.h (create_tmp_reg_fn): Declare.
5007 * gimple-expr.c (create_tmp_reg_fn): New function.
5008 * gimple-low.c (record_vars_into): Don't change cfun.
5009 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
5010 code generation without cfun.
5012 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
5015 * Makefile.in (install-driver): Fix shell scripting.
5017 2014-04-03 Cong Hou <congh@google.com>
5019 PR tree-optimization/60505
5020 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
5021 threshold of number of iterations below which no vectorization
5023 * tree-vect-loop.c (new_loop_vec_info):
5024 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
5025 * tree-vect-loop.c (vect_analyze_loop_operations):
5026 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
5027 * tree-vect-loop.c (vect_transform_loop):
5028 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
5029 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
5030 of iterations of the loop and see if we should build the epilogue.
5032 2014-04-03 Richard Biener <rguenther@suse.de>
5034 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
5035 (streamer_tree_cache_create): Adjust.
5036 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
5037 to allow optional nodes array.
5038 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
5039 (streamer_tree_cache_append): Likewise.
5040 (streamer_tree_cache_create): Create nodes array optionally
5041 as specified by parameter.
5042 * lto-streamer-out.c (create_output_block): Avoid maintaining
5043 the node array in the writer cache.
5044 (DFS_write_tree): Remove assertion.
5045 (produce_asm_for_decls): Free the out decl state hash table early.
5046 * lto-streamer-in.c (lto_data_in_create): Adjust for
5047 streamer_tree_cache_create prototype change.
5049 2014-04-03 Richard Biener <rguenther@suse.de>
5051 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
5052 set TREE_CHAIN to NULL_TREE.
5054 2014-04-03 Richard Biener <rguenther@suse.de>
5056 PR tree-optimization/60740
5057 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
5058 over all GIMPLE_COND operands.
5060 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
5062 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
5063 (Weffc++): Remove Scott's numbering, merge lists and reference
5066 2014-04-03 Nick Clifton <nickc@redhat.com>
5068 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
5071 2014-04-03 Martin Jambor <mjambor@suse.cz>
5073 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
5074 mention gcc_unreachable before failing.
5075 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
5078 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
5081 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
5082 inconsistent code and instead mark the context inconsistent.
5083 (possible_polymorphic_call_targets): For inconsistent contexts
5084 return empty complete list.
5086 2014-04-02 Anthony Green <green@moxielogic.com>
5088 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
5089 (extendqisi2, extendhisi2): Define.
5090 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
5091 (WCHAR_TYPE): Change to unsigned int.
5093 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5095 PR tree-optimization/60733
5096 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
5097 insertion point for PHI candidates to be the end of the feeding
5098 block for the PHI argument.
5100 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
5102 PR rtl-optimization/60650
5103 * lra-constraints.c (process_alt_operands): Decrease reject for
5104 earlyclobber matching.
5106 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5108 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
5110 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
5112 * config/spu/spu.c (pad_bb): Do not crash when the last
5113 insn is CODE_FOR_blockage.
5115 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
5117 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
5118 lies outside the target mode.
5120 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
5123 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
5124 software floating point or no floating point registers, do not
5125 allow any type in the FPRs. Eliminate a test for SPE SIMD types
5126 in GPRs that occurs after we tested for GPRs that would never be
5129 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
5130 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
5131 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
5132 specifically allow DDmode, since that does not use the SPE SIMD
5135 2014-04-02 Richard Biener <rguenther@suse.de>
5138 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
5139 MODE_INTs. Properly use negv_optab.
5140 (expand_abs): Likewise.
5142 2014-04-02 Richard Biener <rguenther@suse.de>
5145 * Makefile.in (install-driver): Guard extra installs with special
5148 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
5150 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
5153 2014-04-01 Richard Henderson <rth@redhat.com>
5156 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
5157 alternative enabled before register allocation.
5159 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
5161 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
5162 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
5164 (nios2_large_got_address): Remove unneeded 'sym' parameter.
5165 (nios2_got_address): Update nios2_large_got_address call site.
5166 (nios2_delegitimize_address): New function.
5167 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
5168 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
5169 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
5171 2014-04-01 Martin Husemann <martin@duskware.de>
5173 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
5176 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
5178 PR rtl-optimization/60604
5179 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
5180 check from register_operand.
5181 (register_operand): Redefine in terms of general_operand.
5182 (nonmemory_operand): Use register_operand for the non-constant cases.
5184 2014-04-01 Richard Biener <rguenther@suse.de>
5186 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
5188 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
5190 * doc/invoke.texi (mapp-regs): Clarify.
5192 2014-03-31 Ulrich Drepper <drepper@gmail.com>
5194 * config/i386/avx512fintrin.h (__v32hi): Define type.
5195 (__v64qi): Likewise.
5196 (_mm512_set1_epi8): Define.
5197 (_mm512_set1_epi16): Define.
5198 (_mm512_set4_epi32): Define.
5199 (_mm512_set4_epi64): Define.
5200 (_mm512_set4_pd): Define.
5201 (_mm512_set4_ps): Define.
5202 (_mm512_setr4_epi64): Define.
5203 (_mm512_setr4_epi32): Define.
5204 (_mm512_setr4_pd): Define.
5205 (_mm512_setr4_ps): Define.
5206 (_mm512_setzero_epi32): Define.
5208 2014-03-31 Martin Jambor <mjambor@suse.cz>
5211 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
5212 callsite_arguments_match_p. Updated all callers. Also check types of
5213 corresponding formal parameters and actual arguments.
5214 (not_all_callers_have_enough_arguments_p) Renamed to
5215 some_callers_have_mismatched_arguments_p.
5217 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
5219 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
5221 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
5224 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
5227 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
5229 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
5230 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
5232 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
5233 Use FMAMODE_NOVF512 mode iterator.
5234 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
5235 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
5236 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
5237 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
5239 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
5240 Use VF_128_256 mode iterator.
5241 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
5244 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
5246 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
5247 static chain if needed.
5249 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
5252 * lra-constraints.c (index_part_to_reg): New.
5253 (process_address): Use it.
5255 2014-03-27 Jeff Law <law@redhat.com>
5256 Jakub Jelinek <jakub@redhat.com>
5259 * expr.c (do_tablejump): Use simplify_gen_binary rather than
5260 gen_rtx_{PLUS,MULT} to build up the address expression.
5262 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
5263 creating non-canonical RTL.
5265 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
5268 * ipa-inline.c (want_inline_small_function_p): Short circuit large
5269 functions; reorganize to make cheap checks first.
5270 (inline_small_functions): Do not estimate growth when dumping;
5272 * ipa-inline.h (inline_summary): Add min_size.
5273 (growth_likely_positive): New function.
5274 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
5275 (set_cond_stmt_execution_predicate): Cleanup.
5276 (estimate_edge_size_and_time): Compute min_size.
5277 (estimate_calls_size_and_time): Likewise.
5278 (estimate_node_size_and_time): Likewise.
5279 (inline_update_overall_summary): Update min_size.
5280 (do_estimate_edge_time): Likewise.
5281 (do_estimate_edge_size): Update.
5282 (do_estimate_edge_hints): Update.
5283 (growth_likely_positive): New function.
5285 2014-03-28 Jakub Jelinek <jakub@redhat.com>
5288 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
5289 also if addr has VOIDmode.
5291 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5293 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
5294 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
5296 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
5297 instructions as well as AdvancedSIMD loads.
5299 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5301 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
5302 Use crypto_aese type.
5303 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
5304 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
5305 crypto_aese, crypto_aesmc. Move to types.md.
5306 * config/arm/types.md (crypto_aes): Split into crypto_aese,
5308 * config/arm/iterators.md (crypto_type): Likewise.
5310 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
5312 * cgraph.c: Include expr.h and tree-dfa.h.
5313 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
5316 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
5319 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
5320 regs from checking multi-reg pseudos.
5322 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5324 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
5326 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
5328 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
5329 if it would clobber the stack pointer, even temporarily.
5331 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
5333 * mode-switching.c: Make small adjustments to the top comment.
5335 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
5337 * config/rs6000/constraints.md (wD constraint): New constraint to
5338 match the constant integer to get the top DImode/DFmode out of a
5339 vector in a VSX register.
5341 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
5342 match the constant integer to get the top DImode/DFmode out of a
5343 vector in a VSX register.
5345 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
5348 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
5351 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
5352 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
5354 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
5355 Optimize vec_extract of 64-bit values, where the value being
5356 extracted is in the top word, where we can use scalar
5357 instructions. Add direct move and store support. Combine the big
5358 endian/little endian vector select load support into a single insn.
5359 (vsx_extract_<mode>_internal1): Likewise.
5360 (vsx_extract_<mode>_internal2): Likewise.
5361 (vsx_extract_<mode>_load): Likewise.
5362 (vsx_extract_<mode>_store): Likewise.
5363 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
5364 combined into vsx_extract_<mode>_load.
5365 (vsx_extract_<mode>_one_le): Likewise.
5367 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
5368 define the top 64-bit vector element.
5370 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
5373 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
5374 Document vec_vbpermq builtin.
5377 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
5378 enable use of xxsldwi and xxpermdi builtin functions.
5379 (vec_xxpermdi): Likewise.
5381 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
5382 Document use of vec_xxsldwi and vec_xxpermdi builtins.
5384 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
5386 PR rtl-optimization/60650
5387 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
5389 (find_spills_for): New.
5390 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
5391 Spill all pseudos on the second iteration.
5393 2014-03-27 Marek Polacek <polacek@redhat.com>
5396 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
5399 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5401 * config/s390/s390.c (s390_can_use_return_insn): Check for
5402 call-saved FPRs on 31 bit.
5404 2014-03-27 Jakub Jelinek <jakub@redhat.com>
5407 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
5408 if they need regimplification, just drop them instead of
5409 calling gimple_regimplify_operands on them.
5411 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
5414 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
5415 (aarch64_frame_pointer_required): Adjust logic.
5416 (aarch64_can_eliminate): Adjust logic.
5417 (aarch64_override_options_after_change): Adjust logic.
5419 2014-03-27 Dehao Chen <dehao@google.com>
5421 * ipa-inline.c (early_inliner): Update node's inline info.
5423 2014-03-26 Dehao Chen <dehao@google.com>
5425 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
5426 compiler inserted conditional jumps for NAN float check.
5428 2014-03-26 Jakub Jelinek <jakub@redhat.com>
5430 * ubsan.h (ubsan_create_data): Change second argument's type
5431 to const location_t *.
5432 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
5434 (ubsan_create_data): Change second argument to const location_t *PLOC.
5435 Create Loc field whenever PLOC is non-NULL.
5436 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
5437 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
5441 * real.c (real_to_integer2): Change type of low to UHWI.
5443 2014-03-26 Tobias Burnus <burnus@net-b.de>
5445 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
5446 (CILK_SELF_SPECS): New define.
5447 (driver_self_specs): Use it.
5449 2014-03-26 Richard Biener <rguenther@suse.de>
5451 * tree-pretty-print.c (percent_K_format): Implement special
5452 case for LTO and its stripped down BLOCK tree.
5454 2014-03-26 Jakub Jelinek <jakub@redhat.com>
5457 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
5459 * tree-vrp.c (simplify_internal_call_using_ranges): If only
5460 one range is range_int_cst_p, but not both, at least optimize
5461 addition/subtraction of 0 and multiplication by 0 or 1.
5462 * gimple-fold.c (gimple_fold_call): Fold
5463 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
5464 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
5465 INTEGER_CSTs, try to fold at least x * 0 and y - y.
5467 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
5469 PR rtl-optimization/60452
5470 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
5471 <case REG>: Return 1 for invalid offsets from the frame pointer.
5473 2014-03-26 Marek Polacek <polacek@redhat.com>
5476 * doc/extend.texi (C Extensions): Mention variable-length arrays in
5479 2014-03-26 Marek Polacek <polacek@redhat.com>
5482 * doc/extend.texi (Designated Inits): Describe what happens to omitted
5485 2014-03-26 Marek Polacek <polacek@redhat.com>
5488 * ira-color.c (update_conflict_hard_regno_costs): Perform the
5489 multiplication in unsigned type.
5491 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
5493 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
5495 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
5497 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
5499 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
5502 * cif-code.def (UNREACHABLE) New code.
5503 * ipa-inline.c (inline_small_functions): Skip edges to
5504 __builtlin_unreachable.
5505 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
5506 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
5507 predicate to __bulitin_unreachable.
5508 (set_cond_stmt_execution_predicate): Fix issue when
5509 invert_tree_comparison returns ERROR_MARK.
5510 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
5511 propagate to inline clones.
5512 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
5514 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
5515 * cgraphclones.c (cgraph_clone_node): If call destination is already
5516 ureachable, do not redirect it back.
5517 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
5520 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
5522 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
5523 Do not modify inline clones.
5525 2014-03-25 Jakub Jelinek <jakub@redhat.com>
5527 * config/i386/i386.md (general_sext_operand): New mode attr.
5528 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
5529 don't generate (sign_extend (const_int)).
5530 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
5531 operands[2]. Use We constraint instead of <i> and
5532 <general_sext_operand> predicate instead of <general_operand>.
5533 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
5534 * config/i386/constraints.md (We): New constraint.
5535 * config/i386/predicates.md (x86_64_sext_operand,
5536 sext_operand): New predicates.
5538 2014-03-25 Martin Jambor <mjambor@suse.cz>
5541 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
5542 inconsistent devirtualizations to __builtin_unreachable.
5544 2014-03-25 Marek Polacek <polacek@redhat.com>
5547 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
5549 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
5551 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
5552 order of elements for big-endian.
5554 2014-03-25 Richard Biener <rguenther@suse.de>
5557 * gimplify-me.c (gimple_regimplify_operands): Update the
5560 2014-03-25 Martin Jambor <mjambor@suse.cz>
5563 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
5564 (lto_output_varpool_node): Likewise.
5565 (input_overwrite_node): Likewise.
5566 (input_varpool_node): Likewise.
5568 2014-03-25 Richard Biener <rguenther@suse.de>
5570 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
5571 (run_gcc): Likewise.
5573 2014-03-25 Jakub Jelinek <jakub@redhat.com>
5575 * combine.c (simplify_compare_const): Add MODE argument.
5576 Handle mode_width 0 as very large mode_width.
5577 (try_combine, simplify_comparison): Adjust callers.
5579 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
5580 type to avoid signed integer overflow.
5581 * explow.c (plus_constant): Likewise.
5583 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
5585 * doc/generic.texi: Correct typos.
5587 2014-03-24 Tobias Burnus <burnus@net-b.de>
5589 * doc/invoke.texi (-flto): Expand section about
5590 using static libraries with LTO.
5592 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5594 PR rtl-optimization/60501
5595 * optabs.def (addptr3_optab): New optab.
5596 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
5597 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
5598 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
5600 * lra.c (emit_add3_insn): Use the addptr pattern if available.
5602 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
5604 2014-03-24 Ulrich Drepper <drepper@gmail.com>
5606 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
5609 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
5610 (_mm256_undefined_ps): Define.
5611 (_mm256_undefined_pd): Define.
5612 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
5613 (_mm_undefined_pd): Define.
5614 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
5615 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
5616 (_mm512_undefined_ps): Define.
5617 (_mm512_undefined_pd): Define.
5618 Use _mm*_undefined_*.
5619 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
5621 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
5623 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
5624 (lshr_simd): DI mode added.
5625 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
5626 (aarch64_ushr_simddi): Likewise.
5627 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
5628 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
5629 (vshrd_n_u64): Likewise.
5631 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5633 * Makefile.in (s-macro_list): Depend on cc1.
5635 2014-03-23 Teresa Johnson <tejohnson@google.com>
5637 * ipa-utils.c (ipa_print_order): Use specified dump file.
5639 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
5641 PR rtl-optimization/60601
5642 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
5644 * gcc.c (eval_spec_function): Initialize save_growing_value.
5646 2014-03-22 Jakub Jelinek <jakub@redhat.com>
5649 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
5650 code == MINUS_EXPR, never swap op0 with op1.
5652 * toplev.c (init_local_tick): Avoid signed integer multiplication
5654 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
5655 shift by first operand's bitsize.
5657 2014-03-21 Jakub Jelinek <jakub@redhat.com>
5660 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
5662 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
5663 TARGET_ISA_64BIT_P(x).
5665 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5667 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
5668 pattern for vector nor instead of subtract from splat(-1).
5669 (altivec_expand_vec_perm_const_le): Likewise.
5671 2014-03-21 Richard Henderson <rth@twiddle.net>
5674 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
5675 related insns after epilogue_completed.
5677 2014-03-21 Martin Jambor <mjambor@suse.cz>
5680 * cgraph.h (symtab_node): New flag body_removed.
5681 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
5682 when removing bodies.
5683 * symtab.c (dump_symtab_base): Dump body_removed flag.
5684 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
5685 had their bodies removed.
5687 2014-03-21 Martin Jambor <mjambor@suse.cz>
5690 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
5693 2014-03-21 Richard Biener <rguenther@suse.de>
5695 PR tree-optimization/60577
5696 * tree-core.h (struct tree_base): Document nothrow_flag use
5697 in VAR_DECL_NONALIASED.
5698 * tree.h (VAR_DECL_NONALIASED): New.
5699 (may_be_aliased): Adjust.
5700 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
5702 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
5704 * expr.c (expand_expr_real_1): Remove outdated comment.
5706 2014-03-20 Jakub Jelinek <jakub@redhat.com>
5709 * ira.c (adjust_cleared_regs): Call copy_rtx on
5710 *reg_equiv[REGNO (loc)].src_p before passing it to
5711 simplify_replace_fn_rtx.
5714 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
5715 into CONST, put pic register as first operand of PLUS. Use
5716 gen_const_mem for both 32-bit and 64-bit PIC got loads.
5718 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5720 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
5722 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
5724 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
5725 around for store forwarding issue in the FPU on the UT699.
5726 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
5727 loads and operations if -mfix-ut699 is specified.
5728 (divtf3_hq): Tweak attribute.
5729 (sqrttf2_hq): Likewise.
5731 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
5733 * calls.c (store_one_arg): Remove incorrect const qualification on the
5734 type of the temporary.
5735 * cfgexpand.c (expand_return): Likewise.
5736 * expr.c (expand_constructor): Likewise.
5737 (expand_expr_real_1): Likewise.
5739 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
5741 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
5744 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
5747 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
5749 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
5751 * config/arm/aarch-common-protos.h
5752 (alu_cost_table): Fix spelling of "extend".
5753 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
5755 2014-03-19 Richard Biener <rguenther@suse.de>
5758 * tree-core.h (tree_type_common): Re-order pointer members
5759 to reduce recursion depth during GC walks.
5761 2014-03-19 Marek Polacek <polacek@redhat.com>
5764 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
5765 before accessing it.
5767 2014-03-19 Richard Biener <rguenther@suse.de>
5770 * lto-streamer-in.c (input_function): In WPA stage do not drop
5773 2014-03-19 Jakub Jelinek <jakub@redhat.com>
5775 PR tree-optimization/60559
5776 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
5777 with build_zero_cst assignment.
5779 2014-03-18 Kai Tietz <ktietz@redhat.com>
5781 PR rtl-optimization/56356
5782 * sdbout.c (sdbout_parms): Verify that parms'
5783 incoming argument is valid.
5784 (sdbout_reg_parms): Likewise.
5786 2014-03-18 Richard Henderson <rth@redhat.com>
5789 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
5790 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
5791 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
5793 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
5795 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
5796 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
5797 Italicize plugin event names in description. Explain that
5798 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
5799 Remind that no GCC functions should be called after PLUGIN_FINISH.
5800 Explain what pragmas with expansion are.
5802 2014-03-18 Martin Liska <mliska@suse.cz>
5804 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
5805 gimple call statement is update.
5806 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
5807 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
5809 2014-03-18 Jakub Jelinek <jakub@redhat.com>
5812 * ubsan.c (ubsan_instrument_unreachable): Call
5813 initialize_sanitizer_builtins.
5814 (ubsan_pass): Likewise.
5817 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
5818 varpool_finalize_decl instead of rest_of_decl_compilation.
5820 2014-03-18 Richard Biener <rguenther@suse.de>
5822 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
5823 by using bitmap_and_compl instead of bitmap_and_compl_into.
5824 (df_rd_transfer_function): Likewise.
5826 2014-03-18 Richard Biener <rguenther@suse.de>
5828 * doc/lto.texi (fresolution): Fix typo.
5830 2014-03-18 Richard Biener <rguenther@suse.de>
5832 * doc/invoke.texi (flto): Update for changes in 4.9.
5834 2014-03-18 Richard Biener <rguenther@suse.de>
5836 * doc/loop.texi: Remove section on the removed lambda framework.
5837 Update loop docs with recent changes in preserving loop structure.
5839 2014-03-18 Richard Biener <rguenther@suse.de>
5841 * doc/lto.texi (-fresolution): Document.
5843 2014-03-18 Richard Biener <rguenther@suse.de>
5845 * doc/contrib.texi: Adjust my name.
5847 2014-03-18 Jakub Jelinek <jakub@redhat.com>
5850 * internal-fn.c: Include diagnostic-core.h.
5851 (expand_BUILTIN_EXPECT): New function.
5852 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
5853 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
5854 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
5855 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
5857 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
5858 Revert 3 argument __builtin_expect code.
5859 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
5860 * gimple-fold.c (gimple_fold_call): Likewise.
5861 * tree.h (fold_builtin_expect): New prototype.
5862 * builtins.c (build_builtin_expect_predicate): Add predictor
5863 argument, if non-NULL, create 3 argument __builtin_expect.
5864 (fold_builtin_expect): No longer static. Add ARG2 argument,
5865 pass it through to build_builtin_expect_predicate.
5866 (fold_builtin_2): Adjust caller.
5867 (fold_builtin_3): Handle BUILT_IN_EXPECT.
5868 * internal-fn.def (BUILTIN_EXPECT): New.
5870 2014-03-18 Tobias Burnus <burnus@net-b.de>
5873 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
5874 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
5875 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
5877 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
5880 * predict.c (combine_predictions_for_bb): Fix up formatting.
5881 (expr_expected_value_1, expr_expected_value): Add predictor argument,
5882 fill what it points to if non-NULL.
5883 (tree_predict_by_opcode): Adjust caller, use the predictor.
5884 * predict.def (PRED_COMPARE_AND_SWAP): Add.
5886 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
5888 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
5889 proper constant for the store mode.
5891 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
5893 * symtab.c (change_decl_assembler_name): Fix transparent alias
5896 2014-03-16 Renlin Li <Renlin.Li@arm.com>
5898 * config/aarch64/aarch64.c: Correct the comments about the
5899 aarch64 stack layout.
5901 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
5903 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
5904 check for GF_OMP_FOR_KIND_FOR.
5906 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
5908 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
5909 ymm and zmm register names.
5911 2014-03-17 Jakub Jelinek <jakub@redhat.com>
5914 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
5915 note creation for the 2010-08-31 changes.
5917 2014-03-17 Marek Polacek <polacek@redhat.com>
5920 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
5921 as -fno-tree-loop-vectorize.
5922 (expand_omp_simd): Likewise.
5924 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
5926 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
5927 (eligible_for_call_delay): New prototype.
5928 * config/sparc/sparc.c (tls_call_delay): Rename into...
5929 (eligible_for_call_delay): ...this. Return false if the instruction
5930 cannot be put in the delay slot of a branch.
5931 (eligible_for_restore_insn): Simplify.
5932 (eligible_for_return_delay): Return false if the instruction cannot be
5933 put in the delay slot of a branch and simplify.
5934 (eligible_for_sibcall_delay): Return false if the instruction cannot be
5935 put in the delay slot of a branch.
5936 * config/sparc/sparc.md (fix_ut699): New attribute.
5937 (tls_call_delay): Delete.
5938 (in_call_delay): Reimplement.
5939 (eligible_for_sibcall_delay): Rename into...
5940 (in_sibcall_delay): ...this.
5941 (eligible_for_return_delay): Rename into...
5942 (in_return_delay): ...this.
5943 (in_branch_delay): Reimplement.
5944 (in_uncond_branch_delay): Delete.
5945 (in_annul_branch_delay): Delete.
5947 2014-03-14 Richard Henderson <rth@redhat.com>
5950 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
5951 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
5952 (*floathi<X87MODEF>2_i387_with_temp): Remove.
5953 (floathi splitters): Remove.
5954 (float<SWI48x>xf2): New pattern.
5955 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
5956 code that tried to handle DImode for 32-bit, but which was excluded
5957 by the pattern's condition. Drop allocation of stack temporary.
5958 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
5959 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
5960 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
5961 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
5962 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
5963 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
5964 (*float<SWI48><MODEF>2_sse_interunit): Remove.
5965 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
5966 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
5967 (*float<SWI48x><X87MODEF>2_i387): Remove.
5968 (all float _with_temp splitters): Remove.
5969 (*float<SWI48x><MODEF>2_i387): New pattern.
5970 (*float<SWI48><MODEF>2_sse): New pattern.
5971 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
5972 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
5974 2014-03-14 Jakub Jelinek <jakub@redhat.com>
5975 Marek Polacek <polacek@redhat.com>
5978 * common.opt (dump_base_name_prefixed): New Variable.
5979 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
5980 if x_dump_base_name_prefixed is already set, set it at the end.
5982 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
5984 PR rtl-optimization/60508
5985 * lra-constraints.c (get_reload_reg): Add new parameter
5987 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
5988 Pass the new parameter values.
5990 2014-03-14 Richard Biener <rguenther@suse.de>
5992 * common.opt: Revert unintented changes from r205065.
5995 2014-03-14 Richard Biener <rguenther@suse.de>
5998 * cfghooks.c (split_block): Properly adjust all loops the
5999 block was a latch of.
6001 2014-03-14 Martin Jambor <mjambor@suse.cz>
6004 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
6007 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
6010 * config/avr/avr.c (avr_set_current_function): Pass function name
6011 through default_strip_name_encoding before sanity checking instead
6012 of skipping the first char of the assembler name.
6014 2014-03-13 Richard Henderson <rth@redhat.com>
6017 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
6018 (ix86_force_to_memory, ix86_free_from_memory): Remove.
6019 * config/i386/i386-protos.h: Likewise.
6020 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
6021 in the expander instead of a splitter.
6022 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
6023 any possibility of requiring a memory.
6024 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
6025 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
6026 (fp branch splitters): Update for ix86_split_fp_branch.
6027 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
6028 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
6029 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
6030 (*fop_<MODEF>_2_i387): Remove f/r alternative.
6031 (*fop_<MODEF>_3_i387): Likewise.
6032 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
6033 (splitters for the fop_* register patterns): Remove.
6034 (fscalexf4_i387): Rename from *fscalexf4_i387.
6035 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
6037 2014-03-13 Jakub Jelinek <jakub@redhat.com>
6039 PR tree-optimization/59779
6040 * tree-dfa.c (get_ref_base_and_extent): Use double_int
6041 type for bitsize and maxsize instead of HOST_WIDE_INT.
6043 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
6045 PR rtl-optimization/57320
6046 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
6047 the CFG after thread_prologue_and_epilogue_insns.
6049 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
6051 PR rtl-optimization/57189
6052 * lra-constraints.c (process_alt_operands): Disfavor spilling
6055 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
6057 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
6059 2014-03-13 Jakub Jelinek <jakub@redhat.com>
6061 PR tree-optimization/59025
6063 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
6064 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
6066 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
6069 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
6070 calls of avr_out_plus_1.
6072 2014-03-13 Bin Cheng <bin.cheng@arm.com>
6074 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
6075 BB's single pred and update the father loop's latch info later.
6077 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
6079 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
6083 (VEC_base): Likewise.
6084 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
6085 registers, we need to swap double words in little endian mode.
6087 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
6088 to be a container mode for 128-bit integer operations added in ISA
6089 2.07. Unlike TImode and PTImode, the preferred register set is
6090 the Altivec/VMX registers for the 128-bit operations.
6092 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
6094 (rs6000_split_128bit_ok_p): Likewise.
6096 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
6097 macros for creating ISA 2.07 normal and overloaded builtin
6098 functions with 3 arguments.
6099 (BU_P8V_OVERLOAD_3): Likewise.
6100 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
6101 for use as overloaded functions.
6102 (VPERM_1TI_UNS): Likewise.
6103 (VSEL_1TI): Likewise.
6104 (VSEL_1TI_UNS): Likewise.
6105 (ST_INTERNAL_1ti): Likewise.
6106 (LD_INTERNAL_1ti): Likewise.
6107 (XXSEL_1TI): Likewise.
6108 (XXSEL_1TI_UNS): Likewise.
6109 (VPERM_1TI): Likewise.
6110 (VPERM_1TI_UNS): Likewise.
6111 (XXPERMDI_1TI): Likewise.
6112 (SET_1TI): Likewise.
6113 (LXVD2X_V1TI): Likewise.
6114 (STXVD2X_V1TI): Likewise.
6115 (VEC_INIT_V1TI): Likewise.
6116 (VEC_SET_V1TI): Likewise.
6117 (VEC_EXT_V1TI): Likewise.
6118 (EQV_V1TI): Likewise.
6119 (NAND_V1TI): Likewise.
6120 (ORC_V1TI): Likewise.
6121 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
6122 added in ISA 2.07. Add both normal 'altivec' builtins, and the
6124 (VADDUQM): Likewise.
6125 (VSUBCUQ): Likewise.
6126 (VADDEUQM): Likewise.
6127 (VADDECUQ): Likewise.
6128 (VSUBEUQM): Likewise.
6129 (VSUBECUQ): Likewise.
6131 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
6132 __int128_t and __uint128_t types.
6133 (__uint128_type): Likewise.
6134 (altivec_categorize_keyword): Add support for vector __int128_t,
6135 vector __uint128_t, vector __int128, and vector unsigned __int128
6136 as a container type for TImode operations that need to be done in
6137 VSX/Altivec registers.
6138 (rs6000_macro_to_expand): Likewise.
6139 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
6140 to support 128-bit integer instructions vaddcuq, vadduqm,
6141 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
6142 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
6144 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
6145 for V1TImode, and set up preferences to use VSX/Altivec registers.
6146 Setup VSX reload handlers.
6147 (rs6000_debug_reg_global): Likewise.
6148 (rs6000_init_hard_regno_mode_ok): Likewise.
6149 (rs6000_preferred_simd_mode): Likewise.
6150 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
6151 (easy_altivec_constant): Likewise.
6152 (output_vec_const_move): Likewise.
6153 (rs6000_expand_vector_set): Convert V1TImode set and extract to
6155 (rs6000_expand_vector_extract): Likewise.
6156 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
6158 (rs6000_const_vec): Add support for V1TImode.
6159 (rs6000_emit_le_vsx_load): Swap double words when loading or
6160 storing TImode/V1TImode.
6161 (rs6000_emit_le_vsx_store): Likewise.
6162 (rs6000_emit_le_vsx_move): Likewise.
6163 (rs6000_emit_move): Add support for V1TImode.
6164 (altivec_expand_ld_builtin): Likewise.
6165 (altivec_expand_st_builtin): Likewise.
6166 (altivec_expand_vec_init_builtin): Likewise.
6167 (altivec_expand_builtin): Likewise.
6168 (rs6000_init_builtins): Add support for V1TImode type. Add
6169 support for ISA 2.07 128-bit integer builtins. Define type names
6170 for the VSX/Altivec vector types.
6171 (altivec_init_builtins): Add support for overloaded vector
6172 functions with V1TImode type.
6173 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
6174 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
6176 (rs6000_split_128bit_ok_p): Likewise.
6177 (rs6000_handle_altivec_attribute): Create V1TImode from vector
6178 __int128_t and vector __uint128_t.
6180 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
6181 and mode attributes.
6188 (VS_scalar): Likewise.
6189 (VS_double): Likewise.
6190 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
6192 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
6193 we support the ISA 2.07 128-bit integer arithmetic instructions.
6194 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
6195 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
6196 and TImode types for use with the builtin functions.
6197 (V1TI_type_node): Likewise.
6198 (unsigned_V1TI_type_node): Likewise.
6199 (intTI_type_internal_node): Likewise.
6200 (uintTI_type_internal_node): Likewise.
6202 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
6203 128-bit builtin functions.
6204 (UNSPEC_VADDEUQM): Likewise.
6205 (UNSPEC_VADDECUQ): Likewise.
6206 (UNSPEC_VSUBCUQ): Likewise.
6207 (UNSPEC_VSUBEUQM): Likewise.
6208 (UNSPEC_VSUBECUQ): Likewise.
6209 (VM): Add V1TImode to vector mode iterators.
6211 (VI_unit): Likewise.
6212 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
6213 (altivec_vaddcuq): Likewise.
6214 (altivec_vsubuqm): Likewise.
6215 (altivec_vsubcuq): Likewise.
6216 (altivec_vaddeuqm): Likewise.
6217 (altivec_vaddecuq): Likewise.
6218 (altivec_vsubeuqm): Likewise.
6219 (altivec_vsubecuq): Likewise.
6221 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
6223 (BOOL_128): Likewise.
6224 (BOOL_REGS_OUTPUT): Likewise.
6225 (BOOL_REGS_OP1): Likewise.
6226 (BOOL_REGS_OP2): Likewise.
6227 (BOOL_REGS_UNARY): Likewise.
6228 (BOOL_REGS_AND_CR0): Likewise.
6230 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
6231 128-bit integer builtin support.
6232 (vec_vadduqm): Likewise.
6233 (vec_vaddecuq): Likewise.
6234 (vec_vaddeuqm): Likewise.
6235 (vec_vsubecuq): Likewise.
6236 (vec_vsubeuqm): Likewise.
6237 (vec_vsubcuq): Likewise.
6238 (vec_vsubuqm): Likewise.
6240 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
6241 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
6242 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
6243 128-bit integer add/subtract to ISA 2.07.
6245 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
6247 * config/arc/arc.c (arc_predicate_delay_insns):
6248 Fix third argument passed to conditionalize_nonjump.
6250 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
6252 * config/aarch64/aarch64-builtins.c
6253 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
6254 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
6255 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
6256 instead of __builtin_lfloor.
6257 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
6259 2014-03-12 Jakub Jelinek <jakub@redhat.com>
6261 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
6262 (tree_ssa_ifcombine_bb_1): New function.
6263 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
6264 is an empty forwarder block to then_bb or vice versa and then_bb
6265 and else_bb are effectively swapped.
6267 2014-03-12 Christian Bruel <christian.bruel@st.com>
6270 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
6271 REG_CFA_DEF_CFA note.
6272 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
6273 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
6275 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
6277 PR tree-optimization/60454
6278 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
6280 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6282 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
6283 Do not define target_cpu_default2 to generic.
6284 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
6285 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
6286 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
6288 2014-03-12 Jakub Jelinek <jakub@redhat.com>
6289 Marc Glisse <marc.glisse@inria.fr>
6291 PR tree-optimization/60502
6292 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
6293 instead of build_low_bits_mask.
6295 2014-03-12 Jakub Jelinek <jakub@redhat.com>
6298 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
6299 if there are multiple uses, but op doesn't live on E edge.
6300 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
6301 clobber stmts before __builtin_unreachable.
6303 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
6305 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
6306 hard_frame_pointer_rtx.
6307 * cse.c (cse_insn): Remove volatile check.
6308 * cselib.c (cselib_process_insn): Likewise.
6309 * dse.c (scan_insn): Likewise.
6311 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
6313 * config/arc/arc.c (conditionalize_nonjump): New function,
6315 (arc_ifcvt): ... this.
6316 (arc_predicate_delay_insns): Use it.
6318 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
6320 * config/arc/predicates.md (extend_operand): During/after reload,
6321 allow const_int_operand.
6322 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
6323 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
6324 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
6326 (umulsi3_highpart_i): Likewise.
6328 2014-03-11 Richard Biener <rguenther@suse.de>
6330 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
6331 Add asserts to guard possible wrong-code bugs.
6333 2014-03-11 Richard Biener <rguenther@suse.de>
6335 PR tree-optimization/60429
6336 PR tree-optimization/60485
6337 * tree-ssa-structalias.c (set_union_with_increment): Properly
6338 take into account all fields that overlap the shifted vars.
6339 (do_sd_constraint): Likewise.
6340 (do_ds_constraint): Likewise.
6341 (get_constraint_for_ptr_offset): Likewise.
6343 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
6345 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
6346 (nios2_compute_frame_layout):
6347 Add calculation of cfun->machine->fp_save_offset.
6348 (nios2_expand_prologue): Correct setting of frame pointer register
6350 (nios2_expand_epilogue): Update recovery of stack pointer from
6351 frame pointer accordingly.
6352 (nios2_initial_elimination_offset): Update calculation of offset
6353 for eliminating to HARD_FRAME_POINTER_REGNUM.
6355 2014-03-10 Jakub Jelinek <jakub@redhat.com>
6358 * ipa.c (symtab_remove_unreachable_nodes): Don't call
6359 cgraph_get_create_node on VAR_DECLs.
6361 2014-03-10 Richard Biener <rguenther@suse.de>
6364 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
6366 2014-03-08 Douglas B Rupp <rupp@gnat.com>
6368 * config/vms/vms.opt (vms_float_format): New variable.
6370 2014-03-08 Tobias Burnus <burnus@net-b.de>
6372 * doc/invoke.texi (-fcilkplus): Update implementation status.
6374 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
6375 Richard Biener <rguenther@suse.de>
6377 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
6378 consistently accross all TUs.
6379 (run_gcc): Enable -fshort-double automatically at link at link-time
6380 and disallow override.
6382 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
6385 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
6386 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
6387 if they can't be used.
6389 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6391 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
6392 for Solaris 11/x86 ld.
6393 * configure: Regenerate.
6395 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6397 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
6398 (LIB_TLS_SPEC): Save as ld_tls_libs.
6399 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
6400 (HAVE_AS_IX86_TLSLDM): New test.
6401 * configure, config.in: Regenerate.
6402 * config/i386/i386.c (legitimize_tls_address): Fall back to
6403 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
6404 cannot support TLS_MODEL_LOCAL_DYNAMIC.
6405 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
6406 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
6408 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
6410 * common.opt (fira-loop-pressure): Mark as optimization.
6412 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
6414 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
6415 an OpenMP mappable type.
6417 2014-03-06 Matthias Klose <doko@ubuntu.com>
6419 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
6420 MULTILIB_OSDIRNAMES is not defined.
6422 2014-03-06 Jakub Jelinek <jakub@redhat.com>
6423 Meador Inge <meadori@codesourcery.com>
6426 * config/arm/arm.c (arm_tls_symbol_p): Remove.
6427 (arm_legitimize_address): Call legitimize_tls_address for any
6428 arm_tls_referenced_p expression, handle constant addend. Call it
6429 before testing for !TARGET_ARM.
6430 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
6432 2014-03-06 Richard Biener <rguenther@suse.de>
6438 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
6440 * tree-streamer.c (record_common_node): Assert we don't record
6441 nodes with type double.
6442 (preload_common_node): Skip type double, complex double and double
6443 pointer since it is now frontend dependent due to fshort-double option.
6445 2014-03-06 Richard Biener <rguenther@suse.de>
6447 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
6448 or -fno-lto is specified and the linker has full plugin support.
6449 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
6450 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
6451 * lto-wrapper.c (merge_and_complain): Merge compile-time
6452 optimization levels.
6453 (run_gcc): And pass it through to the link options.
6455 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
6459 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
6461 * cselib.c (remove_useless_values): Skip to avoid quadratic
6462 behavior if the condition moved from...
6463 (cselib_process_insn): ... here holds.
6465 2014-03-05 Jakub Jelinek <jakub@redhat.com>
6468 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
6469 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
6472 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
6473 (TM_H): Add x86-tune.def.
6475 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6477 * config/aarch64/aarch64.c (generic_tunings):
6478 Use cortexa57_extra_costs.
6480 2014-03-05 Jakub Jelinek <jakub@redhat.com>
6483 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
6484 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
6485 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
6488 2014-03-04 Heiher <r@hev.cc>
6490 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
6491 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
6493 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
6495 * config/i386/predicates.md (const2356_operand): Change to ...
6496 (const2367_operand): ... this.
6497 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
6499 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6500 (*avx512pf_scatterpf<mode>sf): Ditto.
6501 (avx512pf_scatterpf<mode>df): Ditto.
6502 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6503 (*avx512pf_scatterpf<mode>df): Ditto.
6504 * config/i386/i386.c (ix86_expand_builtin): Update
6505 incorrect hint operand error message.
6507 2014-03-04 Richard Biener <rguenther@suse.de>
6509 * lto-section-in.c (lto_get_section_data): Fix const cast.
6511 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
6513 * tree-streamer.c (record_common_node): Assert we don't record
6514 nodes with type double.
6515 (preload_common_node): Skip type double, complex double and double
6516 pointer since it is now frontend dependent due to fshort-double option.
6518 2014-03-04 Richard Biener <rguenther@suse.de>
6521 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
6522 (lto_input_toplevel_asms): Likewise.
6523 * lto-section-in.c (lto_get_section_data): Instead do it here
6526 2014-03-04 Richard Biener <rguenther@suse.de>
6528 PR tree-optimization/60382
6529 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
6530 dead PHIs a reduction.
6532 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
6534 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
6536 (_mm_prefetch): Move out of GCC target("sse") pragma.
6537 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
6538 GCC target("prfchw") pragma.
6539 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
6541 * config/i386/i386.c (ix86_option_override_internal): Enable
6542 -mprfchw with -mprefetchwt1.
6544 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
6546 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
6549 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
6551 * opts.h (CL_PCH_IGNORE): Define.
6552 * targhooks.c (option_affects_pch_p):
6553 Return false for options that have CL_PCH_IGNORE set.
6554 * opt-functions.awk: Process PchIgnore.
6555 * doc/options.texi: Document PchIgnore.
6557 * config/arc/arc.opt (misize): Add PchIgnore property.
6559 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6561 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
6562 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
6563 constraint on constants to permit them being loaded into
6564 GENERAL_REGS or BASE_REGS.
6566 2014-03-03 Nick Clifton <nickc@redhat.com>
6568 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
6569 anti-cacnonical alternatives.
6570 (negandhi3_real): New pattern.
6571 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
6573 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
6575 * config/avr/avr-mcus.def: Remove atxmega16x1.
6576 * config/avr/avr-tables.opt: Regenerate.
6577 * config/avr/t-multilib: Regenerate.
6578 * doc/avr-mmcu.texi: Regenerate.
6580 2014-03-03 Tobias Grosser <tobias@grosser.es>
6581 Mircea Namolaru <mircea.namolaru@inria.fr>
6583 PR tree-optimization/58028
6584 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
6587 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6589 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
6590 not handled by recognizers.
6592 2014-03-03 Jakub Jelinek <jakub@redhat.com>
6595 * function.c (expand_function_end): Don't emit
6596 clobber_return_register sequence if clobber_after is a BARRIER.
6597 * cfgexpand.c (construct_exit_block): Append instructions before
6598 return_label to prev_bb.
6600 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6602 * config/rs6000/constraints.md: Document reserved use of "wc".
6604 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
6607 * ipa.c (function_and_variable_visibility): When dissolving comdat
6608 group, also set all symbols to local.
6610 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
6615 2013-12-14 Jan Hubicka <jh@suse.cz>
6617 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
6619 2014-03-02 Jon Beniston <jon@beniston.com>
6625 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
6626 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
6627 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
6628 (simple_return, *simple_return): New patterns
6629 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
6630 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
6632 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
6634 * dwarf2out.c (gen_subprogram_die): Tidy.
6636 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
6639 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
6640 (*mov_t_msb_neg_negc): ... this new insn.
6642 2014-02-28 Jason Merrill <jason@redhat.com>
6645 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
6648 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
6651 * dwarf2out.c (decltype_auto_die): New static.
6652 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
6653 (gen_type_die_with_usage): Handle 'decltype(auto)'.
6654 (is_cxx_auto): Likewise.
6656 2014-02-28 Ian Bolton <ian.bolton@arm.com>
6658 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
6659 we are not using general regs only.
6661 2014-02-28 Richard Biener <rguenther@suse.de>
6664 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
6665 previous fix and only allow to remove trivial pre-headers
6666 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
6667 (remove_forwarder_block): Properly update the latch of a loop.
6669 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
6672 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
6673 (cselib_preserved_hash_table): New.
6674 (preserve_constants_and_equivs): Move preserved vals to it.
6675 (cselib_find_slot): Look it up first.
6676 (cselib_init): Initialize it.
6677 (cselib_finish): Release it.
6678 (dump_cselib_table): Dump it.
6680 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
6683 * cselib.c (remove_useless_values): Skip to avoid quadratic
6684 behavior if the condition moved from...
6685 (cselib_process_insn): ... here holds.
6687 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
6690 * var-tracking.c (vt_initialize): Apply the same condition to
6691 preserve the CFA base value.
6693 2014-02-28 Joey Ye <joey.ye@arm.com>
6696 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
6697 if reload in progress or completed.
6699 2014-02-28 Tobias Burnus <burnus@net-b.de>
6702 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
6705 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
6707 * doc/tm.texi.in (Condition Code Status): Update documention for
6708 relative locations of cc0-setter and cc0-user.
6710 2014-02-27 Jeff Law <law@redhat.com>
6712 PR rtl-optimization/52714
6713 * combine.c (try_combine): When splitting an unrecognized PARALLEL
6714 into two independent simple sets, if I3 is a jump, ensure the
6715 pattern we place into I3 is a (set (pc) ...).
6717 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
6718 Jeff Law <law@redhat.com>
6720 PR rtl-optimization/49847
6721 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
6722 are in different blocks.
6723 * doc/tm.texi (Condition Code Status): Update documention for
6724 relative locations of cc0-setter and cc0-user.
6726 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
6729 * lra.c (lra_emit_add): Check SUBREG too.
6731 2014-02-27 Andreas Schwab <schwab@suse.de>
6733 * config/m68k/m68k.c (m68k_option_override): Disable
6734 -flive-range-shrinkage for classic m68k.
6735 (m68k_override_options_after_change): Likewise.
6737 2014-02-27 Marek Polacek <polacek@redhat.com>
6740 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
6741 -Wmaybe-uninitialized.
6743 2014-02-27 Alan Modra <amodra@gmail.com>
6746 * reload1.c (emit_input_reload_insns): When reload_override_in,
6747 set old to rl->in_reg when rl->in_reg is a subreg.
6749 2014-02-26 Richard Biener <rguenther@suse.de>
6752 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
6754 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
6756 * common/config/i386/predicates.md (const1256_operand): Remove.
6757 (const2356_operand): New.
6758 (const_1_to_2_operand): Remove.
6759 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
6760 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
6761 (*avx512pf_gatherpf<mode>sf): Ditto.
6762 (avx512pf_gatherpf<mode>df): Ditto.
6763 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6764 (*avx512pf_gatherpf<mode>df): Ditto.
6765 (avx512pf_scatterpf<mode>sf): Ditto.
6766 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6767 (*avx512pf_scatterpf<mode>sf): Ditto.
6768 (avx512pf_scatterpf<mode>df): Ditto.
6769 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6770 (*avx512pf_scatterpf<mode>df): Ditto.
6771 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
6773 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
6775 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
6776 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
6777 (_mm512_mask_testn_epi64_mask): Move to ...
6778 * config/i386/avx512cdintrin.h: Here.
6779 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
6780 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
6781 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
6782 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
6783 TARGET_AVX512F from TARGET_AVX512CD.
6785 2014-02-26 Richard Biener <rguenther@suse.de>
6788 * ipa.c (walk_polymorphic_call_targets): Properly guard
6789 call to inline_update_overall_summary.
6791 2014-02-26 Bin Cheng <bin.cheng@arm.com>
6794 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
6795 and latches only if requested. Fix latch if it is removed.
6796 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
6797 LOOPS_HAVE_PREHEADERS.
6799 2014-02-25 Andrew Pinski <apinski@cavium.com>
6801 * builtins.c (expand_builtin_thread_pointer): Create a new target
6802 when the target is NULL.
6804 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
6806 PR rtl-optimization/60317
6807 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
6808 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
6809 * lra-assigns.c: Include params.h.
6810 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
6811 other reload pseudos considerations.
6813 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6815 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
6816 to use canonical form for nor<mode>3.
6818 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6821 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
6824 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
6826 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
6827 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
6828 (ix86_handle_option): Handle OPT_mprefetchwt1.
6829 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
6830 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
6832 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
6833 OPTION_MASK_ISA_PREFETCHWT1.
6834 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
6835 (PTA_PREFETCHWT1): New.
6836 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
6837 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
6838 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
6839 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
6840 (*prefetch_avx512pf_<mode>_: Change into ...
6841 (*prefetch_prefetchwt1_<mode>: This.
6842 * config/i386/i386.opt (mprefetchwt1): New.
6843 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
6844 (_mm_prefetch): Handle intent to write.
6845 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
6847 2014-02-25 Richard Biener <rguenther@suse.de>
6850 * emit-rtl.c (mem_attrs_htab): Remove.
6851 (mem_attrs_htab_hash): Likewise.
6852 (mem_attrs_htab_eq): Likewise.
6853 (set_mem_attrs): Always allocate new mem-attrs when something changed.
6854 (init_emit_once): Do not allocate mem_attrs_htab.
6856 2014-02-25 Richard Biener <rguenther@suse.de>
6859 * lto-opts.c (lto_write_options): Output non-explicit conservative
6860 -fwrapv, -fno-trapv and -fno-strict-overflow.
6861 * lto-wrapper.c (merge_and_complain): Handle merging those options.
6862 (run_gcc): And pass them through.
6864 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
6866 * sel-sched.c (calculate_new_fences): New parameter ptime.
6867 Calculate it as a maximum over all fence cycles.
6868 (sel_sched_region_2): Adjust the call to calculate_new_fences.
6869 Print the final schedule timing when sched_verbose.
6871 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
6873 PR rtl-optimization/60292
6874 * sel-sched.c (fill_vec_av_set): Do not reset target availability
6875 bit fot the fence instruction.
6877 2014-02-24 Alangi Derick <alangiderick@gmail.com>
6879 * calls.h: Fix typo in comment.
6881 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
6883 * config/pa/pa.c (pa_output_move_double): Don't valididate when
6884 adjusting offsetable addresses.
6886 2014-02-24 Guozhi Wei <carrot@google.com>
6888 * sparseset.h (sparseset_pop): Fix the wrong index.
6890 2014-02-24 Walter Lee <walt@tilera.com>
6892 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
6893 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
6895 * common/config/tilegx/tilegx-common.c
6896 (TARGET_DEFAULT_TARGET_FLAGS): Define.
6897 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
6899 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
6900 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
6901 (tilegx_gimplify_va_arg_expr): Handle big endian.
6902 (tilegx_expand_unaligned_load): Ditto.
6903 (tilegx_expand_unaligned_store): Ditto.
6904 (TARGET_RETURN_IN_MSB): New.
6905 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
6906 (TARGET_ENDIAN_DEFAULT): New.
6907 (TARGET_BIG_ENDIAN): Handle big endian.
6908 (BYTES_BIG_ENDIAN): Ditto.
6909 (WORDS_BIG_ENDIAN): Ditto.
6910 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
6913 * config/tilegx/tilegx.md (extv): Handle big endian.
6915 (insn_st<n>): Ditto.
6916 (insn_st<n>_add<bitsuffix>): Ditto.
6917 (insn_stnt<n>): Ditto.
6918 (insn_stnt<n>_add<bitsuffix>):Ditto.
6919 (vec_interleave_highv8qi): Handle big endian.
6920 (vec_interleave_highv8qi_be): New.
6921 (vec_interleave_highv8qi_le): New.
6922 (insn_v1int_h): Handle big endian.
6923 (vec_interleave_lowv8qi): Handle big endian.
6924 (vec_interleave_lowv8qi_be): New.
6925 (vec_interleave_lowv8qi_le): New.
6926 (insn_v1int_l): Handle big endian.
6927 (vec_interleave_highv4hi): Handle big endian.
6928 (vec_interleave_highv4hi_be): New.
6929 (vec_interleave_highv4hi_le): New.
6930 (insn_v2int_h): Handle big endian.
6931 (vec_interleave_lowv4hi): Handle big endian.
6932 (vec_interleave_lowv4hi_be): New.
6933 (vec_interleave_lowv4hi_le): New.
6934 (insn_v2int_l): Handle big endian.
6935 (vec_interleave_highv2si): Handle big endian.
6936 (vec_interleave_highv2si_be): New.
6937 (vec_interleave_highv2si_le): New.
6938 (insn_v4int_h): Handle big endian.
6939 (vec_interleave_lowv2si): Handle big endian.
6940 (vec_interleave_lowv2si_be): New.
6941 (vec_interleave_lowv2si_le): New.
6942 (insn_v4int_l): Handle big endian.
6943 * config/tilegx/tilegx.opt (mbig-endian): New option.
6944 (mlittle-endian): New option.
6945 * doc/install.texi: Document tilegxbe-linux.
6946 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
6948 2014-02-24 Martin Jambor <mjambor@suse.cz>
6951 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
6952 there are no parameter descriptors.
6954 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
6956 PR rtl-optimization/60268
6957 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
6958 initialization to ...
6959 (sched_rgn_init): ... here.
6960 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
6962 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
6964 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
6967 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
6969 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
6972 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
6974 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
6975 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
6977 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
6979 * config/microblaze/predicates.md: Add cmp_op predicate.
6980 * config/microblaze/microblaze.md: Add branch_compare instruction
6981 which uses cmp_op predicate and emits cmp insn before branch.
6982 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
6983 to microblaze_expand_conditional_branch and consolidate logic.
6984 (microblaze_expand_conditional_branch): emit branch_compare
6985 insn instead of handling cmp op separate from branch insn.
6987 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6989 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
6992 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6994 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
6995 define_insn with define_expand and new define_insn
6996 *altivec_lve<VI_char>x_internal.
6997 (altivec_stve<VI_char>x): Replace define_insn with define_expand
6998 and new define_insn *altivec_stve<VI_char>x_internal.
6999 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
7001 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
7003 (altivec_expand_stvex_be): New function.
7005 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
7007 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
7008 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
7009 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
7010 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
7012 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
7015 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
7016 instead of emit_move_insn.
7018 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7020 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
7022 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
7023 gen_altivec_vsumsws.
7025 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7027 * config/rs6000/altivec.md (altivec_lvxl): Rename as
7028 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
7029 (altivec_lvxl_<mode>): New define_expand incorporating
7030 -maltivec=be semantics where needed.
7031 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
7032 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
7033 semantics where needed.
7034 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
7035 (altivec_stvx_<mode>): New define_expand incorporating
7036 -maltivec=be semantics where needed.
7037 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
7038 VM2 iterator instead of V4SI.
7039 (altivec_stvxl_<mode>): New define_expand incorporating
7040 -maltivec=be semantics where needed.
7041 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
7042 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
7043 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
7044 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
7045 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
7046 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
7047 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
7048 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
7049 ALTIVEC_BUILTIN_STVXL.
7050 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
7051 (altivec_expand_stvx_be): Likewise.
7052 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
7053 (altivec_expand_lvx_be): Likewise.
7054 (altivec_expand_stvx_be): Likewise.
7055 (altivec_expand_builtin): Add cases for
7056 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
7057 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
7058 (altivec_init_builtins): Add definitions for
7059 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
7060 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
7062 2014-02-21 Catherine Moore <clm@codesourcery.com>
7064 * doc/invoke.texi (mvirt, mno-virt): Document.
7065 * config/mips/mips.opt (mvirt): New option.
7066 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
7068 2014-02-21 Richard Biener <rguenther@suse.de>
7070 PR tree-optimization/60276
7071 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
7072 (STMT_VINFO_MIN_NEG_DIST): New macro.
7073 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
7074 STMT_VINFO_MIN_NEG_DIST.
7075 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
7076 made for negative dependence distances still hold.
7078 2014-02-21 Richard Biener <rguenther@suse.de>
7081 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
7082 DECL_INITIAL for globals not in the current function context.
7084 2014-02-21 Jakub Jelinek <jakub@redhat.com>
7086 PR tree-optimization/56490
7087 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
7088 * tree-ssa-uninit.c: Include params.h.
7089 (compute_control_dep_chain): Add num_calls argument, return false
7090 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
7091 num_calls to recursive call.
7092 (find_predicates): Change dep_chain into normal array,
7093 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
7094 variable and adjust compute_control_dep_chain caller.
7095 (find_def_preds): Likewise.
7097 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
7099 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
7100 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
7102 2014-02-21 Nick Clifton <nickc@redhat.com>
7104 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
7105 (pushhi1): Likewise.
7106 (popqi1): Add mode to pre_dec.
7109 2014-02-21 Jakub Jelinek <jakub@redhat.com>
7111 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
7112 mode for mask of V8SFmode permutation.
7114 2014-02-20 Richard Henderson <rth@redhat.com>
7117 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
7118 a new pseudo for OLDVAL.
7120 2014-02-20 Jakub Jelinek <jakub@redhat.com>
7123 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
7124 gen_reg_rtx if d->testing_p.
7125 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
7126 if d->testing_p and we will certainly return true.
7127 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
7130 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
7132 * emit-rtl.c (gen_reg_rtx): Assert that
7133 crtl->emit.regno_pointer_align_length is non-zero.
7135 2014-02-20 Richard Henderson <rth@redhat.com>
7138 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
7139 on failure the store back into EXPECT.
7141 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
7142 Sandra Loosemore <sandra@codesourcery.com>
7144 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
7145 * config/nios2/nios2.c (nios2_function_profiler): Add
7146 -fPIC (flag_pic == 2) support.
7147 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
7148 (nios2_large_offset_p): New function.
7149 (nios2_unspec_reloc_p): Move up position, update to use
7150 nios2_large_offset_p.
7151 (nios2_unspec_address): Remove function.
7152 (nios2_unspec_offset): New function.
7153 (nios2_large_got_address): New function.
7154 (nios2_got_address): Add large offset support.
7155 (nios2_legitimize_tls_address): Update usage of removed and new
7157 (nios2_symbol_binds_local_p): New function.
7158 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
7159 (nios2_legitimize_address): Update to use nios2_large_offset_p.
7160 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
7161 (nios2_print_operand): Merge H/L processing, add hiadj/lo
7162 processing for (const (unspec ...)).
7163 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
7165 2014-02-20 Richard Biener <rguenther@suse.de>
7167 * tree-cfg.c (replace_uses_by): Mark altered BBs before
7168 doing the substitution.
7169 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
7171 2014-02-20 Martin Jambor <mjambor@suse.cz>
7174 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
7175 info when checking whether lattices are bottom.
7177 2014-02-20 Richard Biener <rguenther@suse.de>
7180 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
7183 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
7186 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
7187 parameter specifying the scaling.
7188 (inline_call): Update.
7189 (want_inline_recursively): Guard division by zero.
7190 (recursive_inlining): Update.
7191 * ipa-inline.h (clone_inlined_nodes): Update.
7193 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
7196 * config/i386/i386.c (classify_argument): Pass structures of size
7197 64 bytes or less in register.
7199 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
7200 Kirill Yukhin <kirill.yukhin@intel.com>
7202 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
7203 (_mm_rcp28_round_ss): Ditto.
7204 (_mm_rsqrt28_round_sd): Ditto.
7205 (_mm_rsqrt28_round_ss): Ditto.
7206 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
7207 (_mm_rcp14_round_ss): Ditto.
7208 (_mm_rsqrt14_round_sd): Ditto.
7209 (_mm_rsqrt14_round_ss): Ditto.
7210 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
7211 the first input operand, get rid of match_dup.
7212 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
7214 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
7216 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
7217 operand as the first input operand, set type attribute.
7218 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
7220 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
7221 operand as the first input operand, set type attribute.
7223 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7225 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
7228 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
7231 * config/i386/i386.c (construct_container): Remove TFmode check
7232 for X86_64_INTEGER_CLASS.
7234 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
7237 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
7238 only when -Wpsabi is enabled.
7240 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
7243 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
7244 passing arrays in registers are the same as for structs, so remove the
7245 special case for them.
7247 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
7249 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
7250 destination type, extract only the valid bits if the source type is not
7251 integral and has a different mode.
7253 2014-02-19 Richard Biener <rguenther@suse.de>
7256 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
7259 2014-02-19 Richard Biener <rguenther@suse.de>
7262 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
7263 (ipa_modify_call_arguments): Emit an argument load explicitely and
7264 preserve virtual SSA form there and for the replacement call.
7265 Do not update SSA form nor free dominance info.
7267 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
7269 * ipa.c (function_and_variable_visibility): Also clear WEAK
7270 flag when disolving COMDAT_GROUP.
7272 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
7274 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
7275 * ipa-prop.c (ipa_set_jf_known_type): Return early when
7277 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
7278 do more sanity checks.
7279 (detect_type_change): Return true when giving up early.
7280 (compute_complex_assign_jump_func): Fix type parameter of
7281 ipa_set_ancestor_jf.
7282 (compute_complex_ancestor_jump_func): Likewise.
7283 (update_jump_functions_after_inlining): Fix updating of
7285 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
7287 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
7289 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
7290 inline clones when edge disappears.
7292 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
7295 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
7296 Split 64-bit moves into 2 patterns. Do not allow the use of
7297 direct move for TDmode in little endian, since the decimal value
7298 has little endian bytes within a word, but the 64-bit pieces are
7299 ordered in a big endian fashion, and normal subreg's of TDmode are
7301 (mov<mode>_64bit_dm): Likewise.
7302 (movtd_64bit_nodm): Likewise.
7304 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
7306 PR tree-optimization/60174
7307 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
7308 statement of an SSA_NAME that occurs in an abnormal PHI node.
7310 2014-02-18 Jakub Jelinek <jakub@redhat.com>
7313 * final.c (SEEN_BB): Remove.
7314 (SEEN_NOTE, SEEN_EMITTED): Renumber.
7315 (final_scan_insn): Don't force_source_line on second
7316 NOTE_INSN_BASIC_BLOCK.
7318 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
7321 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
7322 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
7323 (type_natural_mode): Warn ABI change when %zmm register is not
7324 available for AVX512F vector value passing.
7326 2014-02-18 Kai Tietz <ktietz@redhat.com>
7329 * config/i386/i386.c (ix86_expand_prologue): Use value in
7330 rax register as displacement when restoring %r10 or %rax.
7331 Fix wrong offset when restoring both registers.
7333 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
7335 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
7336 assertion with conditional return.
7338 2014-02-18 Jakub Jelinek <jakub@redhat.com>
7339 Uros Bizjak <ubizjak@gmail.com>
7342 * config/i386/driver-i386.c (host_detect_local_cpu): If
7343 YMM state is not saved by the OS, also clear has_f16c. Move
7344 CPUID 0x80000001 handling before YMM state saving checking.
7346 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
7348 PR rtl-optimization/58960
7349 * haifa-sched.c (alloc_global_sched_pressure_data): New,
7350 factored out from ...
7351 (sched_init): ... here.
7352 (free_global_sched_pressure_data): New, factored out from ...
7353 (sched_finish): ... here.
7354 * sched-int.h (free_global_sched_pressure_data): Declare.
7355 * sched-rgn.c (nr_regions_initial): New static global.
7356 (haifa_find_rgns): Initialize it.
7357 (schedule_region): Disable sched-pressure for the newly
7360 2014-02-17 Richard Biener <rguenther@suse.de>
7362 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
7363 release SSA defs of pattern stmts.
7365 2014-02-17 Richard Biener <rguenther@suse.de>
7367 * tree-inline.c (expand_call_inline): Release the virtual
7368 operand defined by the call we are about to inline.
7370 2014-02-17 Richard Biener <rguenther@suse.de>
7372 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
7374 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
7375 Ilya Tocar <ilya.tocar@intel.com>
7377 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
7378 arguments order in builtin.
7379 (_mm512_permutexvar_epi64): Ditto.
7380 (_mm512_mask_permutexvar_epi64): Ditto
7381 (_mm512_maskz_permutexvar_epi32): Ditto
7382 (_mm512_permutexvar_epi32): Ditto
7383 (_mm512_mask_permutexvar_epi32): Ditto
7385 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7387 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
7388 (p8_vmrgow): Likewise.
7390 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7392 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
7395 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
7398 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
7399 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
7400 into 64-bit and 32-bit moves. On 64-bit moves, add support for
7401 using direct move instructions on ISA 2.07. Also adjust
7402 instruction length for 64-bit.
7403 (mov<mode>_64bit, TFmode/TDmode): Likewise.
7404 (mov<mode>_32bit, TFmode/TDmode): Likewise.
7406 2014-02-15 Alan Modra <amodra@gmail.com>
7410 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
7411 find_replacement on parts of insn rtl that might be reloaded.
7413 2014-02-15 Richard Biener <rguenther@suse.de>
7415 PR tree-optimization/60183
7416 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
7417 (tree_ssa_phiprop): Calculate and free post-dominators.
7419 2014-02-14 Jeff Law <law@redhat.com>
7421 PR rtl-optimization/60131
7422 * ree.c (get_extended_src_reg): New function.
7423 (combine_reaching_defs): Use it rather than assuming location of REG.
7424 (find_and_remove_re): Verify first operand of extension is
7425 a REG before adding the insns to the copy list.
7427 2014-02-14 Roland McGrath <mcgrathr@google.com>
7429 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
7430 * configure: Regenerated.
7431 * config.in: Regenerated.
7432 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
7433 instead of ASM_SHORT.
7435 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
7436 Richard Earnshaw <rearnsha@arm.com>
7438 PR rtl-optimization/59535
7439 * lra-constraints.c (process_alt_operands): Encourage alternative
7440 when unassigned pseudo class is superset of the alternative class.
7441 (inherit_reload_reg): Don't inherit when optimizing for code size.
7442 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
7443 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
7444 modes not less than 4 for Thumb1.
7446 2014-02-14 Kyle McMartin <kyle@redhat.com>
7449 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
7451 2014-02-14 Richard Biener <rguenther@suse.de>
7453 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
7454 (get_frame_arg): Drop the assert with langhook types_compatible_p.
7455 Do not strip INDIRECT_REFs.
7457 2014-02-14 Richard Biener <rguenther@suse.de>
7460 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
7461 DECL_FUNCTION_SPECIFIC_TARGET.
7462 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
7463 * tree-streamer-out.c (pack_ts_target_option): Remove.
7464 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
7465 (write_ts_function_decl_tree_pointers): Do not stream
7466 DECL_FUNCTION_SPECIFIC_TARGET.
7467 * tree-streamer-in.c (unpack_ts_target_option): Remove.
7468 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
7469 (lto_input_ts_function_decl_tree_pointers): Do not stream
7470 DECL_FUNCTION_SPECIFIC_TARGET.
7472 2014-02-14 Jakub Jelinek <jakub@redhat.com>
7474 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
7475 (get_initial_def_for_induction, vectorizable_induction): Ignore
7476 debug stmts when looking for exit_phi.
7477 (vectorizable_live_operation): Fix up condition.
7479 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
7481 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
7482 nreverse() because it changes the content of original tree list.
7484 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
7486 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
7487 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
7489 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
7491 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
7492 GNU coding standards.
7494 2014-02-13 Jakub Jelinek <jakub@redhat.com>
7497 * dwarf2out.c (gen_subprogram_die): Don't call
7498 add_calling_convention_attribute if subr_die is old_die.
7500 2014-02-13 Sharad Singhai <singhai@google.com>
7502 * doc/optinfo.texi: Fix order of nodes.
7504 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
7506 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
7507 operands[2], not operands[3].
7509 2014-02-13 Richard Biener <rguenther@suse.de>
7512 * doc/install.texi (ISL): Update recommended version to 0.12.2,
7513 mention the possibility of an in-tree build.
7514 (CLooG): Update recommended version to 0.18.1, mention the
7515 possibility of an in-tree build and clarify that the ISL
7516 bundled with CLooG does not work.
7518 2014-02-13 Jakub Jelinek <jakub@redhat.com>
7521 * expr.c (compress_float_constant): If x is a hard register,
7522 extend into a pseudo and then move to x.
7524 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
7526 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
7527 caused by bad second argument to warning_at() with -mhotpatch and
7528 nested functions (e.g. with gfortran).
7530 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
7532 * opts.c (option_name): Remove "enabled by default" rider.
7534 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
7536 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
7538 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
7539 Uros Bizjak <ubizjak@gmail.com>
7542 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
7543 * configure: Regenerated.
7545 2014-02-12 Richard Biener <rguenther@suse.de>
7547 * vec.c (vec_prefix::calculate_allocation): Move as
7548 inline variant to vec.h.
7549 (vec_prefix::calculate_allocation_1): New out-of-line version.
7550 * vec.h (vec_prefix::calculate_allocation_1): Declare.
7551 (vec_prefix::m_has_auto_buf): Rename to ...
7552 (vec_prefix::m_using_auto_storage): ... this.
7553 (vec_prefix::calculate_allocation): Inline the easy cases
7554 and dispatch to calculate_allocation_1 which doesn't need the
7556 (va_heap::reserve): Use gcc_checking_assert.
7557 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
7558 m_using_auto_storage.
7559 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
7561 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
7562 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
7563 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
7565 2014-02-12 Richard Biener <rguenther@suse.de>
7567 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
7568 when we found a dependence.
7570 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
7572 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
7574 (maybe_fold_stmt): ... into this new function.
7575 * omp-low.c (lower_omp): Update comment.
7577 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
7580 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
7583 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
7585 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
7586 identifiers in comments.
7587 (cortexa53_extra_costs): Likewise.
7588 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
7589 (cortexa7_extra_costs): Likewise.
7590 (cortexa12_extra_costs): Likewise.
7591 (cortexa15_extra_costs): Likewise.
7592 (v7m_extra_costs): Likewise.
7594 2014-02-12 Richard Biener <rguenther@suse.de>
7597 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
7598 of posix_memalign being successful.
7599 (lower_stmt): Restrict lowering of posix_memalign to when
7600 -ftree-bit-ccp is enabled.
7602 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
7604 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
7606 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
7608 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
7610 PR rtl-optimization/60116
7611 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
7612 other_insn once the combination has been validated.
7614 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
7617 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
7619 * ipa-devirt.c: Include demangle.h
7620 (odr_violation_reported): New static variable.
7621 (add_type_duplicate): Update odr_violations.
7622 (maybe_record_node): Add completep parameter; update it.
7623 (record_target_from_binfo): Add COMPLETEP parameter;
7624 update it as needed.
7625 (possible_polymorphic_call_targets_1): Likewise.
7626 (struct polymorphic_call_target_d): Add nonconstruction_targets;
7627 rename FINAL to COMPLETE.
7628 (record_targets_from_bases): Sanity check we found the binfo;
7629 fix COMPLETEP updating.
7630 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
7631 parameter, fix computing of COMPLETEP.
7632 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
7633 at LTO time do demangling.
7634 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
7635 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
7637 (gimple_get_virt_method_for_binfo): Likewise.
7638 * gimple-fold.h (gimple_get_virt_method_for_binfo,
7639 gimple_get_virt_method_for_vtable): Update prototypes.
7641 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
7644 * genautomata.c (add_presence_absence): Fix typo with
7645 {final_}presence_list.
7647 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
7650 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
7651 for VSX/Altivec vectors that land in GPR registers.
7653 2014-02-11 Richard Henderson <rth@redhat.com>
7654 Jakub Jelinek <jakub@redhat.com>
7657 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
7658 around drhs if type conversion to lacc->type is not useless.
7660 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7662 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
7664 (cortex-a57.cortex-a53): Likewise.
7665 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
7667 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7669 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
7672 2014-02-11 Renlin Li <Renlin.Li@arm.com>
7674 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
7675 add_options_for_arm_vfp3.
7677 2014-02-11 Jeff Law <law@redhat.com>
7680 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
7681 object with an undesirable mode.
7683 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7686 * config/i386/sol2-9.h: New file.
7687 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
7688 *-*-solaris2.9*): Use it.
7690 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
7692 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
7693 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
7695 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
7697 * config/microblaze/microblaze.c: Extend mcpu version format
7699 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
7701 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
7703 2014-02-10 Richard Henderson <rth@redhat.com>
7706 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
7707 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
7708 ms-abi vs -mno-accumulate-outgoing-args.
7709 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
7710 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
7713 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
7716 * cfgexpand.c (expand_asm_operands): Attach source location to
7717 ASM_INPUT rtx objects.
7718 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
7720 2014-02-10 Nick Clifton <nickc@redhat.com>
7722 * config/mn10300/mn10300.c (popcount): New function.
7723 (mn10300_expand_prologue): Include saved registers in stack usage
7726 2014-02-10 Jeff Law <law@redhat.com>
7729 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
7730 when changing the SET_DEST of a prior insn to avoid an input reload.
7732 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
7734 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
7735 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
7736 -mcall-openbsd, or -mcall-linux.
7737 (CC1_ENDIAN_BIG_SPEC): Remove.
7738 (CC1_ENDIAN_LITTLE_SPEC): Remove.
7739 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
7740 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
7741 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
7742 and %cc1_endian_default.
7743 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
7745 2014-02-10 Richard Biener <rguenther@suse.de>
7747 PR tree-optimization/60115
7748 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
7749 MEM_REF handling. Properly verify that the accesses are not
7750 out of the objects bound.
7752 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7754 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
7757 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
7759 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
7760 proper constants and fix formatting.
7761 (possible_polymorphic_call_targets): Fix formatting.
7763 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
7764 Ilya Tocar <ilya.tocar@intel.com>
7766 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
7767 (_mm512_loadu_epi32): Renamed into...
7768 (_mm512_loadu_si512): This.
7769 (_mm512_storeu_epi32): Renamed into...
7770 (_mm512_storeu_si512): This.
7771 (_mm512_maskz_ceil_ps): Removed.
7772 (_mm512_maskz_ceil_pd): Ditto.
7773 (_mm512_maskz_floor_ps): Ditto.
7774 (_mm512_maskz_floor_pd): Ditto.
7775 (_mm512_floor_round_ps): Ditto.
7776 (_mm512_floor_round_pd): Ditto.
7777 (_mm512_ceil_round_ps): Ditto.
7778 (_mm512_ceil_round_pd): Ditto.
7779 (_mm512_mask_floor_round_ps): Ditto.
7780 (_mm512_mask_floor_round_pd): Ditto.
7781 (_mm512_mask_ceil_round_ps): Ditto.
7782 (_mm512_mask_ceil_round_pd): Ditto.
7783 (_mm512_maskz_floor_round_ps): Ditto.
7784 (_mm512_maskz_floor_round_pd): Ditto.
7785 (_mm512_maskz_ceil_round_ps): Ditto.
7786 (_mm512_maskz_ceil_round_pd): Ditto.
7787 (_mm512_expand_pd): Ditto.
7788 (_mm512_expand_ps): Ditto.
7789 * config/i386/i386.c (ix86_builtins): Remove
7790 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
7791 (bdesc_args): Ditto.
7792 * config/i386/predicates.md (const1256_operand): New.
7793 (const_1_to_2_operand): Ditto.
7794 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
7795 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
7796 (*avx512pf_gatherpf<mode>sf): Ditto.
7797 (avx512pf_gatherpf<mode>df): Ditto.
7798 (*avx512pf_gatherpf<mode>df_mask): Ditto.
7799 (*avx512pf_gatherpf<mode>df): Ditto.
7800 (avx512pf_scatterpf<mode>sf): Ditto.
7801 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
7802 (*avx512pf_scatterpf<mode>sf): Ditto.
7803 (avx512pf_scatterpf<mode>df): Ditto.
7804 (*avx512pf_scatterpf<mode>df_mask): Ditto.
7805 (*avx512pf_scatterpf<mode>df): Ditto.
7806 (avx512f_expand<mode>): Removed.
7807 (<shift_insn><mode>3<mask_name>): Change predicate type.
7809 2014-02-08 Jakub Jelinek <jakub@redhat.com>
7811 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
7812 not at the end of datarefs vector use ordered_remove to avoid
7813 reordering datarefs vector.
7816 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
7817 mark local addressable non-static vars as GOVD_PRIVATE
7818 instead of GOVD_LOCAL.
7819 * omp-low.c (lower_omp_for): Move gimple_bind_vars
7820 and BLOCK_VARS of gimple_bind_block to new_stmt rather
7824 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
7825 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
7826 assume_aligned or alloc_align attributes.
7827 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
7828 arguments. Handle also assume_aligned and alloc_align attributes.
7829 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
7830 calls to functions with assume_aligned or alloc_align attributes.
7831 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
7833 2014-02-08 Terry Guo <terry.guo@arm.com>
7835 * doc/invoke.texi: Document ARM -march=armv7e-m.
7837 2014-02-08 Jakub Jelinek <jakub@redhat.com>
7839 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
7840 flag on __cilkrts_rethrow builtin.
7843 * ipa-cp.c (determine_versionability): Fail at -O0
7844 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
7845 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
7848 2014-02-04 Jakub Jelinek <jakub@redhat.com>
7851 * tree-inline.c (copy_forbidden): Fail for
7852 __attribute__((optimize (0))) functions.
7854 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
7856 * varpool.c: Include pointer-set.h.
7857 (varpool_remove_unreferenced_decls): Variables in other partitions
7858 will not be output; be however careful to not lose information
7861 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
7863 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
7864 lookup in the vtable constructor.
7866 2014-02-07 Jeff Law <law@redhat.com>
7869 * config/m68k/m68k.md (ashldi_extsi): Turn into a
7870 define_insn_and_split.
7872 * ipa-inline.c (inline_small_functions): Fix typos.
7874 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7876 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
7877 (s390_can_use_return_insn): Declare.
7878 * config/s390/s390.h (EPILOGUE_USES): Define.
7879 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
7881 (s390_chunkify_start): Handle return JUMP_LABELs.
7882 (s390_early_mach): Emit a main_pool instruction on the entry edge.
7883 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
7884 (s390_can_use_return_insn): New functions.
7885 (s390_fix_long_loop_prediction): Handle conditional returns.
7886 (TARGET_SET_UP_BY_PROLOGUE): Define.
7887 * config/s390/s390.md (ANY_RETURN): New code iterator.
7888 (*creturn, *csimple_return, return, simple_return): New patterns.
7890 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7892 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
7893 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
7894 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
7895 REG_CFA_RESTORE list when deciding not to restore a register.
7897 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7899 * config/s390/s390.c: Include tree-pass.h and context.h.
7900 (s390_early_mach): New function, split out from...
7901 (s390_emit_prologue): ...here.
7902 (pass_data_s390_early_mach): New pass structure.
7903 (pass_s390_early_mach): New class.
7904 (s390_option_override): Create and register early_mach pass.
7905 Move to end of file.
7907 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7909 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
7910 to match for the exit block.
7912 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7914 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
7915 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
7916 Reject misaligned operands.
7918 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7920 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
7922 2014-02-07 Richard Biener <rguenther@suse.de>
7925 * gimple-low.c (lower_builtin_posix_memalign): New function.
7926 (lower_stmt): Call it to lower posix_memalign in a way
7927 to make alignment info accessible.
7929 2014-02-07 Jakub Jelinek <jakub@redhat.com>
7932 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
7933 __builtin_setjmp_receiver.
7935 2014-02-07 Richard Biener <rguenther@suse.de>
7938 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
7939 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
7940 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
7941 Handle BUILT_IN_POSIX_MEMALIGN.
7942 (find_func_clobbers): Likewise.
7943 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
7944 (call_may_clobber_ref_p_1): Likewise.
7946 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7949 * ipa-devirt.c (record_target_from_binfo): Remove overactive
7952 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7955 * lto-cgraph.c (lto_output_node): Use
7956 symtab_get_symbol_partitioning_class.
7957 (lto_output_varpool_node): likewise.
7958 (symtab_get_symbol_partitioning_class): Move here from
7960 * cgraph.h (symbol_partitioning_class): Likewise.
7961 (symtab_get_symbol_partitioning_class): Declare.
7963 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7965 * ggc.h (ggc_internal_cleared_alloc): New macro.
7966 * vec.h (vec_safe_copy): Handle memory stats.
7967 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
7968 * target-globals.c (save_target_globals): Likewise.
7970 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7973 * expr.c (emit_move_resolve_push): Export; be bit more selective
7974 on when to clear alias set.
7975 * expr.h (emit_move_resolve_push): Declare.
7976 * function.h (struct function): Add tail_call_marked.
7977 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
7978 * config/i386/i386-protos.h (ix86_expand_push): Remove.
7979 * config/i386/i386.md (TImode move expander): De not call
7981 (FP push expanders): Preserve memory attributes.
7982 * config/i386/sse.md (push<mode>1): Remove.
7983 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
7984 (ix86_expand_push): Remove.
7985 * config/i386/mmx.md (push<mode>1): Remove.
7987 2014-02-06 Jakub Jelinek <jakub@redhat.com>
7989 PR rtl-optimization/60030
7990 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
7991 lopart with paradoxical subreg before shifting it up by hprec.
7993 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7995 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
7996 Remove extra newline at end of file.
7997 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
7998 (arm_issue_rate): Handle cortexa57.
7999 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
8000 (cortex-a57.cortex-a53): Likewise.
8002 2014-02-06 Jakub Jelinek <jakub@redhat.com>
8005 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
8006 don't record in REG_FRAME_RELATED_EXPR registers not set in that
8008 (arm_expand_prologue): Adjust all callers.
8009 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
8010 info, registers also at the lowest numbered registers side. Use
8011 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
8015 * var-tracking.c (adjust_mems): Before adding a SET to
8016 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
8018 2014-02-06 Alan Modra <amodra@gmail.com>
8021 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
8022 change SDmode to DDmode when lra_in_progress.
8024 2014-02-06 Jakub Jelinek <jakub@redhat.com>
8027 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
8028 free_data_ref on the dr first, and before goto again also set dr
8029 to the next dr. For simd_lane_access, free old datarefs[i] before
8030 overwriting it. For get_vectype_for_scalar_type failure, don't
8031 free_data_ref if simd_lane_access.
8033 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
8036 * tree.h (opts_for_fn): New inline function.
8037 (opt_for_fn): Define.
8038 * config/i386/i386.c (ix86_function_regparm): Use
8039 opt_for_fn (decl, optimize) instead of optimize.
8041 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
8043 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
8044 for SYMBOL_REF in large memory model.
8046 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8048 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
8050 (cortex-a57): Likewise.
8051 (cortex-a57.cortex-a53): Likewise.
8053 2014-02-06 Yury Gribov <y.gribov@samsung.com>
8054 Kugan Vivekanandarajah <kuganv@linaro.org>
8056 * config/arm/arm.c (arm_vector_alignment_reachable): Check
8058 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
8060 2014-02-06 Richard Biener <rguenther@suse.de>
8062 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
8063 set_loop_copy and initialize_original_copy_tables.
8065 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
8067 * config/aarch64/aarch64-simd.md
8068 (aarch64_ashr_simddi): Change QI to SI.
8070 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
8071 Jakub Jelinek <jakub@redhat.com>
8074 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
8077 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8079 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
8080 CODE_FOR_altivec_vpku[hw]um to
8081 CODE_FOR_altivec_vpku[hw]um_direct.
8082 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
8083 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
8084 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
8085 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
8087 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8089 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
8090 generation for -maltivec=be.
8091 (altivec_vsumsws): Simplify redundant test.
8093 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8095 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
8096 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
8097 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
8098 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
8099 gen_altivec_vpkuwum.
8100 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
8102 (altivec_vpks<VI_char>ss): Likewise.
8103 (altivec_vpks<VI_char>us): Likewise.
8104 (altivec_vpku<VI_char>us): Likewise.
8105 (altivec_vpku<VI_char>um): Likewise.
8106 (altivec_vpku<VI_char>um_direct): New (copy of
8107 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
8109 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
8110 target is little endian and -maltivec=be is not specified.
8111 (*altivec_vupkhs<VU_char>_direct): New (copy of
8112 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
8113 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
8114 target is little endian and -maltivec=be is not specified.
8115 (*altivec_vupkls<VU_char>_direct): New (copy of
8116 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
8117 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
8118 little endian and -maltivec=be is not specified.
8119 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
8120 little endian and -maltivec=be is not specified.
8122 2014-02-05 Richard Henderson <rth@redhat.com>
8125 * combine-stack-adj.c: Revert r206943.
8126 * sched-int.h (struct deps_desc): Add last_args_size.
8127 * sched-deps.c (init_deps): Initialize it.
8128 (sched_analyze_insn): Add OUTPUT dependencies between insns that
8129 contain REG_ARGS_SIZE notes.
8131 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
8133 * lto-cgraph.c (asm_nodes_output): Make global.
8134 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
8135 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
8136 (driver_handle_option): Handle OPT_fwpa.
8138 2014-02-05 Jakub Jelinek <jakub@redhat.com>
8141 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
8142 a comment typo and formatting issue. If odr_hash hasn't been
8143 created, return vNULL and set *completep to false.
8146 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
8147 bb with no successors.
8149 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
8152 * doc/invoke.texi (-march): Clarify documentation for ARM.
8156 2014-02-05 Richard Biener <rguenther@suse.de>
8158 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
8159 when not vectorizing because of too many alias checks.
8160 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
8161 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
8163 2014-02-05 Nick Clifton <nickc@redhat.com>
8165 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
8166 accept extended registers in any mode when compiling for the MN10300.
8168 2014-02-05 Yury Gribov <y.gribov@samsung.com>
8170 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
8171 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
8172 sanitization attributes.
8173 (can_inline_edge_p): Likewise.
8174 (sanitize_attrs_match_for_inline_p): New function.
8176 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
8178 * ipa-prop.c (detect_type_change): Shor circuit testing of
8179 type changes on THIS pointer.
8181 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
8184 * config/pa/pa.c (legitimize_tls_address): Return original address
8185 if not passed a SYMBOL_REF rtx.
8186 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
8188 (pa_emit_move_sequence): Simplify TLS source operands.
8189 (pa_legitimate_constant_p): Reject all TLS constants.
8190 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
8191 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
8193 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
8195 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
8196 groups when we know they are controlled by LTO.
8197 * varasm.c (default_binds_local_p_1): If object is in other partition,
8198 it will be resolved locally.
8200 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
8202 * config/host-linux.c (linux_gt_pch_use_address): Don't
8203 use SSIZE_MAX because it is not always defined.
8205 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
8208 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
8209 threshold for pseudo splitting.
8210 (update_ebb_live_info): Process call argument hard registers and
8211 hard registers from insn definition too.
8212 (max_small_class_regs_num): New constant.
8213 (inherit_in_ebb): Update live hard regs through EBBs. Update
8214 reloads_num only for small register classes. Don't split for
8217 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
8220 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
8223 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
8225 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
8228 2014-02-04 Marek Polacek <polacek@redhat.com>
8230 * gdbinit.in (pel): Define.
8232 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
8234 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
8237 2014-02-04 Richard Biener <rguenther@suse.de>
8240 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
8241 in function context local.
8242 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
8243 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
8244 similar to LTO_imported_decl_ref.
8246 2014-02-04 Jakub Jelinek <jakub@redhat.com>
8248 PR tree-optimization/60002
8249 * cgraphclones.c (build_function_decl_skip_args): Clear
8252 PR tree-optimization/60023
8253 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
8254 false to gsi_replace.
8255 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
8256 has been in some EH region and vec_stmt could throw, add
8257 vec_stmt into the same EH region.
8258 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
8259 has no lhs, ignore it.
8260 * internal-fn.c (expand_MASK_LOAD): Likewise.
8263 * tree-inline.c (copy_forbidden): Fail for
8264 __attribute__((optimize (0))) functions.
8267 * omp-low.c (simd_clone_struct_copy): If from->inbranch
8268 is set, copy one less argument.
8269 (expand_simd_clones): Don't subtract clone_info->inbranch
8270 from simd_clone_struct_alloc argument.
8272 PR rtl-optimization/57915
8273 * recog.c (simplify_while_replacing): If all unary/binary/relational
8274 operation arguments are constant, attempt to simplify those.
8277 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
8278 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
8280 2014-02-04 Richard Biener <rguenther@suse.de>
8282 PR tree-optimization/60012
8283 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
8284 TBAA disambiguation to all DDRs.
8286 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8289 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
8290 (LINK_SPEC): Use it for -shared, -shared-libgcc.
8292 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
8295 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
8297 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
8299 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
8300 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
8302 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
8305 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
8306 to figure out targets of polymorphic calls with known decl.
8307 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
8308 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
8309 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
8310 (get_polymorphic_call_info): ... here.
8311 (get_polymorphic_call_info_from_invariant): New function.
8313 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
8315 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
8316 lookup via vtable pointer; check for type consistency
8317 and turn inconsitent facts into UNREACHABLE.
8318 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
8319 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
8320 type inconsistent querries; return UNREACHABLE instead.
8322 2014-02-03 Richard Henderson <rth@twiddle.net>
8325 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
8326 already processed this node.
8327 (normalize_one_pred_1): Pass along mark_set.
8328 (normalize_one_pred): Create and destroy a pointer_set_t.
8329 (normalize_one_pred_chain): Likewise.
8331 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
8333 PR gcov-profile/58602
8334 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
8336 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
8339 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
8340 -fno-devirtualize; try to devirtualize by the knowledge of
8341 virtual table pointer given by aggregate propagation.
8342 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
8343 (ipa_print_node_jump_functions): Dump also offset that
8344 is relevant for polymorphic calls.
8345 (determine_known_aggregate_parts): Add arg_type parameter; use it
8346 instead of determining the type from pointer type.
8347 (ipa_compute_jump_functions_for_edge): Update call of
8348 determine_known_aggregate_parts.
8349 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
8350 (gimple_get_virt_method_for_binfo): ... here; simplify using
8351 vtable_pointer_value_to_vtable.
8352 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
8353 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
8354 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
8355 (vtable_pointer_value_to_vtable): Break out from ...; handle also
8357 (vtable_pointer_value_to_binfo): ... here.
8358 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
8360 2014-02-03 Teresa Johnson <tejohnson@google.com>
8362 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
8363 redef of outer loop index variable.
8365 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
8369 * doc/extend.texi (Function Attributes): Typo.
8371 2014-02-03 Cong Hou <congh@google.com>
8373 PR tree-optimization/60000
8374 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
8375 if the vectorized statement is a store. A store statement can only
8376 appear at the end of pattern statements.
8378 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
8380 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
8381 (ix86_option_override_internal): Default long double to 64-bit for
8382 32-bit Bionic and to 128-bit for 64-bit Bionic.
8384 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
8385 TARGET_LONG_DOUBLE_128 is true.
8386 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
8388 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
8389 (mlong-double-64): Negate -mlong-double-128.
8390 (mlong-double-128): New option.
8392 * config/i386/i386-c.c (ix86_target_macros): Define
8393 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
8395 * doc/invoke.texi: Document -mlong-double-128.
8397 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
8399 PR rtl-optimization/60024
8400 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
8402 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
8404 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
8406 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
8408 PR rtl-optimization/57662
8409 * sel-sched.c (code_motion_path_driver): Do not mark already not
8410 existing blocks in the visiting bitmap.
8412 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
8414 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
8415 on the insn being emitted.
8417 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
8418 Will Deacon <will.deacon@arm.com>
8420 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
8422 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8424 * config/arm/arm-tables.opt: Regenerate.
8426 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8428 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
8429 for vector types other than V16QImode.
8430 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
8431 define_expand, and call altivec_expand_vec_perm_le when producing
8432 code with little endian element order.
8433 (*altivec_vperm_<mode>_internal): New insn having previous
8434 behavior of altivec_vperm_<mode>.
8435 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
8436 altivec_expand_vec_perm_le when producing code with little endian
8438 (*altivec_vperm_<mode>_uns_internal): New insn having previous
8439 behavior of altivec_vperm_<mode>_uns.
8441 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8443 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
8444 (altivec_vsumsws): Add handling for -maltivec=be with a little
8446 (altivec_vsumsws_direct): New.
8447 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
8448 gen_altivec_vsumsws.
8450 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
8452 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
8453 vtable_pointer_value_to_binfo): New functions.
8454 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
8455 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
8457 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
8459 * config/nios2/nios2.md (load_got_register): Initialize GOT
8460 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
8461 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
8463 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
8465 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
8466 preserverd by passthrough, do not propagate the type.
8468 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
8470 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
8471 (mips_atomic_assign_expand_fenv): New function.
8472 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
8474 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
8476 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
8477 (__builtin_mips_set_fcsr): Likewise.
8478 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
8479 MIPS_USI_FTYPE_VOID.
8480 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
8481 (mips16_expand_set_fcsr): Likewise.
8482 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
8483 (mips16_set_fcsr_stub): Likewise.
8484 (mips16_get_fcsr_one_only_stub): New class.
8485 (mips16_set_fcsr_one_only_stub): Likewise.
8486 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
8487 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
8488 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
8489 (hard_float): New availability predicate.
8490 (mips_builtins): Add get_fcsr and set_fcsr.
8491 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
8492 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
8493 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
8494 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
8495 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
8498 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
8500 * config/mips/mips.c (mips_one_only_stub): New class.
8501 (mips_need_mips16_rdhwr_p): Replace with...
8502 (mips16_rdhwr_stub): ...this new variable.
8503 (mips16_stub_call_address): New function.
8504 (mips16_rdhwr_one_only_stub): New class.
8505 (mips_expand_thread_pointer): Use mips16_stub_call_address.
8506 (mips_output_mips16_rdhwr): Delete.
8507 (mips_finish_stub): New function.
8508 (mips_code_end): Use it to handle rdhwr stubs.
8510 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
8513 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
8514 when calculating size of integer atomic types.
8516 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
8518 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
8520 2014-02-01 Jakub Jelinek <jakub@redhat.com>
8522 PR tree-optimization/60003
8523 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
8524 * profile.c (branch_prob): Use gimple_call_builtin_p
8525 to check for BUILT_IN_SETJMP_RECEIVER.
8526 * tree-inline.c (copy_bb): Call notice_special_calls.
8528 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
8531 * lra-constraints.c (process_alt_operands): Update reload_sum only
8534 2014-01-31 Richard Henderson <rth@redhat.com>
8537 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
8538 until after else_eh is processed.
8540 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
8542 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
8543 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
8544 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
8545 in smmintrin.h, remove them.
8546 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
8547 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
8548 * config/i386/i386.md (ROUND_SAE): Fix value.
8549 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
8550 (const48_operand): New.
8551 * config/i386/subst.md (round), (round_expand): Use
8552 const_4_or_8_to_11_operand.
8553 (round_saeonly), (round_saeonly_expand): Use const48_operand.
8555 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
8557 * config/i386/constraints.md (Yk): Swap meaning with k.
8558 * config/i386/i386.md (movhi_internal): Change Yk to k.
8559 (movqi_internal): Ditto.
8560 (*k<logic><mode>): Ditto.
8563 (kandn<mode>): Ditto.
8564 (*<code>hi_1): Ditto.
8565 (*<code>qi_1): Ditto.
8566 (kxnor<mode>): Ditto.
8567 (kortestzhi): Ditto.
8568 (kortestchi): Ditto.
8570 (*one_cmplhi2_1): Ditto.
8571 (*one_cmplqi2_1): Ditto.
8572 * config/i386/sse.md (): Change k to Yk.
8573 (avx512f_load<mode>_mask): Ditto.
8574 (avx512f_blendm<mode>): Ditto.
8575 (avx512f_store<mode>_mask): Ditto.
8576 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
8577 (avx512f_storedqu<mode>_mask): Ditto.
8578 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
8580 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
8581 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
8582 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
8583 (avx512f_maskcmp<mode>3): Ditto.
8584 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
8585 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
8586 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
8587 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
8588 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
8589 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
8590 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
8591 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
8592 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
8593 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
8594 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
8595 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
8596 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
8597 (vec_extract_lo_<mode>_maskm): Ditto.
8598 (vec_extract_hi_<mode>_maskm): Ditto.
8599 (avx512f_vternlog<mode>_mask): Ditto.
8600 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
8601 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
8602 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
8603 (avx512f_<code>v8div16qi2_mask): Ditto.
8604 (avx512f_<code>v8div16qi2_mask_store): Ditto.
8605 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
8606 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
8607 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
8608 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
8609 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
8610 (*avx512pf_gatherpf<mode>df_mask): Ditto.
8611 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
8612 (*avx512pf_scatterpf<mode>df_mask): Ditto.
8613 (avx512cd_maskb_vec_dupv8di): Ditto.
8614 (avx512cd_maskw_vec_dupv16si): Ditto.
8615 (avx512f_vpermi2var<mode>3_maskz): Ditto.
8616 (avx512f_vpermi2var<mode>3_mask): Ditto.
8617 (avx512f_vpermi2var<mode>3_mask): Ditto.
8618 (avx512f_vpermt2var<mode>3_maskz): Ditto.
8619 (*avx512f_gathersi<mode>): Ditto.
8620 (*avx512f_gathersi<mode>_2): Ditto.
8621 (*avx512f_gatherdi<mode>): Ditto.
8622 (*avx512f_gatherdi<mode>_2): Ditto.
8623 (*avx512f_scattersi<mode>): Ditto.
8624 (*avx512f_scatterdi<mode>): Ditto.
8625 (avx512f_compress<mode>_mask): Ditto.
8626 (avx512f_compressstore<mode>_mask): Ditto.
8627 (avx512f_expand<mode>_mask): Ditto.
8628 * config/i386/subst.md (mask): Change k to Yk.
8629 (mask_scalar_merge): Ditto.
8632 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
8634 * doc/extend.texi (Vector Extensions): Document ?: in C++.
8636 2014-01-31 Richard Biener <rguenther@suse.de>
8639 * builtins.c (fold_builtin_memory_op): Make sure to not
8640 use a floating-point mode or a boolean or enumeral type for
8643 2014-01-30 DJ Delorie <dj@redhat.com>
8645 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
8646 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
8647 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
8648 whenever main() has an epilogue.
8650 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8652 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
8653 unused variable "field".
8654 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
8655 (vsx_mergeh_<mode>): Likewise.
8656 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
8657 (altivec_vmrghh): Likewise.
8658 (altivec_vmrghw): Likewise.
8659 (altivec_vmrglb): Likewise.
8660 (altivec_vmrglh): Likewise.
8661 (altivec_vmrglw): Likewise.
8662 (altivec_vspltb): Add missing uses.
8663 (altivec_vsplth): Likewise.
8664 (altivec_vspltw): Likewise.
8665 (altivec_vspltsf): Likewise.
8667 2014-01-30 Jakub Jelinek <jakub@redhat.com>
8670 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
8671 frame related instructions.
8673 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
8675 PR rtl-optimization/59959
8676 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
8677 any reload of register whose subreg is invalid.
8679 2014-01-30 Jakub Jelinek <jakub@redhat.com>
8681 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
8682 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
8683 Add missing return type - void.
8685 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8687 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
8688 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
8689 remove element index adjustment for endian (now handled in vsx.md
8691 (altivec_expand_vec_perm_const): Use
8692 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
8693 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
8694 (vsx_xxspltw_<mode>): Adjust element index for little endian.
8695 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
8696 define_expand and a new define_insn *altivec_vspltb_internal;
8697 adjust for -maltivec=be on a little endian target.
8698 (altivec_vspltb_direct): New.
8699 (altivec_vsplth): Divide into a define_expand and a new
8700 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
8701 little endian target.
8702 (altivec_vsplth_direct): New.
8703 (altivec_vspltw): Divide into a define_expand and a new
8704 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
8705 little endian target.
8706 (altivec_vspltw_direct): New.
8707 (altivec_vspltsf): Divide into a define_expand and a new
8708 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
8709 a little endian target.
8711 2014-01-30 Richard Biener <rguenther@suse.de>
8713 PR tree-optimization/59993
8714 * tree-ssa-forwprop.c (associate_pointerplus): Check we
8715 can propagate form the earlier stmt and avoid the transform
8716 when the intermediate result is needed.
8718 2014-01-30 Alangi Derick <alangiderick@gmail.com>
8720 * README.Portability: Fix typo.
8722 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
8724 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
8725 comparison_operator with ordered_comparison_operator.
8727 2014-01-30 Nick Clifton <nickc@redhat.com>
8729 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
8730 Rename to mn10300_store_multiple_regs.
8731 * config/mn10300/mn10300.c: Likewise.
8732 * config/mn10300/mn10300.md (store_movm): Fix typo: call
8733 store_multiple_regs.
8734 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
8735 Call mn10300_store_multiple_regs.
8737 2014-01-30 Nick Clifton <nickc@redhat.com>
8738 DJ Delorie <dj@redhat.com>
8740 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
8741 %fp 2 to keep registers after it properly word-aligned.
8742 (rl78_alloc_physical_registers_umul): Handle the case where both
8743 input operands are the same.
8745 2014-01-30 Richard Biener <rguenther@suse.de>
8747 PR tree-optimization/59903
8748 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
8751 2014-01-30 Jason Merrill <jason@redhat.com>
8754 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
8757 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
8759 2014-01-30 Richard Biener <rguenther@suse.de>
8761 PR tree-optimization/59951
8762 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
8764 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
8767 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
8768 SFmode to DFmode case.
8770 2014-01-29 DJ Delorie <dj@redhat.com>
8772 * config/msp430/msp430.opt (-minrt): New.
8773 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
8775 (ENDFILE_SPEC): Likewise.
8777 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
8779 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
8780 (estimate_function_body_sizes): Use it.
8782 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
8785 * dwarf2out.c (is_cxx_auto): New.
8786 (is_base_type): Use it.
8787 (gen_type_die_with_usage): Likewise.
8789 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8791 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
8792 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
8793 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
8794 -maltivec=be with LE targets.
8795 (vsx_mergeh_<mode>): Likewise.
8796 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
8797 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
8798 (altivec_vmrghb): Replace with define_expand and new
8799 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
8800 (altivec_vmrghb_direct): New define_insn.
8801 (altivec_vmrghh): Replace with define_expand and new
8802 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
8803 (altivec_vmrghh_direct): New define_insn.
8804 (altivec_vmrghw): Replace with define_expand and new
8805 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
8806 (altivec_vmrghw_direct): New define_insn.
8807 (*altivec_vmrghsf): Adjust for endianness.
8808 (altivec_vmrglb): Replace with define_expand and new
8809 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
8810 (altivec_vmrglb_direct): New define_insn.
8811 (altivec_vmrglh): Replace with define_expand and new
8812 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
8813 (altivec_vmrglh_direct): New define_insn.
8814 (altivec_vmrglw): Replace with define_expand and new
8815 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
8816 (altivec_vmrglw_direct): New define_insn.
8817 (*altivec_vmrglsf): Adjust for endianness.
8818 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
8819 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
8820 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
8821 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
8822 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
8823 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
8824 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
8825 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
8827 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
8829 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
8830 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
8833 2014-01-29 Richard Biener <rguenther@suse.de>
8835 PR tree-optimization/58742
8836 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
8837 associate_pointerplus_align.
8838 (associate_pointerplus_diff): New function.
8839 (associate_pointerplus): Likewise. Call associate_pointerplus_align
8840 and associate_pointerplus_diff.
8842 2014-01-29 Richard Biener <rguenther@suse.de>
8844 * lto-streamer.h (LTO_major_version): Bump to 3.
8845 (LTO_minor_version): Reset to 0.
8847 2014-01-29 Renlin Li <Renlin.Li@arm.com>
8849 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
8850 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
8851 (arm_file_start): Generate correct asm header for armv7ve.
8852 * config/arm/bpabi.h: Add multilib support for armv7ve.
8853 * config/arm/driver-arm.c: Change the architectures of cortex-a7
8854 and cortex-a15 to armv7ve.
8855 * config/arm/t-aprofile: Add multilib support for armv7ve.
8856 * doc/invoke.texi: Document -march=armv7ve.
8858 2014-01-29 Richard Biener <rguenther@suse.de>
8860 PR tree-optimization/58742
8861 * tree-ssa-forwprop.c (associate_plusminus): Return true
8862 if we changed sth, defer EH cleanup to ...
8863 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
8864 (simplify_mult): New function.
8866 2014-01-29 Jakub Jelinek <jakub@redhat.com>
8869 PR tree-optimization/59920
8870 * tree.c (build_common_builtin_nodes): Remove
8871 __builtin_setjmp_dispatcher initialization.
8872 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
8873 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
8874 instead of gsi_after_labels + manually skipping debug stmts.
8875 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
8876 ignore bbs with IFN_ABNORMAL_DISPATCHER.
8877 * tree-inline.c (copy_edges_for_bb): Remove
8878 can_make_abnormal_goto argument, instead add abnormal_goto_dest
8879 argument. Ignore computed_goto_p stmts. Don't call
8880 make_abnormal_goto_edges. If a call might need abnormal edges
8881 for non-local gotos, see if it already has an edge to
8882 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
8883 with true argument, don't do anything then, otherwise add
8884 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
8885 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
8887 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
8888 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
8889 (lower_stmt): Don't set data->calls_builtin_setjmp.
8890 (lower_builtin_setjmp): Adjust comment.
8891 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
8892 * tree-cfg.c (found_computed_goto): Remove.
8893 (factor_computed_gotos): Remove.
8894 (make_goto_expr_edges): Return bool, true for computed gotos.
8895 Don't call make_abnormal_goto_edges.
8896 (build_gimple_cfg): Don't set found_computed_goto, don't call
8897 factor_computed_gotos.
8898 (computed_goto_p): No longer static.
8899 (make_blocks): Don't set found_computed_goto.
8900 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
8901 (make_edges): If make_goto_expr_edges returns true, push bb
8902 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
8903 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
8904 vector. Record mapping between bbs and OpenMP regions if there
8905 are any, adjust make_gimple_omp_edges caller. Call
8906 handle_abnormal_edges.
8907 (make_abnormal_goto_edges): Remove.
8908 * tree-cfg.h (make_abnormal_goto_edges): Remove.
8909 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
8910 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
8911 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
8912 * internal-fn.def (ABNORMAL_DISPATCHER): New.
8913 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
8914 filling *region also set *region_idx to (*region)->entry->index.
8917 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
8918 For REGs set ORIGINAL_REGNO.
8920 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
8922 * doc/md.texi: Mention that a target shouldn't implement
8923 vec_widen_(s|u)mul_even/odd pair if it is less efficient
8926 2014-01-29 Jakub Jelinek <jakub@redhat.com>
8928 PR tree-optimization/59594
8929 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
8930 a copy of the datarefs vector rather than the vector itself.
8932 2014-01-28 Jason Merrill <jason@redhat.com>
8935 * dwarf2out.c (auto_die): New static.
8936 (gen_type_die_with_usage): Handle C++1y 'auto'.
8937 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
8940 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
8943 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
8944 (SPEC_X32): Likewise.
8945 (SPEC_64): Likewise.
8946 * config/i386/i386.c (ix86_option_override_internal): Turn off
8947 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
8949 (x86_file_start): Output .code16gcc for TARGET_16BIT.
8950 * config/i386/i386.h (TARGET_16BIT): New macro.
8951 (TARGET_16BIT_P): Likewise.
8952 * config/i386/i386.opt: Add m16.
8953 * doc/invoke.texi: Document -m16.
8955 2014-01-28 Jakub Jelinek <jakub@redhat.com>
8957 PR preprocessor/59935
8958 * input.c (location_get_source_line): Bail out on when line number
8959 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
8961 2014-01-28 Richard Biener <rguenther@suse.de>
8963 PR tree-optimization/58742
8964 * tree-ssa-forwprop.c (associate_plusminus): Handle
8965 pointer subtraction of the form (T)(P + A) - (T)P.
8967 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8969 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
8972 2014-01-28 Richard Biener <rguenther@suse.de>
8975 2014-01-28 Richard Biener <rguenther@suse.de>
8977 PR rtl-optimization/45364
8978 PR rtl-optimization/59890
8979 * var-tracking.c (local_get_addr_clear_given_value): Handle
8980 already cleared slot.
8981 (val_reset): Handle not allocated local_get_addr_cache.
8982 (vt_find_locations): Use post-order on the inverted CFG.
8984 2014-01-28 Richard Biener <rguenther@suse.de>
8986 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
8988 2014-01-28 Richard Biener <rguenther@suse.de>
8990 PR rtl-optimization/45364
8991 PR rtl-optimization/59890
8992 * var-tracking.c (local_get_addr_clear_given_value): Handle
8993 already cleared slot.
8994 (val_reset): Handle not allocated local_get_addr_cache.
8995 (vt_find_locations): Use post-order on the inverted CFG.
8997 2014-01-28 Alan Modra <amodra@gmail.com>
8999 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
9000 * configure.ac <recursive call for build != host>: Define
9001 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
9002 and LD_FOR_BUILD too.
9003 * configure: Regenerate.
9005 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
9007 * config/i386/i386.c (get_builtin_code_for_version): Separate
9008 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
9009 Broadwell from Haswell.
9011 2014-01-27 Steve Ellcey <sellcey@mips.com>
9013 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
9014 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
9015 * config/mips/mips.c (mips_option_override): Change setting
9017 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
9018 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
9019 Change from Mask to Var.
9021 2014-01-27 Jeff Law <law@redhat.com>
9023 * ipa-inline.c (inline_small_functions): Fix typo.
9025 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
9027 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
9028 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
9029 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
9030 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
9031 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
9032 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
9033 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
9034 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
9035 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
9036 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
9037 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
9038 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
9039 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
9040 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
9041 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
9042 (_mm512_storeu_epi64): Ditto.
9043 (_mm512_cmpge_epi32_mask): Ditto.
9044 (_mm512_cmpge_epu32_mask): Ditto.
9045 (_mm512_cmpge_epi64_mask): Ditto.
9046 (_mm512_cmpge_epu64_mask): Ditto.
9047 (_mm512_cmple_epi32_mask): Ditto.
9048 (_mm512_cmple_epu32_mask): Ditto.
9049 (_mm512_cmple_epi64_mask): Ditto.
9050 (_mm512_cmple_epu64_mask): Ditto.
9051 (_mm512_cmplt_epi32_mask): Ditto.
9052 (_mm512_cmplt_epu32_mask): Ditto.
9053 (_mm512_cmplt_epi64_mask): Ditto.
9054 (_mm512_cmplt_epu64_mask): Ditto.
9055 (_mm512_cmpneq_epi32_mask): Ditto.
9056 (_mm512_cmpneq_epu32_mask): Ditto.
9057 (_mm512_cmpneq_epi64_mask): Ditto.
9058 (_mm512_cmpneq_epu64_mask): Ditto.
9059 (_mm512_expand_pd): Ditto.
9060 (_mm512_expand_ps): Ditto.
9061 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
9062 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
9063 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
9064 * config/i386/i386.c (ix86_builtins): Add
9065 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
9066 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
9067 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
9068 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
9069 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
9070 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
9071 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
9072 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
9073 IX86_BUILTIN_PMOVUSQW512_MEM.
9074 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
9075 __builtin_ia32_pmovsqd512mem_mask,
9076 __builtin_ia32_pmovqd512mem_mask,
9077 __builtin_ia32_pmovusqw512mem_mask,
9078 __builtin_ia32_pmovsqw512mem_mask,
9079 __builtin_ia32_pmovqw512mem_mask,
9080 __builtin_ia32_pmovusdw512mem_mask,
9081 __builtin_ia32_pmovsdw512mem_mask,
9082 __builtin_ia32_pmovdw512mem_mask,
9083 __builtin_ia32_pmovqb512mem_mask,
9084 __builtin_ia32_pmovusqb512mem_mask,
9085 __builtin_ia32_pmovsqb512mem_mask,
9086 __builtin_ia32_pmovusdb512mem_mask,
9087 __builtin_ia32_pmovsdb512mem_mask,
9088 __builtin_ia32_pmovdb512mem_mask.
9089 (bdesc_args): Add __builtin_ia32_expanddf512,
9090 __builtin_ia32_expandsf512.
9091 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
9092 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
9093 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
9094 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
9095 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
9096 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
9097 (avx512f_<code>v8div16qi2_mask_store): This.
9098 (avx512f_expand<mode>): New.
9100 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
9102 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
9104 (_mm512_mask_prefetch_i64gather_pd): Ditto.
9105 (_mm512_prefetch_i32scatter_pd): Ditto.
9106 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
9107 (_mm512_prefetch_i64scatter_pd): Ditto.
9108 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
9109 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
9110 (_mm512_mask_prefetch_i64gather_ps): Ditto.
9111 (_mm512_prefetch_i32scatter_ps): Ditto.
9112 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
9113 (_mm512_prefetch_i64scatter_ps): Ditto.
9114 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
9115 * config/i386/i386-builtin-types.def: Define
9116 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
9117 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
9118 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
9119 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
9120 IX86_BUILTIN_SCATTERPFQPD.
9121 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
9122 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
9123 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
9124 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
9125 __builtin_ia32_scatterpfqps.
9126 (ix86_expand_builtin): Expand new built-ins.
9127 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
9128 fix memory access data type.
9129 (*avx512pf_gatherpf<mode>_mask): Ditto.
9130 (*avx512pf_gatherpf<mode>): Ditto.
9131 (avx512pf_scatterpf<mode>): Ditto.
9132 (*avx512pf_scatterpf<mode>_mask): Ditto.
9133 (*avx512pf_scatterpf<mode>): Ditto.
9134 (GATHER_SCATTER_SF_MEM_MODE): New.
9135 (avx512pf_gatherpf<mode>df): Ditto.
9136 (*avx512pf_gatherpf<mode>df_mask): Ditto.
9137 (*avx512pf_scatterpf<mode>df): Ditto.
9139 2014-01-27 Jakub Jelinek <jakub@redhat.com>
9142 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
9143 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
9146 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
9148 * common/config/arm/arm-common.c
9149 (arm_rewrite_mcpu): Handle multiple names.
9151 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
9153 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
9155 * gimple-builder.h (create_gimple_tmp): Delete.
9157 2014-01-27 Christian Bruel <christian.bruel@st.com>
9159 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
9162 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
9164 * config/pa/pa.md (call): Generate indirect long calls to non-local
9165 functions when outputing 32-bit code.
9166 (call_value): Likewise except for special call to buggy powf function.
9168 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
9169 portable runtime and PIC indirect calls.
9170 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
9171 and PIC call sequences. Use ldo instead of blr to set return register
9172 in PIC call sequence.
9174 2014-01-25 Walter Lee <walt@tilera.com>
9176 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
9177 avoid clobbering a live register.
9179 2014-01-25 Walter Lee <walt@tilera.com>
9181 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
9182 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
9183 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
9184 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
9186 2014-01-25 Walter Lee <walt@tilera.com>
9188 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
9189 arguments on even registers.
9190 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
9192 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
9193 (BIGGEST_ALIGNMENT): Ditto.
9194 (BIGGEST_FIELD_ALIGNMENT): Ditto.
9196 2014-01-25 Walter Lee <walt@tilera.com>
9198 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
9199 insns before bundling.
9200 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
9202 2014-01-25 Walter Lee <walt@tilera.com>
9204 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
9205 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
9206 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
9208 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
9210 * config/mips/constraints.md (kl): Delete.
9211 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
9212 define expands, using...
9213 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
9214 instructions for MIPS16.
9215 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
9216 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
9218 2014-01-25 Walter Lee <walt@tilera.com>
9220 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
9224 2014-01-25 Walter Lee <walt@tilera.com>
9226 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
9227 (TARGET_EXPAND_TO_RTL_HOOK): Define.
9229 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
9231 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
9234 2014-01-25 Jakub Jelinek <jakub@redhat.com>
9236 * print-rtl.c (in_call_function_usage): New var.
9237 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
9238 EXPR_LIST mode as mode and not as reg note name.
9241 * cfgloopmanip.c (copy_loop_info): If
9242 loop->warned_aggressive_loop_optimizations, make sure
9243 the flag is set in target loop too.
9245 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
9247 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
9249 * builtins.def: Likewise.
9250 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
9251 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
9252 * ira.c (ira_setup_eliminable_regset): Likewise.
9253 * omp-low.c (gate_expand_omp): Likewise.
9254 (execute_lower_omp): Likewise.
9255 (diagnose_sb_0): Likewise.
9256 (gate_diagnose_omp_blocks): Likewise.
9257 (simd_clone_clauses_extract): Likewise.
9260 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9262 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
9263 correction for little endian...
9264 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
9267 2014-01-24 Jeff Law <law@redhat.com>
9269 PR tree-optimization/59919
9270 * tree-vrp.c (find_assert_locations_1): Do not register asserts
9271 for non-returning calls.
9273 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
9275 * common/config/aarch64/aarch64-common.c
9276 (aarch64_rewrite_mcpu): Handle multiple names.
9277 * config/aarch64/aarch64.h
9278 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
9280 2014-01-24 Dodji Seketeli <dodji@redhat.com>
9282 * input.c (add_file_to_cache_tab): Handle the case where fopen
9285 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
9288 * config/i386/i386.md (pushsf splitter): Get stack adjustment
9289 from push operand if code of push isn't PRE_DEC.
9291 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
9294 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
9295 -mquad-memory-atomic. Update -mquad-memory documentation to say
9296 it is only used for non-atomic loads/stores.
9298 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
9299 -mquad-memory or -mquad-memory-atomic switches.
9301 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
9302 -mquad-memory-atomic to ISA 2.07 support.
9304 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
9305 to separate support of normal quad word memory operations (ldq, stq)
9306 from the atomic quad word memory operations.
9308 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
9309 support to separate non-atomic quad word operations from atomic
9310 quad word operations. Disable non-atomic quad word operations in
9311 little endian mode so that we don't have to swap words after the
9312 load and before the store.
9313 (quad_load_store_p): Add comment about atomic quad word support.
9314 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
9315 options printed with -mdebug=reg.
9317 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
9318 -mquad-memory-atomic as the test for whether we have quad word
9319 atomic instructions.
9320 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
9321 or -mp8-vector are used, allow byte/half-word atomic operations.
9323 * config/rs6000/sync.md (load_lockedti): Insure that the address
9324 is a proper indexed or indirect address for the lqarx instruction.
9325 On little endian systems, swap the hi/lo registers after the lqarx
9327 (load_lockedpti): Use indexed_or_indirect_operand predicate to
9328 insure the address is valid for the lqarx instruction.
9329 (store_conditionalti): Insure that the address is a proper indexed
9330 or indirect address for the stqcrx. instruction. On little endian
9331 systems, swap the hi/lo registers before doing the stqcrx.
9333 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
9334 insure the address is valid for the stqcrx. instruction.
9336 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
9337 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
9338 type of quad memory support is available.
9340 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
9343 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
9344 there is a danger of looping.
9346 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
9348 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
9349 force flag_ira_loop_pressure if set via command line.
9351 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
9353 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
9354 (ashr_simd): New builtin handling DI mode.
9355 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
9356 (aarch64_sshr_simddi): New match pattern.
9357 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
9358 (vshrd_n_s64): Likewise.
9359 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
9361 2014-01-23 Nick Clifton <nickc@redhat.com>
9363 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
9364 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
9365 favour of mcu specific scripts.
9366 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
9369 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
9370 Alex Velenko <Alex.Velenko@arm.com>
9372 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
9373 (vaddv_s16): Likewise.
9374 (vaddv_s32): Likewise.
9375 (vaddv_u8): Likewise.
9376 (vaddv_u16): Likewise.
9377 (vaddv_u32): Likewise.
9378 (vaddvq_s8): Likewise.
9379 (vaddvq_s16): Likewise.
9380 (vaddvq_s32): Likewise.
9381 (vaddvq_s64): Likewise.
9382 (vaddvq_u8): Likewise.
9383 (vaddvq_u16): Likewise.
9384 (vaddvq_u32): Likewise.
9385 (vaddvq_u64): Likewise.
9386 (vaddv_f32): Likewise.
9387 (vaddvq_f32): Likewise.
9388 (vaddvq_f64): Likewise.
9389 (vmaxv_f32): Likewise.
9390 (vmaxv_s8): Likewise.
9391 (vmaxv_s16): Likewise.
9392 (vmaxv_s32): Likewise.
9393 (vmaxv_u8): Likewise.
9394 (vmaxv_u16): Likewise.
9395 (vmaxv_u32): Likewise.
9396 (vmaxvq_f32): Likewise.
9397 (vmaxvq_f64): Likewise.
9398 (vmaxvq_s8): Likewise.
9399 (vmaxvq_s16): Likewise.
9400 (vmaxvq_s32): Likewise.
9401 (vmaxvq_u8): Likewise.
9402 (vmaxvq_u16): Likewise.
9403 (vmaxvq_u32): Likewise.
9404 (vmaxnmv_f32): Likewise.
9405 (vmaxnmvq_f32): Likewise.
9406 (vmaxnmvq_f64): Likewise.
9407 (vminv_f32): Likewise.
9408 (vminv_s8): Likewise.
9409 (vminv_s16): Likewise.
9410 (vminv_s32): Likewise.
9411 (vminv_u8): Likewise.
9412 (vminv_u16): Likewise.
9413 (vminv_u32): Likewise.
9414 (vminvq_f32): Likewise.
9415 (vminvq_f64): Likewise.
9416 (vminvq_s8): Likewise.
9417 (vminvq_s16): Likewise.
9418 (vminvq_s32): Likewise.
9419 (vminvq_u8): Likewise.
9420 (vminvq_u16): Likewise.
9421 (vminvq_u32): Likewise.
9422 (vminnmv_f32): Likewise.
9423 (vminnmvq_f32): Likewise.
9424 (vminnmvq_f64): Likewise.
9426 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
9428 * config/aarch64/aarch64-simd.md
9429 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
9430 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
9431 (*aarch64_mul3_elt<mode>): Likewise.
9432 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
9433 (*aarch64_mul3_elt_to_64v2df): Likewise.
9434 (*aarch64_mla_elt<mode>): Likewise.
9435 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
9436 (*aarch64_mls_elt<mode>): Likewise.
9437 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
9438 (*aarch64_fma4_elt<mode>): Likewise.
9439 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
9440 (*aarch64_fma4_elt_to_64v2df): Likewise.
9441 (*aarch64_fnma4_elt<mode>): Likewise.
9442 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
9443 (*aarch64_fnma4_elt_to_64v2df): Likewise.
9444 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
9445 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
9446 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
9447 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
9448 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
9449 (aarch64_sqdmull_lane<mode>_internal): Likewise.
9450 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
9452 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
9454 * config/aarch64/aarch64-simd.md
9455 (aarch64_be_checked_get_lane<mode>): New define_expand.
9456 * config/aarch64/aarch64-simd-builtins.def
9457 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
9458 New builtin definition.
9459 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
9460 Use new safe be builtin.
9462 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
9464 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
9466 (aarch64_be_st1<mode>): Likewise.
9467 (aarch_ld1<VALL:mode>): Define_expand modified.
9468 (aarch_st1<VALL:mode>): Likewise.
9469 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
9470 (UNSPEC_ST1): Likewise.
9472 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
9474 * config/microblaze/microblaze.md: Add trap insn and attribute
9476 2014-01-23 Dodji Seketeli <dodji@redhat.com>
9478 PR preprocessor/58580
9479 * input.h (location_get_source_line): Take an additional line_size
9481 (void diagnostics_file_cache_fini): Declare new function.
9482 * input.c (struct fcache): New type.
9483 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
9484 New static constants.
9485 (diagnostic_file_cache_init, total_lines_num)
9486 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
9487 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
9488 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
9489 (get_next_line, read_next_line, goto_next_line, read_line_num):
9490 New static function definitions.
9491 (diagnostic_file_cache_fini): New function.
9492 (location_get_source_line): Take an additional output line_len
9493 parameter. Re-write using lookup_or_add_file_to_cache_tab and
9495 * diagnostic.c (diagnostic_finish): Call
9496 diagnostic_file_cache_fini.
9497 (adjust_line): Take an additional input parameter for the length
9498 of the line, rather than calculating it with strlen.
9499 (diagnostic_show_locus): Adjust the use of
9500 location_get_source_line and adjust_line with respect to their new
9501 signature. While displaying a line now, do not stop at the first
9502 null byte. Rather, display the zero byte as a space and keep
9503 going until we reach the size of the line.
9504 * Makefile.in: Add vec.o to OBJS-libcommon
9506 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
9507 Ilya Tocar <ilya.tocar@intel.com>
9509 * config/i386/avx512fintrin.h (_mm512_kmov): New.
9510 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
9511 (__builtin_ia32_kmov16): Ditto.
9512 * config/i386/i386.md (UNSPEC_KMOV): New.
9515 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
9517 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
9518 (_mm512_storeu_si512): Ditto.
9520 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
9523 * rtl.h (get_referenced_operands): Declare.
9524 * recog.c (get_referenced_operands): New function.
9525 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
9526 operands have been referenced when recording LO_SUM references.
9528 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
9530 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
9532 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
9534 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
9535 Enable for generic and recent AMD targets.
9537 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
9539 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
9540 ARG_SIZE note when adjustment was eliminated.
9542 2014-01-22 Jeff Law <law@redhat.com>
9544 PR tree-optimization/59597
9545 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
9546 in file. Accept new argument REGISTERING and use it to modify
9547 dump output appropriately.
9548 (register_jump_thread): Corresponding changes.
9549 (mark_threaded_blocks): Reinstate code to cancel unprofitable
9550 thread paths involving joiner blocks. Add code to dump cancelled
9551 jump threading paths.
9553 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
9555 PR rtl-optimization/59477
9556 * lra-constraints.c (inherit_in_ebb): Process call for living hard
9557 regs. Update reloads_num and potential_reload_hard_regs for all insns.
9559 2014-01-22 Tom Tromey <tromey@redhat.com>
9561 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
9563 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
9565 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
9567 PR rtl-optimization/59896
9568 * lra-constraints.c (process_alt_operands): Check unused note for
9569 matched operands of insn with no output reloads.
9571 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
9573 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
9574 (mips_move_from_gpr_cost): Likewise.
9576 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
9578 PR rtl-optimization/59858
9579 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
9580 ira_class_hard_regs_num.
9581 (process_alt_operands): Increase reject for dying matched operand.
9583 2014-01-21 Jakub Jelinek <jakub@redhat.com>
9586 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
9587 smaller than size, perform several stores or loads and stores
9588 at dst + count - size to store or copy all of size bytes, rather
9589 than just last modesize bytes.
9591 2014-01-20 DJ Delorie <dj@redhat.com>
9593 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
9594 that CLOBBERs are REGs before propogating their values.
9596 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
9599 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
9600 (cgraph_inline_failed_type): New function.
9601 * cgraph.h (DEFCIFCODE): Add type.
9602 (cgraph_inline_failed_type_t): New enum.
9603 (cgraph_inline_failed_type): New prototype.
9604 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
9605 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
9606 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
9607 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
9608 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
9609 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
9610 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
9611 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
9612 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
9613 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
9614 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
9615 OPTIMIZATION_MISMATCH.
9616 * tree-inline.c (expand_call_inline): Emit errors during
9617 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
9619 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
9622 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
9623 mode attribute in insn output.
9625 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
9627 * output.h (output_constant): Delete.
9628 * varasm.c (output_constant): Make private.
9630 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
9632 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
9634 2014-01-20 Jakub Jelinek <jakub@redhat.com>
9637 * tree.h (fold_builtin_strcat): New prototype.
9638 * builtins.c (fold_builtin_strcat): No longer static. Add len
9639 argument, if non-NULL, don't call c_strlen. Optimize
9640 directly into __builtin_memcpy instead of __builtin_strcpy.
9641 (fold_builtin_2): Adjust fold_builtin_strcat caller.
9642 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
9644 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
9646 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
9647 for SImode_address_operand operands, having only a REG argument.
9649 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
9651 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
9652 loader name using mbig-endian.
9653 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
9655 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
9657 * doc/invoke.texi (-march): Clarify documentation for AArch64.
9661 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
9663 * config/aarch64/aarch64-protos.h
9664 (aarch64_cannot_change_mode_class_ptr): Declare.
9665 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
9666 aarch64_cannot_change_mode_class_ptr): New.
9667 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
9668 backend hook aarch64_cannot_change_mode_class.
9670 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
9672 * common/config/aarch64/aarch64-common.c
9673 (aarch64_handle_option): Don't handle any option order logic here.
9674 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
9675 selected_cpu, warn on architecture version mismatch.
9676 (aarch64_override_options): Fix parsing order for option strings.
9678 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
9679 Iain Sandoe <iain@codesourcery.com>
9682 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
9683 warning. Amend comment to reflect current functionality.
9685 2014-01-20 Richard Biener <rguenther@suse.de>
9688 * builtins.c (fold_builtin_strcat): Remove case better handled
9689 by tree-ssa-strlen.c.
9691 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
9693 * config/aarch64/aarch64.opt
9694 (mcpu, march, mtune): Make case-insensitive.
9696 2014-01-20 Jakub Jelinek <jakub@redhat.com>
9699 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
9700 if operands[1] is a REG or ZERO_EXTEND of a REG.
9702 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
9704 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
9706 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
9708 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
9709 long non-pic millicode calls.
9711 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
9713 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
9715 2014-01-19 Kito Cheng <kito@0xlab.org>
9717 * builtins.c (expand_movstr): Check movstr expand done or fail.
9719 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
9720 H.J. Lu <hongjiu.lu@intel.com>
9723 * config/i386/i386.md (*lea<mode>): Zero-extend return register
9724 to DImode for zero-extended addresses.
9726 2014-01-19 Jakub Jelinek <jakub@redhat.com>
9728 PR rtl-optimization/57763
9729 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
9730 on the new indirect jump_insn and increment LABEL_NUSES (label).
9732 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
9736 * config.gcc (x86_archs): New variable.
9737 (x86_64_archs): Likewise.
9738 (x86_cpus): Likewise.
9739 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
9740 --with-arch/--with-cpu= options.
9741 Support --with-arch=/--with-cpu={nehalem,westmere,
9742 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
9744 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
9746 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
9747 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
9749 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
9751 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
9753 2014-01-18 Jakub Jelinek <jakub@redhat.com>
9756 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
9757 clear cpp_get_options (parse_in)->warn_unused_macros for
9758 ix86_target_macros_internal with cpp_define.
9760 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
9762 * jump.c (delete_related_insns): Keep (use (insn))s.
9763 * reorg.c (redundant_insn): Check for barriers too.
9765 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
9767 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
9769 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
9771 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
9772 call to $$dyncall when TARGET_LONG_CALLS is true.
9774 2014-01-17 Jeff Law <law@redhat.com>
9776 * ree.c (combine_set_extension): Temporarily disable test for
9777 changing number of hard registers.
9779 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
9782 * ipa-inline-analysis.c (inline_free_summary):
9783 Do not free summary of aliases.
9785 2014-01-17 Jakub Jelinek <jakub@redhat.com>
9788 * gimplify.c (gimplify_expr): Use create_tmp_var
9789 instead of create_tmp_var_raw. If cond doesn't have
9790 integral type, don't add the IFN_ANNOTATE builtin at all.
9792 2014-01-17 Martin Jambor <mjambor@suse.cz>
9795 * ipa-cp.c (prev_edge_clone): New variable.
9796 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
9797 Also resize prev_edge_clone vector.
9798 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
9799 (ipcp_edge_removal_hook): New function.
9800 (ipcp_driver): Register ipcp_edge_removal_hook.
9802 2014-01-17 Andrew Pinski <apinski@cavium.com>
9803 Steve Ellcey <sellcey@mips.com>
9806 * config/mips/mips.c (mips_print_operand): Check operand mode instead
9809 2014-01-17 Jeff Law <law@redhat.com>
9812 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
9813 so that pass_ccp runs first.
9815 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
9817 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
9818 (ix86_adjust_cost): Use !TARGET_XXX.
9819 (do_reorder_for_imul): Likewise.
9820 (swap_top_of_ready_list): Likewise.
9821 (ix86_sched_reorder): Likewise.
9823 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
9825 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
9826 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
9827 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
9828 (intel_memset): New. Duplicate slm_memset.
9829 (intel_cost): New. Duplicate slm_cost.
9830 (m_INTEL): New macro.
9831 (processor_target_table): Add "intel".
9832 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
9833 with PROCESSOR_INTEL for "intel".
9834 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
9835 PROCESSOR_SILVERMONT.
9836 (ix86_issue_rate): Likewise.
9837 (ix86_adjust_cost): Likewise.
9838 (ia32_multipass_dfa_lookahead): Likewise.
9839 (swap_top_of_ready_list): Likewise.
9840 (ix86_sched_reorder): Likewise.
9841 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
9842 instead of TARGET_OPT_AGU.
9843 * config/i386/i386.h (TARGET_INTEL): New.
9844 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
9845 (processor_type): Add PROCESSOR_INTEL.
9846 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
9847 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
9849 2014-01-17 Marek Polacek <polacek@redhat.com>
9852 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
9855 2014-01-17 Richard Biener <rguenther@suse.de>
9857 PR tree-optimization/46590
9858 * opts.c (default_options_table): Add entries for
9859 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
9860 all enabled at -O1 but not for -Og.
9861 * common.opt (fbranch-count-reg): Remove Init(1).
9862 (fmove-loop-invariants): Likewise.
9863 (ftree-pta): Likewise.
9865 2014-01-17 Jakub Jelinek <jakub@redhat.com>
9867 * config/i386/i386.c (ix86_data_alignment): For compatibility with
9868 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
9869 decls to at least the GCC 4.8 used alignments.
9872 * tree-nested.c (convert_nonlocal_reference_stmt,
9873 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
9874 of GIMPLE_BIND stmts, adjust associated decls.
9876 2014-01-17 Richard Biener <rguenther@suse.de>
9878 PR tree-optimization/46590
9879 * vec.h (vec<>::bseach): New member function implementing
9880 binary search according to C89 bsearch.
9881 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
9882 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
9883 bitmap pointer again. Make accesses_in_loop a flat array.
9884 (mem_ref_obstack): New global.
9885 (outermost_indep_loop): Adjust for mem_ref->stored changes.
9886 (mark_ref_stored): Likewise.
9887 (ref_indep_loop_p_2): Likewise.
9888 (set_ref_stored_in_loop): New helper function.
9889 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
9890 (memref_free): Adjust.
9891 (record_mem_ref_loc): Simplify.
9892 (gather_mem_refs_stmt): Adjust.
9893 (sort_locs_in_loop_postorder_cmp): New function.
9894 (analyze_memory_references): Sort accesses_in_loop after
9895 loop postorder number.
9896 (find_ref_loc_in_loop_cmp): New function.
9897 (for_all_locs_in_loop): Find relevant cluster of locs in
9898 accesses_in_loop and iterate without recursion.
9899 (execute_sm): Avoid uninit warning.
9900 (struct ref_always_accessed): Simplify.
9901 (ref_always_accessed::operator ()): Likewise.
9902 (ref_always_accessed_p): Likewise.
9903 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
9904 loop postorder numbers here.
9905 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
9908 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
9911 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
9912 on decls for which assemble_alias has been called.
9914 2014-01-17 Nick Clifton <nickc@redhat.com>
9916 * config/msp430/msp430.opt: (mcpu): New option.
9917 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
9918 (msp430_option_override): Parse target_cpu. If the MCU name
9919 matches a generic string, clear target_mcu.
9920 (msp430_attr): Allow numeric interrupt values up to 63.
9921 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
9922 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
9924 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
9926 * config/msp430/msp430.md (popm): Use %J rather than %I.
9927 (addsi3): Use msp430_nonimmediate_operand for operand 2.
9928 (addhi_cy_i): Use immediate_operand for operand 2.
9929 * doc/invoke.texi: Document -mcpu option.
9931 2014-01-17 Richard Biener <rguenther@suse.de>
9933 PR rtl-optimization/38518
9934 * df.h (df_analyze_loop): Declare.
9935 * df-core.c: Include cfgloop.h.
9936 (df_analyze_1): Split out main part of df_analyze.
9937 (df_analyze): Adjust.
9938 (loop_inverted_post_order_compute): New function.
9939 (loop_post_order_compute): Likewise.
9940 (df_analyze_loop): New function avoiding whole-function
9942 * loop-invariant.c (find_defs): Use df_analyze_loop.
9943 (find_invariants): Adjust.
9944 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
9946 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
9948 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
9949 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
9951 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
9953 * ipa-ref.c (ipa_remove_stmt_references): Fix references
9954 traversal when removing references.
9956 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
9959 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
9961 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
9964 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
9965 pushing a reload for an autoinc when we had previously reloaded an
9966 inner part of the address.
9968 2014-01-16 Jakub Jelinek <jakub@redhat.com>
9970 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
9972 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
9973 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
9974 when not giving up or versioning for alias only because of
9976 (vect_analyze_data_ref_dependences): Set to true.
9977 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
9979 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
9980 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
9984 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
9987 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
9988 operand 0 predicate for gathers, use a new pseudo as subtarget.
9990 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
9993 * lra-constraints.c (process_alt_operands): Add printing debug info.
9994 Check absence of input/output reloads for matched operands too.
9996 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
9998 PR rtl-optimization/59835
9999 * ira.c (ira_init_register_move_cost): Increase cost for
10002 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
10004 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
10006 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
10009 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
10010 non-register objects. Use gen_(high/low)part more consistently.
10013 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
10016 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
10017 endian support, remove tests for WORDS_BIG_ENDIAN.
10018 (p8_mfvsrd_3_<mode>): Likewise.
10019 (reload_gpr_from_vsx<mode>): Likewise.
10020 (reload_gpr_from_vsxsf): Likewise.
10021 (p8_mfvsrd_4_disf): Likewise.
10023 2014-01-16 Richard Biener <rguenther@suse.de>
10025 PR rtl-optimization/46590
10026 * lcm.c (compute_antinout_edge): Use postorder iteration.
10027 (compute_laterin): Use inverted postorder iteration.
10029 2014-01-16 Nick Clifton <nickc@redhat.com>
10031 PR middle-end/28865
10032 * varasm.c (output_constant): Return the number of bytes actually
10034 (output_constructor_array_range): Update the field size with the
10035 number of bytes emitted by output_constant.
10036 (output_constructor_regular_field): Likewise. Also do not
10037 complain if the total number of bytes emitted is now greater
10038 than the expected fieldpos.
10039 * output.h (output_constant): Update prototype and descriptive comment.
10041 2014-01-16 Marek Polacek <polacek@redhat.com>
10043 PR middle-end/59827
10044 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
10045 it is error_mark_node.
10047 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
10049 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
10050 VALID_AVX256_REG_OR_OI_MODE.
10052 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
10054 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
10055 current procedure should be profiled.
10057 2014-01-15 Andrew Pinski <apinski@cavium.com>
10059 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
10060 of moving from/to the STACK_REG register class.
10062 2014-01-15 Richard Henderson <rth@redhat.com>
10065 * reginfo.c (global_regs_decl): Globalize.
10066 * rtl.h (global_regs_decl): Declare.
10067 * ira.c (do_reload): Diagnose frame_pointer_needed and it
10068 reserved via global_regs.
10070 2014-01-15 Teresa Johnson <tejohnson@google.com>
10072 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
10074 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
10076 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
10077 and vmulosh rather than call gen_vec_widen_smult_*.
10078 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
10079 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
10080 (vec_widen_smult_even_v16qi): Likewise.
10081 (vec_widen_umult_even_v8hi): Likewise.
10082 (vec_widen_smult_even_v8hi): Likewise.
10083 (vec_widen_umult_odd_v16qi): Likewise.
10084 (vec_widen_smult_odd_v16qi): Likewise.
10085 (vec_widen_umult_odd_v8hi): Likewise.
10086 (vec_widen_smult_odd_v8hi): Likewise.
10087 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
10088 vmuloub rather than call gen_vec_widen_umult_*.
10089 (vec_widen_umult_lo_v16qi): Likewise.
10090 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
10091 vmulosb rather than call gen_vec_widen_smult_*.
10092 (vec_widen_smult_lo_v16qi): Likewise.
10093 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
10094 rather than call gen_vec_widen_umult_*.
10095 (vec_widen_umult_lo_v8hi): Likewise.
10096 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
10097 rather than call gen_vec_widen_smult_*.
10098 (vec_widen_smult_lo_v8hi): Likewise.
10100 2014-01-15 Jeff Law <law@redhat.com>
10102 PR tree-optimization/59747
10103 * ree.c (find_and_remove_re): Properly handle case where a second
10104 eliminated extension requires widening a copy created for elimination
10105 of a prior extension.
10106 (combine_set_extension): Ensure that the number of hard regs needed
10107 for a destination register does not change when we widen it.
10109 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
10111 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
10112 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
10113 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
10114 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
10115 (avr-*-rtems*): Likewise.
10116 (bfin*-rtems*): Likewise.
10117 (moxie-*-rtems*): Likewise.
10118 (h8300-*-rtems*): Likewise.
10119 (i[34567]86-*-rtems*): Likewise.
10120 (lm32-*-rtems*): Likewise.
10121 (m32r-*-rtems*): Likewise.
10122 (m68k-*-rtems*): Likewise.
10123 (microblaze*-*-rtems*): Likewise.
10124 (mips*-*-rtems*): Likewise.
10125 (powerpc-*-rtems*): Likewise.
10126 (sh-*-rtems*): Likewise.
10127 (sparc-*-rtems*): Likewise.
10128 (sparc64-*-rtems*): Likewise.
10129 (v850-*-rtems*): Likewise.
10130 (m32c-*-rtems*): Likewise.
10132 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
10134 PR rtl-optimization/59511
10135 * ira.c (ira_init_register_move_cost): Use memory costs for some
10136 cases of register move cost calculations.
10137 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
10138 instead of BB frequency.
10139 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
10140 * lra-assigns.c (find_hard_regno_for): Ditto.
10142 2014-01-15 Richard Biener <rguenther@suse.de>
10144 PR tree-optimization/59822
10145 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
10146 (vectorizable_load): Use it to hoist defs of uses of invariant
10147 loads out of the loop.
10149 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
10150 Kugan Vivekanandarajah <kuganv@linaro.org>
10153 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
10156 2014-01-15 Richard Biener <rguenther@suse.de>
10158 PR rtl-optimization/59802
10159 * lcm.c (compute_available): Use inverted postorder to seed
10160 the initial worklist.
10162 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10165 * config/s390/s390.c (s390_preferred_reload_class): Don't return
10166 ADDR_REGS for invalid symrefs in non-PIC code.
10168 2014-01-15 Jakub Jelinek <jakub@redhat.com>
10171 * builtins.c (determine_block_size): Initialize *probable_max_size
10172 even if len_rtx is CONST_INT.
10174 2014-01-14 Andrew Pinski <apinski@cavium.com>
10176 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
10177 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
10178 (cortexa53_tunings): Likewise.
10179 (aarch64_sched_issue_rate): New function.
10180 (TARGET_SCHED_ISSUE_RATE): Define.
10182 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
10184 * ira-costs.c (find_costs_and_classes): Add missed
10185 ira_init_register_move_cost_if_necessary.
10187 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
10190 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
10192 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
10195 * config/i386/i386.c (type_natural_mode): Add a bool parameter
10196 to indicate if type is used for function return value. Warn ABI
10197 change if the vector mode isn't available for function return value.
10198 (ix86_function_arg_advance): Pass false to type_natural_mode.
10199 (ix86_function_arg): Likewise.
10200 (ix86_gimplify_va_arg): Likewise.
10201 (function_arg_32): Don't warn ABI change.
10202 (ix86_function_value): Pass true to type_natural_mode.
10203 (ix86_return_in_memory): Likewise.
10204 (ix86_struct_value_rtx): Removed.
10205 (TARGET_STRUCT_VALUE_RTX): Likewise.
10207 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
10209 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
10210 converting a conditional jump into a conditional return.
10212 2014-01-14 Richard Biener <rguenther@suse.de>
10214 PR tree-optimization/58921
10215 PR tree-optimization/59006
10216 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
10217 hoisting invariant stmts.
10218 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
10219 invariant loads on the preheader edge if possible.
10221 2014-01-14 Joey Ye <joey.ye@arm.com>
10223 * doc/plugin.texi (Building GCC plugins): Update to C++.
10225 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
10227 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
10228 (_mm_rcp28_round_ss): Ditto.
10229 (_mm_rsqrt28_round_sd): Ditto.
10230 (_mm_rsqrt28_round_ss): Ditto.
10231 (_mm_rcp28_sd): Ditto.
10232 (_mm_rcp28_ss): Ditto.
10233 (_mm_rsqrt28_sd): Ditto.
10234 (_mm_rsqrt28_ss): Ditto.
10235 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
10236 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
10237 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
10238 (IX86_BUILTIN_RCP28SD): Ditto.
10239 (IX86_BUILTIN_RCP28SS): Ditto.
10240 (IX86_BUILTIN_RSQRT28SD): Ditto.
10241 (IX86_BUILTIN_RSQRT28SS): Ditto.
10242 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
10243 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
10244 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
10245 (ix86_expand_special_args_builtin): Expand new FTYPE.
10246 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
10247 (srcp14<mode>): Make insn unary.
10248 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
10249 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
10250 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
10251 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
10252 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
10253 Fix rounding: make it SAE only.
10254 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
10256 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
10258 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
10259 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
10260 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
10261 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
10262 (round_saeonly_mask_scalar_operand4): Ditto.
10263 (round_saeonly_mask_scalar_op3): Ditto.
10264 (round_saeonly_mask_scalar_op4): Ditto.
10266 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10268 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
10269 Implement -maltivec=be for vec_insert and vec_extract.
10271 2014-01-10 DJ Delorie <dj@redhat.com>
10273 * config/msp430/msp430.md (call_internal): Don't allow memory
10274 references with SP as the base register.
10275 (call_value_internal): Likewise.
10276 * config/msp430/constraints.md (Yc): New. For memory references
10277 that don't use SP as a base register.
10279 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
10280 "an integer without a # prefix"
10281 * config/msp430/msp430.md (epilogue_helper): Use it.
10283 2014-01-13 Jakub Jelinek <jakub@redhat.com>
10286 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
10287 AVX512F gather builtins.
10288 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
10289 on gather decls with INTEGER_TYPE masktype.
10290 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
10291 directly into the builtin rather than hoisting it before loop.
10293 PR tree-optimization/59387
10294 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
10295 (scev_const_prop): If folded_casts and type has undefined overflow,
10296 use force_gimple_operand instead of force_gimple_operand_gsi and
10297 for each added stmt if it is assign with
10298 arith_code_with_undefined_signed_overflow, call
10299 rewrite_to_defined_overflow.
10300 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
10301 gimple-fold.h instead.
10302 (arith_code_with_undefined_signed_overflow,
10303 rewrite_to_defined_overflow): Moved to ...
10304 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
10305 rewrite_to_defined_overflow): ... here. No longer static.
10306 Include gimplify-me.h.
10307 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
10308 rewrite_to_defined_overflow): New prototypes.
10310 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10312 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
10314 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
10316 * builtins.c (get_object_alignment_2): Minor tweak.
10317 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
10319 2014-01-13 Christian Bruel <christian.bruel@st.com>
10321 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
10322 optimized non constant lengths.
10324 2014-01-13 Jakub Jelinek <jakub@redhat.com>
10327 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
10328 load as __atomic_load_N if possible.
10330 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
10332 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
10334 (rs6000_expand_builtin): Adjust call.
10336 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
10339 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
10340 * config/rs6000/rs6000.c: Include target-globals.h.
10341 (rs6000_set_current_function): Instead of doing target_reinit
10342 unconditionally, use save_target_globals_default_opts and
10343 restore_target_globals.
10345 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
10347 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
10348 (rs6000_expand_builtin): Handle mffs and mtfsf.
10349 (rs6000_init_builtins): Define mffs and mtfsf.
10350 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
10351 (rs6000_mffs): New pattern.
10352 (rs6000_mtfsf): New pattern.
10354 2014-01-11 Bin Cheng <bin.cheng@arm.com>
10356 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
10357 Start narrowing with START. Apply candidate-use pair
10358 and check overall cost in narrowing.
10359 (iv_ca_prune): Pass new argument.
10361 2014-01-10 Jeff Law <law@redhat.com>
10363 PR middle-end/59743
10364 * ree.c (combine_reaching_defs): Ensure the defining statement
10365 occurs before the extension when optimizing extensions with
10366 different source and destination hard registers.
10368 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
10371 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
10372 vtables into the type inheritance graph.
10374 2014-01-10 Jakub Jelinek <jakub@redhat.com>
10376 PR rtl-optimization/59754
10377 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
10378 modes in the REGNO != REGNO case.
10380 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10382 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
10384 2014-01-10 Jakub Jelinek <jakub@redhat.com>
10386 PR tree-optimization/59745
10387 * tree-predcom.c (tree_predictive_commoning_loop): Call
10388 free_affine_expand_cache if giving up because components is NULL.
10390 * target-globals.c (save_target_globals): Allocate < 4KB structs using
10391 GC in payload of target_globals struct instead of allocating them on
10392 the heap and the larger structs separately using GC.
10393 * target-globals.h (struct target_globals): Make regs, hard_regs,
10394 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
10395 of GTY((skip)) and change type to void *.
10396 (reset_target_globals): Cast loads from those fields to corresponding
10399 2014-01-10 Steve Ellcey <sellcey@mips.com>
10402 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
10403 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
10404 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
10406 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
10409 * aarch64-modes.def (CC_Zmode): New flags mode.
10410 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
10411 represents an equality.
10412 (aarch64_get_condition_code): Handle CC_Zmode.
10413 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
10415 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10417 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
10418 extraction in good case.
10420 2014-01-10 Richard Biener <rguenther@suse.de>
10422 PR tree-optimization/59374
10423 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
10424 checking after SLP discovery. Mark stmts not participating
10425 in any SLP instance properly.
10427 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10429 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
10430 when handling a SET rtx.
10432 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10434 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
10435 (cortex-a57): Likewise.
10436 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
10438 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10440 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
10441 non-iwmmxt builtins.
10443 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
10447 * ipa-devirt.c record_target_from_binfo): Take as argument
10448 stack of binfos and lookup matching one for virtual inheritance.
10449 (possible_polymorphic_call_targets_1): Update.
10451 2014-01-10 Huacai Chen <chenhc@lemote.com>
10453 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
10454 kernel strings for Loongson-2E/2F/3A.
10456 2014-01-10 Jakub Jelinek <jakub@redhat.com>
10458 PR middle-end/59670
10459 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
10460 is_gimple_call before calling gimple_call_internal_p.
10462 2014-01-09 Steve Ellcey <sellcey@mips.com>
10464 * Makefile.in (TREE_FLOW_H): Remove.
10465 (TREE_SSA_H): Add file names from tree-flow.h.
10466 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
10467 * tree.h: Remove tree-flow.h reference.
10468 * hash-table.h: Remove tree-flow.h reference.
10469 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
10470 reference with tree-ssa-loop.h.
10472 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10474 * doc/invoke.texi: Add -maltivec={be,le} options, and document
10475 default element-order behavior for -maltivec.
10476 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
10477 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
10478 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
10479 when targeting big endian, at least for now.
10480 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
10482 2014-01-09 Jakub Jelinek <jakub@redhat.com>
10484 PR middle-end/47735
10485 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
10486 var satisfies use_register_for_decl, just take into account type
10487 alignment, rather than decl alignment.
10489 PR tree-optimization/59622
10490 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
10491 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
10492 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
10493 Don't devirtualize for inplace at all. For targets.length () == 1,
10494 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
10496 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
10498 * config/i386/i386.md (cpu): Remove the unused btver1.
10500 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
10502 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
10504 2014-01-09 Jakub Jelinek <jakub@redhat.com>
10507 * tree-core.h (struct target_globals): New forward declaration.
10508 (struct tree_target_option): Add globals field.
10509 * tree.h (TREE_TARGET_GLOBALS): Define.
10510 (prepare_target_option_nodes_for_pch): New prototype.
10511 * target-globals.h (struct target_globals): Define even if
10512 !SWITCHABLE_TARGET.
10513 * tree.c (prepare_target_option_node_for_pch,
10514 prepare_target_option_nodes_for_pch): New functions.
10515 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
10516 * config/i386/i386.c: Include target-globals.h.
10517 (ix86_set_current_function): Instead of doing target_reinit
10518 unconditionally, use save_target_globals_default_opts and
10519 restore_target_globals.
10521 2014-01-09 Richard Biener <rguenther@suse.de>
10523 PR tree-optimization/59715
10524 * tree-cfg.h (split_critical_edges): Declare.
10525 * tree-cfg.c (split_critical_edges): Export.
10526 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
10528 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
10530 * cfgexpand.c (expand_stack_vars): Optionally disable
10531 asan stack protection.
10532 (expand_used_vars): Likewise.
10533 (partition_stack_vars): Likewise.
10534 * asan.c (asan_emit_stack_protection): Optionally disable
10535 after return stack usage.
10536 (instrument_derefs): Optionally disable memory access instrumentation.
10537 (instrument_builtin_call): Likewise.
10538 (instrument_strlen_call): Likewise.
10539 (asan_protect_global): Optionally disable global variables protection.
10540 * doc/invoke.texi: Added doc for new options.
10541 * params.def: Added new options.
10542 * params.h: Likewise.
10544 2014-01-09 Jakub Jelinek <jakub@redhat.com>
10546 PR rtl-optimization/59724
10547 * ifcvt.c (cond_exec_process_if_block): Don't call
10548 flow_find_head_matching_sequence with 0 longest_match.
10549 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
10550 non-active insns if !stop_after.
10551 (try_head_merge_bb): Revert 2014-01-07 changes.
10553 2014-01-08 Jeff Law <law@redhat.com>
10555 * ree.c (get_sub_rtx): New function, extracted from...
10556 (merge_def_and_ext): Here.
10557 (combine_reaching_defs): Use get_sub_rtx.
10559 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
10561 * cgraph.h (varpool_variable_node): Do not choke on null node.
10563 2014-01-08 Catherine Moore <clm@codesourcery.com>
10565 * config/mips/mips.md (simple_return): Attempt to use JRC
10567 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
10569 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
10571 PR rtl-optimization/59137
10572 * reorg.c (steal_delay_list_from_target): Call update_block for
10574 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
10576 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10578 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
10579 two duplicate entries.
10581 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
10584 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
10586 * config/mips/mips.c (mips_truncated_op_cost): New function.
10587 (mips_rtx_costs): Adjust test for BADDU.
10588 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
10590 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
10592 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
10593 (*baddu_si): ...this new pattern.
10595 2014-01-08 Jakub Jelinek <jakub@redhat.com>
10598 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
10600 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
10602 PR middle-end/57748
10603 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
10605 (expand_expr, expand_normal): Adjust.
10606 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
10607 inner_reference_p. Use inner_reference_p to expand inner references.
10608 (store_expr): Adjust.
10609 * cfgexpand.c (expand_call_stmt): Adjust.
10611 2014-01-08 Rong Xu <xur@google.com>
10613 * gcov-io.c (gcov_var): Move from gcov-io.h.
10614 (gcov_position): Ditto.
10615 (gcov_is_error): Ditto.
10616 (gcov_rewrite): Ditto.
10617 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
10618 only part to libgcc/libgcov.h.
10620 2014-01-08 Marek Polacek <polacek@redhat.com>
10622 PR middle-end/59669
10623 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
10625 2014-01-08 Marek Polacek <polacek@redhat.com>
10628 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
10630 2014-01-08 Jakub Jelinek <jakub@redhat.com>
10632 PR rtl-optimization/59649
10633 * stor-layout.c (get_mode_bounds): For BImode return
10634 0 and STORE_FLAG_VALUE.
10636 2014-01-08 Richard Biener <rguenther@suse.de>
10638 PR middle-end/59630
10639 * gimple.h (is_gimple_builtin_call): Remove.
10640 (gimple_builtin_call_types_compatible_p): New.
10641 (gimple_call_builtin_p): New overload.
10642 * gimple.c (is_gimple_builtin_call): Remove.
10643 (validate_call): Rename to ...
10644 (gimple_builtin_call_types_compatible_p): ... this and export. Also
10645 check return types.
10646 (validate_type): New static function.
10647 (gimple_call_builtin_p): New overload and adjust.
10648 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
10649 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
10650 (gimple_fold_stmt_to_constant_1): Likewise.
10651 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
10653 2014-01-08 Richard Biener <rguenther@suse.de>
10655 PR middle-end/59471
10656 * gimplify.c (gimplify_expr): Gimplify register-register type
10657 VIEW_CONVERT_EXPRs to separate stmts.
10659 2014-01-07 Jeff Law <law@redhat.com>
10661 PR middle-end/53623
10662 * ree.c (combine_set_extension): Handle case where source
10663 and destination registers in an extension insn are different.
10664 (combine_reaching_defs): Allow source and destination registers
10665 in extension to be different under limited circumstances.
10666 (add_removable_extension): Remove restriction that the
10667 source and destination registers in the extension are the same.
10668 (find_and_remove_re): Emit a copy from the extension's
10669 destination to its source after the defining insn if
10670 the source and destination registers are different.
10672 PR middle-end/59285
10673 * ifcvt.c (merge_if_block): If we are merging a block with more than
10674 one successor with a block with no successors, remove any BARRIER
10675 after the second block.
10677 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
10679 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
10681 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
10684 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
10685 for 14-bit register offsets when INT14_OK_STRICT is false.
10687 2014-01-07 Roland Stigge <stigge@antcom.de>
10688 Michael Meissner <meissner@linux.vnet.ibm.com>
10691 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
10692 Only check TFmode for SPE constants. Don't check TImode or TDmode.
10694 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
10696 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
10699 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
10701 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
10702 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
10703 rtx is const0_rtx or not.
10705 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
10708 * target-globals.c (save_target_globals): Remove this_fn_optab
10710 * toplev.c: Include optabs.h.
10711 (target_reinit): Temporarily restore the global options if another
10712 set of options are in force.
10714 2014-01-07 Jakub Jelinek <jakub@redhat.com>
10716 PR rtl-optimization/58668
10717 * cfgcleanup.c (flow_find_cross_jump): Don't count
10718 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
10719 to determine what is counted.
10720 (flow_find_head_matching_sequence): Use active_insn_p to determine
10722 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
10724 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
10725 determine what is counted.
10727 PR tree-optimization/59643
10728 * tree-predcom.c (split_data_refs_to_components): If one dr is
10729 read and one write, determine_offset fails and the write isn't
10730 in the bad component, just put the read into the bad component.
10732 2014-01-07 Mike Stump <mikestump@comcast.net>
10733 Jakub Jelinek <jakub@redhat.com>
10736 * tree-core.h (struct tree_optimization_option): Change optabs
10737 type from unsigned char * to void *.
10738 * optabs.c (init_tree_optimization_optabs): Adjust
10739 TREE_OPTIMIZATION_OPTABS initialization.
10741 2014-01-06 Jakub Jelinek <jakub@redhat.com>
10744 * config/i386/i386.h (struct machine_function): Add
10745 no_drap_save_restore field.
10746 * config/i386/i386.c (ix86_save_reg): Use
10747 !cfun->machine->no_drap_save_restore instead of
10748 crtl->stack_realign_needed.
10749 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
10750 this function clears frame_pointer_needed. Set
10751 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
10752 and DRAP reg is needed.
10754 2014-01-06 Marek Polacek <polacek@redhat.com>
10757 * doc/implement-c.texi: Mention that other integer types are
10758 permitted as bit-field types in strictly conforming mode.
10760 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
10762 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
10763 is newly allocated.
10765 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
10767 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
10769 2014-01-06 Martin Jambor <mjambor@suse.cz>
10772 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
10774 * ipa-prop.c (ipa_print_node_params): Fix indentation.
10776 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
10780 * var-tracking.c (add_stores): Preserve the value of the source even if
10781 we don't record the store.
10783 2014-01-06 Terry Guo <terry.guo@arm.com>
10785 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
10787 2014-01-05 Iain Sandoe <iain@codesourcery.com>
10790 * config/darwin.c (darwin_function_section): Adjust return values to
10791 correspond to optimisation changes made in r206070.
10793 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
10795 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
10796 from prefetch_block tune setting.
10797 (nocona_cost): Correct size of prefetch block to 64.
10799 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
10801 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
10802 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
10803 used to save the static chain register in the computation of the offset
10804 from which the FP registers need to be restored.
10806 2014-01-04 Jakub Jelinek <jakub@redhat.com>
10808 PR tree-optimization/59519
10809 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
10810 ICE if get_current_def (current_new_name) is already non-NULL, as long
10811 as it is a phi result of some other phi in *new_exit_bb that has
10814 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
10815 or vmovdqu* for misaligned_operand.
10816 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
10817 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
10818 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
10819 aligned_mem for AVX512F masked aligned load and store builtins and for
10820 non-temporal moves.
10822 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
10824 PR tree-optimization/59651
10825 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
10826 Address range for negative step should be added by TYPE_SIZE_UNIT.
10828 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
10830 * config/m68k/m68k.c (handle_move_double): Handle pushes with
10831 overlapping registers also for registers other than the stack pointer.
10833 2014-01-03 Marek Polacek <polacek@redhat.com>
10836 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
10839 2014-01-03 Jakub Jelinek <jakub@redhat.com>
10842 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
10845 * config/i386/i386.md (MODE_SIZE): New mode attribute.
10846 (push splitter): Use <P:MODE_SIZE> instead of
10847 GET_MODE_SIZE (<P:MODE>mode).
10848 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
10849 (mov -1, reg peephole2): Likewise.
10850 * config/i386/sse.md (*mov<mode>_internal,
10851 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
10852 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
10853 *<code><mode>3, *andnot<mode>3<mask_name>,
10854 <mask_codefor><code><mode>3<mask_name>): Likewise.
10855 * config/i386/subst.md (mask_mode512bit_condition,
10856 sd_mask_mode512bit_condition): Likewise.
10858 2014-01-02 Xinliang David Li <davidxl@google.com>
10860 PR tree-optimization/59303
10861 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
10862 (dump_predicates): Better output format.
10863 (pred_equal_p): New function.
10864 (is_neq_relop_p): Ditto.
10865 (is_neq_zero_form_p): Ditto.
10866 (pred_expr_equal_p): Ditto.
10867 (pred_neg_p): Ditto.
10868 (simplify_pred): Ditto.
10869 (simplify_preds_2): Ditto.
10870 (simplify_preds_3): Ditto.
10871 (simplify_preds_4): Ditto.
10872 (simplify_preds): Ditto.
10873 (push_pred): Ditto.
10874 (push_to_worklist): Ditto.
10875 (get_pred_info_from_cmp): Ditto.
10876 (is_degenerated_phi): Ditto.
10877 (normalize_one_pred_1): Ditto.
10878 (normalize_one_pred): Ditto.
10879 (normalize_one_pred_chain): Ditto.
10880 (normalize_preds): Ditto.
10881 (normalize_cond_1): Remove function.
10882 (normalize_cond): Ditto.
10883 (is_gcond_subset_of): Ditto.
10884 (is_subset_of_any): Ditto.
10885 (is_or_set_subset_of): Ditto.
10886 (is_and_set_subset_of): Ditto.
10887 (is_norm_cond_subset_of): Ditto.
10888 (pred_chain_length_cmp): Ditto.
10889 (convert_control_dep_chain_into_preds): Type change.
10890 (find_predicates): Ditto.
10891 (find_def_preds): Ditto.
10892 (destroy_predicates_vecs): Ditto.
10893 (find_matching_predicates_in_rest_chains): Ditto.
10894 (use_pred_not_overlap_with_undef_path_pred): Ditto.
10895 (is_pred_expr_subset): Ditto.
10896 (is_pred_chain_subset_of): Ditto.
10897 (is_included_in): Ditto.
10898 (is_superset_of): Ditto.
10900 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
10902 Update copyright years.
10904 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
10906 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
10907 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
10908 config/arc/arc.md, config/arc/arc.opt,
10909 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
10910 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
10911 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
10912 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
10913 config/linux-protos.h, config/linux.c, config/winnt-c.c,
10914 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
10915 vtable-verify.c, vtable-verify.h: Use the standard form for the
10918 2014-01-02 Tobias Burnus <burnus@net-b.de>
10920 * gcc.c (process_command): Update copyright notice dates.
10921 * gcov-dump.c: Ditto.
10923 * doc/cpp.texi: Bump @copying's copyright year.
10924 * doc/cppinternals.texi: Ditto.
10925 * doc/gcc.texi: Ditto.
10926 * doc/gccint.texi: Ditto.
10927 * doc/gcov.texi: Ditto.
10928 * doc/install.texi: Ditto.
10929 * doc/invoke.texi: Ditto.
10931 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
10933 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
10935 2014-01-01 Jakub Jelinek <jakub@redhat.com>
10937 * config/i386/sse.md (*mov<mode>_internal): Guard
10938 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
10940 PR rtl-optimization/59647
10941 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
10942 new_rtx into UNSIGNED_FLOAT rtxes.
10944 Copyright (C) 2014 Free Software Foundation, Inc.
10946 Copying and distribution of this file, with or without modification,
10947 are permitted in any medium without royalty provided the copyright
10948 notice and this notice are preserved.