1 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
4 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost
5 ZERO_EXTEND and SIGN_EXTEND better.
7 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
8 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
10 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve cost for
13 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
14 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
16 * config/aarch64/aarch64.c (aarch64_rtx_costs): Use address
17 costs when costing loads and stores to memory.
19 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
20 Philip Tomsich <philipp.tomsich@theobroma-systems.com>
22 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costing
25 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
27 * config/aarch64/aarch64.c (aarch64_rtx_costs): Set default costs.
29 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
30 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
32 * config/aarch64/aarch64.c (aarch64_strip_shift_or_extend): Rename
34 (aarch64_strip_extend): ...this, don't strip shifts, check RTX is
36 (aarch64_rtx_mult_cost): New.
37 (aarch64_rtx_costs): Use it, refactor as appropriate.
39 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
40 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
42 * config/aarch64/aarch64.c (aarch64_build_constant): Conditionally
43 emit instructions, return number of instructions which would
45 (aarch64_add_constant): Update call to aarch64_build_constant.
46 (aarch64_output_mi_thunk): Likewise.
47 (aarch64_rtx_costs): Estimate cost of a CONST_INT, cost
50 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
52 * config/aarch64/aarch64.c (aarch64_rtx_costs_wrapper): New.
53 (TARGET_RTX_COSTS): Call it.
55 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
57 * config/aarch64/aarch64.c (cortexa57_addrcost_table): New.
58 (cortexa57_vector_cost): Likewise.
59 (cortexa57_tunings): Use them.
61 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
63 * config/aarch64/aarch64-protos.h (scale_addr_mode_cost): New.
64 (cpu_addrcost_table): Use it.
65 * config/aarch64/aarch64.c (generic_addrcost_table): Initialize it.
66 (aarch64_address_cost): Rewrite using aarch64_classify_address,
69 2014-05-16 Richard Biener <rguenther@suse.de>
71 * tree-ssa-sccvn.c: Include tree-cfg.h and domwalk.h.
72 (set_ssa_val_to): Handle unexpected sets to VN_TOP.
73 (visit_phi): Ignore edges marked as not executable.
74 (class cond_dom_walker): New.
75 (cond_dom_walker::before_dom_children): Value-number
76 control statements and mark successor edges as not
77 executable if possible.
78 (run_scc_vn): First walk all control statements in
79 dominator order, marking edges as not executable.
80 * tree-inline.c (copy_edges_for_bb): Be not confused
81 about random edge flags.
83 2014-05-16 Richard Biener <rguenther@suse.de>
85 * tree-ssa-sccvn.c (visit_use): Also constant-fold calls.
87 2014-05-15 Peter Bergner <bergner@vnet.ibm.com>
90 * config/rs6000/htmxlintrin.h (_HTM_TBEGIN_STARTED): New define.
91 (__TM_simple_begin): Use it.
92 (__TM_begin): Likewise.
94 2014-05-15 Martin Jambor <mjambor@suse.cz>
97 * ipa-prop.c (update_indirect_edges_after_inlining): Check
98 type_preserved flag when the indirect edge is polymorphic.
100 2014-05-15 Martin Jambor <mjambor@suse.cz>
102 PR tree-optimization/61090
103 * tree-sra.c (sra_modify_expr): Pass the current gsi to
106 2014-05-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
108 * config/arm/arm.c (arm_option_override): Use the SCHED_PRESSURE_MODEL
109 enum name for PARAM_SCHED_PRESSURE_ALGORITHM.
111 2014-05-15 Jakub Jelinek <jakub@redhat.com>
113 PR tree-optimization/61158
114 * fold-const.c (fold_binary_loc): If X is zero-extended and
115 shiftc >= prec, make sure zerobits is all ones instead of
116 invoking undefined behavior.
118 2014-05-15 Zhenqiang Chen <zhenqiang.chen@linaro.org>
120 * regcprop.h: New file.
121 * regcprop.c (skip_debug_insn_p): New decl.
122 (replace_oldest_value_reg): Check skip_debug_insn_p.
123 (copyprop_hardreg_forward_bb_without_debug_insn): New function.
124 * shrink-wrap.c: Include regcprop.h.
125 (prepare_shrink_wrap): Call
126 copyprop_hardreg_forward_bb_without_debug_insn.
128 2014-05-15 Zhenqiang Chen <zhenqiang.chen@linaro.org>
130 * shrink-wrap.h: Update comment.
131 * shrink-wrap.c: Update comment.
132 (next_block_for_reg): Rename to live_edge_for_reg.
133 (live_edge_for_reg): Allow live_edge->dest has two predecessors.
134 (move_insn_for_shrink_wrap): Split live_edge.
135 (prepre_shrink_wrap): One more parameter for move_insn_for_shrink_wrap.
137 2014-05-14 Eric Botcazou <ebotcazou@adacore.com>
139 * config/sparc/sparc-protos.h (sparc_absnegfloat_split_legitimate):
141 * config/sparc/sparc.c (sparc_absnegfloat_split_legitimate): Likewise.
142 * config/sparc/sparc.md (fptype_ut699): New attribute.
143 (in_branch_delay): Return false if -mfix-ut699 is specified and
144 fptype_ut699 is set to single.
145 (truncdfsf2): Add fptype_ut699 attribute.
146 (fix_truncdfsi2): Likewise.
147 (floatsisf2): Change fptype attribute.
148 (fix_truncsfsi2): Likewise.
149 (negtf2_notv9): Delete.
150 (negtf2_v9): Likewise.
151 (negtf2_hq): New instruction.
152 (negtf2): New instruction and splitter.
153 (negdf2_notv9): Rewrite.
154 (abstf2_notv9): Delete.
155 (abstf2_hq_v9): Likewise.
156 (abstf2_v9): Likewise.
157 (abstf2_hq): New instruction.
158 (abstf2): New instruction and splitter.
159 (absdf2_notv9): Rewrite.
161 2014-05-14 Cary Coutant <ccoutant@google.com>
164 * opts.c (common_handle_option): Don't special-case "-g".
165 (set_debug_level): Default to at least level 2 with "-g".
167 2014-05-14 DJ Delorie <dj@redhat.com>
169 * config/msp430/msp430.c (msp430_builtin): Add
170 MSP430_BUILTIN_DELAY_CYCLES.
171 (msp430_init_builtins): Register void __delay_cycles(long long).
172 (msp430_builtin_decl): Add it.
173 (cg_magic_constant): New.
174 (msp430_expand_delay_cycles): New.
175 (msp430_expand_builtin): Call it.
176 (msp430_print_operand_raw): Change integer printing from "int" to
178 * config/msp430/msp430.md (define_constants): Add delay_cycles tags.
179 (delay_cycles_start): New.
180 (delay_cycles_end): New.
181 (delay_cycles_32): New.
182 (delay_cycles_32x): New.
183 (delay_cycles_16): New.
184 (delay_cycles_16x): New.
185 (delay_cycles_2): New.
186 (delay_cycles_1): New.
187 * doc/extend.texi: Document __delay_cycles().
189 2014-05-14 Sandra Loosemore <sandra@codesourcery.com>
191 * config/nios2/nios2.md (nios2_cbranch): Fix paste-o in
192 length attribute computation.
194 2014-05-14 Richard Sandiford <rdsandiford@googlemail.com>
197 * print-rtl.c (print_rtx): Suppress uids if flag_dump_unnumbered.
199 2014-05-14 Richard Sandiford <r.sandiford@uk.ibm.com>
202 * config/sparc/sparc.md: Fix types of low and high in DI constant
203 splitter. Use gen_int_mode in some other splitters.
205 2014-05-14 Martin Jambor <mjambor@suse.cz>
208 * ipa-prop.c (ipa_modify_formal_parameters): Reset DECL_LANG_SPECIFIC.
210 2014-05-14 James Norris <jnorris@codesourcery.com>
212 * omp-low.c (expand_parallel_call): Remove shadow variable.
213 (expand_omp_taskreg): Likewise.
215 2014-05-14 Ilya Tocar <ilya.tocar@intel.com>
217 * common/config/i386/i386-common.c
218 (OPTION_MASK_ISA_CLFLUSHOPT_SET): Define.
219 (OPTION_MASK_ISA_XSAVES_SET): Ditto.
220 (OPTION_MASK_ISA_XSAVEC_SET): Ditto.
221 (OPTION_MASK_ISA_CLFLUSHOPT_UNSET): Ditto.
222 (OPTION_MASK_ISA_XSAVES_UNSET): Ditto.
223 (OPTION_MASK_ISA_XSAVEC_UNSET): Ditto.
224 (ix86_handle_option): Handle OPT_mxsavec, OPT_mxsaves,
226 * config.gcc (i[34567]86-*-*): Add clflushoptintrin.h,
227 xsavecintrin.h, xsavesintrin.h.
229 * config/i386/clflushoptintrin.h: New.
230 * config/i386/xsavecintrin.h: Ditto.
231 * config/i386/xsavesintrin.h: Ditto.
232 * config/i386/cpuid.h (bit_CLFLUSHOPT): Define.
235 * config/i386/driver-i386.c (host_detect_local_cpu): Handle
236 -mclflushopt, -mxsavec, -mxsaves, -mno-xsaves, -mno-xsavec,
238 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
239 OPTION_MASK_ISA_CLFLUSHOPT, OPTION_MASK_ISA_XSAVEC,
240 OPTION_MASK_ISA_XSAVES.
241 * config/i386/i386.c (ix86_target_string): Handle -mclflushopt,
243 (PTA_CLFLUSHOPT) Define.
246 (ix86_option_override_internal): Handle new options.
247 (ix86_valid_target_attribute_inner_p): Ditto.
248 (ix86_builtins): Add IX86_BUILTIN_XSAVEC, IX86_BUILTIN_XSAVEC64,
249 IX86_BUILTIN_XSAVES, IX86_BUILTIN_XRSTORS, IX86_BUILTIN_XSAVES64,
250 IX86_BUILTIN_XRSTORS64, IX86_BUILTIN_CLFLUSHOPT.
251 (bdesc_special_args): Add __builtin_ia32_xsaves, __builtin_ia32_xrstors,
252 __builtin_ia32_xsavec, __builtin_ia32_xsaves64, __builtin_ia32_xrstors64,
253 __builtin_ia32_xsavec64.
254 (ix86_init_mmx_sse_builtins): Add __builtin_ia32_clflushopt.
255 (ix86_expand_builtin): Handle new builtins.
256 * config/i386/i386.h (TARGET_CLFLUSHOPT) Define.
257 (TARGET_CLFLUSHOPT_P): Ditto.
258 (TARGET_XSAVEC): Ditto.
259 (TARGET_XSAVEC_P): Ditto.
260 (TARGET_XSAVES): Ditto.
261 (TARGET_XSAVES_P): Ditto.
262 * config/i386/i386.md (ANY_XSAVE): Add UNSPECV_XSAVEC, UNSPECV_XSAVES.
263 (ANY_XSAVE64)" Add UNSPECV_XSAVEC64, UNSPECV_XSAVES64.
264 (attr xsave): Add xsavec, xsavec64, xsaves, xsaves64.
266 (ANY_XRSTOR64): Ditto.
268 (xrstor): Change into <xrstor>.
269 (xrstor_rex64): Change into <xrstor>_rex64.
270 (xrstor64): Change into <xrstor>64
272 * config/i386/i386.opt (mclflushopt): New.
275 * config/i386/x86intrin.h: Add clflushoptintrin.h, xsavesintrin.h,
277 * doc/invoke.texi: Document new options.
279 2014-05-14 Andrey Belevantsev <abel@ispras.ru>
281 PR rtl-optimization/60866
282 * sel-sched-ir (sel_init_new_insn): New parameter old_seqno.
283 Default it to -1. Pass it down to init_simplejump_data.
284 (init_simplejump_data): New parameter old_seqno. Pass it down
285 to get_seqno_for_a_jump.
286 (get_seqno_for_a_jump): New parameter old_seqno. Use it for
287 initializing new jump seqno as a last resort. Add comment.
288 (sel_redirect_edge_and_branch): Save old seqno of the conditional
289 jump and pass it down to sel_init_new_insn.
290 (sel_redirect_edge_and_branch_force): Likewise.
292 2014-05-14 Georg-Johann Lay <avr@gjlay.de>
294 * config/avr/avr.h (REG_CLASS_CONTENTS): Use unsigned suffix for
295 shifted values to avoid build warning.
297 2014-05-14 Eric Botcazou <ebotcazou@adacore.com>
299 * cfgcleanup.c (try_forward_edges): Use location_t for locations.
300 * cfgrtl.c (rtl_merge_blocks): Fix comment.
301 (cfg_layout_merge_blocks): Likewise.
302 * except.c (emit_to_new_bb_before): Remove prev_bb local variable.
304 2014-05-14 Andrey Belevantsev <abel@ispras.ru>
306 PR rtl-optimization/60901
307 * config/i386/i386.c (ix86_dependencies_evaluation_hook): Check that
308 bb predecessor belongs to the same scheduling region. Adjust comment.
310 2014-05-13 Peter Bergner <bergner@vnet.ibm.com>
312 * doc/sourcebuild.texi: (dfp_hw): Document.
313 (p8vector_hw): Likewise.
314 (powerpc_eabi_ok): Likewise.
315 (powerpc_elfv2): Likewise.
316 (powerpc_htm_ok): Likewise.
317 (ppc_recip_hw): Likewise.
320 2014-05-13 Cary Coutant <ccoutant@google.com>
322 * opts.c (finish_options): Use -ggnu-pubnames with -gsplit-dwarf.
324 2014-05-13 David Malcolm <dmalcolm@redhat.com>
326 * gengtype-parse.c (require3): Eliminate in favor of...
328 (require_template_declaration): Update to support optional single *
331 * gengtype.c (get_ultimate_base_class): Add a non-const overload.
332 (create_user_defined_type): Handle a single level of explicit
333 pointerness within template arguments.
334 (struct write_types_data): Add field "kind".
335 (filter_type_name): Handle "*" character.
336 (write_user_func_for_structure_ptr): Require a write_types_data
337 rather than just a prefix string, so that we can look up the kind
338 of the wtd and use it as an index into wrote_user_func_for_ptr,
339 ensuring that such functions are written at most once. Support
340 subclasses by invoking the marking function of the ultimate base
342 (write_user_func_for_structure_body): Require a write_types_data
343 rather than just a prefix string, so that we can pass this to
344 write_user_func_for_structure_ptr.
345 (write_func_for_structure): Likewise.
346 (ggc_wtd): Add initializer of new "kind" field.
349 * gengtype.h (enum write_types_kinds): New.
350 (struct type): Add field wrote_user_func_for_ptr to the "s"
353 2014-05-13 Richard Sandiford <r.sandiford@uk.ibm.com>
355 * fold-const.c (optimize_bit_field_compare): Use wi:: operations
356 instead of const_binop.
357 (fold_binary_loc): Likewise.
359 2014-05-13 Richard Sandiford <r.sandiford@uk.ibm.com>
361 * tree-dfa.h (get_addr_base_and_unit_offset_1): Update array index
362 calculation to match get_ref_base_and_extent.
364 2014-05-13 Catherine Moore <clm@codesourcery.com>
365 Sandra Loosemore <sandra@codesourcery.com>
367 * configure.ac: Fix assembly for explicit JALR relocation check.
368 * configure: Regenerate.
370 2014-05-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
372 * config/arm/arm.c (neon_itype): Remove NEON_RESULTPAIR.
373 (arm_init_neon_builtins): Remove handling of NEON_RESULTPAIR.
374 Remove associated type declarations and initialisations.
375 (arm_expand_neon_builtin): Likewise.
376 (neon_emit_pair_result_insn): Delete.
377 * config/arm/arm_neon_builtins (vtrn, vzip, vuzp): Delete.
378 * config/arm/neon.md (neon_vtrn<mode>): Delete.
379 (neon_vzip<mode>): Likewise.
380 (neon_vuzp<mode>): Likewise.
382 2014-05-13 Richard Biener <rguenther@suse.de>
385 * tree-inline.c (remap_gimple_stmt): Clear tail call flag,
386 it needs revisiting whether the call still may be tail-called.
388 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
390 * rtl.def (SYMBOL_REF): Remove middle "0" field.
391 * rtl.h (block_symbol): Reduce number of fields to 2.
392 (rtx_def): Add u2.symbol_ref_flags.
393 (SYMBOL_REF_FLAGS): Use it.
394 (SYMBOL_REF_DATA, SET_SYMBOL_REF_DECL, SYMBOL_REF_DECL)
395 (SET_SYMBOL_REF_CONSTANT, SYMBOL_REF_CONSTANT): Lower index.
396 * gengtype.c (adjust_field_rtx_def): Remove SYMBOL_REF_FLAGS handling.
397 Lower index of SYMBOL_REF_DATA.
398 * print-rtl.c (print_rtx): Lower index for SYMBOL_REF_DATA.
399 Print SYMBOL_REF_FLAGS at the same time.
400 * genattrtab.c (attr_rtx_1): Only initialize 1 "0" SYMBOL_REF field.
402 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
404 * rtl.def (VAR_LOCATION): Remove "i" field.
405 * rtl.h (rtx_def): Add u2.var_location_status.
406 (PAT_VAR_LOCATION_STATUS): Use it.
407 (gen_rtx_VAR_LOCATION): Declare.
408 * gengenrtl.c (excluded_rtx): Add VAR_LOCATION.
409 * emit-rtl.c (gen_rtx_VAR_LOCATION): New function.
410 * var-tracking.c (emit_note_insn_var_location): Remove casts.
412 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
414 * rtl.def (scratch): Fix outdated comment and remove "0" field.
415 * gengtype.c (adjust_field_rtx_def): Update accordingly.
417 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
419 * rtl.def (DEBUG_INSN, INSN, JUMP_INSN, CALL_INSN, JUMP_TABLE_DATA)
420 (BARRIER, CODE_LABEL, NOTE): Remove first "i" field.
421 * rtl.h (rtx_def): Add insn_uid to u2 field.
422 (RTX_FLAG_CHECK8): Delete in favor of...
423 (RTL_INSN_CHAIN_FLAG_CHECK): ...this new macro.
424 (INSN_DELETED_P): Update accordingly.
425 (INSN_UID): Use u2.insn_uid.
426 (INSN_CHAIN_CODE_P): Define.
427 (PREV_INSN, NEXT_INSN, BLOCK_FOR_INSN, PATTERN, INSN_LOCATION)
428 (INSN_CODE, REG_NOTES, CALL_INSN_FUNCTION_USAGE, CODE_LABEL_NUMBER)
429 (NOTE_DATA, NOTE_DELETED_LABEL_NAME, NOTE_BLOCK, NOTE_EH_HANDLER)
430 (NOTE_BASIC_BLOCK, NOTE_VAR_LOCATION, NOTE_CFI, NOTE_LABEL_NUMBER)
431 (NOTE_KIND, LABEL_NAME, LABEL_NUSES, JUMP_LABEL, LABEL_REFS): Lower
433 * print-rtl.c (print_rtx): Print INSN_UIDs before the main loop.
434 Update indices for insn-chain rtxes.
435 * gengtype.c (gen_rtx_next): Adjust test for insn-chain rtxes.
436 (adjust_field_rtx_def): Lower '0' indices for all insn-chain rtxes.
437 * emit-rtl.c (gen_label_rtx): Update gen_rtx_LABEL call.
438 * caller-save.c (init_caller_save): Update gen_rtx_INSN calls.
439 * combine.c (try_combine): Likewise.
440 * ira.c (setup_prohibited_mode_move_regs): Likewise.
442 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
444 * rtl.def (REG): Remove middle field.
445 * rtl.h (rtx_def): Add orignal_regno to u2.
446 (ORIGINAL_REGNO): Use it instead of field 1.
447 (REG_ATTRS): Lower field index accordingly.
448 * gengtype.c (adjust_field_rtx_def): Remove handling of
449 ORIGINAL_REGNO. Move REG_ATTRS index down.
450 * print-rtl.c (print_rtx): Move ORIGINAL_REGNO handling to the
451 code that prints the REGNO.
453 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
455 * print-rtl.c (print_rtx): Guard whole '0' block with ifndef
458 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
460 * rtl.h (rtx_def): Mark u2 as GTY ((skip)).
462 2014-05-13 Bin Cheng <bin.cheng@arm.com>
464 * tree-ssa-loop-ivopts.c (contain_complex_addr_expr): New.
465 (alloc_iv): Lower base expressions containing ADDR_EXPR.
467 2014-05-13 Ian Bolton <ian.bolton@arm.com>
469 * config/aarch64/aarch64-protos.h
470 (aarch64_hard_regno_caller_save_mode): New prototype.
471 * config/aarch64/aarch64.c (aarch64_hard_regno_caller_save_mode):
473 * config/aarch64/aarch64.h (HARD_REGNO_CALLER_SAVE_MODE): New macro.
475 2014-05-13 Christian Bruel <christian.bruel@st.com>
477 * target.def (mode_switching): New hook vector.
478 (mode_emit, mode_needed, mode_after, mode_entry): New hooks.
479 (mode_exit, modepriority_to_mode): Likewise.
480 * mode-switching.c (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Hookify.
481 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
482 * target.h: Include tm.h and hard-reg-set.h.
483 * doc/tm.texi.in (EMIT_MODE_SET, MODE_NEEDED, MODE_AFTER, MODE_ENTRY)
484 (MODE_EXIT, MODE_PRIORITY_TO_MODE): Delete and hookify.
485 * doc/tm.texi Regenerate.
486 * config/sh/sh.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
487 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
488 * config/sh/sh.c (sh_emit_mode_set, sh_mode_priority): Hookify.
489 (sh_mode_needed, sh_mode_after, sh_mode_entry, sh_mode_exit): Likewise.
490 * config/i386/i386.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
491 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
492 * config/i386/i386-protos.h (ix86_mode_needed, ix86_mode_after)
493 (ix86_mode_entrym, ix86_emit_mode_set): Remove external declaration.
494 * config/i386/i386.c (ix86_mode_needed, ix86_mode_after, ix86_mode_exit,
495 (ix86_mode_entry, ix86_mode_priority, ix86_emit_mode_set): Hookify.
496 * config/epiphany/epiphany.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY):
498 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
499 * config/epiphany/epiphany-protos.h (epiphany_mode_needed)
500 (emit_set_fp_mode, epiphany_mode_entry_exit, epiphany_mode_after)
501 (epiphany_mode_priority_to_mode): Remove declaration.
502 * config/epiphany/epiphany.c (emit_set_fp_mode): Hookify.
503 (epiphany_mode_needed, epiphany_mode_priority_to_mode): Likewise.
504 (epiphany_mode_entry, epiphany_mode_exit, epiphany_mode_after):
506 (epiphany_mode_priority_to_mode): Change priority type. Hookify.
507 (epiphany_mode_needed, epiphany_mode_entry_exit): Hookify.
508 (epiphany_mode_after, epiphany_mode_entry, emit_set_fp_mode): Hookify.
510 2014-05-13 Jakub Jelinek <jakub@redhat.com>
513 * config/i386/i386.c (ix86_expand_set_or_movmem): If count_exp
514 is const0_rtx, return immediately. Don't test count == 0 when
517 2014-05-13 Zhenqiang Chen <zhenqiang.chen@linaro.org>
519 * Makefile.in: add shrink-wrap.o.
520 * config/i386/i386.c: include "shrink-wrap.h"
521 * function.c: Likewise.
522 (requires_stack_frame_p, next_block_for_reg,
523 move_insn_for_shrink_wrap, prepare_shrink_wrap,
524 dup_block_and_redirect): Move to shrink-wrap.c
525 (thread_prologue_and_epilogue_insns): Extract three code segments
526 as functions in shrink-wrap.c
527 * function.h: Move #ifdef HAVE_simple_return ... #endif block to
529 * shrink-wrap.c: New file.
530 * shrink-wrap.h: New file.
532 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
534 * doc/extend.texi: Reflect current numbers of pragmas. Remove
535 reference to Solaris.
537 2014-05-12 Mike Stump <mikestump@comcast.net>
540 * genattrtab.c (filename): Add.
541 (convert_set_attr_alternative): Improve error message.
542 (check_defs): Restore read_md_filename for error messages.
543 (gen_insn): Save filename.
545 2014-05-12 Dimitris Papavasiliou <dpapavas@gmail.com>
547 * doc/invoke.texi: Document new switches -Wno-shadow-ivar,
548 -fno-local-ivars and -fivar-visibility.
549 * c-family/c.opt: Make -Wshadow also implicitly enable
552 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
554 * doc/tm.texi: Remove reference to deleted macro.
555 * doc/tm.texi.in: Likewise.
557 2014-05-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
560 * config/avr/avr.c (avr_out_store_psi): Use correct constant
563 2014-05-12 Georg-Johann Lay <avr@gjlay.de>
565 * config/arm/arm.h (License): Add GCC Runtime Library Exception.
566 * config/arm/aout.h (License): Same.
567 * config/arm/bpabi.h (License): Same.
568 * config/arm/elf.h (License): Same.
569 * config/arm/linux-elf.h (License): Same.
570 * config/arm/linux-gas.h (License): Same.
571 * config/arm/netbsd-elf.h (License): Same.
572 * config/arm/uclinux-eabi.h (License): Same.
573 * config/arm/uclinux-elf.h (License): Same.
574 * config/arm/vxworks.h (License): Same.
576 2014-05-11 Jakub Jelinek <jakub@redhat.com>
578 * tree.h (OMP_CLAUSE_LINEAR_STMT): Define.
579 * tree.c (omp_clause_num_ops): Increase OMP_CLAUSE_LINEAR
580 number of operands to 3.
581 (walk_tree_1): Walk all operands of OMP_CLAUSE_LINEAR.
582 * tree-nested.c (convert_nonlocal_omp_clauses,
583 convert_local_omp_clauses): Handle OMP_CLAUSE_DEPEND.
584 * gimplify.c (gimplify_scan_omp_clauses): Handle
585 OMP_CLAUSE_LINEAR_STMT.
586 * omp-low.c (lower_rec_input_clauses): Fix typo.
587 (maybe_add_implicit_barrier_cancel, lower_omp_1): Add
588 cast between Fortran boolean_type_node and C _Bool if
591 2014-05-11 Richard Sandiford <rdsandiford@googlemail.com>
593 PR tree-optimization/61136
594 * wide-int.h (multiple_of_p): Define a version that doesn't return
596 * fold-const.c (extract_muldiv_1): Use wi::multiple_of_p instead of an
597 integer_zerop/const_binop pair.
598 (multiple_of_p): Likewise, converting both operands to widest_int
601 2014-05-09 Teresa Johnson <tejohnson@google.com>
603 * cgraphunit.c (analyze_functions): Use correct dump file.
605 2014-05-09 Florian Weimer <fweimer@redhat.com>
607 * cfgexpand.c (stack_protect_decl_p): New function, extracted from
609 (stack_protect_return_slot_p): New function.
610 (expand_used_vars): Call stack_protect_decl_p and
611 stack_protect_return_slot_p for -fstack-protector-strong.
613 2014-05-09 Kenneth Zadeck <zadeck@naturalbridge.com>
616 * fold-const.c (fold_binary_loc): Changed width of mask.
618 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
620 * config/avr/avr-fixed.md (round<mode>3): Use -1U instead of -1 in
621 unsigned int initializers for regno_in, regno_out.
623 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
626 * config/avr/avr.md (cc): Add new attribute set_vzn.
627 (addqi3, addqq3, adduqq3, subqi3, subqq3, subuqq3, negqi2) [cc]:
628 Set cc insn attribute to set_vzn instead of set_zn for alternatives
629 with INC, DEC or NEG.
630 * config/avr/avr.c (avr_notice_update_cc): Handle SET_VZN.
631 (avr_out_plus_1): ADIW sets cc0 to CC_SET_CZN.
632 INC, DEC and ADD+ADC set cc0 to CC_CLOBBER.
634 2014-05-09 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
637 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
639 * wide-int.cc (UTItype): Define.
640 (UDWtype): Define for appropriate W_TYPE_SIZE.
642 2014-05-09 Richard Biener <rguenther@suse.de>
644 * Makefile.in (GTFILES): Remove tree-ssa-propagate.c.
645 * tree-ssa-propagate.c: Do not include gt-tree-ssa-propagate.h.
646 (interesting_ssa_edges, varying_ssa_edges): Move out of GC space.
647 (add_ssa_edge, process_ssa_edge_worklist, ssa_prop_init,
648 ssa_propagate): Adjust.
650 2014-05-08 Jeff Law <law@redhat.com>
652 PR tree-optimization/61009
653 * tree-ssa-threadedge.c (thread_through_normal_block): Return a
654 tri-state rather than a boolean. When a block is too big to
655 thread through, inform caller via negative return value.
656 (thread_across_edge): If a block was too big for normal threading,
657 then it's too big for a joiner too, so remove temporary equivalences
658 and return immediately.
660 2014-05-08 Manuel López-Ibáñez <manu@gcc.gnu.org>
661 Matthias Klose <doko@ubuntu.com>
664 * optc-gen.awk: Fix option handling for -Wunused-parameter.
666 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
669 * config/i386/i386.c (PTA_HASWELL): Remove PTA_RTM.
671 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
674 * config/alpha/alpha.c: Include gimple-iterator.h.
675 (alpha_gimple_fold_builtin): New function. Move
676 ALPHA_BUILTIN_UMULH folding from ...
677 (alpha_fold_builtin): ... here.
678 (TARGET_GIMPLE_FOLD_BUILTIN): New define.
680 2014-05-08 Wei Mi <wmi@google.com>
683 * config/i386/i386.c (ix86_compute_frame_layout): Update
684 preferred_stack_boundary for call, expanded from tls descriptor.
685 * config/i386/i386.md (*tls_global_dynamic_32_gnu): Update RTX
686 to depend on SP register.
687 (*tls_local_dynamic_base_32_gnu): Ditto.
688 (*tls_local_dynamic_32_once): Ditto.
689 (tls_global_dynamic_64_<mode>): Set
690 ix86_tls_descriptor_calls_expanded_in_cfun.
691 (tls_local_dynamic_base_64_<mode>): Ditto.
692 (tls_global_dynamic_32): Set
693 ix86_tls_descriptor_calls_expanded_in_cfun. Update RTX
694 to depend on SP register.
695 (tls_local_dynamic_base_32): Ditto.
697 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
699 * config/arm/arm_neon.h: Update comment.
700 * config/arm/neon-docgen.ml: Delete.
701 * config/arm/neon-gen.ml: Delete.
702 * doc/arm-neon-intrinsics.texi: Update comment.
704 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
706 * config/arm/arm_neon_builtins.def (vadd, vsub): Only define the v2sf
708 (vand, vorr, veor, vorn, vbic): Remove.
709 * config/arm/neon.md (neon_vadd, neon_vsub, neon_vadd_unspec): Adjust
711 (neon_vsub_unspec): Likewise.
712 (neon_vorr, neon_vand, neon_vbic, neon_veor, neon_vorn): Remove.
714 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
716 * config/arm/arm_neon.h (vadd_s8): GNU C implementation
717 (vadd_s16): Likewise.
718 (vadd_s32): Likewise.
719 (vadd_f32): Likewise.
721 (vadd_u16): Likewise.
722 (vadd_u32): Likewise.
723 (vadd_s64): Likewise.
724 (vadd_u64): Likewise.
725 (vaddq_s8): Likewise.
726 (vaddq_s16): Likewise.
727 (vaddq_s32): Likewise.
728 (vaddq_s64): Likewise.
729 (vaddq_f32): Likewise.
730 (vaddq_u8): Likewise.
731 (vaddq_u16): Likewise.
732 (vaddq_u32): Likewise.
733 (vaddq_u64): Likewise.
735 (vmul_s16): Likewise.
736 (vmul_s32): Likewise.
737 (vmul_f32): Likewise.
739 (vmul_u16): Likewise.
740 (vmul_u32): Likewise.
742 (vmulq_s8): Likewise.
743 (vmulq_s16): Likewise.
744 (vmulq_s32): Likewise.
745 (vmulq_f32): Likewise.
746 (vmulq_u8): Likewise.
747 (vmulq_u16): Likewise.
748 (vmulq_u32): Likewise.
750 (vsub_s16): Likewise.
751 (vsub_s32): Likewise.
752 (vsub_f32): Likewise.
754 (vsub_u16): Likewise.
755 (vsub_u32): Likewise.
756 (vsub_s64): Likewise.
757 (vsub_u64): Likewise.
758 (vsubq_s8): Likewise.
759 (vsubq_s16): Likewise.
760 (vsubq_s32): Likewise.
761 (vsubq_s64): Likewise.
762 (vsubq_f32): Likewise.
763 (vsubq_u8): Likewise.
764 (vsubq_u16): Likewise.
765 (vsubq_u32): Likewise.
766 (vsubq_u64): Likewise.
768 (vand_s16): Likewise.
769 (vand_s32): Likewise.
771 (vand_u16): Likewise.
772 (vand_u32): Likewise.
773 (vand_s64): Likewise.
774 (vand_u64): Likewise.
775 (vandq_s8): Likewise.
776 (vandq_s16): Likewise.
777 (vandq_s32): Likewise.
778 (vandq_s64): Likewise.
779 (vandq_u8): Likewise.
780 (vandq_u16): Likewise.
781 (vandq_u32): Likewise.
782 (vandq_u64): Likewise.
784 (vorr_s16): Likewise.
785 (vorr_s32): Likewise.
787 (vorr_u16): Likewise.
788 (vorr_u32): Likewise.
789 (vorr_s64): Likewise.
790 (vorr_u64): Likewise.
791 (vorrq_s8): Likewise.
792 (vorrq_s16): Likewise.
793 (vorrq_s32): Likewise.
794 (vorrq_s64): Likewise.
795 (vorrq_u8): Likewise.
796 (vorrq_u16): Likewise.
797 (vorrq_u32): Likewise.
798 (vorrq_u64): Likewise.
800 (veor_s16): Likewise.
801 (veor_s32): Likewise.
803 (veor_u16): Likewise.
804 (veor_u32): Likewise.
805 (veor_s64): Likewise.
806 (veor_u64): Likewise.
807 (veorq_s8): Likewise.
808 (veorq_s16): Likewise.
809 (veorq_s32): Likewise.
810 (veorq_s64): Likewise.
811 (veorq_u8): Likewise.
812 (veorq_u16): Likewise.
813 (veorq_u32): Likewise.
814 (veorq_u64): Likewise.
816 (vbic_s16): Likewise.
817 (vbic_s32): Likewise.
819 (vbic_u16): Likewise.
820 (vbic_u32): Likewise.
821 (vbic_s64): Likewise.
822 (vbic_u64): Likewise.
823 (vbicq_s8): Likewise.
824 (vbicq_s16): Likewise.
825 (vbicq_s32): Likewise.
826 (vbicq_s64): Likewise.
827 (vbicq_u8): Likewise.
828 (vbicq_u16): Likewise.
829 (vbicq_u32): Likewise.
830 (vbicq_u64): Likewise.
832 (vorn_s16): Likewise.
833 (vorn_s32): Likewise.
835 (vorn_u16): Likewise.
836 (vorn_u32): Likewise.
837 (vorn_s64): Likewise.
838 (vorn_u64): Likewise.
839 (vornq_s8): Likewise.
840 (vornq_s16): Likewise.
841 (vornq_s32): Likewise.
842 (vornq_s64): Likewise.
843 (vornq_u8): Likewise.
844 (vornq_u16): Likewise.
845 (vornq_u32): Likewise.
846 (vornq_u64): Likewise.
848 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
850 * wide-int.cc (UTItype): Define.
851 (UDWtype): Define for appropriate W_TYPE_SIZE.
853 2014-05-08 Marc Glisse <marc.glisse@inria.fr>
855 PR tree-optimization/59100
856 * tree-ssa-phiopt.c: Include tree-inline.h.
857 (neutral_element_p, absorbing_element_p): New functions.
858 (value_replacement): Handle conditional binary operations with a
859 neutral or absorbing element.
861 2014-05-08 Richard Biener <rguenther@suse.de>
863 * tree-ssa-sccvn.c (vn_get_expr_for): Valueize operands before
864 folding the expression.
865 (valueize_expr): Remove.
866 (visit_reference_op_load): Do not valueize the result of
868 (simplify_binary_expression): Likewise.
869 (simplify_unary_expression): Likewise.
871 2014-05-08 Richard Biener <rguenther@suse.de>
873 * gimplify.c (gimplify_call_expr): Use saved fnptrtype for
874 looking at TYPE_ARG_TYPES.
876 2014-05-08 Richard Biener <rguenther@suse.de>
878 * gimple-fold.c (gimple_fold_stmt_to_constant_1): Remove
879 pointer propagation special-case.
881 2014-05-08 Bin Cheng <bin.cheng@arm.com>
883 * tree-affine.c (tree_to_aff_combination): Handle MEM_REF for
884 core part of address expressions.
886 2014-05-08 Alan Modra <amodra@gmail.com>
889 * config/rs6000/rs6000.c (expand_block_move): Allow 64-bit
890 loads and stores when -mno-strict-align at any alignment.
891 (expand_block_clear): Similarly. Also correct calculation of
894 2014-05-07 Thomas Preud'homme <thomas.preudhomme@arm.com>
897 * tree-complex.c (expand_complex_move): Keep line info when expanding
899 * tree-ssa-uninit.c (warn_uninit): New argument. Ignore assignment
900 of complex expression. Use new argument to display correct location
901 for values coming from phi statement.
902 (warn_uninitialized_vars): Adapt to new signature of warn_uninit.
903 (warn_uninitialized_phi): Pass location of phi argument to
905 * tree-ssa.c (ssa_undefined_value_p): For SSA_NAME initialized by a
906 COMPLEX_EXPR, recurse on each part of the COMPLEX_EXPR.
908 2014-05-07 Segher Boessenkool <segher@kernel.crashing.org>
910 * config/rs6000/predicates.md (indexed_address_mem): New.
911 * config/rs6000/rs6000.md (type): Remove load_ext, load_ext_u,
912 load_ext_ux, load_ux, load_u, store_ux, store_u, fpload_ux, fpload_u,
913 fpstore_ux, fpstore_u.
914 (sign_extend, indexed, update): New.
915 (cell_micro): Adjust.
916 (*zero_extend<mode>di2_internal1, *zero_extendsidi2_lfiwzx,
917 *extendsidi2_lfiwax, *extendsidi2_nocell, *extendsfdf2_fpr,
918 *movsi_internal1, *movsi_internal1_single, *movhi_internal,
919 *movqi_internal, *movcc_internal1, mov<mode>_hardfloat,
920 *mov<mode>_softfloat, *mov<mode>_hardfloat32, *mov<mode>_hardfloat64,
921 *mov<mode>_softfloat64, *movdi_internal32, *movdi_internal64,
922 *mov<mode>_string, *ldmsi8, *ldmsi7, *ldmsi6, *ldmsi5, *ldmsi4,
923 *ldmsi3, *stmsi8, *stmsi7, *stmsi6, *stmsi5, *stmsi4, *stmsi3,
924 *movdi_update1, movdi_<mode>_update, movdi_<mode>_update_stack,
925 *movsi_update1, *movsi_update2, movsi_update, movsi_update_stack,
926 *movhi_update1, *movhi_update2, *movhi_update3, *movhi_update4,
927 *movqi_update1, *movqi_update2, *movqi_update3, *movsf_update1,
928 *movsf_update2, *movsf_update3, *movsf_update4, *movdf_update1,
929 *movdf_update2, load_toc_aix_si, load_toc_aix_di, probe_stack_<mode>,
930 *stmw, *lmw, as well as 10 anonymous patterns): Adjust.
932 * config/rs6000/dfp.md (movsd_store, movsd_load): Adjust.
933 * config/rs6000/vsx.md (*vsx_movti_32bit, *vsx_extract_<mode>_load,
934 *vsx_extract_<mode>_store): Adjust.
935 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_microcoded_insn,
936 is_cracked_insn, insn_must_be_first_in_group,
937 insn_must_be_last_in_group): Adjust.
939 * config/rs6000/40x.md (ppc403-load, ppc403-store, ppc405-float):
941 * config/rs6000/440.md (ppc440-load, ppc440-store, ppc440-fpload,
942 ppc440-fpstore): Adjust.
943 * config/rs6000/476.md (ppc476-load, ppc476-store, ppc476-fpload,
944 ppc476-fpstore): Adjust.
945 * config/rs6000/601.md (ppc601-load, ppc601-store, ppc601-fpload,
946 ppc601-fpstore): Adjust.
947 * config/rs6000/603.md (ppc603-load, ppc603-store, ppc603-fpload):
949 * config/rs6000/6xx.md (ppc604-load, ppc604-store, ppc604-fpload):
951 * config/rs6000/7450.md (ppc7450-load, ppc7450-store, ppc7450-fpload,
952 ppc7450-fpstore): Adjust.
953 * config/rs6000/7xx.md (ppc750-load, ppc750-store): Adjust.
954 * config/rs6000/8540.md (ppc8540_load, ppc8540_store): Adjust.
955 * config/rs6000/a2.md (ppca2-load, ppca2-fp-load, ppca2-fp-store):
957 * config/rs6000/cell.md (cell-load, cell-load-ux, cell-load-ext,
958 cell-fpload, cell-fpload-update, cell-store, cell-store-update,
959 cell-fpstore, cell-fpstore-update): Adjust.
960 * config/rs6000/e300c2c3.md (ppce300c3_load, ppce300c3_fpload,
961 ppce300c3_store, ppce300c3_fpstore): Adjust.
962 * config/rs6000/e500mc.md (e500mc_load, e500mc_fpload, e500mc_store,
963 e500mc_fpstore): Adjust.
964 * config/rs6000/e500mc64.md (e500mc64_load, e500mc64_fpload,
965 e500mc64_store, e500mc64_fpstore): Adjust.
966 * config/rs6000/e5500.md (e5500_load, e5500_fpload, e5500_store,
967 e5500_fpstore): Adjust.
968 * config/rs6000/e6500.md (e6500_load, e6500_fpload, e6500_store,
969 e6500_fpstore): Adjust.
970 * config/rs6000/mpc.md (mpccore-load, mpccore-store, mpccore-fpload):
972 * config/rs6000/power4.md (power4-load, power4-load-ext,
973 power4-load-ext-update, power4-load-ext-update-indexed,
974 power4-load-update-indexed, power4-load-update, power4-fpload,
975 power4-fpload-update, power4-store, power4-store-update,
976 power4-store-update-indexed, power4-fpstore, power4-fpstore-update):
978 * config/rs6000/power5.md (power5-load, power5-load-ext,
979 power5-load-ext-update, power5-load-ext-update-indexed,
980 power5-load-update-indexed, power5-load-update, power5-fpload,
981 power5-fpload-update, power5-store, power5-store-update,
982 power5-store-update-indexed, power5-fpstore, power5-fpstore-update):
984 * config/rs6000/power6.md (power6-load, power6-load-ext,
985 power6-load-update, power6-load-update-indexed,
986 power6-load-ext-update, power6-load-ext-update-indexed, power6-fpload,
987 power6-fpload-update, power6-store, power6-store-update,
988 power6-store-update-indexed, power6-fpstore, power6-fpstore-update):
990 * config/rs6000/power7.md (power7-load, power7-load-ext,
991 power7-load-update, power7-load-update-indexed,
992 power7-load-ext-update, power7-load-ext-update-indexed, power7-fpload,
993 power7-fpload-update, power7-store, power7-store-update,
994 power7-store-update-indexed, power7-fpstore, power7-fpstore-update):
996 * config/rs6000/power8.md (power8-load, power8-load-update,
997 power8-load-ext, power8-load-ext-update, power8-fpload,
998 power8-fpload-update, power8-store, power8-store-update-indexed,
999 power8-fpstore, power8-fpstore-update): Adjust.
1000 * config/rs6000/rs64.md (rs64a-load, rs64a-store, rs64a-fpload):
1002 * config/rs6000/titan.md (titan_lsu_load, titan_lsu_fpload,
1003 titan_lsu_store, titan_lsu_fpstore): Adjust.
1004 * config/rs6000/xfpu.md (fp-load, fp-store): Adjust.
1006 2014-05-07 Oleg Endo <olegendo@gcc.gnu.org>
1009 * config/sh/sh-mem.cc (sh_expand_strlen): Use loop when emitting
1010 unrolled byte insns. Emit address increments after move insns.
1012 2014-05-07 David Malcolm <dmalcolm@redhat.com>
1014 * gimple.h (gimple_builtin_call_types_compatible_p): Accept a
1015 const_gimple, rather than a gimple.
1016 (gimple_call_builtin_p): Likewise, for the three variants.
1018 * gimple.c (gimple_builtin_call_types_compatible_p): Likewise.
1019 (gimple_call_builtin_p): Likewise, for the three variants.
1021 2014-05-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
1023 PR tree-optimization/61095
1024 * tree-ssanames.c (get_nonzero_bits): Fix type extension in wi::shwi.
1026 2014-05-07 Richard Biener <rguenther@suse.de>
1028 PR tree-optimization/61034
1029 * tree-ssa-alias.c (call_may_clobber_ref_p_1): Export.
1030 (maybe_skip_until): Use translate to take into account
1031 lattices when trying to do disambiguations.
1032 (get_continuation_for_phi_1): Likewise.
1033 (get_continuation_for_phi): Adjust for added translate arguments.
1034 (walk_non_aliased_vuses): Likewise.
1035 * tree-ssa-alias.h (get_continuation_for_phi): Adjust prototype.
1036 (walk_non_aliased_vuses): Likewise.
1037 (call_may_clobber_ref_p_1): Declare.
1038 * tree-ssa-sccvn.c (vn_reference_lookup_3): Also disambiguate against
1039 calls. Stop early if we are only supposed to disambiguate.
1040 * tree-ssa-pre.c (translate_vuse_through_block): Adjust.
1042 2014-05-07 Joern Rennecke <joern.rennecke@embecosm.com>
1044 * config/epiphany/epiphany.c (epiphany_handle_interrupt_attribute):
1045 Emit an error when the function has arguments.
1047 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
1049 * cfgloop.h (unswitch_loops): Remove.
1050 * doc/passes.texi: Remove references to loop-unswitch.c
1051 * timevar.def (TV_LOOP_UNSWITCH): Remove.
1053 2014-05-07 Evgeny Stupachenko <evstupac@gmail.com>
1055 * tree-vect-data-refs.c (vect_grouped_load_supported): New
1056 check for loads group of length 3.
1057 (vect_permute_load_chain): New permutations for loads group of
1059 * tree-vect-stmts.c (vect_model_load_cost): Change cost
1060 of vec_perm_shuffle for the new permutations.
1062 2014-05-07 Alan Lawrence <alan.lawrence@arm.com>
1064 * config/aarch64/arm_neon.h (vtrn1_f32, vtrn1_p8, vtrn1_p16, vtrn1_s8,
1065 vtrn1_s16, vtrn1_s32, vtrn1_u8, vtrn1_u16, vtrn1_u32, vtrn1q_f32,
1066 vtrn1q_f64, vtrn1q_p8, vtrn1q_p16, vtrn1q_s8, vtrn1q_s16, vtrn1q_s32,
1067 vtrn1q_s64, vtrn1q_u8, vtrn1q_u16, vtrn1q_u32, vtrn1q_u64, vtrn2_f32,
1068 vtrn2_p8, vtrn2_p16, vtrn2_s8, vtrn2_s16, vtrn2_s32, vtrn2_u8,
1069 vtrn2_u16, vtrn2_u32, vtrn2q_f32, vtrn2q_f64, vtrn2q_p8, vtrn2q_p16,
1070 vtrn2q_s8, vtrn2q_s16, vtrn2q_s32, vtrn2q_s64, vtrn2q_u8, vtrn2q_u16,
1071 vtrn2q_u32, vtrn2q_u64): Replace temporary asm with __builtin_shuffle.
1073 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
1075 * loop-unswitch.c: Delete.
1077 2014-05-07 Richard Biener <rguenther@suse.de>
1079 * config.gcc: Always set need_64bit_hwint to yes.
1081 2014-05-07 Chung-Ju Wu <jasonwucj@gmail.com>
1083 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Have it in favor
1084 of using optimize_size.
1086 2014-05-06 Mike Stump <mikestump@comcast.net>
1088 * wide-int.h (wi::int_traits <HOST_WIDE_INT>): Always define.
1090 2014-05-06 Joseph Myers <joseph@codesourcery.com>
1092 * config/i386/sse.md (*mov<mode>_internal)
1093 (*<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>)
1094 (*<sse2_avx_avx512f>_loaddqu<mode><mask_name>)
1095 (<sse>_andnot<mode>3, <code><mode>3, *andnot<mode>3)
1096 (*<code><mode>3, *andnot<mode>3<mask_name>)
1097 (<mask_codefor><code><mode>3<mask_name>): Only consider
1098 TARGET_SSE_PACKED_SINGLE_INSN_OPTIMAL for modes of size 16.
1100 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
1103 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
1105 * lra-constraints.c (valid_address_p): Move earlier in file.
1106 Add a constraint argument to the address_info version.
1107 (satisfies_memory_constraint_p): New function.
1108 (satisfies_address_constraint_p): Likewise.
1109 (process_alt_operands, curr_insn_transform): Use them.
1110 (process_address): Pass the constraint to valid_address_p when
1111 checking address operands.
1113 2014-05-06 Richard Sandiford <r.sandiford@uk.ibm.com>
1115 * lto-cgraph.c (compute_ltrans_boundary): Make node variables local
1116 to their respective blocks. Fix inadvertent use of "node".
1118 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
1120 * emit-rtl.c (init_derived_machine_modes): New functionm, split
1122 (init_emit_once): ...here.
1123 * rtl.h (init_derived_machine_modes): Declare.
1124 * toplev.c (do_compile): Call it even if no_backend.
1126 2014-05-06 Kenneth Zadeck <zadeck@naturalbridge.com>
1127 Mike Stump <mikestump@comcast.net>
1128 Richard Sandiford <rdsandiford@googlemail.com>
1129 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1131 * alias.c (ao_ref_from_mem): Use wide-int interfaces.
1132 (rtx_equal_for_memref_p): Update comment.
1133 (adjust_offset_for_component_ref): Use wide-int interfaces.
1134 * builtins.c (get_object_alignment_2): Likewise.
1135 (c_readstr): Likewise.
1136 (target_char_cast): Add comment.
1137 (determine_block_size): Use wide-int interfaces.
1138 (expand_builtin_signbit): Likewise.
1139 (fold_builtin_int_roundingfn): Likewise.
1140 (fold_builtin_bitop): Likewise.
1141 (fold_builtin_bswap): Likewise.
1142 (fold_builtin_logarithm): Use signop.
1143 (fold_builtin_pow): Likewise.
1144 (fold_builtin_memory_op): Use wide-int interfaces.
1145 (fold_builtin_object_size): Likewise.
1146 * cfgloop.c (alloc_loop): Initialize nb_iterations_upper_bound and
1147 nb_iterations_estimate.
1148 (record_niter_bound): Use wide-int interfaces.
1149 (get_estimated_loop_iterations_int): Likewise.
1150 (get_estimated_loop_iterations): Likewise.
1151 (get_max_loop_iterations): Likewise.
1152 * cfgloop.h: Include wide-int.h.
1153 (struct nb_iter_bound): Change bound to widest_int.
1154 (struct loop): Change nb_iterations_upper_bound and
1155 nb_iterations_estimate to widest_int.
1156 (record_niter_bound): Switch to use widest_int.
1157 (get_estimated_loop_iterations): Likewise.
1158 (get_max_loop_iterations): Likewise.
1159 (gcov_type_to_double_int): Rename to gcov_type_to_wide_int and
1160 update for wide-int.
1161 * cgraph.c (cgraph_add_thunk): Use wide-int interfaces.
1162 * combine.c (try_combine): Likewise.
1163 (subst): Use CONST_SCALAR_INT_P rather than CONST_INT_P.
1164 * config/aarch64/aarch64.c (aapcs_vfp_sub_candidate): Use wide-int
1166 (aarch64_float_const_representable_p): Likewise.
1167 * config/arc/arc.c: Include wide-int.h.
1168 (arc_can_use_doloop_p): Use wide-int interfaces.
1169 * config/arm/arm.c (aapcs_vfp_sub_candidate): Likewise.
1170 (vfp3_const_double_index): Likewise.
1171 * config/avr/avr.c (avr_out_round): Likewise.
1172 (avr_fold_builtin): Likewise.
1173 * config/bfin/bfin.c (bfin_local_alignment): Likewise.
1174 (bfin_can_use_doloop_p): Likewise.
1175 * config/darwin.c (darwin_mergeable_constant_section): Likewise.
1176 (machopic_select_rtx_section): Update to handle CONST_WIDE_INT.
1177 * config/i386/i386.c: Include wide-int.h.
1178 (ix86_data_alignment): Use wide-int interfaces.
1179 (ix86_local_alignment): Likewise.
1180 (ix86_emit_swsqrtsf): Update real_from_integer.
1181 * config/msp430/msp430.c (msp430_attr): Use wide-int interfaces.
1182 * config/nds32/nds32.c (nds32_insert_attributes): Likewise.
1183 * config/rs6000/predicates.md (any_operand): Add const_wide_int.
1184 (zero_constant): Likewise.
1185 (input_operand): Likewise.
1186 (splat_input_operand): Likewise.
1187 (non_logical_cint_operand): Change const_double to const_wide_int.
1188 * config/rs6000/rs6000.c (num_insns_constant): Handle CONST_WIDE_INT.
1189 (easy_altivec_constant): Remove comment.
1190 (paired_expand_vector_init): Use CONSTANT_P.
1191 (rs6000_legitimize_address): Handle CONST_WIDE_INT.
1192 (rs6000_emit_move): Update checks.
1193 (rs6000_aggregate_candidate): Use wide-int interfaces.
1194 (rs6000_expand_ternop_builtin): Likewise.
1195 (rs6000_output_move_128bit): Handle CONST_WIDE_INT.
1196 (rs6000_assemble_integer): Likewise.
1197 (rs6000_hash_constant): Likewise.
1198 (output_toc): Likewise.
1199 (rs6000_rtx_costs): Likewise.
1200 (rs6000_emit_swrsqrt); Update call to real_from_integer.
1201 * config/rs6000/rs6000-c.c: Include wide-int.h.
1202 (altivec_resolve_overloaded_builtin): Use wide-int interfaces.
1203 * config/rs6000/rs6000.h (TARGET_SUPPORTS_WIDE_INT): New.
1204 * config/rs6000/rs6000.md: Use const_scalar_int_operand.
1205 Handle CONST_WIDE_INT.
1206 * config/sol2-c.c (solaris_pragma_align): Change low to unsigned HWI.
1207 Use tree_fits_uhwi_p.
1208 * config/sparc/sparc.c: Include wide-int.h.
1209 (sparc_fold_builtin): Use wide-int interfaces.
1210 * config/vax/vax.c: Include wide-int.h.
1211 (vax_float_literal): Use real_from_integer.
1212 * coretypes.h (struct hwivec_def): New.
1214 (const_hwivec): New.
1215 * cse.c (hash_rtx_cb): Handle CONST_WIDE_INT.
1216 (equiv_constant): Handle CONST_WIDE_INT.
1217 * cselib.c (rtx_equal_for_cselib_1): Use CASE_CONST_UNIQUE.
1218 (cselib_hash_rtx): Handle CONST_WIDE_INT.
1219 * dbxout.c (stabstr_U): Use wide-int interfaces.
1220 (dbxout_type): Update to use cst_fits_shwi_p.
1221 * defaults.h (LOG2_BITS_PER_UNIT): Define.
1222 (TARGET_SUPPORTS_WIDE_INT): Add default.
1223 * dfp.c: Include wide-int.h.
1224 (decimal_real_to_integer2): Use wide-int interfaces and rename to
1225 decimal_real_to_integer.
1226 * dfp.h (decimal_real_to_integer2): Return a wide_int and rename to
1227 decimal_real_to_integer.
1228 * doc/generic.texi (Constant expressions): Update for wide_int.
1229 * doc/rtl.texi (const_double): Likewise.
1230 (const_wide_int, CONST_WIDE_INT, CONST_WIDE_INT_VEC): New.
1231 (CONST_WIDE_INT_NUNITS, CONST_WIDE_INT_ELT): New.
1232 * doc/tm.texi.in (REAL_VALUE_TO_INT): Remove.
1233 (REAL_VALUE_FROM_INT): Remove.
1234 (TARGET_SUPPORTS_WIDE_INT): New.
1235 * doc/tm.texi: Regenerate.
1236 * dojump.c (prefer_and_bit_test): Use wide-int interfaces.
1237 * double-int.h: Include wide-int.h.
1238 (struct wi::int_traits): New.
1239 * dwarf2out.c (get_full_len): New.
1240 (dw_val_equal_p): Add case dw_val_class_wide_int.
1241 (size_of_loc_descr): Likewise.
1242 (output_loc_operands): Likewise.
1243 (insert_double): Remove.
1244 (insert_wide_int): New.
1246 (print_die): Add case dw_val_class_wide_int.
1247 (attr_checksum): Likewise.
1248 (attr_checksum_ordered): Likewise.
1249 (same_dw_val_p): Likewise.
1250 (size_of_die): Likewise.
1251 (value_format): Likewise.
1252 (output_die): Likewise.
1253 (double_int_type_size_in_bits): Rename to offset_int_type_size_in_bits.
1255 (clz_loc_descriptor): Use wide-int interfaces.
1256 (mem_loc_descriptor): Likewise. Handle CONST_WIDE_INT.
1257 (loc_descriptor): Use wide-int interfaces. Handle CONST_WIDE_INT.
1258 (round_up_to_align): Use wide-int interfaces.
1259 (field_byte_offset): Likewise.
1260 (insert_double): Rename to insert_wide_int. Use wide-int interfaces.
1261 (add_const_value_attribute): Handle CONST_WIDE_INT. Update
1262 CONST_DOUBLE handling. Use wide-int interfaces.
1263 (add_bound_info): Use tree_fits_uhwi_p. Use wide-int interfaces.
1264 (gen_enumeration_type_die): Use add_AT_wide.
1265 (hash_loc_operands): Add case dw_val_class_wide_int.
1266 (compare_loc_operands): Likewise.
1267 * dwarf2out.h: Include wide-int.h.
1268 (wide_int_ptr): New.
1269 (enum dw_val_class): Add dw_val_class_wide_int.
1270 (struct dw_val_struct): Add val_wide.
1271 * emit-rtl.c (const_wide_int_htab): New.
1272 (const_wide_int_htab_hash): New.
1273 (const_wide_int_htab_eq): New.
1274 (lookup_const_wide_int): New.
1275 (const_double_htab_hash): Use wide-int interfaces.
1276 (const_double_htab_eq): Likewise.
1277 (rtx_to_double_int): Conditionally compile for wide-int.
1278 (immed_double_int_const): Rename to immed_wide_int_const and
1279 update for wide-int.
1280 (immed_double_const): Conditionally compile for wide-int.
1281 (init_emit_once): Use wide-int interfaces.
1282 * explow.c (plus_constant): Likewise.
1283 * expmed.c (mask_rtx): Move further up file. Use wide-int interfaces.
1284 (lshift_value): Use wide-int interfaces.
1285 (expand_mult): Likewise.
1286 (choose_multiplier): Likewise.
1287 (expand_smod_pow2): Likewise.
1288 (make_tree): Likewise.
1289 * expr.c (convert_modes): Consolidate handling of constants.
1290 Use wide-int interfaces.
1291 (emit_group_load_1): Add note.
1292 (store_expr): Update comment.
1293 (get_inner_reference): Use wide-int interfaces.
1294 (expand_constructor): Update comment.
1295 (expand_expr_real_2): Use wide-int interfaces.
1296 (expand_expr_real_1): Likewise.
1297 (reduce_to_bit_field_precision): Likewise.
1298 (const_vector_from_tree): Likewise.
1299 * final.c: Include wide-int-print.h.
1300 (output_addr_const): Handle CONST_WIDE_INT. Use CONST_DOUBLE_AS_INT_P.
1301 * fixed-value.c: Include wide-int.h.
1302 (fixed_from_string): Use wide-int interfaces.
1303 (fixed_to_decimal): Likewise.
1304 (fixed_convert_from_real): Likewise.
1305 (real_convert_from_fixed): Likewise.
1306 * fold-const.h (mem_ref_offset): Return an offset_int.
1307 (div_if_zero_remainder): Remove code parameter.
1308 * fold-const.c (div_if_zero_remainder): Remove code parameter.
1309 Use wide-int interfaces.
1310 (may_negate_without_overflow_p): Use wide-int interfaces.
1311 (negate_expr_p): Likewise.
1312 (fold_negate_expr): Likewise.
1313 (int_const_binop_1): Likewise.
1314 (const_binop): Likewise.
1315 (fold_convert_const_int_from_int): Likewise.
1316 (fold_convert_const_int_from_real): Likewise.
1317 (fold_convert_const_int_from_fixed): Likewise.
1318 (fold_convert_const_fixed_from_int): Likewise.
1319 (all_ones_mask_p): Take an unsigned size. Use wide-int interfaces.
1320 (sign_bit_p): Use wide-int interfaces.
1321 (make_range_step): Likewise.
1322 (build_range_check): Likewise. Pass an integer of the correct type
1323 instead of using integer_one_node.
1324 (range_predecessor): Pass an integer of the correct type instead
1325 of using integer_one_node.
1326 (range_successor): Likewise.
1327 (merge_ranges): Likewise.
1328 (unextend): Use wide-int interfaces.
1329 (extract_muldiv_1): Likewise.
1330 (fold_div_compare): Likewise.
1331 (fold_single_bit_test): Likewise.
1332 (fold_sign_changed_comparison): Likewise.
1333 (try_move_mult_to_index): Update calls to div_if_zero_remainder.
1334 (fold_plusminus_mult_expr): Use wide-int interfaces.
1335 (native_encode_int): Likewise.
1336 (native_interpret_int): Likewise.
1337 (fold_unary_loc): Likewise.
1338 (pointer_may_wrap_p): Likewise.
1339 (size_low_cst): Likewise.
1340 (mask_with_tz): Likewise.
1341 (fold_binary_loc): Likewise.
1342 (fold_ternary_loc): Likewise.
1343 (multiple_of_p): Likewise.
1344 (tree_call_nonnegative_warnv_p): Update calls to
1345 tree_int_cst_min_precision and real_from_integer.
1346 (fold_negate_const): Use wide-int interfaces.
1347 (fold_abs_const): Likewise.
1348 (fold_relational_const): Use tree_int_cst_lt.
1349 (round_up_loc): Use wide-int interfaces.
1350 * genemit.c (gen_exp): Add CONST_WIDE_INT case.
1351 * gengenrtl.c (excluded_rtx): Add CONST_WIDE_INT case.
1352 * gengtype.c: Remove include of double-int.h.
1353 (do_typedef): Use wide-int interfaces.
1354 (open_base_files): Add wide-int.h.
1355 (main): Add offset_int and widest_int typedefs.
1356 * gengtype-lex.l: Handle "^".
1357 (CXX_KEYWORD): Add "static".
1358 * gengtype-parse.c (require3): New.
1359 (require_template_declaration): Handle constant template arguments
1360 and nested templates.
1361 * gengtype-state.c: Don't include "double-int.h".
1362 * genpreds.c (write_one_predicate_function): Update comment.
1363 (write_tm_constrs_h): Add check for hval and lval use in
1365 * genrecog.c (validate_pattern): Add CONST_WIDE_INT case.
1366 (add_to_sequence): Likewise.
1367 * gensupport.c (struct std_pred_table): Add const_scalar_int_operand
1368 and const_double_operand.
1369 * gimple.c (preprocess_case_label_vec_for_gimple): Use wide-int
1371 * gimple-fold.c (get_base_constructor): Likewise.
1372 (fold_array_ctor_reference): Likewise.
1373 (fold_nonarray_ctor_reference): Likewise.
1374 (fold_const_aggregate_ref_1): Likewise.
1375 (gimple_val_nonnegative_real_p): Likewise.
1376 (gimple_fold_indirect_ref): Likewise.
1377 * gimple-pretty-print.c (dump_ssaname_info): Likewise.
1378 * gimple-ssa-strength-reduction.c: Include wide-int-print.h.
1379 (struct slsr_cand_d): Change index to be widest_int.
1380 (struct incr_info_d): Change incr to be widest_int.
1381 (alloc_cand_and_find_basis): Use wide-int interfaces.
1382 (slsr_process_phi): Likewise.
1383 (backtrace_base_for_ref): Likewise. Return a widest_int.
1384 (restructure_reference): Take a widest_int instead of a double_int.
1385 (slsr_process_ref): Use wide-int interfaces.
1386 (create_mul_ssa_cand): Likewise.
1387 (create_mul_imm_cand): Likewise.
1388 (create_add_ssa_cand): Likewise.
1389 (create_add_imm_cand): Take a widest_int instead of a double_int.
1390 (slsr_process_add): Use wide-int interfaces.
1391 (slsr_process_cast): Likewise.
1392 (slsr_process_copy): Likewise.
1393 (dump_candidate): Likewise.
1394 (dump_incr_vec): Likewise.
1395 (replace_ref): Likewise.
1396 (cand_increment): Likewise. Return a widest_int.
1397 (cand_abs_increment): Likewise.
1398 (replace_mult_candidate): Take a widest_int instead of a double_int.
1399 (replace_unconditional_candidate): Use wide-int interfaces.
1400 (incr_vec_index): Take a widest_int instead of a double_int.
1401 (create_add_on_incoming_edge): Likewise.
1402 (create_phi_basis): Use wide-int interfaces.
1403 (replace_conditional_candidate): Likewise.
1404 (record_increment): Take a widest_int instead of a double_int.
1405 (record_phi_increments): Use wide-int interfaces.
1406 (phi_incr_cost): Take a widest_int instead of a double_int.
1407 (lowest_cost_path): Likewise.
1408 (total_savings): Likewise.
1409 (analyze_increments): Use wide-int interfaces.
1410 (ncd_with_phi): Take a widest_int instead of a double_int.
1411 (ncd_of_cand_and_phis): Likewise.
1412 (nearest_common_dominator_for_cands): Likewise.
1413 (insert_initializers): Use wide-int interfaces.
1414 (all_phi_incrs_profitable): Likewise.
1415 (replace_one_candidate): Likewise.
1416 (replace_profitable_candidates): Likewise.
1417 * godump.c: Include wide-int-print.h.
1418 (go_output_typedef): Use wide-int interfaces.
1419 * graphite-clast-to-gimple.c (gmp_cst_to_tree): Likewise.
1420 * graphite-sese-to-poly.c (tree_int_to_gmp): Likewise.
1421 (build_loop_iteration_domains): Likewise.
1422 * hooks.h: Include wide-int.h rather than double-int.h.
1423 (hook_bool_dint_dint_uint_bool_true): Delete.
1424 (hook_bool_wint_wint_uint_bool_true): Declare.
1425 * hooks.c (hook_bool_dint_dint_uint_bool_true): Removed.
1426 (hook_bool_wint_wint_uint_bool_true): New.
1427 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): Use wide-int
1429 (ubsan_expand_si_overflow_mul_check): Likewise.
1430 * ipa-devirt.c (get_polymorphic_call_info): Likewise.
1431 * ipa-prop.c (compute_complex_assign_jump_func): Likewise.
1432 (get_ancestor_addr_info): Likewise.
1433 (ipa_modify_call_arguments): Likewise.
1434 * loop-doloop.c (doloop_modify): Likewise.
1435 (doloop_optimize): Likewise.
1436 * loop-iv.c (iv_number_of_iterations): Likewise.
1437 * loop-unroll.c (decide_unroll_constant_iterations): Likewise.
1438 (unroll_loop_constant_iterations): Likewise.
1439 (decide_unroll_runtime_iterations): Likewise.
1440 (unroll_loop_runtime_iterations): Likewise.
1441 (decide_peel_simple): Likewise.
1442 (decide_unroll_stupid): Likewise.
1443 * lto-streamer-in.c (streamer_read_wi): Add.
1444 (input_cfg): Use wide-int interfaces.
1445 (lto_input_tree_1): Likewise.
1446 * lto-streamer-out.c (streamer_write_wi): Add.
1447 (hash_tree): Use wide-int interfaces.
1448 (output_cfg): Likewise.
1449 * Makefile.in (OBJS): Add wide-int.o and wide-int-print.o.
1450 (GTFILES): Add wide-int.h and signop.h.
1451 (TAGS): Look for .cc files too.
1452 * omp-low.c (scan_omp_1_op): Use wide-int interfaces.
1453 * optabs.c (expand_subword_shift): Likewise.
1454 (expand_doubleword_shift): Likewise.
1455 (expand_absneg_bit): Likewise.
1456 (expand_copysign_absneg): Likewise.
1457 (expand_copysign_bit): Likewise.
1458 * postreload.c (reload_cse_simplify_set): Likewise.
1459 * predict.c (predict_iv_comparison): Likewise.
1460 * pretty-print.h: Include wide-int-print.h.
1462 * print-rtl.c (print_rtx): Add CONST_WIDE_INT case.
1463 * print-tree.c: Include wide-int-print.h.
1464 (print_node_brief): Use wide-int interfaces.
1465 (print_node): Likewise.
1466 * read-rtl.c (validate_const_wide_int): New.
1467 (read_rtx_code): Add CONST_WIDE_INT case.
1468 * real.c: Include wide-int.h.
1469 (real_to_integer2): Delete.
1470 (real_to_integer): New function, returning a wide_int.
1471 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
1472 (ten_to_ptwo): Update call to real_from_integer.
1473 (real_digit): Likewise.
1474 * real.h: Include signop.h, wide-int.h and insn-modes.h.
1475 (real_to_integer2, REAL_VALUE_FROM_INT, REAL_VALUE_FROM_UNSIGNED_INT)
1476 (REAL_VALUE_TO_INT): Delete.
1477 (real_to_integer): Declare a wide-int form.
1478 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
1479 * recog.c (const_int_operand): Improve comment.
1480 (const_scalar_int_operand): New.
1481 (const_double_operand): Add a separate definition for CONST_WIDE_INT.
1482 * rtlanal.c (commutative_operand_precedence): Handle CONST_WIDE_INT.
1483 (split_double): Likewise.
1484 * rtl.c (DEF_RTL_EXPR): Handle CONST_WIDE_INT.
1485 (rtx_size): Likewise.
1486 (rtx_alloc_stat_v): New.
1487 (rtx_alloc_stat): Now calls rtx_alloc_stat_v.
1488 (cwi_output_hex): New.
1489 (iterative_hash_rtx): Handle CONST_WIDE_INT.
1490 (cwi_check_failed_bounds): New.
1491 * rtl.def (CONST_WIDE_INT): New.
1492 * rtl.h: Include <utility> and wide-int.h.
1493 (struct hwivec_def): New.
1494 (CWI_GET_NUM_ELEM): New.
1495 (CWI_PUT_NUM_ELEM): New.
1496 (struct rtx_def): Add num_elem and hwiv.
1497 (CASE_CONST_SCALAR_INT): Modify for TARGET_SUPPORTS_WIDE_INT.
1498 (CASE_CONST_UNIQUE): Likewise.
1499 (CASE_CONST_ANY): Likewise.
1500 (CONST_SCALAR_INT_P): Likewise.
1501 (CONST_WIDE_INT_P): New.
1503 (HWIVEC_CHECK): New.
1504 (cwi_check_failed_bounds): New.
1506 (HWIVEC_CHECK): New.
1507 (CONST_WIDE_INT_VEC) New.
1508 (CONST_WIDE_INT_NUNITS) New.
1509 (CONST_WIDE_INT_ELT) New.
1510 (rtx_mode_t): New type.
1511 (wi::int_traits <rtx_mode_t>): New.
1513 (wi::min_value): New.
1514 (wi::max_value): New.
1516 (const_wide_int_alloc): New.
1517 (immed_wide_int_const): New.
1518 * sched-vis.c (print_value): Handle CONST_WIDE_INT.
1519 * sel-sched-ir.c (lhs_and_rhs_separable_p): Update comment.
1520 * signop.h: New file.
1521 * simplify-rtx.c (mode_signbit_p): Handle CONST_WIDE_INT.
1522 (simplify_const_unary_operation): Use wide-int interfaces.
1523 (simplify_binary_operation_1): Likewise.
1524 (simplify_const_binary_operation): Likewise.
1525 (simplify_const_relational_operation): Likewise.
1526 (simplify_immed_subreg): Likewise.
1527 * stmt.c (expand_case): Likewise.
1528 * stor-layout.h (set_min_and_max_values_for_integral_type): Take a
1529 signop rather than a bool.
1530 * stor-layout.c (layout_type): Use wide-int interfaces.
1531 (initialize_sizetypes): Update calls to
1532 set_min_and_max_values_for_integral_type.
1533 (set_min_and_max_values_for_integral_type): Take a signop rather
1534 than a bool. Use wide-int interfaces.
1535 (fixup_signed_type): Update accordingly. Remove
1536 HOST_BITS_PER_DOUBLE_INT limit.
1537 (fixup_unsigned_type): Likewise.
1538 * system.h (STATIC_CONSTANT_P): New.
1539 (STATIC_ASSERT): New.
1540 * target.def (can_use_doloop_p): Take widest_ints rather than
1542 * target.h: Include wide-int.h rather than double-int.h.
1543 * targhooks.h (can_use_doloop_if_innermost): Take widest_ints rather
1545 * targhooks.c (default_cxx_get_cookie_size): Use tree_int_cst_lt
1546 rather than INT_CST_LT_UNSIGNED.
1547 (can_use_doloop_if_innermost): Take widest_ints rather than
1549 * tree-affine.c: Include wide-int-print.h.
1550 (double_int_ext_for_comb): Delete.
1551 (wide_int_ext_for_comb): New.
1552 (aff_combination_zero): Use wide-int interfaces.
1553 (aff_combination_const): Take a widest_int instead of a double_int.
1554 (aff_combination_elt): Use wide-int interfaces.
1555 (aff_combination_scale): Take a widest_int instead of a double_int.
1556 (aff_combination_add_elt): Likewise.
1557 (aff_combination_add_cst): Likewise.
1558 (aff_combination_add): Use wide-int interfaces.
1559 (aff_combination_convert): Likewise.
1560 (tree_to_aff_combination): Likewise.
1561 (add_elt_to_tree): Take a widest_int instead of a double_int.
1562 (aff_combination_to_tree): Use wide-int interfaces.
1563 (aff_combination_remove_elt): Likewise.
1564 (aff_combination_add_product): Take a widest_int instead of
1566 (aff_combination_mult): Use wide-int interfaces.
1567 (aff_combination_expand): Likewise.
1568 (double_int_constant_multiple_p): Delete.
1569 (wide_int_constant_multiple_p): New.
1570 (aff_combination_constant_multiple_p): Take a widest_int pointer
1571 instead of a double_int pointer.
1572 (print_aff): Use wide-int interfaces.
1573 (get_inner_reference_aff): Take a widest_int pointer
1574 instead of a double_int pointer.
1575 (aff_comb_cannot_overlap_p): Take widest_ints instead of double_ints.
1576 * tree-affine.h: Include wide-int.h.
1577 (struct aff_comb_elt): Change type of coef to widest_int.
1578 (struct affine_tree_combination): Change type of offset to widest_int.
1579 (double_int_ext_for_comb): Delete.
1580 (wide_int_ext_for_comb): New.
1581 (aff_combination_const): Use widest_int instead of double_int.
1582 (aff_combination_scale): Likewise.
1583 (aff_combination_add_elt): Likewise.
1584 (aff_combination_constant_multiple_p): Likewise.
1585 (get_inner_reference_aff): Likewise.
1586 (aff_comb_cannot_overlap_p): Likewise.
1587 (aff_combination_zero_p): Use wide-int interfaces.
1588 * tree.c: Include tree.h.
1589 (init_ttree): Use make_int_cst.
1590 (tree_code_size): Removed code for INTEGER_CST case.
1591 (tree_size): Add INTEGER_CST case.
1592 (make_node_stat): Update comment.
1593 (get_int_cst_ext_nunits, build_new_int_cst, build_int_cstu): New.
1594 (build_int_cst_type): Use wide-int interfaces.
1595 (double_int_to_tree): Likewise.
1596 (double_int_fits_to_tree_p): Delete.
1597 (force_fit_type_double): Delete.
1598 (force_fit_type): New.
1599 (int_cst_hash_hash): Use wide-int interfaces.
1600 (int_cst_hash_eq): Likewise.
1601 (build_int_cst_wide): Delete.
1602 (wide_int_to_tree): New.
1603 (cache_integer_cst): Use wide-int interfaces.
1604 (build_low_bits_mask): Likewise.
1605 (cst_and_fits_in_hwi): Likewise.
1606 (real_value_from_int_cst): Likewise.
1607 (make_int_cst_stat): New.
1608 (integer_zerop): Use wide_int interfaces.
1609 (integer_onep): Likewise.
1610 (integer_all_onesp): Likewise.
1611 (integer_pow2p): Likewise.
1612 (integer_nonzerop): Likewise.
1613 (tree_log2): Likewise.
1614 (tree_floor_log2): Likewise.
1615 (tree_ctz): Likewise.
1616 (int_size_in_bytes): Likewise.
1617 (mem_ref_offset): Return an offset_int rather than a double_int.
1618 (build_type_attribute_qual_variant): Use wide_int interfaces.
1619 (type_hash_eq): Likewise
1620 (tree_int_cst_equal): Likewise.
1621 (tree_int_cst_lt): Delete.
1622 (tree_int_cst_compare): Likewise.
1623 (tree_fits_shwi_p): Use wide_int interfaces.
1624 (tree_fits_uhwi_p): Likewise.
1625 (tree_int_cst_sign_bit): Likewise.
1626 (tree_int_cst_sgn): Likewise.
1627 (tree_int_cst_min_precision): Take a signop rather than a bool.
1628 (simple_cst_equal): Use wide_int interfaces.
1629 (compare_tree_int): Likewise.
1630 (iterative_hash_expr): Likewise.
1631 (int_fits_type_p): Likewise. Use tree_int_cst_lt rather than
1633 (get_type_static_bounds): Use wide_int interfaces.
1634 (tree_int_cst_elt_check_failed): New.
1635 (build_common_tree_nodes): Reordered to set prec before filling in
1637 (int_cst_value): Check cst_and_fits_in_hwi.
1638 (widest_int_cst_value): Use wide_int interfaces.
1639 (upper_bound_in_type): Likewise.
1640 (lower_bound_in_type): Likewise.
1641 (num_ending_zeros): Likewise.
1642 (drop_tree_overflow): Likewise.
1643 * tree-call-cdce.c (check_pow): Update call to real_from_integer.
1644 (gen_conditions_for_pow_cst_base): Likewise.
1645 * tree-cfg.c: Include wide-int.h and wide-int-print.h.
1646 (group_case_labels_stmt): Use wide-int interfaces.
1647 (verify_gimple_assign_binary): Likewise.
1648 (print_loop): Likewise.
1649 * tree-chrec.c (tree_fold_binomial): Likewise.
1650 * tree-core.h (struct tree_base): Add int_length.
1651 (struct tree_int_cst): Change rep of value.
1652 * tree-data-ref.c (dr_analyze_innermost): Use wide-int interfaces.
1653 (dr_may_alias_p): Likewise.
1654 (max_stmt_executions_tree): Likewise.
1655 * tree.def (INTEGER_CST): Update comment.
1656 * tree-dfa.c (get_ref_base_and_extent): Use wide-int interfaces.
1657 * tree-dfa.h (get_addr_base_and_unit_offset_1): Likewise.
1658 * tree-dump.c: Include wide-int.h and wide-int-print.h.
1659 (dequeue_and_dump): Use wide-int interfaces.
1660 * tree.h: Include wide-int.h.
1661 (NULL_TREE): Moved to earlier loc in file.
1662 (TREE_INT_CST_ELT_CHECK): New.
1663 (tree_int_cst_elt_check_failed): New.
1665 (TREE_INT_CST): Delete.
1666 (TREE_INT_CST_LOW): Use wide-int interfaces.
1667 (TREE_INT_CST_HIGH): Delete.
1668 (TREE_INT_CST_NUNITS): New.
1669 (TREE_INT_CST_EXT_NUNITS): Likewise.
1670 (TREE_INT_CST_OFFSET_NUNITS): Likewise.
1671 (TREE_INT_CST_ELT): Likewise.
1672 (INT_CST_LT): Delete.
1673 (tree_int_cst_elt_check): New (two forms).
1674 (type_code_size): Update comment.
1675 (make_int_cst_stat, make_int_cst): New.
1676 (tree_to_double_int): Delete.
1677 (double_int_fits_to_tree_p): Delete.
1678 (force_fit_type_double): Delete.
1679 (build_int_cstu): Replace with out-of-line function.
1680 (build_int_cst_wide): Delete.
1681 (tree_int_cst_lt): Define inline.
1682 (tree_int_cst_le): New.
1683 (tree_int_cst_compare): Define inline.
1684 (tree_int_cst_min_precision): Take a signop rather than a bool.
1685 (wi::int_traits <const_tree>): New.
1686 (wi::int_traits <tree>): New.
1687 (wi::extended_tree): New.
1688 (wi::int_traits <wi::extended_tree>): New.
1689 (wi::to_widest): New.
1690 (wi::to_offset): New.
1691 (wi::fits_to_tree_p): New.
1692 (wi::min_value): New.
1693 (wi::max_value): New.
1694 * tree-inline.c (remap_gimple_op_r): Use wide-int interfaces.
1695 (copy_tree_body_r): Likewise.
1696 * tree-object-size.c (compute_object_offset): Likewise.
1697 (addr_object_size): Likewise.
1698 * tree-predcom.c: Include wide-int-print.h.
1699 (struct dref_d): Change type of offset to widest_int.
1700 (dump_dref): Call wide-int printer.
1701 (aff_combination_dr_offset): Use wide-int interfaces.
1702 (determine_offset): Take a widest_int pointer rather than a
1704 (split_data_refs_to_components): Use wide-int interfaces.
1705 (suitable_component_p): Likewise.
1706 (order_drefs): Likewise.
1707 (add_ref_to_chain): Likewise.
1708 (valid_initializer_p): Likewise.
1709 (determine_roots_comp): Likewise.
1710 * tree-pretty-print.c: Include wide-int-print.h.
1711 (dump_generic_node): Use wide-int interfaces.
1712 * tree-sra.c (sra_ipa_modify_expr): Likewise.
1713 * tree-ssa-address.c (addr_for_mem_ref): Likewise.
1714 (move_fixed_address_to_symbol): Likewise.
1715 (move_hint_to_base): Likewise.
1716 (move_pointer_to_base): Likewise.
1717 (move_variant_to_index): Likewise.
1718 (most_expensive_mult_to_index): Likewise.
1719 (addr_to_parts): Likewise.
1720 (copy_ref_info): Likewise.
1721 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
1722 (indirect_refs_may_alias_p): Likewise.
1723 (stmt_kills_ref_p_1): Likewise.
1724 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
1725 * tree-ssa-ccp.c: Update comment at top of file. Include
1727 (struct prop_value_d): Change type of mask to widest_int.
1728 (extend_mask): New function.
1729 (dump_lattice_value): Use wide-int interfaces.
1730 (get_default_value): Likewise.
1731 (set_constant_value): Likewise.
1732 (set_value_varying): Likewise.
1733 (valid_lattice_transition): Likewise.
1734 (set_lattice_value): Likewise.
1735 (value_to_double_int): Delete.
1736 (value_to_wide_int): New.
1737 (get_value_from_alignment): Use wide-int interfaces.
1738 (get_value_for_expr): Likewise.
1739 (do_dbg_cnt): Likewise.
1740 (ccp_finalize): Likewise.
1741 (ccp_lattice_meet): Likewise.
1742 (bit_value_unop_1): Use widest_ints rather than double_ints.
1743 (bit_value_binop_1): Likewise.
1744 (bit_value_unop): Use wide-int interfaces.
1745 (bit_value_binop): Likewise.
1746 (bit_value_assume_aligned): Likewise.
1747 (evaluate_stmt): Likewise.
1748 (ccp_fold_stmt): Likewise.
1749 (visit_cond_stmt): Likewise.
1750 (ccp_visit_stmt): Likewise.
1751 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Likewise.
1752 (constant_pointer_difference): Likewise.
1753 (associate_pointerplus): Likewise.
1754 (combine_conversions): Likewise.
1755 * tree-ssa-loop.h: Include wide-int.h.
1756 (struct tree_niter_desc): Change type of max to widest_int.
1757 * tree-ssa-loop-im.c (mem_refs_may_alias_p): Use wide-int interfaces.
1758 * tree-ssa-loop-ivcanon.c (remove_exits_and_undefined_stmts): Likewise.
1759 (remove_redundant_iv_tests): Likewise.
1760 (canonicalize_loop_induction_variables): Likewise.
1761 * tree-ssa-loop-ivopts.c (alloc_iv): Likewise.
1762 (constant_multiple_of): Take a widest_int pointer instead of
1763 a double_int pointer.
1764 (get_computation_aff): Use wide-int interfaces.
1765 (ptr_difference_cost): Likewise.
1766 (difference_cost): Likewise.
1767 (get_loop_invariant_expr_id): Likewise.
1768 (get_computation_cost_at): Likewise.
1769 (iv_elimination_compare_lt): Likewise.
1770 (may_eliminate_iv): Likewise.
1771 * tree-ssa-loop-niter.h (estimated_loop_iterations): Use widest_int
1772 instead of double_int.
1773 (max_loop_iterations): Likewise.
1774 (max_stmt_executions): Likewise.
1775 (estimated_stmt_executions): Likewise.
1776 * tree-ssa-loop-niter.c: Include wide-int-print.h.
1777 (split_to_var_and_offset): Use wide-int interfaces.
1778 (determine_value_range): Likewise.
1779 (bound_difference_of_offsetted_base): Likewise.
1780 (bounds_add): Take a widest_int instead of a double_int.
1781 (number_of_iterations_ne_max): Use wide-int interfaces.
1782 (number_of_iterations_ne): Likewise.
1783 (number_of_iterations_lt_to_ne): Likewise.
1784 (assert_loop_rolls_lt): Likewise.
1785 (number_of_iterations_lt): Likewise.
1786 (number_of_iterations_le): Likewise.
1787 (number_of_iterations_cond): Likewise.
1788 (number_of_iterations_exit): Likewise.
1789 (finite_loop_p): Likewise.
1790 (derive_constant_upper_bound_assign): Likewise.
1791 (derive_constant_upper_bound): Return a widest_int.
1792 (derive_constant_upper_bound_ops): Likewise.
1793 (do_warn_aggressive_loop_optimizations): Use wide-int interfaces.
1794 (record_estimate): Take a widest_int rather than a double_int.
1795 (record_nonwrapping_iv): Use wide-int interfaces.
1796 (double_int_cmp): Delete.
1797 (wide_int_cmp): New.
1798 (bound_index): Take a widest_int rather than a double_int.
1799 (discover_iteration_bound_by_body_walk): Use wide-int interfaces.
1800 (maybe_lower_iteration_bound): Likewise.
1801 (estimate_numbers_of_iterations_loop): Likewise.
1802 (estimated_loop_iterations): Take a widest_int pointer than than
1803 a double_int pointer.
1804 (estimated_loop_iterations_int): Use wide-int interfaces.
1805 (max_loop_iterations): Take a widest_int pointer than than
1806 a double_int pointer.
1807 (max_loop_iterations_int): Use wide-int interfaces.
1808 (max_stmt_executions): Take a widest_int pointer than than
1809 a double_int pointer.
1810 (estimated_stmt_executions): Likewise.
1811 (n_of_executions_at_most): Use wide-int interfaces.
1812 (scev_probably_wraps_p): Likewise.
1813 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Update calls
1815 * tree-scalar-evolution.c (simplify_peeled_chrec): Use wide-int
1817 * tree-ssanames.c (set_range_info): Use wide_int_refs rather than
1818 double_ints. Adjust for trailing_wide_ints <3> representation.
1819 (set_nonzero_bits): Likewise.
1820 (get_range_info): Return wide_ints rather than double_ints.
1821 Adjust for trailing_wide_ints <3> representation.
1822 (get_nonzero_bits): Likewise.
1823 (duplicate_ssa_name_range_info): Adjust for trailing_wide_ints <3>
1825 * tree-ssanames.h (struct range_info_def): Replace min, max and
1826 nonzero_bits with a trailing_wide_ints <3>.
1827 (set_range_info): Use wide_int_refs rather than double_ints.
1828 (set_nonzero_bits): Likewise.
1829 (get_range_info): Return wide_ints rather than double_ints.
1830 (get_nonzero_bits): Likewise.
1831 * tree-ssa-phiopt.c (jump_function_from_stmt): Use wide-int interfaces.
1832 * tree-ssa-pre.c (phi_translate_1): Likewise.
1833 * tree-ssa-reassoc.c (decrement_power): Use calls to real_from_integer.
1834 (acceptable_pow_call): Likewise.
1835 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Use wide-int
1837 (vn_reference_fold_indirect): Likewise.
1838 (vn_reference_maybe_forwprop_address): Likewise.
1839 (valueize_refs_1): Likewise.
1840 * tree-ssa-structalias.c (get_constraint_for_ptr_offset): Likewise.
1841 * tree-ssa-uninit.c (is_value_included_in): Use wide-int interfaces,
1842 tree_int_cst_lt and tree_int_cst_le.
1843 * tree-streamer-in.c (unpack_ts_base_value_fields): Use wide-int
1845 (streamer_alloc_tree): Likewise.
1846 * tree-streamer-out.c (pack_ts_int_cst_value_fields): Likewise.
1847 (streamer_write_tree_header): Likewise.
1848 (streamer_write_integer_cst): Likewise.
1849 * tree-switch-conversion.c (emit_case_bit_tests): Likewise.
1850 (build_constructors): Likewise.
1851 (array_value_type): Likewise.
1852 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list): Likewise.
1853 (vect_check_gather): Likewise.
1854 * tree-vect-generic.c (build_replicated_const): Likewise.
1855 (expand_vector_divmod): Likewise.
1856 * tree-vect-loop.c (vect_transform_loop): Likewise.
1857 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
1858 (vect_do_peeling_for_alignment): Likewise.
1859 * tree-vect-patterns.c (vect_recog_divmod_pattern): Likewise.
1860 * tree-vrp.c: Include wide-int.h.
1861 (operand_less_p): Use wide-int interfaces and tree_int_cst_lt.
1862 (extract_range_from_assert): Use wide-int interfaces.
1863 (vrp_int_const_binop): Likewise.
1864 (zero_nonzero_bits_from_vr): Take wide_int pointers rather than
1865 double_int pointers.
1866 (ranges_from_anti_range): Use wide-int interfaces.
1867 (quad_int_cmp): Delete.
1868 (quad_int_pair_sort): Likewise.
1869 (extract_range_from_binary_expr_1): Use wide-int interfaces.
1870 (extract_range_from_unary_expr_1): Likewise.
1871 (adjust_range_with_scev): Likewise.
1872 (masked_increment): Take and return wide_ints rather than double_ints.
1873 (register_edge_assert_for_2): Use wide-int interfaces.
1874 (check_array_ref): Likewise.
1875 (search_for_addr_array): Likewise.
1876 (maybe_set_nonzero_bits): Likewise.
1877 (union_ranges): Pass an integer of the correct type instead of
1878 using integer_one_node.
1879 (intersect_ranges): Likewise.
1880 (simplify_truth_ops_using_ranges): Likewise.
1881 (simplify_bit_ops_using_ranges): Use wide-int interfaces.
1882 (range_fits_type_p): Likewise.
1883 (simplify_cond_using_ranges): Likewise. Take a signop rather than
1885 (simplify_conversion_using_ranges): Use wide-int interfaces.
1886 (simplify_float_conversion_using_ranges): Likewise.
1887 (vrp_finalize): Likewise.
1888 * value-prof.c (gimple_divmod_fixed_value_transform): Likewise.
1889 (gimple_stringops_transform): Likewise.
1890 * varasm.c (decode_addr_const): Likewise.
1891 (const_hash_1): Likewise.
1892 (const_rtx_hash_1): Likewise
1893 (output_constant): Likewise.
1894 (array_size_for_constructor): Likewise.
1895 (output_constructor_regular_field): Likewise.
1896 (output_constructor_bitfield): Likewise.
1897 * var-tracking.c (loc_cmp): Handle CONST_WIDE_INT.
1898 * mkconfig.sh: Include machmode.h to pick up BITS_PER_UNIT for
1900 * gencheck.c: Define BITS_PER_UNIT.
1903 * wide-int-print.cc: New.
1904 * wide-int-print.h: New.
1906 2014-05-06 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1908 * config/avr/avr.c (avr_can_eliminate): Mark unused argument.
1910 2014-05-06 Richard Biener <rguenther@suse.de>
1912 * tree-pass.h (TODO_verify_ssa, TODO_verify_flow,
1913 TODO_verify_stmts, TODO_verify_rtl_sharing): Remove.
1914 (TODO_verify_all): Adjust.
1915 * asan.c: Remove references to TODO_verify_ssa, TODO_verify_flow,
1916 TODO_verify_stmts and TODO_verify_rtl_sharing.
1917 * bb-reorder.c: Likewise.
1918 * cfgexpand.c: Likewise.
1919 * cprop.c: Likewise.
1921 * function.c: Likewise.
1922 * fwprop.c: Likewise.
1924 * gimple-ssa-isolate-paths.c: Likewise.
1925 * gimple-ssa-strength-reduction.c: Likewise.
1926 * ipa-split.c: Likewise.
1927 * loop-init.c: Likewise.
1928 * loop-unroll.c: Likewise.
1929 * lower-subreg.c: Likewise.
1930 * modulo-sched.c: Likewise.
1931 * postreload-gcse.c: Likewise.
1932 * predict.c: Likewise.
1933 * recog.c: Likewise.
1934 * sched-rgn.c: Likewise.
1935 * store-motion.c: Likewise.
1936 * tracer.c: Likewise.
1937 * trans-mem.c: Likewise.
1938 * tree-call-cdce.c: Likewise.
1939 * tree-cfg.c: Likewise.
1940 * tree-cfgcleanup.c: Likewise.
1941 * tree-complex.c: Likewise.
1942 * tree-eh.c: Likewise.
1943 * tree-emutls.c: Likewise.
1944 * tree-if-conv.c: Likewise.
1945 * tree-into-ssa.c: Likewise.
1946 * tree-loop-distribution.c: Likewise.
1947 * tree-object-size.c: Likewise.
1948 * tree-parloops.c: Likewise.
1949 * tree-pass.h: Likewise.
1950 * tree-sra.c: Likewise.
1951 * tree-ssa-ccp.c: Likewise.
1952 * tree-ssa-copy.c: Likewise.
1953 * tree-ssa-copyrename.c: Likewise.
1954 * tree-ssa-dce.c: Likewise.
1955 * tree-ssa-dom.c: Likewise.
1956 * tree-ssa-dse.c: Likewise.
1957 * tree-ssa-forwprop.c: Likewise.
1958 * tree-ssa-ifcombine.c: Likewise.
1959 * tree-ssa-loop-ch.c: Likewise.
1960 * tree-ssa-loop-ivcanon.c: Likewise.
1961 * tree-ssa-loop.c: Likewise.
1962 * tree-ssa-math-opts.c: Likewise.
1963 * tree-ssa-phiopt.c: Likewise.
1964 * tree-ssa-phiprop.c: Likewise.
1965 * tree-ssa-pre.c: Likewise.
1966 * tree-ssa-reassoc.c: Likewise.
1967 * tree-ssa-sink.c: Likewise.
1968 * tree-ssa-strlen.c: Likewise.
1969 * tree-ssa-tail-merge.c: Likewise.
1970 * tree-ssa-uncprop.c: Likewise.
1971 * tree-switch-conversion.c: Likewise.
1972 * tree-tailcall.c: Likewise.
1973 * tree-vect-generic.c: Likewise.
1974 * tree-vectorizer.c: Likewise.
1975 * tree-vrp.c: Likewise.
1977 * var-tracking.c: Likewise.
1978 * bt-load.c: Likewise.
1979 * cfgcleanup.c: Likewise.
1980 * combine-stack-adj.c: Likewise.
1981 * combine.c: Likewise.
1982 * compare-elim.c: Likewise.
1983 * config/epiphany/resolve-sw-modes.c: Likewise.
1984 * config/i386/i386.c: Likewise.
1985 * config/mips/mips.c: Likewise.
1986 * config/s390/s390.c: Likewise.
1987 * config/sh/sh_treg_combine.cc: Likewise.
1988 * config/sparc/sparc.c: Likewise.
1991 * final.c: Likewise.
1992 * ifcvt.c: Likewise.
1993 * mode-switching.c: Likewise.
1994 * passes.c: Likewise.
1995 * postreload.c: Likewise.
1997 * reg-stack.c: Likewise.
1998 * regcprop.c: Likewise.
1999 * regrename.c: Likewise.
2002 2014-05-06 Richard Biener <rguenther@suse.de>
2005 * bitmap.c (debug_bitmap): Dump to stderr, not stdout.
2006 * tree-ssa-structalias.c (dump_solution_for_var): Likewise.
2008 2014-05-05 Jan Hubicka <hubicka@ucw.cz>
2011 * ipa-devirt.c (get_class_context): Allow POD to change to non-POD.
2013 2014-05-05 Radovan Obradovic <robradovic@mips.com>
2014 Tom de Vries <tom@codesourcery.com>
2016 * target.def (call_fusage_contains_non_callee_clobbers): New
2018 * doc/tm.texi.in (@node Stack and Calling): Add Miscellaneous Register
2020 (@node Miscellaneous Register Hooks): New node.
2021 (@hook TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS): New hook.
2022 * doc/tm.texi: Regenerate.
2024 2014-05-05 Marek Polacek <polacek@redhat.com>
2027 * opts.c (common_handle_option): Call error_at instead of warning_at.
2029 2014-05-05 Richard Biener <rguenther@suse.de>
2031 * passes.c (execute_function_todo): Don't reset TODO_verify_ssa
2032 from last_verified if update_ssa ran. Move TODO_verify_rtl_sharing
2033 under the TODO_verify_il umbrella.
2035 2014-05-05 Richard Biener <rguenther@suse.de>
2037 * passes.c (execute_function_todo): Move TODO_verify_flow under
2038 the TODO_verify_ul umbrella.
2040 2014-05-05 Richard Biener <rguenther@suse.de>
2043 * fold-const.c (fold_binary_loc): Consistently avoid canonicalizing
2044 X & CST away from a CST that is the mask of a mode.
2046 2014-05-05 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2048 * config/picochip/picochip-protos.h (picochip_regno_nregs): Change
2049 int argument to enum machine_mode.
2050 (picochip_class_max_nregs): Ditto.
2051 * config/picochip/picochip.c (picochip_regno_nregs): Ditto.
2052 (picochip_class_max_nregs): Ditto.
2054 2014-05-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2056 * target.def: Add new target hook.
2057 * doc/tm.texi: Regenerate.
2058 * targhooks.h (default_keep_leaf_when_profiled): Add prototype.
2059 * targhooks.c (default_keep_leaf_when_profiled): New function.
2061 * config/s390/s390.c (s390_keep_leaf_when_profiled): New function.
2062 (TARGET_KEEP_LEAF_WHEN_PROFILED): Define.
2064 2014-05-05 Bin Cheng <bin.cheng@arm.com>
2066 PR tree-optimization/60363
2067 * gcc/tree-ssa-threadupdate.c (get_value_locus_in_path): New.
2068 (copy_phi_args): New parameters. Call get_value_locus_in_path.
2069 (update_destination_phis): New parameter.
2070 (create_edge_and_update_destination_phis): Ditto.
2071 (ssa_fix_duplicate_block_edges): Pass new arguments.
2072 (thread_single_edge): Ditto.
2074 2014-05-04 Peter Bergner <bergner@vnet.ibm.com>
2076 * config/rs6000/rs6000.h (RS6000_BTM_HARD_FLOAT): New define.
2077 (RS6000_BTM_COMMON): Add RS6000_BTM_HARD_FLOAT.
2078 (TARGET_EXTRA_BUILTINS): Add TARGET_HARD_FLOAT.
2079 * config/rs6000/rs6000-builtin.def (BU_MISC_1):
2080 Use RS6000_BTM_HARD_FLOAT.
2081 (BU_MISC_2): Likewise.
2082 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Handle
2083 RS6000_BTM_HARD_FLOAT.
2084 (rs6000_option_override_internal): Enforce -mhard-float if -mhard-dfp
2086 (rs6000_invalid_builtin): Add hard floating builtin support.
2087 (rs6000_expand_builtin): Relax the gcc_assert to allow the new
2088 hard float builtins.
2089 (rs6000_builtin_mask_names): Add RS6000_BTM_HARD_FLOAT.
2091 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
2093 * config/sh/sh_optimize_sett_clrt.cc (sh_optimize_sett_clrt::execute):
2094 Add missing function* argument.
2096 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
2098 * lra-constraints.c (valid_address_p): Move earlier in file.
2099 Add a constraint argument to the address_info version.
2100 (satisfies_memory_constraint_p): New function.
2101 (satisfies_address_constraint_p): Likewise.
2102 (process_alt_operands, curr_insn_transform): Use them.
2103 (process_address): Pass the constraint to valid_address_p when
2104 checking address operands.
2106 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
2108 * config/mips/mips.c (mips_isa_rev): New variable.
2109 (mips_set_architecture): Set it.
2110 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Set __mips_isa_rev
2112 (ISA_HAS_MUL3, ISA_HAS_FP_CONDMOVE, ISA_HAS_8CC, ISA_HAS_FP4)
2113 (ISA_HAS_PAIRED_SINGLE, ISA_HAS_MADD_MSUB, ISA_HAS_FP_RECIP_RSQRT)
2114 (ISA_HAS_CLZ_CLO, ISA_HAS_ROR, ISA_HAS_WSBH, ISA_HAS_PREFETCH)
2115 (ISA_HAS_SEB_SEH, ISA_HAS_EXT_INS, ISA_HAS_MXHC1)
2116 (ISA_HAS_HILO_INTERLOCKS, ISA_HAS_SYNCI, MIN_FPRS_PER_FMT): Reexpress
2117 conditions in terms of mips_isa_rev.
2118 (mips_isa_rev): Declare.
2120 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
2122 * config/sh/sh-mem.cc: Use tabs instead of spaces.
2123 (prob_unlikely, prob_likely): Make variables const.
2125 2014-05-03 Denis Chertykov <chertykov@gmail.com>
2127 * config/avr/avr.c (avr_adjust_insn_length): Handle JUMP_TABLE_DATA.
2129 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
2131 * config/sh/sh.h (SH_ASM_SPEC): Handle m1, m2*, m3* and m4* cases.
2133 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
2135 * config/sh/sh.h (ROUND_ADVANCE): Delete macro.
2136 (ROUND_REG, PASS_IN_REG_P): Move and rename macros to ...
2137 * config/sh/sh.c (sh_round_reg, sh_pass_in_reg_p): ... these new
2139 (sh_arg_partial_bytes, sh_function_arg, sh_function_arg_advance,
2140 sh_setup_incoming_varargs): Replace usage of PASS_IN_REG_P with
2142 Replace usage of ROUND_REG with sh_round_reg.
2143 Use CEIL instead of ROUND_ADVANCE.
2145 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
2148 * config/sh/sh.c: Include stdlib headers before everything else.
2150 2014-05-02 Jakub Jelinek <jakub@redhat.com>
2152 * gimplify.c (gimplify_adjust_omp_clauses_1): Handle
2153 GOVD_FIRSTPRIVATE | GOVD_LASTPRIVATE.
2154 (gimplify_adjust_omp_clauses): Simd region is never
2155 directly nested in combined parallel. Instead, for linear
2156 with copyin/copyout, if in combined for simd loop, make decl
2157 firstprivate/lastprivate on OMP_FOR.
2158 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
2159 expand_omp_for_static_chunk): When setting endvar, also set
2160 fd->loop.v to the same value.
2162 2014-05-02 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
2164 * hwint.h (zext_hwi): Fix signed overflow for prec == 63.
2166 2014-05-02 Alan Lawrence <alan.lawrence@arm.com>
2168 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_1): Tidy bit-flip
2171 2014-05-02 Marek Polacek <polacek@redhat.com>
2173 * doc/invoke.texi: Describe -fsanitize=float-divide-by-zero.
2175 2014-05-02 Kito Cheng <kito@0xlab.org>
2177 * defaults.h (HONOR_REG_ALLOC_ORDER): Change HONOR_REG_ALLOC_ORDER
2178 to a C expression marco.
2179 * ira-color.c (HONOR_REG_ALLOC_ORDER) : Ditto.
2180 * config/arm/arm.h (HONOR_REG_ALLOC_ORDER): Ditto.
2181 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Ditto.
2182 * doc/tm.texi (HONOR_REG_ALLOC_ORDER): Update document for
2183 HONOR_REG_ALLOC_ORDER.
2184 * doc/tm.texi.in (HONOR_REG_ALLOC_ORDER): Ditto.
2186 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2188 * config/arc/arc.c (TARGET_LRA_P): Undef before redefine.
2190 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2192 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
2194 2014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
2196 * tree-if-conv.c (is_cond_scalar_reduction): New function.
2197 (convert_scalar_cond_reduction): Likewise.
2198 (predicate_scalar_phi): Add recognition and transformation
2199 of simple conditioanl reduction to be vectorizable.
2201 2014-05-01 Marek Polacek <polacek@redhat.com>
2204 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
2206 2014-04-30 Alan Lawrence <alan.lawrence@arm.com>
2208 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
2209 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
2210 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
2211 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
2212 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
2213 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
2214 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
2215 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
2217 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
2219 * config/arc/arc.opt (mlra): Move comment above option name
2220 to avoid mis-parsing as language options.
2222 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2224 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
2225 * config/sol2.h: ... here.
2226 * config/sol2-10.h: Remove.
2228 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
2229 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
2230 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
2231 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
2232 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
2233 * config/sol2.h: ... here.
2234 (SECTION_NAME_FORMAT): Don't redefine.
2235 (STARTFILE_ARCH32_SPEC): Rename to ...
2236 (STARTFILE_ARCH_SPEC): ... this.
2237 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
2238 * config/sparc/sol2.h: ... here.
2239 (SECTION_NAME_FORMAT): Don't undef.
2240 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
2241 (SUBTARGET_EXTRA_SPECS): Remove.
2242 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
2244 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
2245 (MD_STARTFILE_PREFIX): Remove.
2246 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
2247 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
2248 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
2249 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
2250 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
2251 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
2252 * config/i386/sol2.h: ... here.
2253 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
2254 * config/i386/sol2-bi.h: Remove.
2255 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
2256 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
2258 * config/i386/t-sol2-64: Rename to ...
2259 * config/i386/t-sol2: ... this.
2260 * config/sparc/t-sol2-64: Rename to ...
2261 * config/sparc/t-sol2: ... this.
2263 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
2264 sol2_tm_file_head, sol2_tm_file_tail.
2265 Include ${cpu_type}/sol2.h before sol2.h.
2267 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
2268 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
2269 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
2270 Reflect i386/t-sol2-64 renaming.
2271 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
2272 Reflect sparc/t-sol2-64 renaming.
2274 2014-04-30 Richard Biener <rguenther@suse.de>
2276 * passes.c (execute_function_todo): Move TODO_verify_stmts
2277 and TODO_verify_ssa under the TODO_verify_il umbrella.
2278 * tree-ssa.h (verify_ssa): Adjust prototype.
2279 * tree-ssa.c (verify_ssa): Add parameter to tell whether
2280 we should verify SSA operands.
2281 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
2282 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
2283 whether we should verify whether not throwing stmts have EH info.
2284 * graphite-scop-detection.c (create_sese_edges): Adjust.
2285 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
2286 * tree-eh.c (lower_try_finally_switch): Do not add the
2287 default case label twice.
2289 2014-04-30 Marek Polacek <polacek@redhat.com>
2291 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
2292 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
2293 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
2294 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
2296 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
2298 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
2299 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
2300 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
2301 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
2302 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
2303 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
2304 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
2305 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
2307 2014-04-29 David Malcolm <dmalcolm@redhat.com>
2309 * tree-cfg.c (dump_function_to_file): Dump the return type of
2310 functions, in a line to itself before the function body, mimicking
2311 the layout of a C function.
2313 2014-04-29 Jakub Jelinek <jakub@redhat.com>
2315 PR tree-optimization/60971
2316 * tree-tailcall.c (process_assignment): Reject conversions which
2319 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
2321 * calls.c (initialize_argument_information): Always treat
2322 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
2323 (expand_call): Likewise.
2324 (emit_library_call_calue_1): Likewise.
2325 * expr.c (PUSH_ARGS_REVERSED): Do not define.
2326 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
2329 2014-04-29 Nick Clifton <nickc@redhat.com>
2331 * config/msp430/msp430.md (umulsidi): Fix typo.
2332 (mulhisi3): Enable even inside interrupt handlers.
2333 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
2334 bigger return address pushed in large mode.
2336 2014-04-29 Nick Clifton <nickc@redhat.com>
2338 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
2339 (arc_init_reg_tables): Use a machine_mode enum to iterate over
2341 * config/m32r/m32r.c (init_reg_tables): Likewise.
2342 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
2343 enum to hold the modes.
2345 2014-04-29 Richard Biener <rguenther@suse.de>
2347 * dominance.c (free_dominance_info): Add overload with
2349 (dom_info_state): Likewise.
2350 (dom_info_available_p): Likewise.
2351 * basic-block.h (free_dominance_info, dom_info_state,
2352 dom_info_available_p): Declare overloads.
2353 * passes.c (execute_function_todo): Verify that verifiers
2354 don't change dominator info state. Drop dominator info
2355 for IPA pass invocations.
2356 * cgraph.c (release_function_body): Restore asserts that
2357 dominator information is released.
2359 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
2361 * doc/invoke.texi: Fix typo.
2362 * tree-vrp.c: Fix typos.
2363 * gimple.c (infer_nonnull_range): Reorder operands of an && condition.
2365 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2367 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
2369 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
2371 * config/aarch64/aarch64-builtins.c
2372 (aarch64_types_storestruct_lane_qualifiers): New.
2373 (TYPES_STORESTRUCT_LANE): Likewise.
2374 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
2375 (st3_lane): Likewise.
2376 (st4_lane): Likewise.
2377 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
2378 (vec_store_lanesci_lane<mode>): Likewise.
2379 (vec_store_lanesxi_lane<mode>): Likewise.
2380 (aarch64_st2_lane<VQ:mode>): Likewise.
2381 (aarch64_st3_lane<VQ:mode>): Likewise.
2382 (aarch64_st4_lane<VQ:mode>): Likewise.
2383 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
2384 * config/aarch64/arm_neon.h
2385 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
2386 use new macro arguments.
2387 (__ST3_LANE_FUNC): Likewise.
2388 (__ST4_LANE_FUNC): Likewise.
2389 * config/aarch64/iterators.md (V_TWO_ELEM): New.
2390 (V_THREE_ELEM): Likewise.
2391 (V_FOUR_ELEM): Likewise.
2393 2014-04-28 David Malcolm <dmalcolm@redhat.com>
2395 * doc/gimple.texi: Replace the description of the now-defunct
2396 union gimple_statement_d with a diagram showing the
2397 gimple_statement_base class hierarchy and its relationships to
2398 the GSS_ and GIMPLE_ enums.
2400 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
2402 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
2403 * config/aarch64/aarch64.c
2404 (aarch64_cannot_change_mode_class): Weaken conditions.
2405 (aarch64_modes_tieable_p): New.
2406 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
2408 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
2410 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2411 (loadsync_<mode>): Change mode.
2412 (load_quadpti, store_quadpti): New.
2413 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2414 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2416 2014-04-28 Martin Jambor <mjambor@suse.cz>
2418 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
2419 same alias type as the original statement.
2420 (subreplacement_assignment_data): New type.
2421 (handle_unscalarized_data_in_subtree): New type of parameter,
2422 generate new memory accesses with same alias type as the original
2424 (load_assign_lhs_subreplacements): Likewise.
2425 (sra_modify_constructor_assign): Generate new memory accesses with
2426 same alias type as the original statement.
2428 2014-04-28 Richard Biener <rguenther@suse.de>
2430 * tree-pass.h (TODO_verify_il): Define.
2431 (TODO_verify_all): Complete properly.
2432 * passes.c (execute_function_todo): Move existing loop-closed
2433 SSA verification under TODO_verify_il.
2434 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
2435 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
2436 Fix tree sharing issue.
2438 2014-04-28 Richard Biener <rguenther@suse.de>
2441 * builtins.def (DEF_C11_BUILTIN): Add.
2442 (BUILT_IN_ALIGNED_ALLOC): Likewise.
2443 * coretypes.h (enum function_class): Add function_c11_misc.
2444 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
2445 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
2446 (call_may_clobber_ref_p_1): Likewise.
2447 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
2448 (mark_all_reaching_defs_necessary_1): Likewise.
2449 (propagate_necessity): Likewise.
2450 (eliminate_unnecessary_stmts): Likewise.
2451 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
2453 2014-04-28 Richard Biener <rguenther@suse.de>
2455 * tree-vrp.c (vrp_var_may_overflow): Remove.
2456 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
2457 with overflow immediately bump to one before that value and
2458 let iteration figure out overflow status.
2460 2014-04-28 Richard Biener <rguenther@suse.de>
2462 * configure.ac: Do valgrind header checks unconditionally.
2463 Add --enable-valgrind-annotations.
2464 * system.h: Guard valgrind header inclusion with
2465 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
2466 * alloc-pool.c (pool_alloc, pool_free): Use
2467 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
2468 to guard possibly dead code.
2469 * config.in: Regenerated.
2470 * configure: Likewise.
2472 2014-04-28 Jeff Law <law@redhat.com>
2474 PR tree-optimization/60902
2475 * tree-ssa-threadedge.c
2476 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
2477 over real defs when invalidating outputs from statements that do not
2478 produce useful outputs for threading.
2480 2014-04-28 Richard Biener <rguenther@suse.de>
2482 PR tree-optimization/60979
2483 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
2484 SCOPs that end in a block with a successor with abnormal
2487 2014-04-28 Richard Biener <rguenther@suse.de>
2489 * tree-pass.h (execute_pass_list): Adjust prototype.
2490 * passes.c (pass_manager::execute_early_local_passes): Adjust.
2491 (do_per_function): Change callback signature, push all actual
2492 work to the callbals.
2493 (do_per_function_toporder): Likewise.
2494 (execute_function_dump): Adjust.
2495 (execute_function_todo): Likewise.
2496 (clear_last_verified): Likewise.
2497 (verify_curr_properties): Likewise.
2498 (update_properties_after_pass): Likewise.
2499 (execute_pass_list_1): Split out from ...
2500 (execute_pass_list): ... here. Adjust.
2501 (execute_ipa_pass_list): Likewise.
2502 * cgraphunit.c (cgraph_add_new_function): Adjust.
2503 (analyze_function): Likewise.
2504 (expand_function): Likewise.
2505 * cgraph.c (release_function_body): Free dominance info
2506 here instead of asserting it was magically freed elsewhere.
2508 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
2510 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
2511 * configure: Regenerate.
2512 * config/sparc/sparc.opt (muser-mode): New option.
2513 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
2515 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
2516 * doc/invoke.texi (SPARC options): Document -muser-mode.
2518 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
2520 * cselib.c (find_slot_memmode): Delete.
2521 (cselib_hasher): Change compare_type to a struct.
2522 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
2524 (preserve_constants_and_equivs): Adjust for new compare_type.
2525 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
2526 (wrap_constant): Delete.
2527 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
2529 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
2531 * doc/install.texi (Building with profile feedback): Remove
2534 2014-04-26 Tom de Vries <tom@codesourcery.com>
2536 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
2539 2014-04-25 Cary Coutant <ccoutant@google.com>
2542 * dwarf2out.c (should_move_die_to_comdat): A type definition
2543 can contain a subprogram definition, but don't move it to a
2545 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
2546 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
2548 (clone_tree_hash): Rename to...
2549 (clone_tree_partial): ...this; change callers. Copy
2550 DW_TAG_subprogram DIEs as declarations.
2551 (copy_decls_walk): Don't copy children of a declaration into a
2554 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
2557 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
2560 2014-04-25 Jiong Wang <jiong.wang@arm.com>
2562 * config/arm/predicates.md (call_insn_operand): Add long_call check.
2563 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
2565 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
2568 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2570 * config/arm/arm.c (arm_cortex_a8_tune): Initialise T16-related fields.
2572 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2574 PR tree-optimization/60930
2575 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
2576 creating a multiply candidate by folding two constant
2577 multiplicands when the result overflows.
2579 2014-04-25 Jakub Jelinek <jakub@redhat.com>
2581 PR tree-optimization/60960
2582 * tree-vect-generic.c (expand_vector_operation): Only call
2583 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
2585 2014-04-25 Tom de Vries <tom@codesourcery.com>
2587 * expr.c (clobber_reg_mode): New function.
2588 * expr.h (clobber_reg): New function.
2590 2014-04-25 Tom de Vries <tom@codesourcery.com>
2592 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
2595 2014-04-25 Radovan Obradovic <robradovic@mips.com>
2596 Tom de Vries <tom@codesourcery.com>
2598 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
2600 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
2601 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
2602 new argument to find_all_hard_reg_sets call.
2604 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2606 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
2607 Use HOST_WIDE_INT_C for mask literal.
2608 (aarch_rev16_shleft_mask_imm_p): Likewise.
2610 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
2613 * config/sparc/sparc.md (ashlsi3_extend): Delete.
2615 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
2617 PR preprocessor/56540
2618 * config/i386/i386-c.c (ix86_target_macros): Define
2619 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
2621 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2623 * configure.ac (tga_func): Remove.
2624 (LIB_TLS_SPEC): Remove.
2625 * configure: Regenerate.
2626 * config.in: Regenerate.
2627 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
2629 2014-04-25 Richard Biener <rguenther@suse.de>
2632 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
2633 call stmt use/clobber sets during stmt walk instead of
2634 walking the possibly incomplete set of caller edges.
2636 2014-04-25 Richard Biener <rguenther@suse.de>
2639 * passes.c (apply_ipa_transforms): Inline into only caller ...
2640 (execute_one_pass): ... here. Properly bring in function
2641 bodies for nodes we want to apply IPA transforms to.
2643 2014-04-24 Cong Hou <congh@google.com>
2645 PR tree-optimization/60896
2646 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
2647 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
2648 (vect_mark_pattern_stmts): Set the def type of all statements in
2649 PATTERN_DEF_SEQ as vect_internal_def.
2651 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
2653 * doc/extend.texi (PowerPC Built-in Functions): Document new
2654 powerpc extended divide, bcd, pack/unpack 128-bit, builtin functions.
2655 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
2657 * config/rs6000/predicates.md (const_0_to_3_operand): New
2658 predicate to match 0..3 integer constants.
2660 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
2661 to support adding miscellaneous builtin functions.
2662 (BU_DFP_MISC_2): Likewise.
2663 (BU_P7_MISC_1): Likewise.
2664 (BU_P7_MISC_2): Likewise.
2665 (BU_P8V_MISC_3): Likewise.
2666 (BU_MISC_1): Likewise.
2667 (BU_MISC_2): Likewise.
2668 (DIVWE): Add extended divide builtin functions.
2671 (DIVWEUO): Likewise.
2675 (DIVDEUO): Likewise.
2676 (DXEX): Add decimal floating-point builtin functions.
2679 (DDEDPDQ): Likewise.
2681 (DENBCDQ): Likewise.
2688 (CDTBCD): Add new BCD builtin functions.
2692 (BCDADD_LT): Likewise.
2693 (BCDADD_EQ): Likewise.
2694 (BCDADD_GT): Likewise.
2695 (BCDADD_OV): Likewise.
2697 (BCDSUB_LT): Likewise.
2698 (BCDSUB_EQ): Likewise.
2699 (BCDSUB_GT): Likewise.
2700 (BCDSUB_OV): Likewise.
2701 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
2702 (UNPACK_TD): Likewise.
2703 (PACK_TF): Likewise.
2704 (UNPACK_TF): Likewise.
2705 (UNPACK_TF_0): Likewise.
2706 (UNPACK_TF_1): Likewise.
2707 (PACK_V1TI): Likewise.
2708 (UNPACK_V1TI): Likewise.
2710 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
2711 support for decimal floating point builtin functions.
2712 (rs6000_expand_ternop_builtin): Add checks for the new builtin
2713 functions that take constant arguments.
2714 (rs6000_invalid_builtin): Add decimal floating point builtin support.
2715 (rs6000_init_builtins): Setup long double, _Decimal64, and
2716 _Decimal128 types for new builtin functions.
2717 (builtin_function_type): Set the unsigned flags appropriately for
2718 the new builtin functions.
2719 (rs6000_opt_masks): Add support for decimal floating point builtin
2722 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
2723 floating point builtin functions.
2724 (RS6000_BTM_COMMON): Likewise.
2725 (RS6000_BTI_long_double): Likewise.
2726 (RS6000_BTI_dfloat64): Likewise.
2727 (RS6000_BTI_dfloat128): Likewise.
2728 (long_double_type_internal_node): Likewise.
2729 (dfloat64_type_internal_node): Likewise.
2730 (dfloat128_type_internal_node): Likewise.
2732 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
2733 2.07 bcd arithmetic instructions.
2734 (UNSPEC_BCDSUB): Likewise.
2735 (UNSPEC_BCD_OVERFLOW): Likewise.
2736 (UNSPEC_BCD_ADD_SUB): Likewise.
2737 (bcd_add_sub): Likewise.
2738 (BCD_TEST): Likewise.
2739 (bcd<bcd_add_sub>): Likewise.
2740 (bcd<bcd_add_sub>_test): Likewise.
2741 (bcd<bcd_add_sub>_test2): Likewise.
2742 (bcd<bcd_add_sub>_<code>): Likewise.
2743 (peephole2 for combined bcd ops): Likewise.
2745 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
2746 decimal floating point builtin functions.
2747 (UNSPEC_DENBCD): Likewise.
2748 (UNSPEC_DXEX): Likewise.
2749 (UNSPEC_DIEX): Likewise.
2750 (UNSPEC_DSCLI): Likewise.
2751 (UNSPEC_DSCRI): Likewise.
2752 (D64_D128): Likewise.
2753 (dfp_suffix): Likewise.
2754 (dfp_ddedpd_<mode>): Likewise.
2755 (dfp_denbcd_<mode>): Likewise.
2756 (dfp_dxex_<mode>): Likewise.
2757 (dfp_diex_<mode>): Likewise.
2758 (dfp_dscli_<mode>): Likewise.
2759 (dfp_dscri_<mode>): Likewise.
2761 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
2763 (UNSPEC_CDTBCD): Likewise.
2764 (UNSPEC_CBCDTD): Likewise.
2765 (UNSPEC_DIVE): Add support for new extended divide builtin functions.
2766 (UNSPEC_DIVEO): Likewise.
2767 (UNSPEC_DIVEU): Likewise.
2768 (UNSPEC_DIVEUO): Likewise.
2769 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
2770 pack/unpack 128-bit types.
2771 (UNSPEC_PACK_128BIT): Likewise.
2772 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
2773 (udiv<mode>3): Use idiv_ldiv mode attribute.
2774 (div<mode>3): Likewise.
2775 (addg6s): Add new BCD builtin functions.
2778 (UNSPEC_DIV_EXTEND): Add support for new extended divide instructions.
2779 (div_extend): Likewise.
2780 (div<div_extend>_<mode>"): Likewise.
2781 (FP128_64): Add support for new builtin functions to pack/unpack
2783 (unpack<mode>): Likewise.
2784 (unpacktf_0): Likewise.
2785 (unpacktf_1): Likewise.
2786 (unpack<mode>_dm): Likewise.
2787 (unpack<mode>_nodm): Likewise.
2788 (pack<mode>): Likewise.
2789 (unpackv1ti): Likewise.
2790 (packv1ti): Likewise.
2792 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
2794 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
2797 2014-04-24 Jakub Jelinek <jakub@redhat.com>
2799 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
2800 * gimplify.c (omp_is_private): Change last argument's type to int.
2801 Only diagnose lastprivate if the simd argument is 1, only diagnose
2802 linear if the simd argument is 2.
2803 (gimplify_omp_for): Adjust omp_is_private callers. When adding
2804 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
2805 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
2806 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
2807 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
2808 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
2809 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
2810 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
2811 * tree-nested.c (convert_nonlocal_omp_clauses,
2812 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
2814 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
2817 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
2820 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
2822 * flag-types.h (enum ivar_visibility): Add.
2824 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
2826 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
2827 function * argument.
2829 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
2831 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
2833 2014-04-24 Radovan Obradovic <robradovic@mips.com>
2834 Tom de Vries <tom@codesourcery.com>
2836 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
2837 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
2839 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
2840 * emit-rtl.c (try_split): Same.
2842 2014-04-24 Radovan Obradovic <robradovic@mips.com>
2843 Tom de Vries <tom@codesourcery.com>
2845 * common.opt (fuse-caller-save): New option.
2847 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
2849 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of
2850 elements for big-endian.
2852 2014-04-24 Richard Biener <rguenther@suse.de>
2854 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
2855 during TER and instead use the sepops interface for expanding
2856 non-GIMPLE_SINGLE_RHS.
2858 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2860 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
2861 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
2863 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2865 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
2866 assembler 64-bit option.
2867 * configure: Regenerate.
2869 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2871 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
2872 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
2873 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
2874 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
2875 (TARGET_CRYPTO): Take TARGET_SIMD into account.
2877 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2879 * config/aarch64/aarch64-builtins.c
2880 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
2881 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
2882 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
2883 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
2885 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
2886 (Vrevsuff): New mode attribute.
2888 2014-04-24 Terry Guo <terry.guo@arm.com>
2890 * config/arm/arm.h (machine_function): Define variable
2891 after_arm_reorg here.
2892 * config/arm/arm.c (after_arm_reorg): Remove the definition.
2893 (arm_split_constant): Update the way to access variable
2896 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
2898 2014-04-23 Tom de Vries <tom@codesourcery.com>
2900 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
2902 2014-04-23 David Malcolm <dmalcolm@redhat.com>
2904 * is-a.h: Update comments to reflect the following changes to the
2905 "pointerness" of the API, making the template parameter match the
2906 return type, allowing use of is-a.h with typedefs of pointers.
2907 (is_a_helper::cast): Return a T rather then a pointer to a T, so
2908 that the return type matches the parameter to the is_a_helper.
2910 (dyn_cast): Likewise.
2912 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
2913 pointer from the is-a.h API.
2915 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
2916 (is_a_helper <cgraph_node *>::test): ...this, matching change to
2918 (is_a_helper <varpool_node>::test): Likewise, convert to...
2919 (is_a_helper <varpool_node *>::test): ...this.
2921 (varpool_first_variable): Update for removal of implicit pointer
2922 from the is-a.h API.
2923 (varpool_next_variable): Likewise.
2924 (varpool_first_static_initializer): Likewise.
2925 (varpool_next_static_initializer): Likewise.
2926 (varpool_first_defined_variable): Likewise.
2927 (varpool_next_defined_variable): Likewise.
2928 (cgraph_first_defined_function): Likewise.
2929 (cgraph_next_defined_function): Likewise.
2930 (cgraph_first_function): Likewise.
2931 (cgraph_next_function): Likewise.
2932 (cgraph_first_function_with_gimple_body): Likewise.
2933 (cgraph_next_function_with_gimple_body): Likewise.
2934 (cgraph_alias_target): Likewise.
2935 (varpool_alias_target): Likewise.
2936 (cgraph_function_or_thunk_node): Likewise.
2937 (varpool_variable_node): Likewise.
2938 (symtab_real_symbol_p): Likewise.
2939 * cgraphunit.c (referred_to_p): Likewise.
2940 (analyze_functions): Likewise.
2941 (handle_alias_pairs): Likewise.
2942 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
2943 * gimple-ssa.h (gimple_vuse_op): Likewise.
2944 (gimple_vdef_op): Likewise.
2945 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
2946 * gimple.c (gimple_build_asm_1): Likewise.
2947 (gimple_build_try): Likewise.
2948 (gimple_build_resx): Likewise.
2949 (gimple_build_eh_dispatch): Likewise.
2950 (gimple_build_omp_for): Likewise.
2951 (gimple_omp_for_set_clauses): Likewise.
2953 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
2954 (is_a_helper <gimple_statement_asm *>::test): ...this.
2955 (is_a_helper <gimple_statement_bind>::test): Convert to...
2956 (is_a_helper <gimple_statement_bind *>::test): ...this.
2957 (is_a_helper <gimple_statement_call>::test): Convert to...
2958 (is_a_helper <gimple_statement_call *>::test): ...this.
2959 (is_a_helper <gimple_statement_catch>::test): Convert to...
2960 (is_a_helper <gimple_statement_catch *>::test): ...this.
2961 (is_a_helper <gimple_statement_resx>::test): Convert to...
2962 (is_a_helper <gimple_statement_resx *>::test): ...this.
2963 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
2964 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
2965 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
2966 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
2967 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
2968 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
2969 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
2970 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
2971 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
2972 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
2973 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
2974 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
2975 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
2976 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
2977 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
2978 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
2979 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
2980 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
2981 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
2982 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
2983 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
2984 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
2985 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
2986 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
2987 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
2988 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
2989 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
2990 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
2991 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
2992 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
2993 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
2994 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
2995 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
2996 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
2997 (is_a_helper <gimple_statement_phi>::test): Convert to...
2998 (is_a_helper <gimple_statement_phi *>::test): ...this.
2999 (is_a_helper <gimple_statement_transaction>::test): Convert to...
3000 (is_a_helper <gimple_statement_transaction *>::test): ...this.
3001 (is_a_helper <gimple_statement_try>::test): Convert to...
3002 (is_a_helper <gimple_statement_try *>::test): ...this.
3003 (is_a_helper <gimple_statement_wce>::test): Convert to...
3004 (is_a_helper <gimple_statement_wce *>::test): ...this.
3005 (is_a_helper <const gimple_statement_asm>::test): Convert to...
3006 (is_a_helper <const gimple_statement_asm *>::test): ...this.
3007 (is_a_helper <const gimple_statement_bind>::test): Convert to...
3008 (is_a_helper <const gimple_statement_bind *>::test): ...this.
3009 (is_a_helper <const gimple_statement_call>::test): Convert to...
3010 (is_a_helper <const gimple_statement_call *>::test): ...this.
3011 (is_a_helper <const gimple_statement_catch>::test): Convert to...
3012 (is_a_helper <const gimple_statement_catch *>::test): ...this.
3013 (is_a_helper <const gimple_statement_resx>::test): Convert to...
3014 (is_a_helper <const gimple_statement_resx *>::test): ...this.
3015 (is_a_helper <const gimple_statement_eh_dispatch>::test): Convert to...
3016 (is_a_helper <const gimple_statement_eh_dispatch *>::test): ...this.
3017 (is_a_helper <const gimple_statement_eh_filter>::test): Convert to...
3018 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
3019 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
3021 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
3023 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
3025 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
3027 (is_a_helper <const gimple_statement_omp_return>::test): Convert to...
3028 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
3029 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
3031 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
3032 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
3034 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
3035 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
3036 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
3037 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert to...
3038 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
3039 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
3041 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
3042 (is_a_helper <const gimple_statement_omp_target>::test): Convert to...
3043 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
3044 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
3046 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
3047 (is_a_helper <const gimple_statement_omp_single>::test): Convert to...
3048 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
3049 (is_a_helper <const gimple_statement_omp_teams>::test): Convert to...
3050 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
3051 (is_a_helper <const gimple_statement_omp_task>::test): Convert to...
3052 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
3053 (is_a_helper <const gimple_statement_phi>::test): Convert to...
3054 (is_a_helper <const gimple_statement_phi *>::test): ...this.
3055 (is_a_helper <const gimple_statement_transaction>::test): Convert to...
3056 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
3057 (is_a_helper <const gimple_statement_with_ops>::test): Convert to...
3058 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
3059 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
3060 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
3061 (is_a_helper <const gimple_statement_with_memory_ops>::test): Convert
3063 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
3065 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert to...
3066 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
3068 (gimple_use_ops): Update for removal of implicit pointer from the
3070 (gimple_set_use_ops): Likewise.
3071 (gimple_vuse): Likewise.
3072 (gimple_vdef): Likewise.
3073 (gimple_vuse_ptr): Likewise.
3074 (gimple_vdef_ptr): Likewise.
3075 (gimple_set_vuse): Likewise.
3076 (gimple_set_vdef): Likewise.
3077 (gimple_omp_return_set_lhs): Likewise.
3078 (gimple_omp_return_lhs): Likewise.
3079 (gimple_omp_return_lhs_ptr): Likewise.
3080 (gimple_call_fntype): Likewise.
3081 (gimple_call_set_fntype): Likewise.
3082 (gimple_call_set_internal_fn): Likewise.
3083 (gimple_call_use_set): Likewise.
3084 (gimple_call_clobber_set): Likewise.
3085 (gimple_bind_vars): Likewise.
3086 (gimple_bind_set_vars): Likewise.
3087 (gimple_bind_body_ptr): Likewise.
3088 (gimple_bind_set_body): Likewise.
3089 (gimple_bind_add_stmt): Likewise.
3090 (gimple_bind_block): Likewise.
3091 (gimple_bind_set_block): Likewise.
3092 (gimple_asm_ninputs): Likewise.
3093 (gimple_asm_noutputs): Likewise.
3094 (gimple_asm_nclobbers): Likewise.
3095 (gimple_asm_nlabels): Likewise.
3096 (gimple_asm_input_op): Likewise.
3097 (gimple_asm_input_op_ptr): Likewise.
3098 (gimple_asm_output_op): Likewise.
3099 (gimple_asm_output_op_ptr): Likewise.
3100 (gimple_asm_set_output_op): Likewise.
3101 (gimple_asm_clobber_op): Likewise.
3102 (gimple_asm_set_clobber_op): Likewise.
3103 (gimple_asm_label_op): Likewise.
3104 (gimple_asm_set_label_op): Likewise.
3105 (gimple_asm_string): Likewise.
3106 (gimple_catch_types): Likewise.
3107 (gimple_catch_types_ptr): Likewise.
3108 (gimple_catch_handler_ptr): Likewise.
3109 (gimple_catch_set_types): Likewise.
3110 (gimple_catch_set_handler): Likewise.
3111 (gimple_eh_filter_types): Likewise.
3112 (gimple_eh_filter_types_ptr): Likewise.
3113 (gimple_eh_filter_failure_ptr): Likewise.
3114 (gimple_eh_filter_set_types): Likewise.
3115 (gimple_eh_filter_set_failure): Likewise.
3116 (gimple_eh_must_not_throw_fndecl): Likewise.
3117 (gimple_eh_must_not_throw_set_fndecl): Likewise.
3118 (gimple_eh_else_n_body_ptr): Likewise.
3119 (gimple_eh_else_e_body_ptr): Likewise.
3120 (gimple_eh_else_set_n_body): Likewise.
3121 (gimple_eh_else_set_e_body): Likewise.
3122 (gimple_try_eval_ptr): Likewise.
3123 (gimple_try_cleanup_ptr): Likewise.
3124 (gimple_try_set_eval): Likewise.
3125 (gimple_try_set_cleanup): Likewise.
3126 (gimple_wce_cleanup_ptr): Likewise.
3127 (gimple_wce_set_cleanup): Likewise.
3128 (gimple_phi_capacity): Likewise.
3129 (gimple_phi_num_args): Likewise.
3130 (gimple_phi_result): Likewise.
3131 (gimple_phi_result_ptr): Likewise.
3132 (gimple_phi_set_result): Likewise.
3133 (gimple_phi_arg): Likewise.
3134 (gimple_phi_set_arg): Likewise.
3135 (gimple_resx_region): Likewise.
3136 (gimple_resx_set_region): Likewise.
3137 (gimple_eh_dispatch_region): Likewise.
3138 (gimple_eh_dispatch_set_region): Likewise.
3139 (gimple_omp_critical_name): Likewise.
3140 (gimple_omp_critical_name_ptr): Likewise.
3141 (gimple_omp_critical_set_name): Likewise.
3142 (gimple_omp_for_clauses): Likewise.
3143 (gimple_omp_for_clauses_ptr): Likewise.
3144 (gimple_omp_for_set_clauses): Likewise.
3145 (gimple_omp_for_collapse): Likewise.
3146 (gimple_omp_for_index): Likewise.
3147 (gimple_omp_for_index_ptr): Likewise.
3148 (gimple_omp_for_set_index): Likewise.
3149 (gimple_omp_for_initial): Likewise.
3150 (gimple_omp_for_initial_ptr): Likewise.
3151 (gimple_omp_for_set_initial): Likewise.
3152 (gimple_omp_for_final): Likewise.
3153 (gimple_omp_for_final_ptr): Likewise.
3154 (gimple_omp_for_set_final): Likewise.
3155 (gimple_omp_for_incr): Likewise.
3156 (gimple_omp_for_incr_ptr): Likewise.
3157 (gimple_omp_for_set_incr): Likewise.
3158 (gimple_omp_for_pre_body_ptr): Likewise.
3159 (gimple_omp_for_set_pre_body): Likewise.
3160 (gimple_omp_parallel_clauses): Likewise.
3161 (gimple_omp_parallel_clauses_ptr): Likewise.
3162 (gimple_omp_parallel_set_clauses): Likewise.
3163 (gimple_omp_parallel_child_fn): Likewise.
3164 (gimple_omp_parallel_child_fn_ptr): Likewise.
3165 (gimple_omp_parallel_set_child_fn): Likewise.
3166 (gimple_omp_parallel_data_arg): Likewise.
3167 (gimple_omp_parallel_data_arg_ptr): Likewise.
3168 (gimple_omp_parallel_set_data_arg): Likewise.
3169 (gimple_omp_task_clauses): Likewise.
3170 (gimple_omp_task_clauses_ptr): Likewise.
3171 (gimple_omp_task_set_clauses): Likewise.
3172 (gimple_omp_task_child_fn): Likewise.
3173 (gimple_omp_task_child_fn_ptr): Likewise.
3174 (gimple_omp_task_set_child_fn): Likewise.
3175 (gimple_omp_task_data_arg): Likewise.
3176 (gimple_omp_task_data_arg_ptr): Likewise.
3177 (gimple_omp_task_set_data_arg): Likewise.
3178 (gimple_omp_taskreg_clauses): Likewise.
3179 (gimple_omp_taskreg_clauses_ptr): Likewise.
3180 (gimple_omp_taskreg_set_clauses): Likewise.
3181 (gimple_omp_taskreg_child_fn): Likewise.
3182 (gimple_omp_taskreg_child_fn_ptr): Likewise.
3183 (gimple_omp_taskreg_set_child_fn): Likewise.
3184 (gimple_omp_taskreg_data_arg): Likewise.
3185 (gimple_omp_taskreg_data_arg_ptr): Likewise.
3186 (gimple_omp_taskreg_set_data_arg): Likewise.
3187 (gimple_omp_task_copy_fn): Likewise.
3188 (gimple_omp_task_copy_fn_ptr): Likewise.
3189 (gimple_omp_task_set_copy_fn): Likewise.
3190 (gimple_omp_task_arg_size): Likewise.
3191 (gimple_omp_task_arg_size_ptr): Likewise.
3192 (gimple_omp_task_set_arg_size): Likewise.
3193 (gimple_omp_task_arg_align): Likewise.
3194 (gimple_omp_task_arg_align_ptr): Likewise.
3195 (gimple_omp_task_set_arg_align): Likewise.
3196 (gimple_omp_single_clauses): Likewise.
3197 (gimple_omp_single_clauses_ptr): Likewise.
3198 (gimple_omp_single_set_clauses): Likewise.
3199 (gimple_omp_target_clauses): Likewise.
3200 (gimple_omp_target_clauses_ptr): Likewise.
3201 (gimple_omp_target_set_clauses): Likewise.
3202 (gimple_omp_target_child_fn): Likewise.
3203 (gimple_omp_target_child_fn_ptr): Likewise.
3204 (gimple_omp_target_set_child_fn): Likewise.
3205 (gimple_omp_target_data_arg): Likewise.
3206 (gimple_omp_target_data_arg_ptr): Likewise.
3207 (gimple_omp_target_set_data_arg): Likewise.
3208 (gimple_omp_teams_clauses): Likewise.
3209 (gimple_omp_teams_clauses_ptr): Likewise.
3210 (gimple_omp_teams_set_clauses): Likewise.
3211 (gimple_omp_sections_clauses): Likewise.
3212 (gimple_omp_sections_clauses_ptr): Likewise.
3213 (gimple_omp_sections_set_clauses): Likewise.
3214 (gimple_omp_sections_control): Likewise.
3215 (gimple_omp_sections_control_ptr): Likewise.
3216 (gimple_omp_sections_set_control): Likewise.
3217 (gimple_omp_for_set_cond): Likewise.
3218 (gimple_omp_for_cond): Likewise.
3219 (gimple_omp_atomic_store_set_val): Likewise.
3220 (gimple_omp_atomic_store_val): Likewise.
3221 (gimple_omp_atomic_store_val_ptr): Likewise.
3222 (gimple_omp_atomic_load_set_lhs): Likewise.
3223 (gimple_omp_atomic_load_lhs): Likewise.
3224 (gimple_omp_atomic_load_lhs_ptr): Likewise.
3225 (gimple_omp_atomic_load_set_rhs): Likewise.
3226 (gimple_omp_atomic_load_rhs): Likewise.
3227 (gimple_omp_atomic_load_rhs_ptr): Likewise.
3228 (gimple_omp_continue_control_def): Likewise.
3229 (gimple_omp_continue_control_def_ptr): Likewise.
3230 (gimple_omp_continue_set_control_def): Likewise.
3231 (gimple_omp_continue_control_use): Likewise.
3232 (gimple_omp_continue_control_use_ptr): Likewise.
3233 (gimple_omp_continue_set_control_use): Likewise.
3234 (gimple_transaction_body_ptr): Likewise.
3235 (gimple_transaction_label): Likewise.
3236 (gimple_transaction_label_ptr): Likewise.
3237 (gimple_transaction_set_body): Likewise.
3238 (gimple_transaction_set_label): Likewise.
3240 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
3241 * ipa-inline-analysis.c (inline_write_summary): Likewise.
3242 * ipa-ref.c (ipa_record_reference): Likewise.
3243 * ipa-reference.c (analyze_function): Likewise.
3244 (ipa_reference_write_optimization_summary): Likewise.
3245 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
3246 (address_taken_from_non_vtable_p): Likewise.
3247 (comdat_can_be_unshared_p_1): Likewise.
3248 * lto-cgraph.c (lto_output_ref): Likewise.
3249 (add_references): Likewise.
3250 (compute_ltrans_boundary): Likewise.
3251 (output_symtab): Likewise.
3252 (input_ref): Likewise.
3253 (input_cgraph_1): Likewise.
3254 (output_cgraph_opt_summary): Likewise.
3255 * lto-streamer-out.c (lto_output): Likewise.
3256 (output_symbol_p): Likewise.
3257 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
3258 (lsei_start_function_in_partition): Likewise.
3259 (lsei_next_variable_in_partition): Likewise.
3260 (lsei_start_variable_in_partition): Likewise.
3261 * symtab.c (insert_to_assembler_name_hash): Likewise.
3262 (unlink_from_assembler_name_hash): Likewise.
3263 (symtab_unregister_node): Likewise.
3264 (symtab_remove_node): Likewise.
3265 (dump_symtab_node): Likewise.
3266 (verify_symtab_base): Likewise.
3267 (verify_symtab_node): Likewise.
3268 (symtab_make_decl_local): Likewise.
3269 (symtab_alias_ultimate_target): Likewise.
3270 (symtab_resolve_alias): Likewise.
3271 (symtab_get_symbol_partitioning_class): Likewise.
3272 * tree-phinodes.c (allocate_phi_node): Likewise.
3273 (reserve_phi_args_for_new_edge): Likewise.
3274 (remove_phi_args): Likewise.
3275 * varpool.c (varpool_node_for_asm): Likewise.
3276 (varpool_remove_unreferenced_decls): Likewise.
3278 2014-04-23 Jeff Law <law@redhat.com>
3280 PR tree-optimization/60902
3281 * tree-ssa-threadedge.c
3282 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
3283 invalidate outputs from statements that do not produce useful
3284 outputs for threading.
3286 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
3288 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
3289 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
3290 machine descriptions for Stack Smashing Protector.
3292 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
3294 * aarch64.md (<optab>_rol<mode>3): New pattern.
3295 (<optab>_rolsi3_uxtw): Likewise.
3296 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
3298 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
3300 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
3301 (arm_cortex_a12_tune): Likewise.
3303 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3305 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
3307 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3309 * config/arm/arm.md (arm_rev16si2): New pattern.
3310 (arm_rev16si2_alt): Likewise.
3311 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
3313 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3315 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
3316 (rev16<mode>2_alt): Likewise.
3317 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
3318 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
3319 (aarch_rev16_shleft_mask_imm_p): Likewise.
3320 (aarch_rev16_p_1): Likewise.
3321 (aarch_rev16_p): Likewise.
3322 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
3323 (aarch_rev16_shright_mask_imm_p): Likewise.
3324 (aarch_rev16_shleft_mask_imm_p): Likewise.
3326 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3328 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
3329 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
3331 (cortex_a53_extra_costs): Likewise.
3332 (cortex_a57_extra_costs): Likewise.
3333 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
3334 (cortexa7_extra_costs): Likewise.
3335 (cortexa8_extra_costs): Likewise.
3336 (cortexa12_extra_costs): Likewise.
3337 (cortexa15_extra_costs): Likewise.
3338 (v7m_extra_costs): Likewise.
3339 (arm_new_rtx_costs): Handle BSWAP.
3341 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3343 * config/arm/arm.c (cortexa8_extra_costs): New table.
3344 (arm_cortex_a8_tune): New tuning struct.
3345 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
3347 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3349 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
3351 2014-04-23 Richard Biener <rguenther@suse.de>
3353 * Makefile.in (OBJS): Remove loop-unswitch.o.
3354 * tree-pass.h (make_pass_rtl_unswitch): Remove.
3355 * passes.def (pass_rtl_unswitch): Likewise.
3356 * loop-init.c (gate_rtl_unswitch): Likewise.
3357 (rtl_unswitch): Likewise.
3358 (pass_data_rtl_unswitch): Likewise.
3359 (pass_rtl_unswitch): Likewise.
3360 (make_pass_rtl_unswitch): Likewise.
3361 * rtl.h (reversed_condition): Likewise.
3362 (compare_and_jump_seq): Likewise.
3363 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
3365 * loop-unroll.c (compare_and_jump_seq): Likewise.
3367 2014-04-23 Richard Biener <rguenther@suse.de>
3369 PR tree-optimization/60903
3370 * tree-ssa-loop-im.c (analyze_memory_references): Remove
3371 commented code block.
3372 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
3373 loop flags to newly created BBs and edges.
3375 2014-04-23 Nick Clifton <nickc@redhat.com>
3377 * config/msp430/msp430.c (msp430_handle_option): Move function
3379 (msp430_option_override): Simplify mcu and mcpu option handling.
3380 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
3381 support for -mhwmult command line option.
3382 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
3383 -mhwmult command line option.
3384 (msp430_hwmult_enabled): Delete.
3385 (msp43o_output_labelref): Add support for -mhwmult command line option.
3386 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
3387 (umulsidi3): Likewise.
3388 * config/msp430/msp430.opt (mmcu): Add Report attribute.
3389 (mcpu, mlarge, msmall): Likewise.
3390 (mhwmult): New option.
3391 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
3393 (msp430_is_f5_mcu): Remove prototype.
3394 (msp430_use_f5_series_hwmult): Add prototype.
3395 * config/msp430/msp430-opts.h: New file.
3396 * common/config/msp430: New directory.
3397 * common/config/msp430/msp430-common.c: New file.
3398 * config.gcc (msp430): Remove target_has_targetm_common.
3399 * doc/invoke.texi: Document -mhwmult command line option.
3401 2014-04-23 Nick Clifton <nickc@redhat.com>
3403 * config/i386/cygwin.h (ENDFILE_SPEC): Include
3404 default-manifest.o if it can be found in the search path.
3405 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
3407 2014-04-23 Terry Guo <terry.guo@arm.com>
3409 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
3411 2014-04-23 Richard Biener <rguenther@suse.de>
3414 * tree-inline.c (declare_return_variable): Use mark_addressable.
3416 2014-04-23 Richard Biener <rguenther@suse.de>
3419 * loop-init.c (loop_optimizer_init): Make sure to apply
3420 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
3422 2014-04-22 Jakub Jelinek <jakub@redhat.com>
3425 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
3427 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
3428 if flag_sanitize_undefined_trap_on_error.
3429 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
3430 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
3431 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
3432 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
3433 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
3434 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
3435 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
3436 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
3437 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
3438 * ubsan.c (ubsan_instrument_unreachable): Return
3439 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
3440 (ubsan_expand_null_ifn): Emit __builtin_trap ()
3441 if flag_sanitize_undefined_trap_on_error and
3442 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
3443 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
3444 instrument_bool_enum_load): Emit __builtin_trap () if
3445 flag_sanitize_undefined_trap_on_error and
3446 __builtin_handle_*_abort () if !flag_sanitize_recover.
3447 * doc/invoke.texi (-fsanitize-recover,
3448 -fsanitize-undefined-trap-on-error): Document.
3450 2014-04-22 Christian Bruel <christian.bruel@st.com>
3452 * config/sh/sh.md (mov<mode>): Replace movQIHI.
3453 Force immediates to SImode.
3455 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
3457 * config/nios2/nios2.md (UNSPEC_ROUND): New.
3459 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
3460 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
3461 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
3462 (nios2_fpu_insn): Add entry for round.
3463 (N2FPU_NO_ERRNO_P): Define.
3464 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
3466 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
3468 2014-04-22 Richard Henderson <rth@redhat.com>
3470 * config/aarch64/aarch64 (addti3, subti3): New expanders.
3471 (add<GPI>3_compare0): Remove leading * from name.
3472 (add<GPI>3_carryin): Likewise.
3473 (sub<GPI>3_compare0): Likewise.
3474 (sub<GPI>3_carryin): Likewise.
3475 (<su_optab>mulditi3): New expander.
3476 (multi3): New expander.
3477 (madd<GPI>): Remove leading * from name.
3479 2014-04-22 Martin Jambor <mjambor@suse.cz>
3481 * cgraphclones.c (cgraph_function_versioning): Copy
3482 ipa_transforms_to_apply instead of asserting it is empty.
3484 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
3487 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
3488 on count_exp to get mode.
3490 2014-04-22 Andrew Pinski <apinski@cavium.com>
3492 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
3493 Handle TLS for ILP32.
3494 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
3495 (tlsie_small_<mode>): this and handle PTR.
3496 (tlsie_small_sidi): New pattern.
3497 (tlsle_small): Change to an expand to handle ILP32.
3498 (tlsle_small_<mode>): New pattern.
3499 (tlsdesc_small): Rename to ...
3500 (tlsdesc_small_<mode>): this and handle PTR.
3502 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3504 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
3506 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
3508 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
3509 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
3510 (aarch64_types_signed_poly_qualifiers): Likewise.
3511 (aarch64_types_unsigned_signed_qualifiers): Likewise.
3512 (aarch64_types_poly_signed_qualifiers): Likewise.
3513 (TYPES_REINTERP_SS): Type macro added.
3514 (TYPES_REINTERP_SU): Likewise.
3515 (TYPES_REINTERP_SP): Likewise.
3516 (TYPES_REINTERP_US): Likewise.
3517 (TYPES_REINTERP_PS): Likewise.
3518 (aarch64_fold_builtin): New expression folding added.
3519 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
3520 Declarations removed.
3521 (REINTERP_SS): Declarations added.
3522 (REINTERP_US): Likewise.
3523 (REINTERP_PS): Likewise.
3524 (REINTERP_SU): Likewise.
3525 (REINTERP_SP): Likewise.
3526 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
3527 (vreinterpretq_p8_f64): Likewise.
3528 (vreinterpret_p16_f64): Likewise.
3529 (vreinterpretq_p16_f64): Likewise.
3530 (vreinterpret_f32_f64): Likewise.
3531 (vreinterpretq_f32_f64): Likewise.
3532 (vreinterpret_f64_f32): Likewise.
3533 (vreinterpret_f64_p8): Likewise.
3534 (vreinterpret_f64_p16): Likewise.
3535 (vreinterpret_f64_s8): Likewise.
3536 (vreinterpret_f64_s16): Likewise.
3537 (vreinterpret_f64_s32): Likewise.
3538 (vreinterpret_f64_s64): Likewise.
3539 (vreinterpret_f64_u8): Likewise.
3540 (vreinterpret_f64_u16): Likewise.
3541 (vreinterpret_f64_u32): Likewise.
3542 (vreinterpret_f64_u64): Likewise.
3543 (vreinterpretq_f64_f32): Likewise.
3544 (vreinterpretq_f64_p8): Likewise.
3545 (vreinterpretq_f64_p16): Likewise.
3546 (vreinterpretq_f64_s8): Likewise.
3547 (vreinterpretq_f64_s16): Likewise.
3548 (vreinterpretq_f64_s32): Likewise.
3549 (vreinterpretq_f64_s64): Likewise.
3550 (vreinterpretq_f64_u8): Likewise.
3551 (vreinterpretq_f64_u16): Likewise.
3552 (vreinterpretq_f64_u32): Likewise.
3553 (vreinterpretq_f64_u64): Likewise.
3554 (vreinterpret_s64_f64): Likewise.
3555 (vreinterpretq_s64_f64): Likewise.
3556 (vreinterpret_u64_f64): Likewise.
3557 (vreinterpretq_u64_f64): Likewise.
3558 (vreinterpret_s8_f64): Likewise.
3559 (vreinterpretq_s8_f64): Likewise.
3560 (vreinterpret_s16_f64): Likewise.
3561 (vreinterpretq_s16_f64): Likewise.
3562 (vreinterpret_s32_f64): Likewise.
3563 (vreinterpretq_s32_f64): Likewise.
3564 (vreinterpret_u8_f64): Likewise.
3565 (vreinterpretq_u8_f64): Likewise.
3566 (vreinterpret_u16_f64): Likewise.
3567 (vreinterpretq_u16_f64): Likewise.
3568 (vreinterpret_u32_f64): Likewise.
3569 (vreinterpretq_u32_f64): Likewise.
3571 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
3573 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
3574 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
3575 (vreinterpret_p8_s8): Likewise.
3576 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
3577 (vreinterpret_p8_s16): Likewise.
3578 (vreinterpret_p8_s32): Likewise.
3579 (vreinterpret_p8_s64): Likewise.
3580 (vreinterpret_p8_f32): Likewise.
3581 (vreinterpret_p8_u8): Likewise.
3582 (vreinterpret_p8_u16): Likewise.
3583 (vreinterpret_p8_u32): Likewise.
3584 (vreinterpret_p8_u64): Likewise.
3585 (vreinterpret_p8_p16): Likewise.
3586 (vreinterpretq_p8_s8): Likewise.
3587 (vreinterpretq_p8_s16): Likewise.
3588 (vreinterpretq_p8_s32): Likewise.
3589 (vreinterpretq_p8_s64): Likewise.
3590 (vreinterpretq_p8_f32): Likewise.
3591 (vreinterpretq_p8_u8): Likewise.
3592 (vreinterpretq_p8_u16): Likewise.
3593 (vreinterpretq_p8_u32): Likewise.
3594 (vreinterpretq_p8_u64): Likewise.
3595 (vreinterpretq_p8_p16): Likewise.
3596 (vreinterpret_p16_s8): Likewise.
3597 (vreinterpret_p16_s16): Likewise.
3598 (vreinterpret_p16_s32): Likewise.
3599 (vreinterpret_p16_s64): Likewise.
3600 (vreinterpret_p16_f32): Likewise.
3601 (vreinterpret_p16_u8): Likewise.
3602 (vreinterpret_p16_u16): Likewise.
3603 (vreinterpret_p16_u32): Likewise.
3604 (vreinterpret_p16_u64): Likewise.
3605 (vreinterpret_p16_p8): Likewise.
3606 (vreinterpretq_p16_s8): Likewise.
3607 (vreinterpretq_p16_s16): Likewise.
3608 (vreinterpretq_p16_s32): Likewise.
3609 (vreinterpretq_p16_s64): Likewise.
3610 (vreinterpretq_p16_f32): Likewise.
3611 (vreinterpretq_p16_u8): Likewise.
3612 (vreinterpretq_p16_u16): Likewise.
3613 (vreinterpretq_p16_u32): Likewise.
3614 (vreinterpretq_p16_u64): Likewise.
3615 (vreinterpretq_p16_p8): Likewise.
3616 (vreinterpret_f32_s8): Likewise.
3617 (vreinterpret_f32_s16): Likewise.
3618 (vreinterpret_f32_s32): Likewise.
3619 (vreinterpret_f32_s64): Likewise.
3620 (vreinterpret_f32_u8): Likewise.
3621 (vreinterpret_f32_u16): Likewise.
3622 (vreinterpret_f32_u32): Likewise.
3623 (vreinterpret_f32_u64): Likewise.
3624 (vreinterpret_f32_p8): Likewise.
3625 (vreinterpret_f32_p16): Likewise.
3626 (vreinterpretq_f32_s8): Likewise.
3627 (vreinterpretq_f32_s16): Likewise.
3628 (vreinterpretq_f32_s32): Likewise.
3629 (vreinterpretq_f32_s64): Likewise.
3630 (vreinterpretq_f32_u8): Likewise.
3631 (vreinterpretq_f32_u16): Likewise.
3632 (vreinterpretq_f32_u32): Likewise.
3633 (vreinterpretq_f32_u64): Likewise.
3634 (vreinterpretq_f32_p8): Likewise.
3635 (vreinterpretq_f32_p16): Likewise.
3636 (vreinterpret_s64_s8): Likewise.
3637 (vreinterpret_s64_s16): Likewise.
3638 (vreinterpret_s64_s32): Likewise.
3639 (vreinterpret_s64_f32): Likewise.
3640 (vreinterpret_s64_u8): Likewise.
3641 (vreinterpret_s64_u16): Likewise.
3642 (vreinterpret_s64_u32): Likewise.
3643 (vreinterpret_s64_u64): Likewise.
3644 (vreinterpret_s64_p8): Likewise.
3645 (vreinterpret_s64_p16): Likewise.
3646 (vreinterpretq_s64_s8): Likewise.
3647 (vreinterpretq_s64_s16): Likewise.
3648 (vreinterpretq_s64_s32): Likewise.
3649 (vreinterpretq_s64_f32): Likewise.
3650 (vreinterpretq_s64_u8): Likewise.
3651 (vreinterpretq_s64_u16): Likewise.
3652 (vreinterpretq_s64_u32): Likewise.
3653 (vreinterpretq_s64_u64): Likewise.
3654 (vreinterpretq_s64_p8): Likewise.
3655 (vreinterpretq_s64_p16): Likewise.
3656 (vreinterpret_u64_s8): Likewise.
3657 (vreinterpret_u64_s16): Likewise.
3658 (vreinterpret_u64_s32): Likewise.
3659 (vreinterpret_u64_s64): Likewise.
3660 (vreinterpret_u64_f32): Likewise.
3661 (vreinterpret_u64_u8): Likewise.
3662 (vreinterpret_u64_u16): Likewise.
3663 (vreinterpret_u64_u32): Likewise.
3664 (vreinterpret_u64_p8): Likewise.
3665 (vreinterpret_u64_p16): Likewise.
3666 (vreinterpretq_u64_s8): Likewise.
3667 (vreinterpretq_u64_s16): Likewise.
3668 (vreinterpretq_u64_s32): Likewise.
3669 (vreinterpretq_u64_s64): Likewise.
3670 (vreinterpretq_u64_f32): Likewise.
3671 (vreinterpretq_u64_u8): Likewise.
3672 (vreinterpretq_u64_u16): Likewise.
3673 (vreinterpretq_u64_u32): Likewise.
3674 (vreinterpretq_u64_p8): Likewise.
3675 (vreinterpretq_u64_p16): Likewise.
3676 (vreinterpret_s8_s16): Likewise.
3677 (vreinterpret_s8_s32): Likewise.
3678 (vreinterpret_s8_s64): Likewise.
3679 (vreinterpret_s8_f32): Likewise.
3680 (vreinterpret_s8_u8): Likewise.
3681 (vreinterpret_s8_u16): Likewise.
3682 (vreinterpret_s8_u32): Likewise.
3683 (vreinterpret_s8_u64): Likewise.
3684 (vreinterpret_s8_p8): Likewise.
3685 (vreinterpret_s8_p16): Likewise.
3686 (vreinterpretq_s8_s16): Likewise.
3687 (vreinterpretq_s8_s32): Likewise.
3688 (vreinterpretq_s8_s64): Likewise.
3689 (vreinterpretq_s8_f32): Likewise.
3690 (vreinterpretq_s8_u8): Likewise.
3691 (vreinterpretq_s8_u16): Likewise.
3692 (vreinterpretq_s8_u32): Likewise.
3693 (vreinterpretq_s8_u64): Likewise.
3694 (vreinterpretq_s8_p8): Likewise.
3695 (vreinterpretq_s8_p16): Likewise.
3696 (vreinterpret_s16_s8): Likewise.
3697 (vreinterpret_s16_s32): Likewise.
3698 (vreinterpret_s16_s64): Likewise.
3699 (vreinterpret_s16_f32): Likewise.
3700 (vreinterpret_s16_u8): Likewise.
3701 (vreinterpret_s16_u16): Likewise.
3702 (vreinterpret_s16_u32): Likewise.
3703 (vreinterpret_s16_u64): Likewise.
3704 (vreinterpret_s16_p8): Likewise.
3705 (vreinterpret_s16_p16): Likewise.
3706 (vreinterpretq_s16_s8): Likewise.
3707 (vreinterpretq_s16_s32): Likewise.
3708 (vreinterpretq_s16_s64): Likewise.
3709 (vreinterpretq_s16_f32): Likewise.
3710 (vreinterpretq_s16_u8): Likewise.
3711 (vreinterpretq_s16_u16): Likewise.
3712 (vreinterpretq_s16_u32): Likewise.
3713 (vreinterpretq_s16_u64): Likewise.
3714 (vreinterpretq_s16_p8): Likewise.
3715 (vreinterpretq_s16_p16): Likewise.
3716 (vreinterpret_s32_s8): Likewise.
3717 (vreinterpret_s32_s16): Likewise.
3718 (vreinterpret_s32_s64): Likewise.
3719 (vreinterpret_s32_f32): Likewise.
3720 (vreinterpret_s32_u8): Likewise.
3721 (vreinterpret_s32_u16): Likewise.
3722 (vreinterpret_s32_u32): Likewise.
3723 (vreinterpret_s32_u64): Likewise.
3724 (vreinterpret_s32_p8): Likewise.
3725 (vreinterpret_s32_p16): Likewise.
3726 (vreinterpretq_s32_s8): Likewise.
3727 (vreinterpretq_s32_s16): Likewise.
3728 (vreinterpretq_s32_s64): Likewise.
3729 (vreinterpretq_s32_f32): Likewise.
3730 (vreinterpretq_s32_u8): Likewise.
3731 (vreinterpretq_s32_u16): Likewise.
3732 (vreinterpretq_s32_u32): Likewise.
3733 (vreinterpretq_s32_u64): Likewise.
3734 (vreinterpretq_s32_p8): Likewise.
3735 (vreinterpretq_s32_p16): Likewise.
3736 (vreinterpret_u8_s8): Likewise.
3737 (vreinterpret_u8_s16): Likewise.
3738 (vreinterpret_u8_s32): Likewise.
3739 (vreinterpret_u8_s64): Likewise.
3740 (vreinterpret_u8_f32): Likewise.
3741 (vreinterpret_u8_u16): Likewise.
3742 (vreinterpret_u8_u32): Likewise.
3743 (vreinterpret_u8_u64): Likewise.
3744 (vreinterpret_u8_p8): Likewise.
3745 (vreinterpret_u8_p16): Likewise.
3746 (vreinterpretq_u8_s8): Likewise.
3747 (vreinterpretq_u8_s16): Likewise.
3748 (vreinterpretq_u8_s32): Likewise.
3749 (vreinterpretq_u8_s64): Likewise.
3750 (vreinterpretq_u8_f32): Likewise.
3751 (vreinterpretq_u8_u16): Likewise.
3752 (vreinterpretq_u8_u32): Likewise.
3753 (vreinterpretq_u8_u64): Likewise.
3754 (vreinterpretq_u8_p8): Likewise.
3755 (vreinterpretq_u8_p16): Likewise.
3756 (vreinterpret_u16_s8): Likewise.
3757 (vreinterpret_u16_s16): Likewise.
3758 (vreinterpret_u16_s32): Likewise.
3759 (vreinterpret_u16_s64): Likewise.
3760 (vreinterpret_u16_f32): Likewise.
3761 (vreinterpret_u16_u8): Likewise.
3762 (vreinterpret_u16_u32): Likewise.
3763 (vreinterpret_u16_u64): Likewise.
3764 (vreinterpret_u16_p8): Likewise.
3765 (vreinterpret_u16_p16): Likewise.
3766 (vreinterpretq_u16_s8): Likewise.
3767 (vreinterpretq_u16_s16): Likewise.
3768 (vreinterpretq_u16_s32): Likewise.
3769 (vreinterpretq_u16_s64): Likewise.
3770 (vreinterpretq_u16_f32): Likewise.
3771 (vreinterpretq_u16_u8): Likewise.
3772 (vreinterpretq_u16_u32): Likewise.
3773 (vreinterpretq_u16_u64): Likewise.
3774 (vreinterpretq_u16_p8): Likewise.
3775 (vreinterpretq_u16_p16): Likewise.
3776 (vreinterpret_u32_s8): Likewise.
3777 (vreinterpret_u32_s16): Likewise.
3778 (vreinterpret_u32_s32): Likewise.
3779 (vreinterpret_u32_s64): Likewise.
3780 (vreinterpret_u32_f32): Likewise.
3781 (vreinterpret_u32_u8): Likewise.
3782 (vreinterpret_u32_u16): Likewise.
3783 (vreinterpret_u32_u64): Likewise.
3784 (vreinterpret_u32_p8): Likewise.
3785 (vreinterpret_u32_p16): Likewise.
3786 (vreinterpretq_u32_s8): Likewise.
3787 (vreinterpretq_u32_s16): Likewise.
3788 (vreinterpretq_u32_s32): Likewise.
3789 (vreinterpretq_u32_s64): Likewise.
3790 (vreinterpretq_u32_f32): Likewise.
3791 (vreinterpretq_u32_u8): Likewise.
3792 (vreinterpretq_u32_u16): Likewise.
3793 (vreinterpretq_u32_u64): Likewise.
3794 (vreinterpretq_u32_p8): Likewise.
3795 (vreinterpretq_u32_p16): Likewise.
3797 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
3799 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
3801 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator extended.
3803 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
3804 (vqnegd_s64): Likewise.
3805 (vqabs_s64): Likewise.
3806 (vqabsd_s64): Likewise.
3808 2014-04-22 Richard Henderson <rth@redhat.com>
3810 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
3811 computation to the top of the loop.
3813 2014-04-22 Renlin <renlin.li@arm.com>
3814 Jiong Wang <jiong.wang@arm.com>
3816 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
3817 * config/aarch64/aarch64.c (aarch64_layout_frame)
3818 (aarch64_initial_elimination_offset): Likewise.
3820 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
3822 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
3825 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
3827 * machmode.h (bitwise_mode_for_mode): Declare.
3828 * stor-layout.h (bitwise_type_for_mode): Likewise.
3829 * stor-layout.c (bitwise_mode_for_mode): New function.
3830 (bitwise_type_for_mode): Likewise.
3831 * builtins.c (fold_builtin_memory_op): Use it instead of
3832 int_mode_for_mode and build_nonstandard_integer_type.
3834 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3836 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
3837 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
3838 (*-*-solaris2*): Simplify.
3839 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
3840 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
3841 *-*-solaris2.9* handling.
3843 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
3845 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
3846 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
3848 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
3849 * configure: Regenerate.
3851 * config/i386/sol2-9.h: Remove.
3853 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
3854 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
3855 Remove Solaris 9 references.
3857 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
3859 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
3860 (floatuns<GPI:mode><GPF:mode>2): Remove.
3861 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
3862 and floatuns conversions.
3863 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
3864 and floatuns conversions.
3865 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
3866 (w1,w2): New mode attributes for inequal width conversions.
3868 2014-04-22 Renlin Li <Renlin.Li@arm.com>
3870 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
3871 the output asm format.
3873 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
3875 * config/aarch64/aarch64-simd.md
3876 (aarch64_cm<optab>di): Always split.
3877 (*aarch64_cm<optab>di): New.
3878 (aarch64_cmtstdi): Always split.
3879 (*aarch64_cmtstdi): New.
3881 2014-04-22 Jakub Jelinek <jakub@redhat.com>
3883 PR tree-optimization/60823
3884 * omp-low.c (ipa_simd_modify_function_body): Go through
3885 all SSA_NAMEs and for those refering to vector arguments
3886 which are going to be replaced adjust SSA_NAME_VAR and,
3887 if it is a default definition, change it into a non-default
3888 definition assigned at the beginning of function from new_decl.
3889 (ipa_simd_modify_stmt_ops): Rewritten.
3890 * tree-dfa.c (set_ssa_default_def): When removing default def,
3891 check for NULL loc instead of NULL *loc.
3893 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3895 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
3896 restrictions on core registers for DImode values in Thumb2.
3898 2014-04-22 Ian Bolton <ian.bolton@arm.com>
3900 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
3901 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
3903 2014-04-22 Ian Bolton <ian.bolton@arm.com>
3905 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
3906 (*iordi_notzesidi_di): Likewise.
3907 (*iordi_notsesidi_di): Likewise.
3909 2014-04-22 Ian Bolton <ian.bolton@arm.com>
3911 * config/arm/arm-protos.h (tune_params): New struct members.
3912 * config/arm/arm.c: Initialise tune_params per processor.
3913 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
3914 for speed, based on new tune_params.
3916 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
3918 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro added.
3919 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added macro.
3920 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment corrected.
3921 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
3922 * config/aarch64/arm_neon.h (vrnd_f64): Added.
3923 (vrnda_f64): Likewise.
3924 (vrndi_f64): Likewise.
3925 (vrndm_f64): Likewise.
3926 (vrndn_f64): Likewise.
3927 (vrndp_f64): Likewise.
3928 (vrndx_f64): Likewise.
3930 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3932 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
3933 GET_MODE_SIZE argument is enum machine_mode.
3935 2014-04-22 Jakub Jelinek <jakub@redhat.com>
3938 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
3939 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
3941 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
3944 * asan.c (asan_emit_stack_protection): Force the base to align to
3945 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
3946 appropriate bits if STRICT_ALIGNMENT.
3947 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
3949 (expand_used_vars): Leave a space in the stack frame for alignment
3950 if STRICT_ALIGNMENT.
3952 2014-04-21 David Malcolm <dmalcolm@redhat.com>
3954 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
3956 (gimple_store_p): Likewise.
3957 (gimple_assign_load_p): Likewise.
3958 (gimple_assign_cast_p): Likewise.
3959 (gimple_clobber_p): Likewise.
3961 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
3962 rather than a gimple.
3963 (gimple_assign_cast_p): Likewise.
3965 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
3968 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
3969 If mode is DDmode and TARGET_E500_DOUBLE allow move.
3971 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
3972 more debug information for E500 if -mdebug=reg.
3974 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
3977 * config/i386/i386.c (ix86_expand_builtin)
3978 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
3979 register for target RTX.
3980 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
3982 2014-04-18 Cong Hou <congh@google.com>
3984 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
3985 the widen-mult pattern by handling two operands with different sizes,
3986 and operands whose size is smaller than half of the result type.
3988 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
3990 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
3991 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
3992 (do_estimate_edge_time): Compute it.
3993 * ipa-inline.c (want_inline_small_function_p): Bypass
3994 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
3996 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
3998 * ipa-inline.c (spec_rem): New static variable.
3999 (dump_overall_stats): New function.
4000 (dump_inline_stats): New function.
4002 2014-04-18 Richard Henderson <rth@redhat.com>
4004 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
4005 to GET_MODE_SIZE, not a reg_class_t.
4007 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4009 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
4010 (vsx_xxmrglw_<mode>): Likewise.
4012 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
4015 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
4016 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
4017 (rs6000_init_hard_regno_mode_ok): Likewise.
4019 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
4021 * ipa-inline.c (inline_small_functions): Account only non-cold
4023 * doc/invoke.texi (inline-unit-growth): Update documentation.
4025 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
4027 * config/rs6000/rs6000.md (addti3, subti3): New.
4029 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
4032 * config/i386/i386.c (ix86_expand_clear): Remove outdated
4033 comment. Check optimize_insn_for_size_p instead of
4034 optimize_insn_for_speed_p.
4036 2014-04-17 Martin Jambor <mjambor@suse.cz>
4038 * gimple-iterator.c (gsi_start_edge): New function.
4039 * gimple-iterator.h (gsi_start_edge): Declare.
4040 * tree-sra.c (single_non_eh_succ): New function.
4041 (disqualify_ops_if_throwing_stmt): Renamed to
4042 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
4043 having one non-EH successor BB.
4044 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
4045 generate loads into replacements.
4046 (sra_modify_assign): Likewise and and also use the simple path for
4048 (sra_modify_function_body): Commit statements on edges.
4050 2014-04-17 Richard Biener <rguenther@suse.de>
4053 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
4054 comparison results and add clarifying comment.
4056 2014-04-17 Jakub Jelinek <jakub@redhat.com>
4058 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
4059 (blank_mode): Initialize it.
4060 (emit_mode_size_inline, emit_mode_nunits_inline,
4061 emit_mode_inner_inline): New functions.
4062 (emit_insn_modes_h): Call them and surround their output with
4063 #if GCC_VERSION >= 4001 ... #endif.
4064 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
4065 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
4066 mode_* arrays if the argument is __builtin_constant_p.
4067 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
4068 is enum machine_mode.
4070 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
4072 * passes.c (opt_pass::execute): Adjust.
4073 (pass_manager::execute_pass_mode_switching): Likewise.
4074 (early_local_passes::execute): Likewise.
4075 (execute_one_pass): Pass cfun to the pass's execute method.
4076 * tree-pass.h (opt_pass::execute): Add function * argument.
4077 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
4078 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
4079 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
4080 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
4081 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
4082 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
4083 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
4084 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
4085 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
4086 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
4087 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
4088 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
4089 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
4090 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
4091 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
4092 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
4093 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
4094 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
4095 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
4096 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
4097 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
4098 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
4099 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
4100 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
4101 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
4102 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
4103 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
4104 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
4105 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
4108 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
4110 * passes.c (opt_pass::gate): Take function * argument.
4111 (gate_all_early_local_passes): Merge into
4112 (early_local_passes::gate): this.
4113 (gate_all_early_optimizations): Merge into
4114 (all_early_optimizations::gate): this.
4115 (gate_all_optimizations): Mege into
4116 (all_optimizations::gate): this.
4117 (gate_all_optimizations_g): Merge into
4118 (all_optimizations_g::gate): this.
4119 (gate_rest_of_compilation): Mege into
4120 (rest_of_compilation::gate): this.
4121 (gate_postreload): Merge into
4122 (postreload::gate): this.
4123 (dump_one_pass): Pass cfun to the pass's gate method.
4124 (execute_ipa_summary_passes): Likewise.
4125 (execute_one_pass): Likewise.
4126 (ipa_write_summaries_2): Likewise.
4127 (ipa_write_optimization_summaries_1): Likewise.
4128 (ipa_read_summaries_1): Likewise.
4129 (ipa_read_optimization_summaries_1): Likewise.
4130 (execute_ipa_stmt_fixups): Likewise.
4131 * tree-pass.h (opt_pass::gate): Add function * argument.
4132 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
4133 combine-stack-adj.c, combine.c, compare-elim.c,
4134 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
4135 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
4136 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
4137 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
4138 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
4139 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
4140 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
4141 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
4142 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
4143 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
4144 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
4145 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
4146 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
4147 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
4148 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
4149 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
4150 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
4151 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
4152 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
4153 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
4154 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
4155 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
4156 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
4157 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
4158 var-tracking.c, vtable-verify.c, web.c: Adjust.
4160 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
4162 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
4163 * configure: Regenerate.
4165 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
4167 * passes.c (dump_one_pass): don't check pass->has_gate.
4168 (execute_ipa_summary_passes): Likewise.
4169 (execute_one_pass): Likewise.
4170 (ipa_write_summaries_2): Likewise.
4171 (ipa_write_optimization_summaries_1): Likewise.
4172 (ipa_read_optimization_summaries_1): Likewise.
4173 (execute_ipa_stmt_fixups): Likewise.
4174 * tree-pass.h (pass_data::has_gate): Remove.
4175 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
4176 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
4177 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
4178 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
4179 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
4180 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
4181 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
4182 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
4183 gimple-low.c, gimple-ssa-isolate-paths.c,
4184 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
4185 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
4186 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
4187 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
4188 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
4189 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
4190 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
4191 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
4192 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
4193 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
4194 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
4195 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
4196 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
4197 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
4198 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
4199 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
4200 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
4201 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
4202 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
4203 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
4204 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
4205 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
4208 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
4210 * pass_manager.h (pass_manager::register_dump_files_1): Remove
4212 * passes.c (pass_manager::register_dump_files_1): Merge into
4213 (pass_manager::register_dump_files): this, and remove its handling of
4214 properties since the pass always has the properties anyway.
4215 (pass_manager::pass_manager): Adjust.
4217 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
4219 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
4220 * passes.c (pass_manager::register_dump_files_1): Remove dead code
4221 dealing with properties.
4222 (pass_manager::register_dump_files): Adjust.
4224 2014-03-20 Mark Wielaard <mjw@redhat.com>
4226 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
4227 then represent the bound as normal constant value.
4229 2014-04-17 Jakub Jelinek <jakub@redhat.com>
4232 Forward port from 4.8 branch
4233 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
4235 * config/i386/bmiintrin.h (_blsi_u32): New.
4239 (_blsmsk_u32): Ditto.
4240 (_blsmsk_u64): Ditto.
4241 (_tzcnt_u32): Ditto.
4242 (_tzcnt_u64): Ditto.
4244 2014-04-17 Kito Cheng <kito@0xlab.org>
4246 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
4248 2014-04-17 Richard Biener <rguenther@suse.de>
4251 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
4252 boolean results for comparisons.
4254 2014-04-17 Richard Biener <rguenther@suse.de>
4256 PR tree-optimization/60836
4257 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
4258 initial PHI args to be gimple values.
4260 2014-04-17 Richard Biener <rguenther@suse.de>
4262 PR tree-optimization/60841
4263 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
4264 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
4265 of stmts to SLP build.
4266 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
4267 (vect_analyze_slp): Likewise.
4268 (vect_analyze_slp_instance): Likewise.
4269 (vect_build_slp_tree): Limit overall SLP tree growth.
4270 * tree-vectorizer.h (vect_analyze_data_refs,
4271 vect_analyze_slp): Adjust prototypes.
4273 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
4275 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
4278 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
4280 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
4281 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
4282 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
4283 for TARGET_SLOW_PSHUFB
4285 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
4287 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
4288 * config/i386/i386.c (intel_cost): Ditto.
4290 2014-04-17 Joey Ye <joey.ye@arm.com>
4292 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
4294 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
4296 * opts.c (common_handle_option): Disable -fipa-reference coorectly
4299 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
4301 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
4302 (type_all_derivations_known_p): New predicate.
4303 (type_all_ctors_visible_p): New predicate.
4304 (type_possibly_instantiated_p): New predicate.
4305 (get_odr_type): Compute all_derivations_known.
4306 (dump_odr_type): Dump the flag.
4307 (maybe_record_type): Cleanup.
4308 (record_target_from_binfo): Add bases_to_consider array;
4309 record bases for types w/o instances and skip CXX destructor.
4310 (possible_polymorphic_call_targets_1): Add bases_to_consider
4311 and consider_construction parameters; check if type may have instance.
4312 (get_polymorphic_call_info): Set maybe_in_construction to true
4313 when we know nothing.
4314 (record_targets_from_bases): Skip CXX destructors; they are
4315 never called for types in construction.
4316 (possible_polymorphic_call_targets): Do not record target when
4317 type may not have instance.
4319 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
4322 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
4323 external aliases alive, too.
4325 2014-04-16 Andrew Pinski <apinski@cavium.com>
4327 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
4330 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
4332 * final.c (compute_alignments): Do not apply loop alignment to a block
4333 falling through to the exit.
4335 2014-04-16 Catherine Moore <clm@codesourcery.com>
4337 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
4338 Adjust constraints for microMIPS store patterns.
4340 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
4342 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
4344 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
4346 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
4347 (append_use): Run at -O0.
4348 (append_vdef): Likewise.
4349 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
4350 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
4352 2014-04-16 Jakub Jelinek <jakub@redhat.com>
4354 PR tree-optimization/60844
4355 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
4356 (propagate_op_to_single_use, remove_visited_stmt_chain,
4357 linearize_expr, repropagate_negates, reassociate_bb): Use it
4358 instead of gsi_remove.
4360 2014-04-16 Martin Jambor <mjambor@suse.cz>
4362 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
4363 ipa_transforms_to_apply.
4364 (cgraph_function_versioning): Assert that old_node has empty
4365 ipa_transforms_to_apply.
4366 * trans-mem.c (ipa_tm_create_version): Likewise.
4367 * tree-inline.c (tree_function_versioning): Do not duplicate
4368 ipa_transforms_to_apply.
4370 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4373 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
4375 Pass necessary as flags on 64-bit Solaris/x86.
4376 Use lowercase relocs for x86_64-*-*.
4377 * configure: Regenerate.
4379 2014-04-15 Jan Hubicka <jh@suse.cz>
4381 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
4382 (maybe_record_node, likely_target_p): Use it.
4384 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4387 Revert following patch
4389 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
4392 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
4393 software floating point or no floating point registers, do not
4394 allow any type in the FPRs. Eliminate a test for SPE SIMD types
4395 in GPRs that occurs after we tested for GPRs that would never be
4398 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
4399 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
4400 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
4401 specifically allow DDmode, since that does not use the SPE SIMD
4404 2014-03-21 Mark Wielaard <mjw@redhat.com>
4406 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
4407 as unsigned or int depending on type and value used.
4409 2014-04-15 Richard Biener <rguenther@suse.de>
4411 PR rtl-optimization/56965
4412 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
4413 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
4415 * alias.c (true_dependence_1): Do not call
4416 nonoverlapping_component_refs_p.
4417 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
4418 nonoverlapping_component_refs_p.
4419 (indirect_refs_may_alias_p): Likewise.
4421 2014-04-15 Teresa Johnson <tejohnson@google.com>
4423 * cfg.c (dump_bb_info): Fix flags check.
4424 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
4426 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4428 PR rtl-optimization/60663
4429 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
4432 2014-04-15 Richard Biener <rguenther@suse.de>
4434 * lto-streamer.h (LTO_major_version): Bump to 4.
4436 2014-04-15 Richard Biener <rguenther@suse.de>
4438 * common.opt (lto_partition_model): New enum.
4439 (flto-partition=): Merge separate options with a single with argument,
4440 add -flto-partition=one support.
4441 * flag-types.h (enum lto_partition_model): Declare.
4442 * opts.c (finish_options): Remove duplicate -flto-partition=
4444 * lto-wrapper.c (run_gcc): Adjust.
4446 2014-04-15 Richard Biener <rguenther@suse.de>
4448 * alias.c (ncr_compar): New function.
4449 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
4451 2014-04-15 Richard Biener <rguenther@suse.de>
4453 * alias.c (record_component_aliases): Do not walk BINFOs.
4455 2014-04-15 Richard Biener <rguenther@suse.de>
4457 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
4458 Add struct function argument and adjust.
4459 (find_func_aliases_for_call): Likewise.
4460 (find_func_aliases): Likewise.
4461 (find_func_clobbers): Likewise.
4462 (intra_create_variable_infos): Likewise.
4463 (compute_points_to_sets): Likewise.
4464 (ipa_pta_execute): Adjust. Do not push/pop cfun.
4466 2014-04-15 Richard Biener <rguenther@suse.de>
4468 * tree.c (iterative_hash_expr): Use enum tree_code_class
4469 to store TREE_CODE_CLASS.
4470 (tree_block): Likewise.
4471 (tree_set_block): Likewise.
4472 * tree.h (fold_build_pointer_plus_loc): Use
4473 convert_to_ptrofftype_loc.
4475 2014-04-15 Jakub Jelinek <jakub@redhat.com>
4478 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
4481 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
4483 * cfgloop.h (struct loop): Move force_vectorize down.
4484 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
4485 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
4486 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
4487 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
4488 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
4489 * tree-core.h (enum annot_expr_kind): Add new kind values.
4490 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
4491 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
4493 * tree.def (ANNOTATE_EXPR): Tweak comment.
4495 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
4497 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
4500 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
4502 * tree.h (TYPE_IDENTIFIER): Declare.
4503 * tree.c (subrange_type_for_debug_p): Use it.
4504 * godump.c (go_format_type): Likewise.
4505 * dwarf2out.c (is_cxx_auto, modified_type_die,
4506 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
4507 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
4509 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
4512 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
4514 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
4516 * config/i386/i386.c (examine_argument): Return bool. Return true if
4517 parameter should be passed in memory.
4518 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
4519 (construct_container): Update calls to examine_argument.
4520 (function_arg_advance_64): Ditto.
4521 (return_in_memory_32): Merge with ix86_return_in_memory.
4522 (return_in_memory_64): Ditto.
4523 (return_in_memory_ms_64): Ditto.
4525 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
4527 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
4528 * coverage.c (coverage_compute_profile_id): Handle externally visible
4531 2014-04-14 Martin Jambor <mjambor@suse.cz>
4533 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
4534 DECL_DISREGARD_INLINE_LIMITS functions.
4536 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
4539 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
4541 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
4544 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
4545 optimize_insn_for_speed_p instead of
4546 optimize_function_for_speed_p.
4548 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
4550 * doc/invoke.texi (free): Document AArch64.
4552 2014-04-14 Richard Biener <rguenther@suse.de>
4554 PR tree-optimization/60042
4555 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
4556 (insert_into_preds_of_block): Do not prevent PHI insertion
4557 for REFERENCE exprs here ...
4558 (eliminate_dom_walker::before_dom_children): ... but prevent
4559 their use here under similar conditions when applied to the
4560 IL after PRE optimizations.
4562 2014-04-14 Richard Biener <rguenther@suse.de>
4564 * passes.def: Move early points-to after early SRA.
4566 2014-04-14 Richard Biener <rguenther@suse.de>
4568 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
4569 check for which sign-changes we allow when forwarding
4570 a converted value into a switch.
4572 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
4574 * stor-layout.c (place_field): Finalize non-constant offset for the
4577 2014-04-14 Richard Biener <rguenther@suse.de>
4579 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
4581 (expand_switch_using_bit_tests_p): Likewise.
4582 (process_switch): Compute and pass on speed_p based on the
4584 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
4585 optimize_bb_for_speed_p.
4587 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
4589 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
4590 * function.h (struct function): Rename has_force_vect_loops into
4591 has_force_vectorize_loops.
4592 * lto-streamer-in.c (input_cfg): Adjust for renaming.
4593 (input_struct_function_base): Likewise.
4594 * lto-streamer-out.c (output_cfg): Likewise.
4595 (output_struct_function_base): Likewise.
4596 * omp-low.c (expand_omp_simd): Likewise.
4597 * tree-cfg.c (move_sese_region_to_fn): Likewise.
4598 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
4599 (version_loop_for_if_conversion): Likewise.
4600 (tree_if_conversion): Likewise.
4601 (main_tree_if_conversion): Likewise.
4602 (gate_tree_if_conversion): Likewise.
4603 * tree-inline.c (copy_loops): Likewise.
4604 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
4605 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
4606 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
4607 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
4608 * tree-vectorizer.c (vectorize_loops): Likewise.
4609 * tree-vectorizer.h (unlimited_cost_model): Likewise.
4611 2014-04-14 Richard Biener <rguenther@suse.de>
4614 * lto-streamer-out.c (wrap_refs): New function.
4615 (lto_output): Wrap symbol references in global initializes in
4616 type-preserving MEM_REFs.
4618 2014-04-14 Christian Bruel <christian.bruel@st.com>
4620 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
4622 2014-04-14 Christian Bruel <christian.bruel@st.com>
4624 * config/sh/sh.md (setmemqi): New expand pattern.
4625 * config/sh/sh.h (CLEAR_RATIO): Define.
4626 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
4627 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
4629 2014-04-14 Richard Biener <rguenther@suse.de>
4632 * fold-const.c (negate_expr_p): Don't negate directional rounding
4634 (fold_negate_expr): Likewise.
4636 2014-04-14 Richard Biener <rguenther@suse.de>
4638 PR tree-optimization/59817
4639 PR tree-optimization/60453
4640 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
4641 recursion to catch all CHRECs in the scalar evolution and restrict
4642 the predicate for the remains appropriately.
4644 2014-04-12 Catherine Moore <clm@codesourcery.com>
4646 * config/mips/constraints.md: Add new register constraint "kb".
4647 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
4648 (*movhi_internal): Likewise.
4649 (*movqi_internal): Likewise.
4650 * config/mips/mips.h (M16_STORE_REGS): New register class.
4651 (REG_CLASS_NAMES): Add M16_STORE_REGS.
4652 (REG_CLASS_CONTENTS): Likewise.
4653 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
4655 2014-04-11 Tobias Burnus <burnus@net-b.de>
4658 * doc/invoke.texi (-Wformat-signedness): Document it.
4659 (Wformat=2): Mention that this enables -Wformat-signedness.
4661 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
4663 * common/config/epiphany/epiphany-common.c
4664 (epiphany_option_optimization_table): Enable section anchors by
4665 default at -O1 or higher.
4666 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
4667 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
4668 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
4669 carries no extra cost.
4670 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
4671 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
4672 * config/epiphany/predicates.md (memclob_operand): New predicate.
4673 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
4674 Use memclob_operand predicate and X constraint for operand 3.
4676 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
4678 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
4679 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
4682 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
4684 PR rtl-optimization/60651
4685 * mode-switching.c (optimize_mode_switching): Make sure to emit
4686 sets of a lower numbered entity before sets of a higher numbered
4687 entity to a mode of the same or lower priority.
4688 When creating a seginfo for a basic block that starts with a code
4689 label, move the insertion point past the code label.
4690 (new_seginfo): Document and enforce requirement that
4691 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
4692 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
4693 * doc/tm.texi: Regenerate.
4695 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
4698 * config/arc/arc.c (arc_save_restore): Fix assert typo.
4700 2013-04-11 Jakub Jelinek <jakub@redhat.com>
4702 * BASE-VER: Set to 4.10.0.
4704 2014-04-11 Tobias Burnus <burnus@net-b.de>
4707 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
4708 * doc/gcc.texi (Service): Update description in the @menu
4709 * doc/invoke.texi (Option Summary): Remove misplaced and
4712 2014-04-11 Steve Ellcey <sellcey@mips.com>
4713 Jakub Jelinek <jakub@redhat.com>
4716 * expr.c (convert_move): Use emit_store_flag_force instead of
4717 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
4720 2014-04-11 Richard Biener <rguenther@suse.de>
4723 * varasm.c (assemble_alias): Avoid endless error reporting
4724 recursion by setting TREE_ASM_WRITTEN.
4726 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4728 * config/s390/s390.md: Add a splitter for NOT rtx.
4730 2014-04-11 Jakub Jelinek <jakub@redhat.com>
4732 PR rtl-optimization/60663
4733 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
4735 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
4736 Jakub Jelinek <jakub@redhat.com>
4739 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
4740 flag from decl_node to node.
4742 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4745 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
4746 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
4747 ameliorating the cases where it can be.
4749 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
4752 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
4754 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
4755 (loadsync_<mode>): Change mode.
4756 (load_quadpti, store_quadpti): New.
4757 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
4758 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
4759 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
4761 2014-04-09 Cong Hou <congh@google.com>
4764 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
4767 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4769 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
4770 instead of vnor to exploit possible fusion opportunity in the
4772 (altivec_expand_vec_perm_const_le): Likewise.
4774 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
4776 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
4777 (loadsync_<mode>): Change mode.
4778 (load_quadpti, store_quadpti): New.
4779 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
4780 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
4782 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
4785 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
4786 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
4787 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
4789 2014-04-08 Richard Biener <rguenther@suse.de>
4792 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
4793 a 64bit widest int print double-int similar to on HWI64 hosts.
4795 2014-04-08 Richard Biener <rguenther@suse.de>
4797 PR tree-optimization/60785
4798 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
4799 default defs properly.
4801 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
4803 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
4804 (Weffc++): Likewise.
4806 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
4808 * ipa-devirt.c (maybe_record_node): When node is not recorded,
4809 set completep to false rather than true.
4811 2014-04-07 Douglas B Rupp <rupp@adacore.com>
4814 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
4815 ARM_TARGET2_DWARF_FORMAT.
4817 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
4820 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
4821 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
4824 2014-04-07 Richard Biener <rguenther@suse.de>
4826 PR tree-optimization/60766
4827 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
4828 (may_eliminate_iv): Convert cand_value_at result to desired type.
4830 2014-04-07 Jason Merrill <jason@redhat.com>
4833 * common.opt (-fno-gnu-unique): Add.
4834 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
4836 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4838 * haifa-sched.c: Fix outdated function reference and minor
4839 grammar errors in introductory comment.
4841 2014-04-07 Richard Biener <rguenther@suse.de>
4844 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
4846 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
4848 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
4851 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
4852 size accounting for thunks.
4853 (pa_asm_output_mi_thunk): Use final_start_function() and
4854 final_end_function() to output function start and end directives.
4856 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
4858 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
4859 device specific ISA/ feature information. Remove short_sp and
4860 errata_skip ds. Add avr_device_specific_features enum to have device
4862 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
4863 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
4864 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
4865 updated device specific info.
4866 * config/avr/avr-mcus.def: Merge device specific details to
4867 dev_attribute field.
4868 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
4870 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
4871 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
4872 assembler if RMW isa supported by current device.
4873 * config/avr/genmultilib.awk: Update as device info structure changed.
4874 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
4876 2014-04-04 Cong Hou <congh@google.com>
4878 PR tree-optimization/60656
4879 * tree-vect-stmts.c (supportable_widening_operation):
4880 Fix a bug that elements in a vector with vect_used_by_reduction
4881 property are incorrectly reordered when the operation on it is not
4882 consistant with the one in reduction operation.
4884 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
4886 PR rtl-optimization/60155
4887 * gcse.c (record_set_data): New function.
4888 (single_set_gcse): New function.
4889 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
4890 (hoist_code): Likewise.
4891 (get_pressure_class_and_nregs): Likewise.
4893 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
4895 * explow.c (probe_stack_range): Emit a final optimization blockage.
4897 2014-04-04 Anthony Green <green@moxielogic.com>
4899 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
4902 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
4905 * lto-cgraph.c (input_overwrite_node): Check that partitioning
4906 flags are set only during streaming.
4907 * ipa.c (process_references, walk_polymorphic_call_targets,
4908 symtab_remove_unreachable_nodes): Drop bodies of always inline
4909 after early inlining.
4910 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
4912 2014-04-04 Jakub Jelinek <jakub@redhat.com>
4913 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4916 * dwarf2out.c (const_ok_for_output_1): Reject expressions
4919 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4922 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
4924 (cortex_a53_fdivd): Likewise.
4926 2014-04-04 Martin Jambor <mjambor@suse.cz>
4929 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
4931 * cgraph.c (clone_of_p): Also return true if thunks match.
4932 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
4933 cgraph_function_or_thunk_node and an obsolete comment.
4934 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
4936 (build_function_decl_skip_args): Likewise.
4937 (set_new_clone_decl_and_node_flags): New function.
4938 (duplicate_thunk_for_node): Likewise.
4939 (redirect_edge_duplicating_thunks): Likewise.
4940 (cgraph_clone_node): New parameter args_to_skip, pass it to
4941 redirect_edge_duplicating_thunks which is called instead of
4942 cgraph_redirect_edge_callee.
4943 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
4944 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
4946 2014-04-04 Jeff Law <law@redhat.com>
4949 * config/arm/predicates.md (const_int_I_operand): New predicate.
4950 (const_int_M_operand): Similarly.
4951 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
4953 (insv_t2, extv_reg, extzv_t2): Likewise.
4954 (load_multiple_with_writeback): Similarly for const_int_I_operand.
4955 (pop_multiple_with_writeback_and_return): Likewise.
4956 (vfp_pop_multiple_with_writeback): Likewise
4958 2014-04-04 Richard Biener <rguenther@suse.de>
4961 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
4962 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
4964 * gimplify.h (gimple_add_tmp_var_fn): Declare.
4965 * gimplify.c (gimple_add_tmp_var_fn): New function.
4966 * gimple-expr.h (create_tmp_reg_fn): Declare.
4967 * gimple-expr.c (create_tmp_reg_fn): New function.
4968 * gimple-low.c (record_vars_into): Don't change cfun.
4969 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
4970 code generation without cfun.
4972 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
4975 * Makefile.in (install-driver): Fix shell scripting.
4977 2014-04-03 Cong Hou <congh@google.com>
4979 PR tree-optimization/60505
4980 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
4981 threshold of number of iterations below which no vectorization
4983 * tree-vect-loop.c (new_loop_vec_info):
4984 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
4985 * tree-vect-loop.c (vect_analyze_loop_operations):
4986 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
4987 * tree-vect-loop.c (vect_transform_loop):
4988 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
4989 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
4990 of iterations of the loop and see if we should build the epilogue.
4992 2014-04-03 Richard Biener <rguenther@suse.de>
4994 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
4995 (streamer_tree_cache_create): Adjust.
4996 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
4997 to allow optional nodes array.
4998 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
4999 (streamer_tree_cache_append): Likewise.
5000 (streamer_tree_cache_create): Create nodes array optionally
5001 as specified by parameter.
5002 * lto-streamer-out.c (create_output_block): Avoid maintaining
5003 the node array in the writer cache.
5004 (DFS_write_tree): Remove assertion.
5005 (produce_asm_for_decls): Free the out decl state hash table early.
5006 * lto-streamer-in.c (lto_data_in_create): Adjust for
5007 streamer_tree_cache_create prototype change.
5009 2014-04-03 Richard Biener <rguenther@suse.de>
5011 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
5012 set TREE_CHAIN to NULL_TREE.
5014 2014-04-03 Richard Biener <rguenther@suse.de>
5016 PR tree-optimization/60740
5017 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
5018 over all GIMPLE_COND operands.
5020 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
5022 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
5023 (Weffc++): Remove Scott's numbering, merge lists and reference
5026 2014-04-03 Nick Clifton <nickc@redhat.com>
5028 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
5031 2014-04-03 Martin Jambor <mjambor@suse.cz>
5033 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
5034 mention gcc_unreachable before failing.
5035 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
5038 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
5041 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
5042 inconsistent code and instead mark the context inconsistent.
5043 (possible_polymorphic_call_targets): For inconsistent contexts
5044 return empty complete list.
5046 2014-04-02 Anthony Green <green@moxielogic.com>
5048 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
5049 (extendqisi2, extendhisi2): Define.
5050 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
5051 (WCHAR_TYPE): Change to unsigned int.
5053 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5055 PR tree-optimization/60733
5056 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
5057 insertion point for PHI candidates to be the end of the feeding
5058 block for the PHI argument.
5060 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
5062 PR rtl-optimization/60650
5063 * lra-constraints.c (process_alt_operands): Decrease reject for
5064 earlyclobber matching.
5066 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5068 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
5070 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
5072 * config/spu/spu.c (pad_bb): Do not crash when the last
5073 insn is CODE_FOR_blockage.
5075 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
5077 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
5078 lies outside the target mode.
5080 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
5083 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
5084 software floating point or no floating point registers, do not
5085 allow any type in the FPRs. Eliminate a test for SPE SIMD types
5086 in GPRs that occurs after we tested for GPRs that would never be
5089 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
5090 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
5091 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
5092 specifically allow DDmode, since that does not use the SPE SIMD
5095 2014-04-02 Richard Biener <rguenther@suse.de>
5098 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
5099 MODE_INTs. Properly use negv_optab.
5100 (expand_abs): Likewise.
5102 2014-04-02 Richard Biener <rguenther@suse.de>
5105 * Makefile.in (install-driver): Guard extra installs with special
5108 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
5110 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
5113 2014-04-01 Richard Henderson <rth@redhat.com>
5116 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
5117 alternative enabled before register allocation.
5119 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
5121 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
5122 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
5124 (nios2_large_got_address): Remove unneeded 'sym' parameter.
5125 (nios2_got_address): Update nios2_large_got_address call site.
5126 (nios2_delegitimize_address): New function.
5127 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
5128 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
5129 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
5131 2014-04-01 Martin Husemann <martin@duskware.de>
5133 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
5136 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
5138 PR rtl-optimization/60604
5139 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
5140 check from register_operand.
5141 (register_operand): Redefine in terms of general_operand.
5142 (nonmemory_operand): Use register_operand for the non-constant cases.
5144 2014-04-01 Richard Biener <rguenther@suse.de>
5146 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
5148 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
5150 * doc/invoke.texi (mapp-regs): Clarify.
5152 2014-03-31 Ulrich Drepper <drepper@gmail.com>
5154 * config/i386/avx512fintrin.h (__v32hi): Define type.
5155 (__v64qi): Likewise.
5156 (_mm512_set1_epi8): Define.
5157 (_mm512_set1_epi16): Define.
5158 (_mm512_set4_epi32): Define.
5159 (_mm512_set4_epi64): Define.
5160 (_mm512_set4_pd): Define.
5161 (_mm512_set4_ps): Define.
5162 (_mm512_setr4_epi64): Define.
5163 (_mm512_setr4_epi32): Define.
5164 (_mm512_setr4_pd): Define.
5165 (_mm512_setr4_ps): Define.
5166 (_mm512_setzero_epi32): Define.
5168 2014-03-31 Martin Jambor <mjambor@suse.cz>
5171 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
5172 callsite_arguments_match_p. Updated all callers. Also check types of
5173 corresponding formal parameters and actual arguments.
5174 (not_all_callers_have_enough_arguments_p) Renamed to
5175 some_callers_have_mismatched_arguments_p.
5177 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
5179 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
5181 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
5184 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
5187 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
5189 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
5190 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
5192 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
5193 Use FMAMODE_NOVF512 mode iterator.
5194 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
5195 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
5196 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
5197 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
5199 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
5200 Use VF_128_256 mode iterator.
5201 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
5204 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
5206 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
5207 static chain if needed.
5209 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
5212 * lra-constraints.c (index_part_to_reg): New.
5213 (process_address): Use it.
5215 2014-03-27 Jeff Law <law@redhat.com>
5216 Jakub Jelinek <jakub@redhat.com>
5219 * expr.c (do_tablejump): Use simplify_gen_binary rather than
5220 gen_rtx_{PLUS,MULT} to build up the address expression.
5222 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
5223 creating non-canonical RTL.
5225 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
5228 * ipa-inline.c (want_inline_small_function_p): Short circuit large
5229 functions; reorganize to make cheap checks first.
5230 (inline_small_functions): Do not estimate growth when dumping;
5232 * ipa-inline.h (inline_summary): Add min_size.
5233 (growth_likely_positive): New function.
5234 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
5235 (set_cond_stmt_execution_predicate): Cleanup.
5236 (estimate_edge_size_and_time): Compute min_size.
5237 (estimate_calls_size_and_time): Likewise.
5238 (estimate_node_size_and_time): Likewise.
5239 (inline_update_overall_summary): Update min_size.
5240 (do_estimate_edge_time): Likewise.
5241 (do_estimate_edge_size): Update.
5242 (do_estimate_edge_hints): Update.
5243 (growth_likely_positive): New function.
5245 2014-03-28 Jakub Jelinek <jakub@redhat.com>
5248 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
5249 also if addr has VOIDmode.
5251 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5253 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
5254 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
5256 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
5257 instructions as well as AdvancedSIMD loads.
5259 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5261 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
5262 Use crypto_aese type.
5263 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
5264 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
5265 crypto_aese, crypto_aesmc. Move to types.md.
5266 * config/arm/types.md (crypto_aes): Split into crypto_aese,
5268 * config/arm/iterators.md (crypto_type): Likewise.
5270 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
5272 * cgraph.c: Include expr.h and tree-dfa.h.
5273 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
5276 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
5279 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
5280 regs from checking multi-reg pseudos.
5282 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5284 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
5286 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
5288 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
5289 if it would clobber the stack pointer, even temporarily.
5291 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
5293 * mode-switching.c: Make small adjustments to the top comment.
5295 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
5297 * config/rs6000/constraints.md (wD constraint): New constraint to
5298 match the constant integer to get the top DImode/DFmode out of a
5299 vector in a VSX register.
5301 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
5302 match the constant integer to get the top DImode/DFmode out of a
5303 vector in a VSX register.
5305 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
5308 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
5311 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
5312 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
5314 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
5315 Optimize vec_extract of 64-bit values, where the value being
5316 extracted is in the top word, where we can use scalar
5317 instructions. Add direct move and store support. Combine the big
5318 endian/little endian vector select load support into a single insn.
5319 (vsx_extract_<mode>_internal1): Likewise.
5320 (vsx_extract_<mode>_internal2): Likewise.
5321 (vsx_extract_<mode>_load): Likewise.
5322 (vsx_extract_<mode>_store): Likewise.
5323 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
5324 combined into vsx_extract_<mode>_load.
5325 (vsx_extract_<mode>_one_le): Likewise.
5327 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
5328 define the top 64-bit vector element.
5330 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
5333 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
5334 Document vec_vbpermq builtin.
5337 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
5338 enable use of xxsldwi and xxpermdi builtin functions.
5339 (vec_xxpermdi): Likewise.
5341 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
5342 Document use of vec_xxsldwi and vec_xxpermdi builtins.
5344 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
5346 PR rtl-optimization/60650
5347 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
5349 (find_spills_for): New.
5350 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
5351 Spill all pseudos on the second iteration.
5353 2014-03-27 Marek Polacek <polacek@redhat.com>
5356 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
5359 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5361 * config/s390/s390.c (s390_can_use_return_insn): Check for
5362 call-saved FPRs on 31 bit.
5364 2014-03-27 Jakub Jelinek <jakub@redhat.com>
5367 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
5368 if they need regimplification, just drop them instead of
5369 calling gimple_regimplify_operands on them.
5371 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
5374 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
5375 (aarch64_frame_pointer_required): Adjust logic.
5376 (aarch64_can_eliminate): Adjust logic.
5377 (aarch64_override_options_after_change): Adjust logic.
5379 2014-03-27 Dehao Chen <dehao@google.com>
5381 * ipa-inline.c (early_inliner): Update node's inline info.
5383 2014-03-26 Dehao Chen <dehao@google.com>
5385 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
5386 compiler inserted conditional jumps for NAN float check.
5388 2014-03-26 Jakub Jelinek <jakub@redhat.com>
5390 * ubsan.h (ubsan_create_data): Change second argument's type
5391 to const location_t *.
5392 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
5394 (ubsan_create_data): Change second argument to const location_t *PLOC.
5395 Create Loc field whenever PLOC is non-NULL.
5396 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
5397 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
5401 * real.c (real_to_integer2): Change type of low to UHWI.
5403 2014-03-26 Tobias Burnus <burnus@net-b.de>
5405 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
5406 (CILK_SELF_SPECS): New define.
5407 (driver_self_specs): Use it.
5409 2014-03-26 Richard Biener <rguenther@suse.de>
5411 * tree-pretty-print.c (percent_K_format): Implement special
5412 case for LTO and its stripped down BLOCK tree.
5414 2014-03-26 Jakub Jelinek <jakub@redhat.com>
5417 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
5419 * tree-vrp.c (simplify_internal_call_using_ranges): If only
5420 one range is range_int_cst_p, but not both, at least optimize
5421 addition/subtraction of 0 and multiplication by 0 or 1.
5422 * gimple-fold.c (gimple_fold_call): Fold
5423 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
5424 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
5425 INTEGER_CSTs, try to fold at least x * 0 and y - y.
5427 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
5429 PR rtl-optimization/60452
5430 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
5431 <case REG>: Return 1 for invalid offsets from the frame pointer.
5433 2014-03-26 Marek Polacek <polacek@redhat.com>
5436 * doc/extend.texi (C Extensions): Mention variable-length arrays in
5439 2014-03-26 Marek Polacek <polacek@redhat.com>
5442 * doc/extend.texi (Designated Inits): Describe what happens to omitted
5445 2014-03-26 Marek Polacek <polacek@redhat.com>
5448 * ira-color.c (update_conflict_hard_regno_costs): Perform the
5449 multiplication in unsigned type.
5451 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
5453 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
5455 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
5457 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
5459 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
5462 * cif-code.def (UNREACHABLE) New code.
5463 * ipa-inline.c (inline_small_functions): Skip edges to
5464 __builtlin_unreachable.
5465 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
5466 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
5467 predicate to __bulitin_unreachable.
5468 (set_cond_stmt_execution_predicate): Fix issue when
5469 invert_tree_comparison returns ERROR_MARK.
5470 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
5471 propagate to inline clones.
5472 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
5474 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
5475 * cgraphclones.c (cgraph_clone_node): If call destination is already
5476 ureachable, do not redirect it back.
5477 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
5480 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
5482 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
5483 Do not modify inline clones.
5485 2014-03-25 Jakub Jelinek <jakub@redhat.com>
5487 * config/i386/i386.md (general_sext_operand): New mode attr.
5488 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
5489 don't generate (sign_extend (const_int)).
5490 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
5491 operands[2]. Use We constraint instead of <i> and
5492 <general_sext_operand> predicate instead of <general_operand>.
5493 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
5494 * config/i386/constraints.md (We): New constraint.
5495 * config/i386/predicates.md (x86_64_sext_operand,
5496 sext_operand): New predicates.
5498 2014-03-25 Martin Jambor <mjambor@suse.cz>
5501 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
5502 inconsistent devirtualizations to __builtin_unreachable.
5504 2014-03-25 Marek Polacek <polacek@redhat.com>
5507 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
5509 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
5511 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
5512 order of elements for big-endian.
5514 2014-03-25 Richard Biener <rguenther@suse.de>
5517 * gimplify-me.c (gimple_regimplify_operands): Update the
5520 2014-03-25 Martin Jambor <mjambor@suse.cz>
5523 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
5524 (lto_output_varpool_node): Likewise.
5525 (input_overwrite_node): Likewise.
5526 (input_varpool_node): Likewise.
5528 2014-03-25 Richard Biener <rguenther@suse.de>
5530 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
5531 (run_gcc): Likewise.
5533 2014-03-25 Jakub Jelinek <jakub@redhat.com>
5535 * combine.c (simplify_compare_const): Add MODE argument.
5536 Handle mode_width 0 as very large mode_width.
5537 (try_combine, simplify_comparison): Adjust callers.
5539 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
5540 type to avoid signed integer overflow.
5541 * explow.c (plus_constant): Likewise.
5543 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
5545 * doc/generic.texi: Correct typos.
5547 2014-03-24 Tobias Burnus <burnus@net-b.de>
5549 * doc/invoke.texi (-flto): Expand section about
5550 using static libraries with LTO.
5552 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5554 PR rtl-optimization/60501
5555 * optabs.def (addptr3_optab): New optab.
5556 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
5557 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
5558 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
5560 * lra.c (emit_add3_insn): Use the addptr pattern if available.
5562 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
5564 2014-03-24 Ulrich Drepper <drepper@gmail.com>
5566 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
5569 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
5570 (_mm256_undefined_ps): Define.
5571 (_mm256_undefined_pd): Define.
5572 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
5573 (_mm_undefined_pd): Define.
5574 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
5575 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
5576 (_mm512_undefined_ps): Define.
5577 (_mm512_undefined_pd): Define.
5578 Use _mm*_undefined_*.
5579 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
5581 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
5583 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
5584 (lshr_simd): DI mode added.
5585 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
5586 (aarch64_ushr_simddi): Likewise.
5587 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
5588 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
5589 (vshrd_n_u64): Likewise.
5591 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5593 * Makefile.in (s-macro_list): Depend on cc1.
5595 2014-03-23 Teresa Johnson <tejohnson@google.com>
5597 * ipa-utils.c (ipa_print_order): Use specified dump file.
5599 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
5601 PR rtl-optimization/60601
5602 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
5604 * gcc.c (eval_spec_function): Initialize save_growing_value.
5606 2014-03-22 Jakub Jelinek <jakub@redhat.com>
5609 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
5610 code == MINUS_EXPR, never swap op0 with op1.
5612 * toplev.c (init_local_tick): Avoid signed integer multiplication
5614 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
5615 shift by first operand's bitsize.
5617 2014-03-21 Jakub Jelinek <jakub@redhat.com>
5620 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
5622 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
5623 TARGET_ISA_64BIT_P(x).
5625 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5627 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
5628 pattern for vector nor instead of subtract from splat(-1).
5629 (altivec_expand_vec_perm_const_le): Likewise.
5631 2014-03-21 Richard Henderson <rth@twiddle.net>
5634 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
5635 related insns after epilogue_completed.
5637 2014-03-21 Martin Jambor <mjambor@suse.cz>
5640 * cgraph.h (symtab_node): New flag body_removed.
5641 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
5642 when removing bodies.
5643 * symtab.c (dump_symtab_base): Dump body_removed flag.
5644 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
5645 had their bodies removed.
5647 2014-03-21 Martin Jambor <mjambor@suse.cz>
5650 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
5653 2014-03-21 Richard Biener <rguenther@suse.de>
5655 PR tree-optimization/60577
5656 * tree-core.h (struct tree_base): Document nothrow_flag use
5657 in VAR_DECL_NONALIASED.
5658 * tree.h (VAR_DECL_NONALIASED): New.
5659 (may_be_aliased): Adjust.
5660 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
5662 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
5664 * expr.c (expand_expr_real_1): Remove outdated comment.
5666 2014-03-20 Jakub Jelinek <jakub@redhat.com>
5669 * ira.c (adjust_cleared_regs): Call copy_rtx on
5670 *reg_equiv[REGNO (loc)].src_p before passing it to
5671 simplify_replace_fn_rtx.
5674 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
5675 into CONST, put pic register as first operand of PLUS. Use
5676 gen_const_mem for both 32-bit and 64-bit PIC got loads.
5678 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5680 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
5682 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
5684 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
5685 around for store forwarding issue in the FPU on the UT699.
5686 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
5687 loads and operations if -mfix-ut699 is specified.
5688 (divtf3_hq): Tweak attribute.
5689 (sqrttf2_hq): Likewise.
5691 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
5693 * calls.c (store_one_arg): Remove incorrect const qualification on the
5694 type of the temporary.
5695 * cfgexpand.c (expand_return): Likewise.
5696 * expr.c (expand_constructor): Likewise.
5697 (expand_expr_real_1): Likewise.
5699 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
5701 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
5704 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
5707 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
5709 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
5711 * config/arm/aarch-common-protos.h
5712 (alu_cost_table): Fix spelling of "extend".
5713 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
5715 2014-03-19 Richard Biener <rguenther@suse.de>
5718 * tree-core.h (tree_type_common): Re-order pointer members
5719 to reduce recursion depth during GC walks.
5721 2014-03-19 Marek Polacek <polacek@redhat.com>
5724 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
5725 before accessing it.
5727 2014-03-19 Richard Biener <rguenther@suse.de>
5730 * lto-streamer-in.c (input_function): In WPA stage do not drop
5733 2014-03-19 Jakub Jelinek <jakub@redhat.com>
5735 PR tree-optimization/60559
5736 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
5737 with build_zero_cst assignment.
5739 2014-03-18 Kai Tietz <ktietz@redhat.com>
5741 PR rtl-optimization/56356
5742 * sdbout.c (sdbout_parms): Verify that parms'
5743 incoming argument is valid.
5744 (sdbout_reg_parms): Likewise.
5746 2014-03-18 Richard Henderson <rth@redhat.com>
5749 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
5750 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
5751 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
5753 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
5755 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
5756 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
5757 Italicize plugin event names in description. Explain that
5758 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
5759 Remind that no GCC functions should be called after PLUGIN_FINISH.
5760 Explain what pragmas with expansion are.
5762 2014-03-18 Martin Liska <mliska@suse.cz>
5764 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
5765 gimple call statement is update.
5766 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
5767 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
5769 2014-03-18 Jakub Jelinek <jakub@redhat.com>
5772 * ubsan.c (ubsan_instrument_unreachable): Call
5773 initialize_sanitizer_builtins.
5774 (ubsan_pass): Likewise.
5777 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
5778 varpool_finalize_decl instead of rest_of_decl_compilation.
5780 2014-03-18 Richard Biener <rguenther@suse.de>
5782 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
5783 by using bitmap_and_compl instead of bitmap_and_compl_into.
5784 (df_rd_transfer_function): Likewise.
5786 2014-03-18 Richard Biener <rguenther@suse.de>
5788 * doc/lto.texi (fresolution): Fix typo.
5790 2014-03-18 Richard Biener <rguenther@suse.de>
5792 * doc/invoke.texi (flto): Update for changes in 4.9.
5794 2014-03-18 Richard Biener <rguenther@suse.de>
5796 * doc/loop.texi: Remove section on the removed lambda framework.
5797 Update loop docs with recent changes in preserving loop structure.
5799 2014-03-18 Richard Biener <rguenther@suse.de>
5801 * doc/lto.texi (-fresolution): Document.
5803 2014-03-18 Richard Biener <rguenther@suse.de>
5805 * doc/contrib.texi: Adjust my name.
5807 2014-03-18 Jakub Jelinek <jakub@redhat.com>
5810 * internal-fn.c: Include diagnostic-core.h.
5811 (expand_BUILTIN_EXPECT): New function.
5812 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
5813 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
5814 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
5815 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
5817 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
5818 Revert 3 argument __builtin_expect code.
5819 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
5820 * gimple-fold.c (gimple_fold_call): Likewise.
5821 * tree.h (fold_builtin_expect): New prototype.
5822 * builtins.c (build_builtin_expect_predicate): Add predictor
5823 argument, if non-NULL, create 3 argument __builtin_expect.
5824 (fold_builtin_expect): No longer static. Add ARG2 argument,
5825 pass it through to build_builtin_expect_predicate.
5826 (fold_builtin_2): Adjust caller.
5827 (fold_builtin_3): Handle BUILT_IN_EXPECT.
5828 * internal-fn.def (BUILTIN_EXPECT): New.
5830 2014-03-18 Tobias Burnus <burnus@net-b.de>
5833 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
5834 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
5835 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
5837 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
5840 * predict.c (combine_predictions_for_bb): Fix up formatting.
5841 (expr_expected_value_1, expr_expected_value): Add predictor argument,
5842 fill what it points to if non-NULL.
5843 (tree_predict_by_opcode): Adjust caller, use the predictor.
5844 * predict.def (PRED_COMPARE_AND_SWAP): Add.
5846 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
5848 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
5849 proper constant for the store mode.
5851 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
5853 * symtab.c (change_decl_assembler_name): Fix transparent alias
5856 2014-03-16 Renlin Li <Renlin.Li@arm.com>
5858 * config/aarch64/aarch64.c: Correct the comments about the
5859 aarch64 stack layout.
5861 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
5863 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
5864 check for GF_OMP_FOR_KIND_FOR.
5866 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
5868 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
5869 ymm and zmm register names.
5871 2014-03-17 Jakub Jelinek <jakub@redhat.com>
5874 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
5875 note creation for the 2010-08-31 changes.
5877 2014-03-17 Marek Polacek <polacek@redhat.com>
5880 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
5881 as -fno-tree-loop-vectorize.
5882 (expand_omp_simd): Likewise.
5884 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
5886 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
5887 (eligible_for_call_delay): New prototype.
5888 * config/sparc/sparc.c (tls_call_delay): Rename into...
5889 (eligible_for_call_delay): ...this. Return false if the instruction
5890 cannot be put in the delay slot of a branch.
5891 (eligible_for_restore_insn): Simplify.
5892 (eligible_for_return_delay): Return false if the instruction cannot be
5893 put in the delay slot of a branch and simplify.
5894 (eligible_for_sibcall_delay): Return false if the instruction cannot be
5895 put in the delay slot of a branch.
5896 * config/sparc/sparc.md (fix_ut699): New attribute.
5897 (tls_call_delay): Delete.
5898 (in_call_delay): Reimplement.
5899 (eligible_for_sibcall_delay): Rename into...
5900 (in_sibcall_delay): ...this.
5901 (eligible_for_return_delay): Rename into...
5902 (in_return_delay): ...this.
5903 (in_branch_delay): Reimplement.
5904 (in_uncond_branch_delay): Delete.
5905 (in_annul_branch_delay): Delete.
5907 2014-03-14 Richard Henderson <rth@redhat.com>
5910 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
5911 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
5912 (*floathi<X87MODEF>2_i387_with_temp): Remove.
5913 (floathi splitters): Remove.
5914 (float<SWI48x>xf2): New pattern.
5915 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
5916 code that tried to handle DImode for 32-bit, but which was excluded
5917 by the pattern's condition. Drop allocation of stack temporary.
5918 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
5919 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
5920 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
5921 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
5922 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
5923 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
5924 (*float<SWI48><MODEF>2_sse_interunit): Remove.
5925 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
5926 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
5927 (*float<SWI48x><X87MODEF>2_i387): Remove.
5928 (all float _with_temp splitters): Remove.
5929 (*float<SWI48x><MODEF>2_i387): New pattern.
5930 (*float<SWI48><MODEF>2_sse): New pattern.
5931 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
5932 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
5934 2014-03-14 Jakub Jelinek <jakub@redhat.com>
5935 Marek Polacek <polacek@redhat.com>
5938 * common.opt (dump_base_name_prefixed): New Variable.
5939 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
5940 if x_dump_base_name_prefixed is already set, set it at the end.
5942 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
5944 PR rtl-optimization/60508
5945 * lra-constraints.c (get_reload_reg): Add new parameter
5947 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
5948 Pass the new parameter values.
5950 2014-03-14 Richard Biener <rguenther@suse.de>
5952 * common.opt: Revert unintented changes from r205065.
5955 2014-03-14 Richard Biener <rguenther@suse.de>
5958 * cfghooks.c (split_block): Properly adjust all loops the
5959 block was a latch of.
5961 2014-03-14 Martin Jambor <mjambor@suse.cz>
5964 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
5967 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
5970 * config/avr/avr.c (avr_set_current_function): Pass function name
5971 through default_strip_name_encoding before sanity checking instead
5972 of skipping the first char of the assembler name.
5974 2014-03-13 Richard Henderson <rth@redhat.com>
5977 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
5978 (ix86_force_to_memory, ix86_free_from_memory): Remove.
5979 * config/i386/i386-protos.h: Likewise.
5980 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
5981 in the expander instead of a splitter.
5982 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
5983 any possibility of requiring a memory.
5984 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
5985 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
5986 (fp branch splitters): Update for ix86_split_fp_branch.
5987 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
5988 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
5989 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
5990 (*fop_<MODEF>_2_i387): Remove f/r alternative.
5991 (*fop_<MODEF>_3_i387): Likewise.
5992 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
5993 (splitters for the fop_* register patterns): Remove.
5994 (fscalexf4_i387): Rename from *fscalexf4_i387.
5995 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
5997 2014-03-13 Jakub Jelinek <jakub@redhat.com>
5999 PR tree-optimization/59779
6000 * tree-dfa.c (get_ref_base_and_extent): Use double_int
6001 type for bitsize and maxsize instead of HOST_WIDE_INT.
6003 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
6005 PR rtl-optimization/57320
6006 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
6007 the CFG after thread_prologue_and_epilogue_insns.
6009 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
6011 PR rtl-optimization/57189
6012 * lra-constraints.c (process_alt_operands): Disfavor spilling
6015 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
6017 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
6019 2014-03-13 Jakub Jelinek <jakub@redhat.com>
6021 PR tree-optimization/59025
6023 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
6024 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
6026 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
6029 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
6030 calls of avr_out_plus_1.
6032 2014-03-13 Bin Cheng <bin.cheng@arm.com>
6034 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
6035 BB's single pred and update the father loop's latch info later.
6037 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
6039 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
6043 (VEC_base): Likewise.
6044 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
6045 registers, we need to swap double words in little endian mode.
6047 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
6048 to be a container mode for 128-bit integer operations added in ISA
6049 2.07. Unlike TImode and PTImode, the preferred register set is
6050 the Altivec/VMX registers for the 128-bit operations.
6052 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
6054 (rs6000_split_128bit_ok_p): Likewise.
6056 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
6057 macros for creating ISA 2.07 normal and overloaded builtin
6058 functions with 3 arguments.
6059 (BU_P8V_OVERLOAD_3): Likewise.
6060 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
6061 for use as overloaded functions.
6062 (VPERM_1TI_UNS): Likewise.
6063 (VSEL_1TI): Likewise.
6064 (VSEL_1TI_UNS): Likewise.
6065 (ST_INTERNAL_1ti): Likewise.
6066 (LD_INTERNAL_1ti): Likewise.
6067 (XXSEL_1TI): Likewise.
6068 (XXSEL_1TI_UNS): Likewise.
6069 (VPERM_1TI): Likewise.
6070 (VPERM_1TI_UNS): Likewise.
6071 (XXPERMDI_1TI): Likewise.
6072 (SET_1TI): Likewise.
6073 (LXVD2X_V1TI): Likewise.
6074 (STXVD2X_V1TI): Likewise.
6075 (VEC_INIT_V1TI): Likewise.
6076 (VEC_SET_V1TI): Likewise.
6077 (VEC_EXT_V1TI): Likewise.
6078 (EQV_V1TI): Likewise.
6079 (NAND_V1TI): Likewise.
6080 (ORC_V1TI): Likewise.
6081 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
6082 added in ISA 2.07. Add both normal 'altivec' builtins, and the
6084 (VADDUQM): Likewise.
6085 (VSUBCUQ): Likewise.
6086 (VADDEUQM): Likewise.
6087 (VADDECUQ): Likewise.
6088 (VSUBEUQM): Likewise.
6089 (VSUBECUQ): Likewise.
6091 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
6092 __int128_t and __uint128_t types.
6093 (__uint128_type): Likewise.
6094 (altivec_categorize_keyword): Add support for vector __int128_t,
6095 vector __uint128_t, vector __int128, and vector unsigned __int128
6096 as a container type for TImode operations that need to be done in
6097 VSX/Altivec registers.
6098 (rs6000_macro_to_expand): Likewise.
6099 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
6100 to support 128-bit integer instructions vaddcuq, vadduqm,
6101 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
6102 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
6104 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
6105 for V1TImode, and set up preferences to use VSX/Altivec registers.
6106 Setup VSX reload handlers.
6107 (rs6000_debug_reg_global): Likewise.
6108 (rs6000_init_hard_regno_mode_ok): Likewise.
6109 (rs6000_preferred_simd_mode): Likewise.
6110 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
6111 (easy_altivec_constant): Likewise.
6112 (output_vec_const_move): Likewise.
6113 (rs6000_expand_vector_set): Convert V1TImode set and extract to
6115 (rs6000_expand_vector_extract): Likewise.
6116 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
6118 (rs6000_const_vec): Add support for V1TImode.
6119 (rs6000_emit_le_vsx_load): Swap double words when loading or
6120 storing TImode/V1TImode.
6121 (rs6000_emit_le_vsx_store): Likewise.
6122 (rs6000_emit_le_vsx_move): Likewise.
6123 (rs6000_emit_move): Add support for V1TImode.
6124 (altivec_expand_ld_builtin): Likewise.
6125 (altivec_expand_st_builtin): Likewise.
6126 (altivec_expand_vec_init_builtin): Likewise.
6127 (altivec_expand_builtin): Likewise.
6128 (rs6000_init_builtins): Add support for V1TImode type. Add
6129 support for ISA 2.07 128-bit integer builtins. Define type names
6130 for the VSX/Altivec vector types.
6131 (altivec_init_builtins): Add support for overloaded vector
6132 functions with V1TImode type.
6133 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
6134 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
6136 (rs6000_split_128bit_ok_p): Likewise.
6137 (rs6000_handle_altivec_attribute): Create V1TImode from vector
6138 __int128_t and vector __uint128_t.
6140 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
6141 and mode attributes.
6148 (VS_scalar): Likewise.
6149 (VS_double): Likewise.
6150 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
6152 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
6153 we support the ISA 2.07 128-bit integer arithmetic instructions.
6154 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
6155 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
6156 and TImode types for use with the builtin functions.
6157 (V1TI_type_node): Likewise.
6158 (unsigned_V1TI_type_node): Likewise.
6159 (intTI_type_internal_node): Likewise.
6160 (uintTI_type_internal_node): Likewise.
6162 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
6163 128-bit builtin functions.
6164 (UNSPEC_VADDEUQM): Likewise.
6165 (UNSPEC_VADDECUQ): Likewise.
6166 (UNSPEC_VSUBCUQ): Likewise.
6167 (UNSPEC_VSUBEUQM): Likewise.
6168 (UNSPEC_VSUBECUQ): Likewise.
6169 (VM): Add V1TImode to vector mode iterators.
6171 (VI_unit): Likewise.
6172 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
6173 (altivec_vaddcuq): Likewise.
6174 (altivec_vsubuqm): Likewise.
6175 (altivec_vsubcuq): Likewise.
6176 (altivec_vaddeuqm): Likewise.
6177 (altivec_vaddecuq): Likewise.
6178 (altivec_vsubeuqm): Likewise.
6179 (altivec_vsubecuq): Likewise.
6181 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
6183 (BOOL_128): Likewise.
6184 (BOOL_REGS_OUTPUT): Likewise.
6185 (BOOL_REGS_OP1): Likewise.
6186 (BOOL_REGS_OP2): Likewise.
6187 (BOOL_REGS_UNARY): Likewise.
6188 (BOOL_REGS_AND_CR0): Likewise.
6190 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
6191 128-bit integer builtin support.
6192 (vec_vadduqm): Likewise.
6193 (vec_vaddecuq): Likewise.
6194 (vec_vaddeuqm): Likewise.
6195 (vec_vsubecuq): Likewise.
6196 (vec_vsubeuqm): Likewise.
6197 (vec_vsubcuq): Likewise.
6198 (vec_vsubuqm): Likewise.
6200 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
6201 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
6202 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
6203 128-bit integer add/subtract to ISA 2.07.
6205 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
6207 * config/arc/arc.c (arc_predicate_delay_insns):
6208 Fix third argument passed to conditionalize_nonjump.
6210 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
6212 * config/aarch64/aarch64-builtins.c
6213 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
6214 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
6215 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
6216 instead of __builtin_lfloor.
6217 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
6219 2014-03-12 Jakub Jelinek <jakub@redhat.com>
6221 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
6222 (tree_ssa_ifcombine_bb_1): New function.
6223 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
6224 is an empty forwarder block to then_bb or vice versa and then_bb
6225 and else_bb are effectively swapped.
6227 2014-03-12 Christian Bruel <christian.bruel@st.com>
6230 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
6231 REG_CFA_DEF_CFA note.
6232 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
6233 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
6235 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
6237 PR tree-optimization/60454
6238 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
6240 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6242 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
6243 Do not define target_cpu_default2 to generic.
6244 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
6245 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
6246 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
6248 2014-03-12 Jakub Jelinek <jakub@redhat.com>
6249 Marc Glisse <marc.glisse@inria.fr>
6251 PR tree-optimization/60502
6252 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
6253 instead of build_low_bits_mask.
6255 2014-03-12 Jakub Jelinek <jakub@redhat.com>
6258 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
6259 if there are multiple uses, but op doesn't live on E edge.
6260 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
6261 clobber stmts before __builtin_unreachable.
6263 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
6265 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
6266 hard_frame_pointer_rtx.
6267 * cse.c (cse_insn): Remove volatile check.
6268 * cselib.c (cselib_process_insn): Likewise.
6269 * dse.c (scan_insn): Likewise.
6271 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
6273 * config/arc/arc.c (conditionalize_nonjump): New function,
6275 (arc_ifcvt): ... this.
6276 (arc_predicate_delay_insns): Use it.
6278 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
6280 * config/arc/predicates.md (extend_operand): During/after reload,
6281 allow const_int_operand.
6282 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
6283 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
6284 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
6286 (umulsi3_highpart_i): Likewise.
6288 2014-03-11 Richard Biener <rguenther@suse.de>
6290 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
6291 Add asserts to guard possible wrong-code bugs.
6293 2014-03-11 Richard Biener <rguenther@suse.de>
6295 PR tree-optimization/60429
6296 PR tree-optimization/60485
6297 * tree-ssa-structalias.c (set_union_with_increment): Properly
6298 take into account all fields that overlap the shifted vars.
6299 (do_sd_constraint): Likewise.
6300 (do_ds_constraint): Likewise.
6301 (get_constraint_for_ptr_offset): Likewise.
6303 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
6305 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
6306 (nios2_compute_frame_layout):
6307 Add calculation of cfun->machine->fp_save_offset.
6308 (nios2_expand_prologue): Correct setting of frame pointer register
6310 (nios2_expand_epilogue): Update recovery of stack pointer from
6311 frame pointer accordingly.
6312 (nios2_initial_elimination_offset): Update calculation of offset
6313 for eliminating to HARD_FRAME_POINTER_REGNUM.
6315 2014-03-10 Jakub Jelinek <jakub@redhat.com>
6318 * ipa.c (symtab_remove_unreachable_nodes): Don't call
6319 cgraph_get_create_node on VAR_DECLs.
6321 2014-03-10 Richard Biener <rguenther@suse.de>
6324 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
6326 2014-03-08 Douglas B Rupp <rupp@gnat.com>
6328 * config/vms/vms.opt (vms_float_format): New variable.
6330 2014-03-08 Tobias Burnus <burnus@net-b.de>
6332 * doc/invoke.texi (-fcilkplus): Update implementation status.
6334 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
6335 Richard Biener <rguenther@suse.de>
6337 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
6338 consistently accross all TUs.
6339 (run_gcc): Enable -fshort-double automatically at link at link-time
6340 and disallow override.
6342 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
6345 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
6346 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
6347 if they can't be used.
6349 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6351 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
6352 for Solaris 11/x86 ld.
6353 * configure: Regenerate.
6355 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6357 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
6358 (LIB_TLS_SPEC): Save as ld_tls_libs.
6359 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
6360 (HAVE_AS_IX86_TLSLDM): New test.
6361 * configure, config.in: Regenerate.
6362 * config/i386/i386.c (legitimize_tls_address): Fall back to
6363 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
6364 cannot support TLS_MODEL_LOCAL_DYNAMIC.
6365 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
6366 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
6368 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
6370 * common.opt (fira-loop-pressure): Mark as optimization.
6372 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
6374 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
6375 an OpenMP mappable type.
6377 2014-03-06 Matthias Klose <doko@ubuntu.com>
6379 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
6380 MULTILIB_OSDIRNAMES is not defined.
6382 2014-03-06 Jakub Jelinek <jakub@redhat.com>
6383 Meador Inge <meadori@codesourcery.com>
6386 * config/arm/arm.c (arm_tls_symbol_p): Remove.
6387 (arm_legitimize_address): Call legitimize_tls_address for any
6388 arm_tls_referenced_p expression, handle constant addend. Call it
6389 before testing for !TARGET_ARM.
6390 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
6392 2014-03-06 Richard Biener <rguenther@suse.de>
6398 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
6400 * tree-streamer.c (record_common_node): Assert we don't record
6401 nodes with type double.
6402 (preload_common_node): Skip type double, complex double and double
6403 pointer since it is now frontend dependent due to fshort-double option.
6405 2014-03-06 Richard Biener <rguenther@suse.de>
6407 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
6408 or -fno-lto is specified and the linker has full plugin support.
6409 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
6410 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
6411 * lto-wrapper.c (merge_and_complain): Merge compile-time
6412 optimization levels.
6413 (run_gcc): And pass it through to the link options.
6415 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
6419 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
6421 * cselib.c (remove_useless_values): Skip to avoid quadratic
6422 behavior if the condition moved from...
6423 (cselib_process_insn): ... here holds.
6425 2014-03-05 Jakub Jelinek <jakub@redhat.com>
6428 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
6429 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
6432 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
6433 (TM_H): Add x86-tune.def.
6435 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6437 * config/aarch64/aarch64.c (generic_tunings):
6438 Use cortexa57_extra_costs.
6440 2014-03-05 Jakub Jelinek <jakub@redhat.com>
6443 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
6444 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
6445 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
6448 2014-03-04 Heiher <r@hev.cc>
6450 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
6451 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
6453 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
6455 * config/i386/predicates.md (const2356_operand): Change to ...
6456 (const2367_operand): ... this.
6457 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
6459 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6460 (*avx512pf_scatterpf<mode>sf): Ditto.
6461 (avx512pf_scatterpf<mode>df): Ditto.
6462 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6463 (*avx512pf_scatterpf<mode>df): Ditto.
6464 * config/i386/i386.c (ix86_expand_builtin): Update
6465 incorrect hint operand error message.
6467 2014-03-04 Richard Biener <rguenther@suse.de>
6469 * lto-section-in.c (lto_get_section_data): Fix const cast.
6471 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
6473 * tree-streamer.c (record_common_node): Assert we don't record
6474 nodes with type double.
6475 (preload_common_node): Skip type double, complex double and double
6476 pointer since it is now frontend dependent due to fshort-double option.
6478 2014-03-04 Richard Biener <rguenther@suse.de>
6481 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
6482 (lto_input_toplevel_asms): Likewise.
6483 * lto-section-in.c (lto_get_section_data): Instead do it here
6486 2014-03-04 Richard Biener <rguenther@suse.de>
6488 PR tree-optimization/60382
6489 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
6490 dead PHIs a reduction.
6492 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
6494 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
6496 (_mm_prefetch): Move out of GCC target("sse") pragma.
6497 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
6498 GCC target("prfchw") pragma.
6499 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
6501 * config/i386/i386.c (ix86_option_override_internal): Enable
6502 -mprfchw with -mprefetchwt1.
6504 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
6506 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
6509 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
6511 * opts.h (CL_PCH_IGNORE): Define.
6512 * targhooks.c (option_affects_pch_p):
6513 Return false for options that have CL_PCH_IGNORE set.
6514 * opt-functions.awk: Process PchIgnore.
6515 * doc/options.texi: Document PchIgnore.
6517 * config/arc/arc.opt (misize): Add PchIgnore property.
6519 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6521 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
6522 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
6523 constraint on constants to permit them being loaded into
6524 GENERAL_REGS or BASE_REGS.
6526 2014-03-03 Nick Clifton <nickc@redhat.com>
6528 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
6529 anti-cacnonical alternatives.
6530 (negandhi3_real): New pattern.
6531 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
6533 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
6535 * config/avr/avr-mcus.def: Remove atxmega16x1.
6536 * config/avr/avr-tables.opt: Regenerate.
6537 * config/avr/t-multilib: Regenerate.
6538 * doc/avr-mmcu.texi: Regenerate.
6540 2014-03-03 Tobias Grosser <tobias@grosser.es>
6541 Mircea Namolaru <mircea.namolaru@inria.fr>
6543 PR tree-optimization/58028
6544 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
6547 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6549 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
6550 not handled by recognizers.
6552 2014-03-03 Jakub Jelinek <jakub@redhat.com>
6555 * function.c (expand_function_end): Don't emit
6556 clobber_return_register sequence if clobber_after is a BARRIER.
6557 * cfgexpand.c (construct_exit_block): Append instructions before
6558 return_label to prev_bb.
6560 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6562 * config/rs6000/constraints.md: Document reserved use of "wc".
6564 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
6567 * ipa.c (function_and_variable_visibility): When dissolving comdat
6568 group, also set all symbols to local.
6570 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
6575 2013-12-14 Jan Hubicka <jh@suse.cz>
6577 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
6579 2014-03-02 Jon Beniston <jon@beniston.com>
6585 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
6586 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
6587 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
6588 (simple_return, *simple_return): New patterns
6589 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
6590 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
6592 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
6594 * dwarf2out.c (gen_subprogram_die): Tidy.
6596 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
6599 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
6600 (*mov_t_msb_neg_negc): ... this new insn.
6602 2014-02-28 Jason Merrill <jason@redhat.com>
6605 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
6608 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
6611 * dwarf2out.c (decltype_auto_die): New static.
6612 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
6613 (gen_type_die_with_usage): Handle 'decltype(auto)'.
6614 (is_cxx_auto): Likewise.
6616 2014-02-28 Ian Bolton <ian.bolton@arm.com>
6618 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
6619 we are not using general regs only.
6621 2014-02-28 Richard Biener <rguenther@suse.de>
6624 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
6625 previous fix and only allow to remove trivial pre-headers
6626 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
6627 (remove_forwarder_block): Properly update the latch of a loop.
6629 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
6632 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
6633 (cselib_preserved_hash_table): New.
6634 (preserve_constants_and_equivs): Move preserved vals to it.
6635 (cselib_find_slot): Look it up first.
6636 (cselib_init): Initialize it.
6637 (cselib_finish): Release it.
6638 (dump_cselib_table): Dump it.
6640 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
6643 * cselib.c (remove_useless_values): Skip to avoid quadratic
6644 behavior if the condition moved from...
6645 (cselib_process_insn): ... here holds.
6647 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
6650 * var-tracking.c (vt_initialize): Apply the same condition to
6651 preserve the CFA base value.
6653 2014-02-28 Joey Ye <joey.ye@arm.com>
6656 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
6657 if reload in progress or completed.
6659 2014-02-28 Tobias Burnus <burnus@net-b.de>
6662 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
6665 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
6667 * doc/tm.texi.in (Condition Code Status): Update documention for
6668 relative locations of cc0-setter and cc0-user.
6670 2014-02-27 Jeff Law <law@redhat.com>
6672 PR rtl-optimization/52714
6673 * combine.c (try_combine): When splitting an unrecognized PARALLEL
6674 into two independent simple sets, if I3 is a jump, ensure the
6675 pattern we place into I3 is a (set (pc) ...).
6677 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
6678 Jeff Law <law@redhat.com>
6680 PR rtl-optimization/49847
6681 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
6682 are in different blocks.
6683 * doc/tm.texi (Condition Code Status): Update documention for
6684 relative locations of cc0-setter and cc0-user.
6686 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
6689 * lra.c (lra_emit_add): Check SUBREG too.
6691 2014-02-27 Andreas Schwab <schwab@suse.de>
6693 * config/m68k/m68k.c (m68k_option_override): Disable
6694 -flive-range-shrinkage for classic m68k.
6695 (m68k_override_options_after_change): Likewise.
6697 2014-02-27 Marek Polacek <polacek@redhat.com>
6700 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
6701 -Wmaybe-uninitialized.
6703 2014-02-27 Alan Modra <amodra@gmail.com>
6706 * reload1.c (emit_input_reload_insns): When reload_override_in,
6707 set old to rl->in_reg when rl->in_reg is a subreg.
6709 2014-02-26 Richard Biener <rguenther@suse.de>
6712 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
6714 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
6716 * common/config/i386/predicates.md (const1256_operand): Remove.
6717 (const2356_operand): New.
6718 (const_1_to_2_operand): Remove.
6719 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
6720 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
6721 (*avx512pf_gatherpf<mode>sf): Ditto.
6722 (avx512pf_gatherpf<mode>df): Ditto.
6723 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6724 (*avx512pf_gatherpf<mode>df): Ditto.
6725 (avx512pf_scatterpf<mode>sf): Ditto.
6726 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6727 (*avx512pf_scatterpf<mode>sf): Ditto.
6728 (avx512pf_scatterpf<mode>df): Ditto.
6729 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6730 (*avx512pf_scatterpf<mode>df): Ditto.
6731 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
6733 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
6735 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
6736 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
6737 (_mm512_mask_testn_epi64_mask): Move to ...
6738 * config/i386/avx512cdintrin.h: Here.
6739 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
6740 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
6741 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
6742 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
6743 TARGET_AVX512F from TARGET_AVX512CD.
6745 2014-02-26 Richard Biener <rguenther@suse.de>
6748 * ipa.c (walk_polymorphic_call_targets): Properly guard
6749 call to inline_update_overall_summary.
6751 2014-02-26 Bin Cheng <bin.cheng@arm.com>
6754 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
6755 and latches only if requested. Fix latch if it is removed.
6756 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
6757 LOOPS_HAVE_PREHEADERS.
6759 2014-02-25 Andrew Pinski <apinski@cavium.com>
6761 * builtins.c (expand_builtin_thread_pointer): Create a new target
6762 when the target is NULL.
6764 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
6766 PR rtl-optimization/60317
6767 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
6768 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
6769 * lra-assigns.c: Include params.h.
6770 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
6771 other reload pseudos considerations.
6773 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6775 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
6776 to use canonical form for nor<mode>3.
6778 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6781 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
6784 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
6786 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
6787 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
6788 (ix86_handle_option): Handle OPT_mprefetchwt1.
6789 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
6790 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
6792 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
6793 OPTION_MASK_ISA_PREFETCHWT1.
6794 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
6795 (PTA_PREFETCHWT1): New.
6796 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
6797 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
6798 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
6799 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
6800 (*prefetch_avx512pf_<mode>_: Change into ...
6801 (*prefetch_prefetchwt1_<mode>: This.
6802 * config/i386/i386.opt (mprefetchwt1): New.
6803 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
6804 (_mm_prefetch): Handle intent to write.
6805 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
6807 2014-02-25 Richard Biener <rguenther@suse.de>
6810 * emit-rtl.c (mem_attrs_htab): Remove.
6811 (mem_attrs_htab_hash): Likewise.
6812 (mem_attrs_htab_eq): Likewise.
6813 (set_mem_attrs): Always allocate new mem-attrs when something changed.
6814 (init_emit_once): Do not allocate mem_attrs_htab.
6816 2014-02-25 Richard Biener <rguenther@suse.de>
6819 * lto-opts.c (lto_write_options): Output non-explicit conservative
6820 -fwrapv, -fno-trapv and -fno-strict-overflow.
6821 * lto-wrapper.c (merge_and_complain): Handle merging those options.
6822 (run_gcc): And pass them through.
6824 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
6826 * sel-sched.c (calculate_new_fences): New parameter ptime.
6827 Calculate it as a maximum over all fence cycles.
6828 (sel_sched_region_2): Adjust the call to calculate_new_fences.
6829 Print the final schedule timing when sched_verbose.
6831 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
6833 PR rtl-optimization/60292
6834 * sel-sched.c (fill_vec_av_set): Do not reset target availability
6835 bit fot the fence instruction.
6837 2014-02-24 Alangi Derick <alangiderick@gmail.com>
6839 * calls.h: Fix typo in comment.
6841 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
6843 * config/pa/pa.c (pa_output_move_double): Don't valididate when
6844 adjusting offsetable addresses.
6846 2014-02-24 Guozhi Wei <carrot@google.com>
6848 * sparseset.h (sparseset_pop): Fix the wrong index.
6850 2014-02-24 Walter Lee <walt@tilera.com>
6852 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
6853 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
6855 * common/config/tilegx/tilegx-common.c
6856 (TARGET_DEFAULT_TARGET_FLAGS): Define.
6857 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
6859 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
6860 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
6861 (tilegx_gimplify_va_arg_expr): Handle big endian.
6862 (tilegx_expand_unaligned_load): Ditto.
6863 (tilegx_expand_unaligned_store): Ditto.
6864 (TARGET_RETURN_IN_MSB): New.
6865 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
6866 (TARGET_ENDIAN_DEFAULT): New.
6867 (TARGET_BIG_ENDIAN): Handle big endian.
6868 (BYTES_BIG_ENDIAN): Ditto.
6869 (WORDS_BIG_ENDIAN): Ditto.
6870 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
6873 * config/tilegx/tilegx.md (extv): Handle big endian.
6875 (insn_st<n>): Ditto.
6876 (insn_st<n>_add<bitsuffix>): Ditto.
6877 (insn_stnt<n>): Ditto.
6878 (insn_stnt<n>_add<bitsuffix>):Ditto.
6879 (vec_interleave_highv8qi): Handle big endian.
6880 (vec_interleave_highv8qi_be): New.
6881 (vec_interleave_highv8qi_le): New.
6882 (insn_v1int_h): Handle big endian.
6883 (vec_interleave_lowv8qi): Handle big endian.
6884 (vec_interleave_lowv8qi_be): New.
6885 (vec_interleave_lowv8qi_le): New.
6886 (insn_v1int_l): Handle big endian.
6887 (vec_interleave_highv4hi): Handle big endian.
6888 (vec_interleave_highv4hi_be): New.
6889 (vec_interleave_highv4hi_le): New.
6890 (insn_v2int_h): Handle big endian.
6891 (vec_interleave_lowv4hi): Handle big endian.
6892 (vec_interleave_lowv4hi_be): New.
6893 (vec_interleave_lowv4hi_le): New.
6894 (insn_v2int_l): Handle big endian.
6895 (vec_interleave_highv2si): Handle big endian.
6896 (vec_interleave_highv2si_be): New.
6897 (vec_interleave_highv2si_le): New.
6898 (insn_v4int_h): Handle big endian.
6899 (vec_interleave_lowv2si): Handle big endian.
6900 (vec_interleave_lowv2si_be): New.
6901 (vec_interleave_lowv2si_le): New.
6902 (insn_v4int_l): Handle big endian.
6903 * config/tilegx/tilegx.opt (mbig-endian): New option.
6904 (mlittle-endian): New option.
6905 * doc/install.texi: Document tilegxbe-linux.
6906 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
6908 2014-02-24 Martin Jambor <mjambor@suse.cz>
6911 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
6912 there are no parameter descriptors.
6914 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
6916 PR rtl-optimization/60268
6917 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
6918 initialization to ...
6919 (sched_rgn_init): ... here.
6920 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
6922 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
6924 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
6927 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
6929 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
6932 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
6934 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
6935 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
6937 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
6939 * config/microblaze/predicates.md: Add cmp_op predicate.
6940 * config/microblaze/microblaze.md: Add branch_compare instruction
6941 which uses cmp_op predicate and emits cmp insn before branch.
6942 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
6943 to microblaze_expand_conditional_branch and consolidate logic.
6944 (microblaze_expand_conditional_branch): emit branch_compare
6945 insn instead of handling cmp op separate from branch insn.
6947 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6949 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
6952 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6954 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
6955 define_insn with define_expand and new define_insn
6956 *altivec_lve<VI_char>x_internal.
6957 (altivec_stve<VI_char>x): Replace define_insn with define_expand
6958 and new define_insn *altivec_stve<VI_char>x_internal.
6959 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
6961 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
6963 (altivec_expand_stvex_be): New function.
6965 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
6967 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
6968 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
6969 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
6970 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
6972 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
6975 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
6976 instead of emit_move_insn.
6978 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6980 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
6982 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
6983 gen_altivec_vsumsws.
6985 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6987 * config/rs6000/altivec.md (altivec_lvxl): Rename as
6988 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
6989 (altivec_lvxl_<mode>): New define_expand incorporating
6990 -maltivec=be semantics where needed.
6991 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
6992 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
6993 semantics where needed.
6994 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
6995 (altivec_stvx_<mode>): New define_expand incorporating
6996 -maltivec=be semantics where needed.
6997 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
6998 VM2 iterator instead of V4SI.
6999 (altivec_stvxl_<mode>): New define_expand incorporating
7000 -maltivec=be semantics where needed.
7001 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
7002 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
7003 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
7004 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
7005 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
7006 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
7007 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
7008 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
7009 ALTIVEC_BUILTIN_STVXL.
7010 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
7011 (altivec_expand_stvx_be): Likewise.
7012 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
7013 (altivec_expand_lvx_be): Likewise.
7014 (altivec_expand_stvx_be): Likewise.
7015 (altivec_expand_builtin): Add cases for
7016 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
7017 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
7018 (altivec_init_builtins): Add definitions for
7019 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
7020 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
7022 2014-02-21 Catherine Moore <clm@codesourcery.com>
7024 * doc/invoke.texi (mvirt, mno-virt): Document.
7025 * config/mips/mips.opt (mvirt): New option.
7026 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
7028 2014-02-21 Richard Biener <rguenther@suse.de>
7030 PR tree-optimization/60276
7031 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
7032 (STMT_VINFO_MIN_NEG_DIST): New macro.
7033 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
7034 STMT_VINFO_MIN_NEG_DIST.
7035 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
7036 made for negative dependence distances still hold.
7038 2014-02-21 Richard Biener <rguenther@suse.de>
7041 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
7042 DECL_INITIAL for globals not in the current function context.
7044 2014-02-21 Jakub Jelinek <jakub@redhat.com>
7046 PR tree-optimization/56490
7047 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
7048 * tree-ssa-uninit.c: Include params.h.
7049 (compute_control_dep_chain): Add num_calls argument, return false
7050 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
7051 num_calls to recursive call.
7052 (find_predicates): Change dep_chain into normal array,
7053 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
7054 variable and adjust compute_control_dep_chain caller.
7055 (find_def_preds): Likewise.
7057 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
7059 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
7060 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
7062 2014-02-21 Nick Clifton <nickc@redhat.com>
7064 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
7065 (pushhi1): Likewise.
7066 (popqi1): Add mode to pre_dec.
7069 2014-02-21 Jakub Jelinek <jakub@redhat.com>
7071 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
7072 mode for mask of V8SFmode permutation.
7074 2014-02-20 Richard Henderson <rth@redhat.com>
7077 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
7078 a new pseudo for OLDVAL.
7080 2014-02-20 Jakub Jelinek <jakub@redhat.com>
7083 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
7084 gen_reg_rtx if d->testing_p.
7085 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
7086 if d->testing_p and we will certainly return true.
7087 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
7090 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
7092 * emit-rtl.c (gen_reg_rtx): Assert that
7093 crtl->emit.regno_pointer_align_length is non-zero.
7095 2014-02-20 Richard Henderson <rth@redhat.com>
7098 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
7099 on failure the store back into EXPECT.
7101 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
7102 Sandra Loosemore <sandra@codesourcery.com>
7104 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
7105 * config/nios2/nios2.c (nios2_function_profiler): Add
7106 -fPIC (flag_pic == 2) support.
7107 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
7108 (nios2_large_offset_p): New function.
7109 (nios2_unspec_reloc_p): Move up position, update to use
7110 nios2_large_offset_p.
7111 (nios2_unspec_address): Remove function.
7112 (nios2_unspec_offset): New function.
7113 (nios2_large_got_address): New function.
7114 (nios2_got_address): Add large offset support.
7115 (nios2_legitimize_tls_address): Update usage of removed and new
7117 (nios2_symbol_binds_local_p): New function.
7118 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
7119 (nios2_legitimize_address): Update to use nios2_large_offset_p.
7120 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
7121 (nios2_print_operand): Merge H/L processing, add hiadj/lo
7122 processing for (const (unspec ...)).
7123 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
7125 2014-02-20 Richard Biener <rguenther@suse.de>
7127 * tree-cfg.c (replace_uses_by): Mark altered BBs before
7128 doing the substitution.
7129 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
7131 2014-02-20 Martin Jambor <mjambor@suse.cz>
7134 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
7135 info when checking whether lattices are bottom.
7137 2014-02-20 Richard Biener <rguenther@suse.de>
7140 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
7143 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
7146 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
7147 parameter specifying the scaling.
7148 (inline_call): Update.
7149 (want_inline_recursively): Guard division by zero.
7150 (recursive_inlining): Update.
7151 * ipa-inline.h (clone_inlined_nodes): Update.
7153 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
7156 * config/i386/i386.c (classify_argument): Pass structures of size
7157 64 bytes or less in register.
7159 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
7160 Kirill Yukhin <kirill.yukhin@intel.com>
7162 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
7163 (_mm_rcp28_round_ss): Ditto.
7164 (_mm_rsqrt28_round_sd): Ditto.
7165 (_mm_rsqrt28_round_ss): Ditto.
7166 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
7167 (_mm_rcp14_round_ss): Ditto.
7168 (_mm_rsqrt14_round_sd): Ditto.
7169 (_mm_rsqrt14_round_ss): Ditto.
7170 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
7171 the first input operand, get rid of match_dup.
7172 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
7174 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
7176 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
7177 operand as the first input operand, set type attribute.
7178 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
7180 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
7181 operand as the first input operand, set type attribute.
7183 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7185 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
7188 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
7191 * config/i386/i386.c (construct_container): Remove TFmode check
7192 for X86_64_INTEGER_CLASS.
7194 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
7197 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
7198 only when -Wpsabi is enabled.
7200 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
7203 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
7204 passing arrays in registers are the same as for structs, so remove the
7205 special case for them.
7207 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
7209 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
7210 destination type, extract only the valid bits if the source type is not
7211 integral and has a different mode.
7213 2014-02-19 Richard Biener <rguenther@suse.de>
7216 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
7219 2014-02-19 Richard Biener <rguenther@suse.de>
7222 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
7223 (ipa_modify_call_arguments): Emit an argument load explicitely and
7224 preserve virtual SSA form there and for the replacement call.
7225 Do not update SSA form nor free dominance info.
7227 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
7229 * ipa.c (function_and_variable_visibility): Also clear WEAK
7230 flag when disolving COMDAT_GROUP.
7232 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
7234 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
7235 * ipa-prop.c (ipa_set_jf_known_type): Return early when
7237 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
7238 do more sanity checks.
7239 (detect_type_change): Return true when giving up early.
7240 (compute_complex_assign_jump_func): Fix type parameter of
7241 ipa_set_ancestor_jf.
7242 (compute_complex_ancestor_jump_func): Likewise.
7243 (update_jump_functions_after_inlining): Fix updating of
7245 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
7247 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
7249 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
7250 inline clones when edge disappears.
7252 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
7255 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
7256 Split 64-bit moves into 2 patterns. Do not allow the use of
7257 direct move for TDmode in little endian, since the decimal value
7258 has little endian bytes within a word, but the 64-bit pieces are
7259 ordered in a big endian fashion, and normal subreg's of TDmode are
7261 (mov<mode>_64bit_dm): Likewise.
7262 (movtd_64bit_nodm): Likewise.
7264 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
7266 PR tree-optimization/60174
7267 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
7268 statement of an SSA_NAME that occurs in an abnormal PHI node.
7270 2014-02-18 Jakub Jelinek <jakub@redhat.com>
7273 * final.c (SEEN_BB): Remove.
7274 (SEEN_NOTE, SEEN_EMITTED): Renumber.
7275 (final_scan_insn): Don't force_source_line on second
7276 NOTE_INSN_BASIC_BLOCK.
7278 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
7281 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
7282 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
7283 (type_natural_mode): Warn ABI change when %zmm register is not
7284 available for AVX512F vector value passing.
7286 2014-02-18 Kai Tietz <ktietz@redhat.com>
7289 * config/i386/i386.c (ix86_expand_prologue): Use value in
7290 rax register as displacement when restoring %r10 or %rax.
7291 Fix wrong offset when restoring both registers.
7293 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
7295 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
7296 assertion with conditional return.
7298 2014-02-18 Jakub Jelinek <jakub@redhat.com>
7299 Uros Bizjak <ubizjak@gmail.com>
7302 * config/i386/driver-i386.c (host_detect_local_cpu): If
7303 YMM state is not saved by the OS, also clear has_f16c. Move
7304 CPUID 0x80000001 handling before YMM state saving checking.
7306 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
7308 PR rtl-optimization/58960
7309 * haifa-sched.c (alloc_global_sched_pressure_data): New,
7310 factored out from ...
7311 (sched_init): ... here.
7312 (free_global_sched_pressure_data): New, factored out from ...
7313 (sched_finish): ... here.
7314 * sched-int.h (free_global_sched_pressure_data): Declare.
7315 * sched-rgn.c (nr_regions_initial): New static global.
7316 (haifa_find_rgns): Initialize it.
7317 (schedule_region): Disable sched-pressure for the newly
7320 2014-02-17 Richard Biener <rguenther@suse.de>
7322 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
7323 release SSA defs of pattern stmts.
7325 2014-02-17 Richard Biener <rguenther@suse.de>
7327 * tree-inline.c (expand_call_inline): Release the virtual
7328 operand defined by the call we are about to inline.
7330 2014-02-17 Richard Biener <rguenther@suse.de>
7332 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
7334 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
7335 Ilya Tocar <ilya.tocar@intel.com>
7337 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
7338 arguments order in builtin.
7339 (_mm512_permutexvar_epi64): Ditto.
7340 (_mm512_mask_permutexvar_epi64): Ditto
7341 (_mm512_maskz_permutexvar_epi32): Ditto
7342 (_mm512_permutexvar_epi32): Ditto
7343 (_mm512_mask_permutexvar_epi32): Ditto
7345 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7347 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
7348 (p8_vmrgow): Likewise.
7350 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7352 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
7355 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
7358 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
7359 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
7360 into 64-bit and 32-bit moves. On 64-bit moves, add support for
7361 using direct move instructions on ISA 2.07. Also adjust
7362 instruction length for 64-bit.
7363 (mov<mode>_64bit, TFmode/TDmode): Likewise.
7364 (mov<mode>_32bit, TFmode/TDmode): Likewise.
7366 2014-02-15 Alan Modra <amodra@gmail.com>
7370 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
7371 find_replacement on parts of insn rtl that might be reloaded.
7373 2014-02-15 Richard Biener <rguenther@suse.de>
7375 PR tree-optimization/60183
7376 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
7377 (tree_ssa_phiprop): Calculate and free post-dominators.
7379 2014-02-14 Jeff Law <law@redhat.com>
7381 PR rtl-optimization/60131
7382 * ree.c (get_extended_src_reg): New function.
7383 (combine_reaching_defs): Use it rather than assuming location of REG.
7384 (find_and_remove_re): Verify first operand of extension is
7385 a REG before adding the insns to the copy list.
7387 2014-02-14 Roland McGrath <mcgrathr@google.com>
7389 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
7390 * configure: Regenerated.
7391 * config.in: Regenerated.
7392 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
7393 instead of ASM_SHORT.
7395 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
7396 Richard Earnshaw <rearnsha@arm.com>
7398 PR rtl-optimization/59535
7399 * lra-constraints.c (process_alt_operands): Encourage alternative
7400 when unassigned pseudo class is superset of the alternative class.
7401 (inherit_reload_reg): Don't inherit when optimizing for code size.
7402 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
7403 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
7404 modes not less than 4 for Thumb1.
7406 2014-02-14 Kyle McMartin <kyle@redhat.com>
7409 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
7411 2014-02-14 Richard Biener <rguenther@suse.de>
7413 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
7414 (get_frame_arg): Drop the assert with langhook types_compatible_p.
7415 Do not strip INDIRECT_REFs.
7417 2014-02-14 Richard Biener <rguenther@suse.de>
7420 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
7421 DECL_FUNCTION_SPECIFIC_TARGET.
7422 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
7423 * tree-streamer-out.c (pack_ts_target_option): Remove.
7424 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
7425 (write_ts_function_decl_tree_pointers): Do not stream
7426 DECL_FUNCTION_SPECIFIC_TARGET.
7427 * tree-streamer-in.c (unpack_ts_target_option): Remove.
7428 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
7429 (lto_input_ts_function_decl_tree_pointers): Do not stream
7430 DECL_FUNCTION_SPECIFIC_TARGET.
7432 2014-02-14 Jakub Jelinek <jakub@redhat.com>
7434 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
7435 (get_initial_def_for_induction, vectorizable_induction): Ignore
7436 debug stmts when looking for exit_phi.
7437 (vectorizable_live_operation): Fix up condition.
7439 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
7441 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
7442 nreverse() because it changes the content of original tree list.
7444 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
7446 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
7447 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
7449 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
7451 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
7452 GNU coding standards.
7454 2014-02-13 Jakub Jelinek <jakub@redhat.com>
7457 * dwarf2out.c (gen_subprogram_die): Don't call
7458 add_calling_convention_attribute if subr_die is old_die.
7460 2014-02-13 Sharad Singhai <singhai@google.com>
7462 * doc/optinfo.texi: Fix order of nodes.
7464 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
7466 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
7467 operands[2], not operands[3].
7469 2014-02-13 Richard Biener <rguenther@suse.de>
7472 * doc/install.texi (ISL): Update recommended version to 0.12.2,
7473 mention the possibility of an in-tree build.
7474 (CLooG): Update recommended version to 0.18.1, mention the
7475 possibility of an in-tree build and clarify that the ISL
7476 bundled with CLooG does not work.
7478 2014-02-13 Jakub Jelinek <jakub@redhat.com>
7481 * expr.c (compress_float_constant): If x is a hard register,
7482 extend into a pseudo and then move to x.
7484 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
7486 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
7487 caused by bad second argument to warning_at() with -mhotpatch and
7488 nested functions (e.g. with gfortran).
7490 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
7492 * opts.c (option_name): Remove "enabled by default" rider.
7494 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
7496 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
7498 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
7499 Uros Bizjak <ubizjak@gmail.com>
7502 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
7503 * configure: Regenerated.
7505 2014-02-12 Richard Biener <rguenther@suse.de>
7507 * vec.c (vec_prefix::calculate_allocation): Move as
7508 inline variant to vec.h.
7509 (vec_prefix::calculate_allocation_1): New out-of-line version.
7510 * vec.h (vec_prefix::calculate_allocation_1): Declare.
7511 (vec_prefix::m_has_auto_buf): Rename to ...
7512 (vec_prefix::m_using_auto_storage): ... this.
7513 (vec_prefix::calculate_allocation): Inline the easy cases
7514 and dispatch to calculate_allocation_1 which doesn't need the
7516 (va_heap::reserve): Use gcc_checking_assert.
7517 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
7518 m_using_auto_storage.
7519 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
7521 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
7522 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
7523 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
7525 2014-02-12 Richard Biener <rguenther@suse.de>
7527 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
7528 when we found a dependence.
7530 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
7532 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
7534 (maybe_fold_stmt): ... into this new function.
7535 * omp-low.c (lower_omp): Update comment.
7537 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
7540 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
7543 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
7545 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
7546 identifiers in comments.
7547 (cortexa53_extra_costs): Likewise.
7548 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
7549 (cortexa7_extra_costs): Likewise.
7550 (cortexa12_extra_costs): Likewise.
7551 (cortexa15_extra_costs): Likewise.
7552 (v7m_extra_costs): Likewise.
7554 2014-02-12 Richard Biener <rguenther@suse.de>
7557 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
7558 of posix_memalign being successful.
7559 (lower_stmt): Restrict lowering of posix_memalign to when
7560 -ftree-bit-ccp is enabled.
7562 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
7564 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
7566 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
7568 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
7570 PR rtl-optimization/60116
7571 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
7572 other_insn once the combination has been validated.
7574 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
7577 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
7579 * ipa-devirt.c: Include demangle.h
7580 (odr_violation_reported): New static variable.
7581 (add_type_duplicate): Update odr_violations.
7582 (maybe_record_node): Add completep parameter; update it.
7583 (record_target_from_binfo): Add COMPLETEP parameter;
7584 update it as needed.
7585 (possible_polymorphic_call_targets_1): Likewise.
7586 (struct polymorphic_call_target_d): Add nonconstruction_targets;
7587 rename FINAL to COMPLETE.
7588 (record_targets_from_bases): Sanity check we found the binfo;
7589 fix COMPLETEP updating.
7590 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
7591 parameter, fix computing of COMPLETEP.
7592 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
7593 at LTO time do demangling.
7594 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
7595 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
7597 (gimple_get_virt_method_for_binfo): Likewise.
7598 * gimple-fold.h (gimple_get_virt_method_for_binfo,
7599 gimple_get_virt_method_for_vtable): Update prototypes.
7601 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
7604 * genautomata.c (add_presence_absence): Fix typo with
7605 {final_}presence_list.
7607 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
7610 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
7611 for VSX/Altivec vectors that land in GPR registers.
7613 2014-02-11 Richard Henderson <rth@redhat.com>
7614 Jakub Jelinek <jakub@redhat.com>
7617 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
7618 around drhs if type conversion to lacc->type is not useless.
7620 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7622 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
7624 (cortex-a57.cortex-a53): Likewise.
7625 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
7627 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7629 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
7632 2014-02-11 Renlin Li <Renlin.Li@arm.com>
7634 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
7635 add_options_for_arm_vfp3.
7637 2014-02-11 Jeff Law <law@redhat.com>
7640 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
7641 object with an undesirable mode.
7643 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7646 * config/i386/sol2-9.h: New file.
7647 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
7648 *-*-solaris2.9*): Use it.
7650 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
7652 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
7653 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
7655 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
7657 * config/microblaze/microblaze.c: Extend mcpu version format
7659 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
7661 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
7663 2014-02-10 Richard Henderson <rth@redhat.com>
7666 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
7667 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
7668 ms-abi vs -mno-accumulate-outgoing-args.
7669 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
7670 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
7673 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
7676 * cfgexpand.c (expand_asm_operands): Attach source location to
7677 ASM_INPUT rtx objects.
7678 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
7680 2014-02-10 Nick Clifton <nickc@redhat.com>
7682 * config/mn10300/mn10300.c (popcount): New function.
7683 (mn10300_expand_prologue): Include saved registers in stack usage
7686 2014-02-10 Jeff Law <law@redhat.com>
7689 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
7690 when changing the SET_DEST of a prior insn to avoid an input reload.
7692 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
7694 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
7695 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
7696 -mcall-openbsd, or -mcall-linux.
7697 (CC1_ENDIAN_BIG_SPEC): Remove.
7698 (CC1_ENDIAN_LITTLE_SPEC): Remove.
7699 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
7700 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
7701 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
7702 and %cc1_endian_default.
7703 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
7705 2014-02-10 Richard Biener <rguenther@suse.de>
7707 PR tree-optimization/60115
7708 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
7709 MEM_REF handling. Properly verify that the accesses are not
7710 out of the objects bound.
7712 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7714 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
7717 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
7719 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
7720 proper constants and fix formatting.
7721 (possible_polymorphic_call_targets): Fix formatting.
7723 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
7724 Ilya Tocar <ilya.tocar@intel.com>
7726 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
7727 (_mm512_loadu_epi32): Renamed into...
7728 (_mm512_loadu_si512): This.
7729 (_mm512_storeu_epi32): Renamed into...
7730 (_mm512_storeu_si512): This.
7731 (_mm512_maskz_ceil_ps): Removed.
7732 (_mm512_maskz_ceil_pd): Ditto.
7733 (_mm512_maskz_floor_ps): Ditto.
7734 (_mm512_maskz_floor_pd): Ditto.
7735 (_mm512_floor_round_ps): Ditto.
7736 (_mm512_floor_round_pd): Ditto.
7737 (_mm512_ceil_round_ps): Ditto.
7738 (_mm512_ceil_round_pd): Ditto.
7739 (_mm512_mask_floor_round_ps): Ditto.
7740 (_mm512_mask_floor_round_pd): Ditto.
7741 (_mm512_mask_ceil_round_ps): Ditto.
7742 (_mm512_mask_ceil_round_pd): Ditto.
7743 (_mm512_maskz_floor_round_ps): Ditto.
7744 (_mm512_maskz_floor_round_pd): Ditto.
7745 (_mm512_maskz_ceil_round_ps): Ditto.
7746 (_mm512_maskz_ceil_round_pd): Ditto.
7747 (_mm512_expand_pd): Ditto.
7748 (_mm512_expand_ps): Ditto.
7749 * config/i386/i386.c (ix86_builtins): Remove
7750 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
7751 (bdesc_args): Ditto.
7752 * config/i386/predicates.md (const1256_operand): New.
7753 (const_1_to_2_operand): Ditto.
7754 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
7755 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
7756 (*avx512pf_gatherpf<mode>sf): Ditto.
7757 (avx512pf_gatherpf<mode>df): Ditto.
7758 (*avx512pf_gatherpf<mode>df_mask): Ditto.
7759 (*avx512pf_gatherpf<mode>df): Ditto.
7760 (avx512pf_scatterpf<mode>sf): Ditto.
7761 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
7762 (*avx512pf_scatterpf<mode>sf): Ditto.
7763 (avx512pf_scatterpf<mode>df): Ditto.
7764 (*avx512pf_scatterpf<mode>df_mask): Ditto.
7765 (*avx512pf_scatterpf<mode>df): Ditto.
7766 (avx512f_expand<mode>): Removed.
7767 (<shift_insn><mode>3<mask_name>): Change predicate type.
7769 2014-02-08 Jakub Jelinek <jakub@redhat.com>
7771 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
7772 not at the end of datarefs vector use ordered_remove to avoid
7773 reordering datarefs vector.
7776 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
7777 mark local addressable non-static vars as GOVD_PRIVATE
7778 instead of GOVD_LOCAL.
7779 * omp-low.c (lower_omp_for): Move gimple_bind_vars
7780 and BLOCK_VARS of gimple_bind_block to new_stmt rather
7784 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
7785 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
7786 assume_aligned or alloc_align attributes.
7787 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
7788 arguments. Handle also assume_aligned and alloc_align attributes.
7789 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
7790 calls to functions with assume_aligned or alloc_align attributes.
7791 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
7793 2014-02-08 Terry Guo <terry.guo@arm.com>
7795 * doc/invoke.texi: Document ARM -march=armv7e-m.
7797 2014-02-08 Jakub Jelinek <jakub@redhat.com>
7799 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
7800 flag on __cilkrts_rethrow builtin.
7803 * ipa-cp.c (determine_versionability): Fail at -O0
7804 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
7805 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
7808 2014-02-04 Jakub Jelinek <jakub@redhat.com>
7811 * tree-inline.c (copy_forbidden): Fail for
7812 __attribute__((optimize (0))) functions.
7814 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
7816 * varpool.c: Include pointer-set.h.
7817 (varpool_remove_unreferenced_decls): Variables in other partitions
7818 will not be output; be however careful to not lose information
7821 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
7823 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
7824 lookup in the vtable constructor.
7826 2014-02-07 Jeff Law <law@redhat.com>
7829 * config/m68k/m68k.md (ashldi_extsi): Turn into a
7830 define_insn_and_split.
7832 * ipa-inline.c (inline_small_functions): Fix typos.
7834 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7836 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
7837 (s390_can_use_return_insn): Declare.
7838 * config/s390/s390.h (EPILOGUE_USES): Define.
7839 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
7841 (s390_chunkify_start): Handle return JUMP_LABELs.
7842 (s390_early_mach): Emit a main_pool instruction on the entry edge.
7843 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
7844 (s390_can_use_return_insn): New functions.
7845 (s390_fix_long_loop_prediction): Handle conditional returns.
7846 (TARGET_SET_UP_BY_PROLOGUE): Define.
7847 * config/s390/s390.md (ANY_RETURN): New code iterator.
7848 (*creturn, *csimple_return, return, simple_return): New patterns.
7850 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7852 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
7853 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
7854 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
7855 REG_CFA_RESTORE list when deciding not to restore a register.
7857 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7859 * config/s390/s390.c: Include tree-pass.h and context.h.
7860 (s390_early_mach): New function, split out from...
7861 (s390_emit_prologue): ...here.
7862 (pass_data_s390_early_mach): New pass structure.
7863 (pass_s390_early_mach): New class.
7864 (s390_option_override): Create and register early_mach pass.
7865 Move to end of file.
7867 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7869 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
7870 to match for the exit block.
7872 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7874 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
7875 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
7876 Reject misaligned operands.
7878 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7880 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
7882 2014-02-07 Richard Biener <rguenther@suse.de>
7885 * gimple-low.c (lower_builtin_posix_memalign): New function.
7886 (lower_stmt): Call it to lower posix_memalign in a way
7887 to make alignment info accessible.
7889 2014-02-07 Jakub Jelinek <jakub@redhat.com>
7892 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
7893 __builtin_setjmp_receiver.
7895 2014-02-07 Richard Biener <rguenther@suse.de>
7898 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
7899 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
7900 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
7901 Handle BUILT_IN_POSIX_MEMALIGN.
7902 (find_func_clobbers): Likewise.
7903 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
7904 (call_may_clobber_ref_p_1): Likewise.
7906 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7909 * ipa-devirt.c (record_target_from_binfo): Remove overactive
7912 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7915 * lto-cgraph.c (lto_output_node): Use
7916 symtab_get_symbol_partitioning_class.
7917 (lto_output_varpool_node): likewise.
7918 (symtab_get_symbol_partitioning_class): Move here from
7920 * cgraph.h (symbol_partitioning_class): Likewise.
7921 (symtab_get_symbol_partitioning_class): Declare.
7923 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7925 * ggc.h (ggc_internal_cleared_alloc): New macro.
7926 * vec.h (vec_safe_copy): Handle memory stats.
7927 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
7928 * target-globals.c (save_target_globals): Likewise.
7930 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7933 * expr.c (emit_move_resolve_push): Export; be bit more selective
7934 on when to clear alias set.
7935 * expr.h (emit_move_resolve_push): Declare.
7936 * function.h (struct function): Add tail_call_marked.
7937 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
7938 * config/i386/i386-protos.h (ix86_expand_push): Remove.
7939 * config/i386/i386.md (TImode move expander): De not call
7941 (FP push expanders): Preserve memory attributes.
7942 * config/i386/sse.md (push<mode>1): Remove.
7943 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
7944 (ix86_expand_push): Remove.
7945 * config/i386/mmx.md (push<mode>1): Remove.
7947 2014-02-06 Jakub Jelinek <jakub@redhat.com>
7949 PR rtl-optimization/60030
7950 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
7951 lopart with paradoxical subreg before shifting it up by hprec.
7953 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7955 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
7956 Remove extra newline at end of file.
7957 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
7958 (arm_issue_rate): Handle cortexa57.
7959 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
7960 (cortex-a57.cortex-a53): Likewise.
7962 2014-02-06 Jakub Jelinek <jakub@redhat.com>
7965 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
7966 don't record in REG_FRAME_RELATED_EXPR registers not set in that
7968 (arm_expand_prologue): Adjust all callers.
7969 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
7970 info, registers also at the lowest numbered registers side. Use
7971 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
7975 * var-tracking.c (adjust_mems): Before adding a SET to
7976 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
7978 2014-02-06 Alan Modra <amodra@gmail.com>
7981 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
7982 change SDmode to DDmode when lra_in_progress.
7984 2014-02-06 Jakub Jelinek <jakub@redhat.com>
7987 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
7988 free_data_ref on the dr first, and before goto again also set dr
7989 to the next dr. For simd_lane_access, free old datarefs[i] before
7990 overwriting it. For get_vectype_for_scalar_type failure, don't
7991 free_data_ref if simd_lane_access.
7993 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
7996 * tree.h (opts_for_fn): New inline function.
7997 (opt_for_fn): Define.
7998 * config/i386/i386.c (ix86_function_regparm): Use
7999 opt_for_fn (decl, optimize) instead of optimize.
8001 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
8003 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
8004 for SYMBOL_REF in large memory model.
8006 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8008 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
8010 (cortex-a57): Likewise.
8011 (cortex-a57.cortex-a53): Likewise.
8013 2014-02-06 Yury Gribov <y.gribov@samsung.com>
8014 Kugan Vivekanandarajah <kuganv@linaro.org>
8016 * config/arm/arm.c (arm_vector_alignment_reachable): Check
8018 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
8020 2014-02-06 Richard Biener <rguenther@suse.de>
8022 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
8023 set_loop_copy and initialize_original_copy_tables.
8025 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
8027 * config/aarch64/aarch64-simd.md
8028 (aarch64_ashr_simddi): Change QI to SI.
8030 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
8031 Jakub Jelinek <jakub@redhat.com>
8034 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
8037 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8039 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
8040 CODE_FOR_altivec_vpku[hw]um to
8041 CODE_FOR_altivec_vpku[hw]um_direct.
8042 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
8043 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
8044 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
8045 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
8047 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8049 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
8050 generation for -maltivec=be.
8051 (altivec_vsumsws): Simplify redundant test.
8053 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8055 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
8056 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
8057 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
8058 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
8059 gen_altivec_vpkuwum.
8060 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
8062 (altivec_vpks<VI_char>ss): Likewise.
8063 (altivec_vpks<VI_char>us): Likewise.
8064 (altivec_vpku<VI_char>us): Likewise.
8065 (altivec_vpku<VI_char>um): Likewise.
8066 (altivec_vpku<VI_char>um_direct): New (copy of
8067 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
8069 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
8070 target is little endian and -maltivec=be is not specified.
8071 (*altivec_vupkhs<VU_char>_direct): New (copy of
8072 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
8073 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
8074 target is little endian and -maltivec=be is not specified.
8075 (*altivec_vupkls<VU_char>_direct): New (copy of
8076 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
8077 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
8078 little endian and -maltivec=be is not specified.
8079 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
8080 little endian and -maltivec=be is not specified.
8082 2014-02-05 Richard Henderson <rth@redhat.com>
8085 * combine-stack-adj.c: Revert r206943.
8086 * sched-int.h (struct deps_desc): Add last_args_size.
8087 * sched-deps.c (init_deps): Initialize it.
8088 (sched_analyze_insn): Add OUTPUT dependencies between insns that
8089 contain REG_ARGS_SIZE notes.
8091 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
8093 * lto-cgraph.c (asm_nodes_output): Make global.
8094 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
8095 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
8096 (driver_handle_option): Handle OPT_fwpa.
8098 2014-02-05 Jakub Jelinek <jakub@redhat.com>
8101 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
8102 a comment typo and formatting issue. If odr_hash hasn't been
8103 created, return vNULL and set *completep to false.
8106 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
8107 bb with no successors.
8109 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
8112 * doc/invoke.texi (-march): Clarify documentation for ARM.
8116 2014-02-05 Richard Biener <rguenther@suse.de>
8118 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
8119 when not vectorizing because of too many alias checks.
8120 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
8121 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
8123 2014-02-05 Nick Clifton <nickc@redhat.com>
8125 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
8126 accept extended registers in any mode when compiling for the MN10300.
8128 2014-02-05 Yury Gribov <y.gribov@samsung.com>
8130 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
8131 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
8132 sanitization attributes.
8133 (can_inline_edge_p): Likewise.
8134 (sanitize_attrs_match_for_inline_p): New function.
8136 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
8138 * ipa-prop.c (detect_type_change): Shor circuit testing of
8139 type changes on THIS pointer.
8141 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
8144 * config/pa/pa.c (legitimize_tls_address): Return original address
8145 if not passed a SYMBOL_REF rtx.
8146 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
8148 (pa_emit_move_sequence): Simplify TLS source operands.
8149 (pa_legitimate_constant_p): Reject all TLS constants.
8150 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
8151 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
8153 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
8155 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
8156 groups when we know they are controlled by LTO.
8157 * varasm.c (default_binds_local_p_1): If object is in other partition,
8158 it will be resolved locally.
8160 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
8162 * config/host-linux.c (linux_gt_pch_use_address): Don't
8163 use SSIZE_MAX because it is not always defined.
8165 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
8168 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
8169 threshold for pseudo splitting.
8170 (update_ebb_live_info): Process call argument hard registers and
8171 hard registers from insn definition too.
8172 (max_small_class_regs_num): New constant.
8173 (inherit_in_ebb): Update live hard regs through EBBs. Update
8174 reloads_num only for small register classes. Don't split for
8177 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
8180 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
8183 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
8185 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
8188 2014-02-04 Marek Polacek <polacek@redhat.com>
8190 * gdbinit.in (pel): Define.
8192 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
8194 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
8197 2014-02-04 Richard Biener <rguenther@suse.de>
8200 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
8201 in function context local.
8202 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
8203 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
8204 similar to LTO_imported_decl_ref.
8206 2014-02-04 Jakub Jelinek <jakub@redhat.com>
8208 PR tree-optimization/60002
8209 * cgraphclones.c (build_function_decl_skip_args): Clear
8212 PR tree-optimization/60023
8213 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
8214 false to gsi_replace.
8215 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
8216 has been in some EH region and vec_stmt could throw, add
8217 vec_stmt into the same EH region.
8218 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
8219 has no lhs, ignore it.
8220 * internal-fn.c (expand_MASK_LOAD): Likewise.
8223 * tree-inline.c (copy_forbidden): Fail for
8224 __attribute__((optimize (0))) functions.
8227 * omp-low.c (simd_clone_struct_copy): If from->inbranch
8228 is set, copy one less argument.
8229 (expand_simd_clones): Don't subtract clone_info->inbranch
8230 from simd_clone_struct_alloc argument.
8232 PR rtl-optimization/57915
8233 * recog.c (simplify_while_replacing): If all unary/binary/relational
8234 operation arguments are constant, attempt to simplify those.
8237 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
8238 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
8240 2014-02-04 Richard Biener <rguenther@suse.de>
8242 PR tree-optimization/60012
8243 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
8244 TBAA disambiguation to all DDRs.
8246 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8249 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
8250 (LINK_SPEC): Use it for -shared, -shared-libgcc.
8252 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
8255 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
8257 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
8259 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
8260 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
8262 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
8265 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
8266 to figure out targets of polymorphic calls with known decl.
8267 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
8268 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
8269 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
8270 (get_polymorphic_call_info): ... here.
8271 (get_polymorphic_call_info_from_invariant): New function.
8273 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
8275 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
8276 lookup via vtable pointer; check for type consistency
8277 and turn inconsitent facts into UNREACHABLE.
8278 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
8279 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
8280 type inconsistent querries; return UNREACHABLE instead.
8282 2014-02-03 Richard Henderson <rth@twiddle.net>
8285 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
8286 already processed this node.
8287 (normalize_one_pred_1): Pass along mark_set.
8288 (normalize_one_pred): Create and destroy a pointer_set_t.
8289 (normalize_one_pred_chain): Likewise.
8291 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
8293 PR gcov-profile/58602
8294 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
8296 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
8299 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
8300 -fno-devirtualize; try to devirtualize by the knowledge of
8301 virtual table pointer given by aggregate propagation.
8302 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
8303 (ipa_print_node_jump_functions): Dump also offset that
8304 is relevant for polymorphic calls.
8305 (determine_known_aggregate_parts): Add arg_type parameter; use it
8306 instead of determining the type from pointer type.
8307 (ipa_compute_jump_functions_for_edge): Update call of
8308 determine_known_aggregate_parts.
8309 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
8310 (gimple_get_virt_method_for_binfo): ... here; simplify using
8311 vtable_pointer_value_to_vtable.
8312 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
8313 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
8314 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
8315 (vtable_pointer_value_to_vtable): Break out from ...; handle also
8317 (vtable_pointer_value_to_binfo): ... here.
8318 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
8320 2014-02-03 Teresa Johnson <tejohnson@google.com>
8322 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
8323 redef of outer loop index variable.
8325 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
8329 * doc/extend.texi (Function Attributes): Typo.
8331 2014-02-03 Cong Hou <congh@google.com>
8333 PR tree-optimization/60000
8334 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
8335 if the vectorized statement is a store. A store statement can only
8336 appear at the end of pattern statements.
8338 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
8340 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
8341 (ix86_option_override_internal): Default long double to 64-bit for
8342 32-bit Bionic and to 128-bit for 64-bit Bionic.
8344 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
8345 TARGET_LONG_DOUBLE_128 is true.
8346 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
8348 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
8349 (mlong-double-64): Negate -mlong-double-128.
8350 (mlong-double-128): New option.
8352 * config/i386/i386-c.c (ix86_target_macros): Define
8353 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
8355 * doc/invoke.texi: Document -mlong-double-128.
8357 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
8359 PR rtl-optimization/60024
8360 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
8362 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
8364 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
8366 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
8368 PR rtl-optimization/57662
8369 * sel-sched.c (code_motion_path_driver): Do not mark already not
8370 existing blocks in the visiting bitmap.
8372 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
8374 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
8375 on the insn being emitted.
8377 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
8378 Will Deacon <will.deacon@arm.com>
8380 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
8382 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8384 * config/arm/arm-tables.opt: Regenerate.
8386 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8388 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
8389 for vector types other than V16QImode.
8390 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
8391 define_expand, and call altivec_expand_vec_perm_le when producing
8392 code with little endian element order.
8393 (*altivec_vperm_<mode>_internal): New insn having previous
8394 behavior of altivec_vperm_<mode>.
8395 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
8396 altivec_expand_vec_perm_le when producing code with little endian
8398 (*altivec_vperm_<mode>_uns_internal): New insn having previous
8399 behavior of altivec_vperm_<mode>_uns.
8401 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8403 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
8404 (altivec_vsumsws): Add handling for -maltivec=be with a little
8406 (altivec_vsumsws_direct): New.
8407 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
8408 gen_altivec_vsumsws.
8410 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
8412 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
8413 vtable_pointer_value_to_binfo): New functions.
8414 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
8415 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
8417 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
8419 * config/nios2/nios2.md (load_got_register): Initialize GOT
8420 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
8421 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
8423 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
8425 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
8426 preserverd by passthrough, do not propagate the type.
8428 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
8430 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
8431 (mips_atomic_assign_expand_fenv): New function.
8432 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
8434 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
8436 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
8437 (__builtin_mips_set_fcsr): Likewise.
8438 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
8439 MIPS_USI_FTYPE_VOID.
8440 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
8441 (mips16_expand_set_fcsr): Likewise.
8442 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
8443 (mips16_set_fcsr_stub): Likewise.
8444 (mips16_get_fcsr_one_only_stub): New class.
8445 (mips16_set_fcsr_one_only_stub): Likewise.
8446 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
8447 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
8448 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
8449 (hard_float): New availability predicate.
8450 (mips_builtins): Add get_fcsr and set_fcsr.
8451 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
8452 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
8453 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
8454 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
8455 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
8458 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
8460 * config/mips/mips.c (mips_one_only_stub): New class.
8461 (mips_need_mips16_rdhwr_p): Replace with...
8462 (mips16_rdhwr_stub): ...this new variable.
8463 (mips16_stub_call_address): New function.
8464 (mips16_rdhwr_one_only_stub): New class.
8465 (mips_expand_thread_pointer): Use mips16_stub_call_address.
8466 (mips_output_mips16_rdhwr): Delete.
8467 (mips_finish_stub): New function.
8468 (mips_code_end): Use it to handle rdhwr stubs.
8470 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
8473 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
8474 when calculating size of integer atomic types.
8476 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
8478 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
8480 2014-02-01 Jakub Jelinek <jakub@redhat.com>
8482 PR tree-optimization/60003
8483 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
8484 * profile.c (branch_prob): Use gimple_call_builtin_p
8485 to check for BUILT_IN_SETJMP_RECEIVER.
8486 * tree-inline.c (copy_bb): Call notice_special_calls.
8488 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
8491 * lra-constraints.c (process_alt_operands): Update reload_sum only
8494 2014-01-31 Richard Henderson <rth@redhat.com>
8497 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
8498 until after else_eh is processed.
8500 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
8502 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
8503 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
8504 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
8505 in smmintrin.h, remove them.
8506 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
8507 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
8508 * config/i386/i386.md (ROUND_SAE): Fix value.
8509 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
8510 (const48_operand): New.
8511 * config/i386/subst.md (round), (round_expand): Use
8512 const_4_or_8_to_11_operand.
8513 (round_saeonly), (round_saeonly_expand): Use const48_operand.
8515 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
8517 * config/i386/constraints.md (Yk): Swap meaning with k.
8518 * config/i386/i386.md (movhi_internal): Change Yk to k.
8519 (movqi_internal): Ditto.
8520 (*k<logic><mode>): Ditto.
8523 (kandn<mode>): Ditto.
8524 (*<code>hi_1): Ditto.
8525 (*<code>qi_1): Ditto.
8526 (kxnor<mode>): Ditto.
8527 (kortestzhi): Ditto.
8528 (kortestchi): Ditto.
8530 (*one_cmplhi2_1): Ditto.
8531 (*one_cmplqi2_1): Ditto.
8532 * config/i386/sse.md (): Change k to Yk.
8533 (avx512f_load<mode>_mask): Ditto.
8534 (avx512f_blendm<mode>): Ditto.
8535 (avx512f_store<mode>_mask): Ditto.
8536 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
8537 (avx512f_storedqu<mode>_mask): Ditto.
8538 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
8540 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
8541 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
8542 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
8543 (avx512f_maskcmp<mode>3): Ditto.
8544 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
8545 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
8546 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
8547 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
8548 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
8549 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
8550 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
8551 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
8552 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
8553 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
8554 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
8555 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
8556 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
8557 (vec_extract_lo_<mode>_maskm): Ditto.
8558 (vec_extract_hi_<mode>_maskm): Ditto.
8559 (avx512f_vternlog<mode>_mask): Ditto.
8560 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
8561 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
8562 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
8563 (avx512f_<code>v8div16qi2_mask): Ditto.
8564 (avx512f_<code>v8div16qi2_mask_store): Ditto.
8565 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
8566 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
8567 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
8568 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
8569 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
8570 (*avx512pf_gatherpf<mode>df_mask): Ditto.
8571 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
8572 (*avx512pf_scatterpf<mode>df_mask): Ditto.
8573 (avx512cd_maskb_vec_dupv8di): Ditto.
8574 (avx512cd_maskw_vec_dupv16si): Ditto.
8575 (avx512f_vpermi2var<mode>3_maskz): Ditto.
8576 (avx512f_vpermi2var<mode>3_mask): Ditto.
8577 (avx512f_vpermi2var<mode>3_mask): Ditto.
8578 (avx512f_vpermt2var<mode>3_maskz): Ditto.
8579 (*avx512f_gathersi<mode>): Ditto.
8580 (*avx512f_gathersi<mode>_2): Ditto.
8581 (*avx512f_gatherdi<mode>): Ditto.
8582 (*avx512f_gatherdi<mode>_2): Ditto.
8583 (*avx512f_scattersi<mode>): Ditto.
8584 (*avx512f_scatterdi<mode>): Ditto.
8585 (avx512f_compress<mode>_mask): Ditto.
8586 (avx512f_compressstore<mode>_mask): Ditto.
8587 (avx512f_expand<mode>_mask): Ditto.
8588 * config/i386/subst.md (mask): Change k to Yk.
8589 (mask_scalar_merge): Ditto.
8592 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
8594 * doc/extend.texi (Vector Extensions): Document ?: in C++.
8596 2014-01-31 Richard Biener <rguenther@suse.de>
8599 * builtins.c (fold_builtin_memory_op): Make sure to not
8600 use a floating-point mode or a boolean or enumeral type for
8603 2014-01-30 DJ Delorie <dj@redhat.com>
8605 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
8606 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
8607 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
8608 whenever main() has an epilogue.
8610 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8612 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
8613 unused variable "field".
8614 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
8615 (vsx_mergeh_<mode>): Likewise.
8616 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
8617 (altivec_vmrghh): Likewise.
8618 (altivec_vmrghw): Likewise.
8619 (altivec_vmrglb): Likewise.
8620 (altivec_vmrglh): Likewise.
8621 (altivec_vmrglw): Likewise.
8622 (altivec_vspltb): Add missing uses.
8623 (altivec_vsplth): Likewise.
8624 (altivec_vspltw): Likewise.
8625 (altivec_vspltsf): Likewise.
8627 2014-01-30 Jakub Jelinek <jakub@redhat.com>
8630 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
8631 frame related instructions.
8633 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
8635 PR rtl-optimization/59959
8636 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
8637 any reload of register whose subreg is invalid.
8639 2014-01-30 Jakub Jelinek <jakub@redhat.com>
8641 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
8642 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
8643 Add missing return type - void.
8645 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8647 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
8648 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
8649 remove element index adjustment for endian (now handled in vsx.md
8651 (altivec_expand_vec_perm_const): Use
8652 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
8653 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
8654 (vsx_xxspltw_<mode>): Adjust element index for little endian.
8655 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
8656 define_expand and a new define_insn *altivec_vspltb_internal;
8657 adjust for -maltivec=be on a little endian target.
8658 (altivec_vspltb_direct): New.
8659 (altivec_vsplth): Divide into a define_expand and a new
8660 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
8661 little endian target.
8662 (altivec_vsplth_direct): New.
8663 (altivec_vspltw): Divide into a define_expand and a new
8664 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
8665 little endian target.
8666 (altivec_vspltw_direct): New.
8667 (altivec_vspltsf): Divide into a define_expand and a new
8668 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
8669 a little endian target.
8671 2014-01-30 Richard Biener <rguenther@suse.de>
8673 PR tree-optimization/59993
8674 * tree-ssa-forwprop.c (associate_pointerplus): Check we
8675 can propagate form the earlier stmt and avoid the transform
8676 when the intermediate result is needed.
8678 2014-01-30 Alangi Derick <alangiderick@gmail.com>
8680 * README.Portability: Fix typo.
8682 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
8684 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
8685 comparison_operator with ordered_comparison_operator.
8687 2014-01-30 Nick Clifton <nickc@redhat.com>
8689 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
8690 Rename to mn10300_store_multiple_regs.
8691 * config/mn10300/mn10300.c: Likewise.
8692 * config/mn10300/mn10300.md (store_movm): Fix typo: call
8693 store_multiple_regs.
8694 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
8695 Call mn10300_store_multiple_regs.
8697 2014-01-30 Nick Clifton <nickc@redhat.com>
8698 DJ Delorie <dj@redhat.com>
8700 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
8701 %fp 2 to keep registers after it properly word-aligned.
8702 (rl78_alloc_physical_registers_umul): Handle the case where both
8703 input operands are the same.
8705 2014-01-30 Richard Biener <rguenther@suse.de>
8707 PR tree-optimization/59903
8708 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
8711 2014-01-30 Jason Merrill <jason@redhat.com>
8714 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
8717 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
8719 2014-01-30 Richard Biener <rguenther@suse.de>
8721 PR tree-optimization/59951
8722 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
8724 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
8727 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
8728 SFmode to DFmode case.
8730 2014-01-29 DJ Delorie <dj@redhat.com>
8732 * config/msp430/msp430.opt (-minrt): New.
8733 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
8735 (ENDFILE_SPEC): Likewise.
8737 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
8739 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
8740 (estimate_function_body_sizes): Use it.
8742 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
8745 * dwarf2out.c (is_cxx_auto): New.
8746 (is_base_type): Use it.
8747 (gen_type_die_with_usage): Likewise.
8749 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8751 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
8752 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
8753 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
8754 -maltivec=be with LE targets.
8755 (vsx_mergeh_<mode>): Likewise.
8756 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
8757 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
8758 (altivec_vmrghb): Replace with define_expand and new
8759 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
8760 (altivec_vmrghb_direct): New define_insn.
8761 (altivec_vmrghh): Replace with define_expand and new
8762 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
8763 (altivec_vmrghh_direct): New define_insn.
8764 (altivec_vmrghw): Replace with define_expand and new
8765 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
8766 (altivec_vmrghw_direct): New define_insn.
8767 (*altivec_vmrghsf): Adjust for endianness.
8768 (altivec_vmrglb): Replace with define_expand and new
8769 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
8770 (altivec_vmrglb_direct): New define_insn.
8771 (altivec_vmrglh): Replace with define_expand and new
8772 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
8773 (altivec_vmrglh_direct): New define_insn.
8774 (altivec_vmrglw): Replace with define_expand and new
8775 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
8776 (altivec_vmrglw_direct): New define_insn.
8777 (*altivec_vmrglsf): Adjust for endianness.
8778 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
8779 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
8780 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
8781 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
8782 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
8783 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
8784 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
8785 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
8787 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
8789 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
8790 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
8793 2014-01-29 Richard Biener <rguenther@suse.de>
8795 PR tree-optimization/58742
8796 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
8797 associate_pointerplus_align.
8798 (associate_pointerplus_diff): New function.
8799 (associate_pointerplus): Likewise. Call associate_pointerplus_align
8800 and associate_pointerplus_diff.
8802 2014-01-29 Richard Biener <rguenther@suse.de>
8804 * lto-streamer.h (LTO_major_version): Bump to 3.
8805 (LTO_minor_version): Reset to 0.
8807 2014-01-29 Renlin Li <Renlin.Li@arm.com>
8809 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
8810 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
8811 (arm_file_start): Generate correct asm header for armv7ve.
8812 * config/arm/bpabi.h: Add multilib support for armv7ve.
8813 * config/arm/driver-arm.c: Change the architectures of cortex-a7
8814 and cortex-a15 to armv7ve.
8815 * config/arm/t-aprofile: Add multilib support for armv7ve.
8816 * doc/invoke.texi: Document -march=armv7ve.
8818 2014-01-29 Richard Biener <rguenther@suse.de>
8820 PR tree-optimization/58742
8821 * tree-ssa-forwprop.c (associate_plusminus): Return true
8822 if we changed sth, defer EH cleanup to ...
8823 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
8824 (simplify_mult): New function.
8826 2014-01-29 Jakub Jelinek <jakub@redhat.com>
8829 PR tree-optimization/59920
8830 * tree.c (build_common_builtin_nodes): Remove
8831 __builtin_setjmp_dispatcher initialization.
8832 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
8833 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
8834 instead of gsi_after_labels + manually skipping debug stmts.
8835 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
8836 ignore bbs with IFN_ABNORMAL_DISPATCHER.
8837 * tree-inline.c (copy_edges_for_bb): Remove
8838 can_make_abnormal_goto argument, instead add abnormal_goto_dest
8839 argument. Ignore computed_goto_p stmts. Don't call
8840 make_abnormal_goto_edges. If a call might need abnormal edges
8841 for non-local gotos, see if it already has an edge to
8842 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
8843 with true argument, don't do anything then, otherwise add
8844 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
8845 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
8847 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
8848 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
8849 (lower_stmt): Don't set data->calls_builtin_setjmp.
8850 (lower_builtin_setjmp): Adjust comment.
8851 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
8852 * tree-cfg.c (found_computed_goto): Remove.
8853 (factor_computed_gotos): Remove.
8854 (make_goto_expr_edges): Return bool, true for computed gotos.
8855 Don't call make_abnormal_goto_edges.
8856 (build_gimple_cfg): Don't set found_computed_goto, don't call
8857 factor_computed_gotos.
8858 (computed_goto_p): No longer static.
8859 (make_blocks): Don't set found_computed_goto.
8860 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
8861 (make_edges): If make_goto_expr_edges returns true, push bb
8862 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
8863 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
8864 vector. Record mapping between bbs and OpenMP regions if there
8865 are any, adjust make_gimple_omp_edges caller. Call
8866 handle_abnormal_edges.
8867 (make_abnormal_goto_edges): Remove.
8868 * tree-cfg.h (make_abnormal_goto_edges): Remove.
8869 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
8870 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
8871 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
8872 * internal-fn.def (ABNORMAL_DISPATCHER): New.
8873 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
8874 filling *region also set *region_idx to (*region)->entry->index.
8877 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
8878 For REGs set ORIGINAL_REGNO.
8880 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
8882 * doc/md.texi: Mention that a target shouldn't implement
8883 vec_widen_(s|u)mul_even/odd pair if it is less efficient
8886 2014-01-29 Jakub Jelinek <jakub@redhat.com>
8888 PR tree-optimization/59594
8889 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
8890 a copy of the datarefs vector rather than the vector itself.
8892 2014-01-28 Jason Merrill <jason@redhat.com>
8895 * dwarf2out.c (auto_die): New static.
8896 (gen_type_die_with_usage): Handle C++1y 'auto'.
8897 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
8900 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
8903 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
8904 (SPEC_X32): Likewise.
8905 (SPEC_64): Likewise.
8906 * config/i386/i386.c (ix86_option_override_internal): Turn off
8907 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
8909 (x86_file_start): Output .code16gcc for TARGET_16BIT.
8910 * config/i386/i386.h (TARGET_16BIT): New macro.
8911 (TARGET_16BIT_P): Likewise.
8912 * config/i386/i386.opt: Add m16.
8913 * doc/invoke.texi: Document -m16.
8915 2014-01-28 Jakub Jelinek <jakub@redhat.com>
8917 PR preprocessor/59935
8918 * input.c (location_get_source_line): Bail out on when line number
8919 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
8921 2014-01-28 Richard Biener <rguenther@suse.de>
8923 PR tree-optimization/58742
8924 * tree-ssa-forwprop.c (associate_plusminus): Handle
8925 pointer subtraction of the form (T)(P + A) - (T)P.
8927 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8929 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
8932 2014-01-28 Richard Biener <rguenther@suse.de>
8935 2014-01-28 Richard Biener <rguenther@suse.de>
8937 PR rtl-optimization/45364
8938 PR rtl-optimization/59890
8939 * var-tracking.c (local_get_addr_clear_given_value): Handle
8940 already cleared slot.
8941 (val_reset): Handle not allocated local_get_addr_cache.
8942 (vt_find_locations): Use post-order on the inverted CFG.
8944 2014-01-28 Richard Biener <rguenther@suse.de>
8946 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
8948 2014-01-28 Richard Biener <rguenther@suse.de>
8950 PR rtl-optimization/45364
8951 PR rtl-optimization/59890
8952 * var-tracking.c (local_get_addr_clear_given_value): Handle
8953 already cleared slot.
8954 (val_reset): Handle not allocated local_get_addr_cache.
8955 (vt_find_locations): Use post-order on the inverted CFG.
8957 2014-01-28 Alan Modra <amodra@gmail.com>
8959 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
8960 * configure.ac <recursive call for build != host>: Define
8961 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
8962 and LD_FOR_BUILD too.
8963 * configure: Regenerate.
8965 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
8967 * config/i386/i386.c (get_builtin_code_for_version): Separate
8968 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
8969 Broadwell from Haswell.
8971 2014-01-27 Steve Ellcey <sellcey@mips.com>
8973 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
8974 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
8975 * config/mips/mips.c (mips_option_override): Change setting
8977 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
8978 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
8979 Change from Mask to Var.
8981 2014-01-27 Jeff Law <law@redhat.com>
8983 * ipa-inline.c (inline_small_functions): Fix typo.
8985 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
8987 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
8988 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
8989 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
8990 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
8991 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
8992 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
8993 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
8994 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
8995 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
8996 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
8997 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
8998 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
8999 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
9000 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
9001 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
9002 (_mm512_storeu_epi64): Ditto.
9003 (_mm512_cmpge_epi32_mask): Ditto.
9004 (_mm512_cmpge_epu32_mask): Ditto.
9005 (_mm512_cmpge_epi64_mask): Ditto.
9006 (_mm512_cmpge_epu64_mask): Ditto.
9007 (_mm512_cmple_epi32_mask): Ditto.
9008 (_mm512_cmple_epu32_mask): Ditto.
9009 (_mm512_cmple_epi64_mask): Ditto.
9010 (_mm512_cmple_epu64_mask): Ditto.
9011 (_mm512_cmplt_epi32_mask): Ditto.
9012 (_mm512_cmplt_epu32_mask): Ditto.
9013 (_mm512_cmplt_epi64_mask): Ditto.
9014 (_mm512_cmplt_epu64_mask): Ditto.
9015 (_mm512_cmpneq_epi32_mask): Ditto.
9016 (_mm512_cmpneq_epu32_mask): Ditto.
9017 (_mm512_cmpneq_epi64_mask): Ditto.
9018 (_mm512_cmpneq_epu64_mask): Ditto.
9019 (_mm512_expand_pd): Ditto.
9020 (_mm512_expand_ps): Ditto.
9021 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
9022 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
9023 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
9024 * config/i386/i386.c (ix86_builtins): Add
9025 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
9026 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
9027 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
9028 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
9029 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
9030 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
9031 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
9032 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
9033 IX86_BUILTIN_PMOVUSQW512_MEM.
9034 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
9035 __builtin_ia32_pmovsqd512mem_mask,
9036 __builtin_ia32_pmovqd512mem_mask,
9037 __builtin_ia32_pmovusqw512mem_mask,
9038 __builtin_ia32_pmovsqw512mem_mask,
9039 __builtin_ia32_pmovqw512mem_mask,
9040 __builtin_ia32_pmovusdw512mem_mask,
9041 __builtin_ia32_pmovsdw512mem_mask,
9042 __builtin_ia32_pmovdw512mem_mask,
9043 __builtin_ia32_pmovqb512mem_mask,
9044 __builtin_ia32_pmovusqb512mem_mask,
9045 __builtin_ia32_pmovsqb512mem_mask,
9046 __builtin_ia32_pmovusdb512mem_mask,
9047 __builtin_ia32_pmovsdb512mem_mask,
9048 __builtin_ia32_pmovdb512mem_mask.
9049 (bdesc_args): Add __builtin_ia32_expanddf512,
9050 __builtin_ia32_expandsf512.
9051 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
9052 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
9053 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
9054 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
9055 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
9056 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
9057 (avx512f_<code>v8div16qi2_mask_store): This.
9058 (avx512f_expand<mode>): New.
9060 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
9062 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
9064 (_mm512_mask_prefetch_i64gather_pd): Ditto.
9065 (_mm512_prefetch_i32scatter_pd): Ditto.
9066 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
9067 (_mm512_prefetch_i64scatter_pd): Ditto.
9068 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
9069 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
9070 (_mm512_mask_prefetch_i64gather_ps): Ditto.
9071 (_mm512_prefetch_i32scatter_ps): Ditto.
9072 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
9073 (_mm512_prefetch_i64scatter_ps): Ditto.
9074 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
9075 * config/i386/i386-builtin-types.def: Define
9076 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
9077 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
9078 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
9079 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
9080 IX86_BUILTIN_SCATTERPFQPD.
9081 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
9082 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
9083 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
9084 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
9085 __builtin_ia32_scatterpfqps.
9086 (ix86_expand_builtin): Expand new built-ins.
9087 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
9088 fix memory access data type.
9089 (*avx512pf_gatherpf<mode>_mask): Ditto.
9090 (*avx512pf_gatherpf<mode>): Ditto.
9091 (avx512pf_scatterpf<mode>): Ditto.
9092 (*avx512pf_scatterpf<mode>_mask): Ditto.
9093 (*avx512pf_scatterpf<mode>): Ditto.
9094 (GATHER_SCATTER_SF_MEM_MODE): New.
9095 (avx512pf_gatherpf<mode>df): Ditto.
9096 (*avx512pf_gatherpf<mode>df_mask): Ditto.
9097 (*avx512pf_scatterpf<mode>df): Ditto.
9099 2014-01-27 Jakub Jelinek <jakub@redhat.com>
9102 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
9103 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
9106 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
9108 * common/config/arm/arm-common.c
9109 (arm_rewrite_mcpu): Handle multiple names.
9111 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
9113 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
9115 * gimple-builder.h (create_gimple_tmp): Delete.
9117 2014-01-27 Christian Bruel <christian.bruel@st.com>
9119 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
9122 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
9124 * config/pa/pa.md (call): Generate indirect long calls to non-local
9125 functions when outputing 32-bit code.
9126 (call_value): Likewise except for special call to buggy powf function.
9128 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
9129 portable runtime and PIC indirect calls.
9130 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
9131 and PIC call sequences. Use ldo instead of blr to set return register
9132 in PIC call sequence.
9134 2014-01-25 Walter Lee <walt@tilera.com>
9136 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
9137 avoid clobbering a live register.
9139 2014-01-25 Walter Lee <walt@tilera.com>
9141 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
9142 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
9143 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
9144 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
9146 2014-01-25 Walter Lee <walt@tilera.com>
9148 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
9149 arguments on even registers.
9150 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
9152 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
9153 (BIGGEST_ALIGNMENT): Ditto.
9154 (BIGGEST_FIELD_ALIGNMENT): Ditto.
9156 2014-01-25 Walter Lee <walt@tilera.com>
9158 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
9159 insns before bundling.
9160 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
9162 2014-01-25 Walter Lee <walt@tilera.com>
9164 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
9165 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
9166 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
9168 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
9170 * config/mips/constraints.md (kl): Delete.
9171 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
9172 define expands, using...
9173 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
9174 instructions for MIPS16.
9175 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
9176 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
9178 2014-01-25 Walter Lee <walt@tilera.com>
9180 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
9184 2014-01-25 Walter Lee <walt@tilera.com>
9186 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
9187 (TARGET_EXPAND_TO_RTL_HOOK): Define.
9189 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
9191 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
9194 2014-01-25 Jakub Jelinek <jakub@redhat.com>
9196 * print-rtl.c (in_call_function_usage): New var.
9197 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
9198 EXPR_LIST mode as mode and not as reg note name.
9201 * cfgloopmanip.c (copy_loop_info): If
9202 loop->warned_aggressive_loop_optimizations, make sure
9203 the flag is set in target loop too.
9205 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
9207 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
9209 * builtins.def: Likewise.
9210 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
9211 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
9212 * ira.c (ira_setup_eliminable_regset): Likewise.
9213 * omp-low.c (gate_expand_omp): Likewise.
9214 (execute_lower_omp): Likewise.
9215 (diagnose_sb_0): Likewise.
9216 (gate_diagnose_omp_blocks): Likewise.
9217 (simd_clone_clauses_extract): Likewise.
9220 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9222 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
9223 correction for little endian...
9224 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
9227 2014-01-24 Jeff Law <law@redhat.com>
9229 PR tree-optimization/59919
9230 * tree-vrp.c (find_assert_locations_1): Do not register asserts
9231 for non-returning calls.
9233 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
9235 * common/config/aarch64/aarch64-common.c
9236 (aarch64_rewrite_mcpu): Handle multiple names.
9237 * config/aarch64/aarch64.h
9238 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
9240 2014-01-24 Dodji Seketeli <dodji@redhat.com>
9242 * input.c (add_file_to_cache_tab): Handle the case where fopen
9245 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
9248 * config/i386/i386.md (pushsf splitter): Get stack adjustment
9249 from push operand if code of push isn't PRE_DEC.
9251 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
9254 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
9255 -mquad-memory-atomic. Update -mquad-memory documentation to say
9256 it is only used for non-atomic loads/stores.
9258 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
9259 -mquad-memory or -mquad-memory-atomic switches.
9261 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
9262 -mquad-memory-atomic to ISA 2.07 support.
9264 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
9265 to separate support of normal quad word memory operations (ldq, stq)
9266 from the atomic quad word memory operations.
9268 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
9269 support to separate non-atomic quad word operations from atomic
9270 quad word operations. Disable non-atomic quad word operations in
9271 little endian mode so that we don't have to swap words after the
9272 load and before the store.
9273 (quad_load_store_p): Add comment about atomic quad word support.
9274 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
9275 options printed with -mdebug=reg.
9277 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
9278 -mquad-memory-atomic as the test for whether we have quad word
9279 atomic instructions.
9280 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
9281 or -mp8-vector are used, allow byte/half-word atomic operations.
9283 * config/rs6000/sync.md (load_lockedti): Insure that the address
9284 is a proper indexed or indirect address for the lqarx instruction.
9285 On little endian systems, swap the hi/lo registers after the lqarx
9287 (load_lockedpti): Use indexed_or_indirect_operand predicate to
9288 insure the address is valid for the lqarx instruction.
9289 (store_conditionalti): Insure that the address is a proper indexed
9290 or indirect address for the stqcrx. instruction. On little endian
9291 systems, swap the hi/lo registers before doing the stqcrx.
9293 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
9294 insure the address is valid for the stqcrx. instruction.
9296 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
9297 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
9298 type of quad memory support is available.
9300 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
9303 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
9304 there is a danger of looping.
9306 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
9308 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
9309 force flag_ira_loop_pressure if set via command line.
9311 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
9313 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
9314 (ashr_simd): New builtin handling DI mode.
9315 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
9316 (aarch64_sshr_simddi): New match pattern.
9317 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
9318 (vshrd_n_s64): Likewise.
9319 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
9321 2014-01-23 Nick Clifton <nickc@redhat.com>
9323 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
9324 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
9325 favour of mcu specific scripts.
9326 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
9329 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
9330 Alex Velenko <Alex.Velenko@arm.com>
9332 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
9333 (vaddv_s16): Likewise.
9334 (vaddv_s32): Likewise.
9335 (vaddv_u8): Likewise.
9336 (vaddv_u16): Likewise.
9337 (vaddv_u32): Likewise.
9338 (vaddvq_s8): Likewise.
9339 (vaddvq_s16): Likewise.
9340 (vaddvq_s32): Likewise.
9341 (vaddvq_s64): Likewise.
9342 (vaddvq_u8): Likewise.
9343 (vaddvq_u16): Likewise.
9344 (vaddvq_u32): Likewise.
9345 (vaddvq_u64): Likewise.
9346 (vaddv_f32): Likewise.
9347 (vaddvq_f32): Likewise.
9348 (vaddvq_f64): Likewise.
9349 (vmaxv_f32): Likewise.
9350 (vmaxv_s8): Likewise.
9351 (vmaxv_s16): Likewise.
9352 (vmaxv_s32): Likewise.
9353 (vmaxv_u8): Likewise.
9354 (vmaxv_u16): Likewise.
9355 (vmaxv_u32): Likewise.
9356 (vmaxvq_f32): Likewise.
9357 (vmaxvq_f64): Likewise.
9358 (vmaxvq_s8): Likewise.
9359 (vmaxvq_s16): Likewise.
9360 (vmaxvq_s32): Likewise.
9361 (vmaxvq_u8): Likewise.
9362 (vmaxvq_u16): Likewise.
9363 (vmaxvq_u32): Likewise.
9364 (vmaxnmv_f32): Likewise.
9365 (vmaxnmvq_f32): Likewise.
9366 (vmaxnmvq_f64): Likewise.
9367 (vminv_f32): Likewise.
9368 (vminv_s8): Likewise.
9369 (vminv_s16): Likewise.
9370 (vminv_s32): Likewise.
9371 (vminv_u8): Likewise.
9372 (vminv_u16): Likewise.
9373 (vminv_u32): Likewise.
9374 (vminvq_f32): Likewise.
9375 (vminvq_f64): Likewise.
9376 (vminvq_s8): Likewise.
9377 (vminvq_s16): Likewise.
9378 (vminvq_s32): Likewise.
9379 (vminvq_u8): Likewise.
9380 (vminvq_u16): Likewise.
9381 (vminvq_u32): Likewise.
9382 (vminnmv_f32): Likewise.
9383 (vminnmvq_f32): Likewise.
9384 (vminnmvq_f64): Likewise.
9386 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
9388 * config/aarch64/aarch64-simd.md
9389 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
9390 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
9391 (*aarch64_mul3_elt<mode>): Likewise.
9392 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
9393 (*aarch64_mul3_elt_to_64v2df): Likewise.
9394 (*aarch64_mla_elt<mode>): Likewise.
9395 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
9396 (*aarch64_mls_elt<mode>): Likewise.
9397 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
9398 (*aarch64_fma4_elt<mode>): Likewise.
9399 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
9400 (*aarch64_fma4_elt_to_64v2df): Likewise.
9401 (*aarch64_fnma4_elt<mode>): Likewise.
9402 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
9403 (*aarch64_fnma4_elt_to_64v2df): Likewise.
9404 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
9405 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
9406 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
9407 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
9408 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
9409 (aarch64_sqdmull_lane<mode>_internal): Likewise.
9410 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
9412 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
9414 * config/aarch64/aarch64-simd.md
9415 (aarch64_be_checked_get_lane<mode>): New define_expand.
9416 * config/aarch64/aarch64-simd-builtins.def
9417 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
9418 New builtin definition.
9419 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
9420 Use new safe be builtin.
9422 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
9424 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
9426 (aarch64_be_st1<mode>): Likewise.
9427 (aarch_ld1<VALL:mode>): Define_expand modified.
9428 (aarch_st1<VALL:mode>): Likewise.
9429 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
9430 (UNSPEC_ST1): Likewise.
9432 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
9434 * config/microblaze/microblaze.md: Add trap insn and attribute
9436 2014-01-23 Dodji Seketeli <dodji@redhat.com>
9438 PR preprocessor/58580
9439 * input.h (location_get_source_line): Take an additional line_size
9441 (void diagnostics_file_cache_fini): Declare new function.
9442 * input.c (struct fcache): New type.
9443 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
9444 New static constants.
9445 (diagnostic_file_cache_init, total_lines_num)
9446 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
9447 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
9448 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
9449 (get_next_line, read_next_line, goto_next_line, read_line_num):
9450 New static function definitions.
9451 (diagnostic_file_cache_fini): New function.
9452 (location_get_source_line): Take an additional output line_len
9453 parameter. Re-write using lookup_or_add_file_to_cache_tab and
9455 * diagnostic.c (diagnostic_finish): Call
9456 diagnostic_file_cache_fini.
9457 (adjust_line): Take an additional input parameter for the length
9458 of the line, rather than calculating it with strlen.
9459 (diagnostic_show_locus): Adjust the use of
9460 location_get_source_line and adjust_line with respect to their new
9461 signature. While displaying a line now, do not stop at the first
9462 null byte. Rather, display the zero byte as a space and keep
9463 going until we reach the size of the line.
9464 * Makefile.in: Add vec.o to OBJS-libcommon
9466 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
9467 Ilya Tocar <ilya.tocar@intel.com>
9469 * config/i386/avx512fintrin.h (_mm512_kmov): New.
9470 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
9471 (__builtin_ia32_kmov16): Ditto.
9472 * config/i386/i386.md (UNSPEC_KMOV): New.
9475 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
9477 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
9478 (_mm512_storeu_si512): Ditto.
9480 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
9483 * rtl.h (get_referenced_operands): Declare.
9484 * recog.c (get_referenced_operands): New function.
9485 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
9486 operands have been referenced when recording LO_SUM references.
9488 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
9490 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
9492 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
9494 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
9495 Enable for generic and recent AMD targets.
9497 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
9499 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
9500 ARG_SIZE note when adjustment was eliminated.
9502 2014-01-22 Jeff Law <law@redhat.com>
9504 PR tree-optimization/59597
9505 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
9506 in file. Accept new argument REGISTERING and use it to modify
9507 dump output appropriately.
9508 (register_jump_thread): Corresponding changes.
9509 (mark_threaded_blocks): Reinstate code to cancel unprofitable
9510 thread paths involving joiner blocks. Add code to dump cancelled
9511 jump threading paths.
9513 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
9515 PR rtl-optimization/59477
9516 * lra-constraints.c (inherit_in_ebb): Process call for living hard
9517 regs. Update reloads_num and potential_reload_hard_regs for all insns.
9519 2014-01-22 Tom Tromey <tromey@redhat.com>
9521 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
9523 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
9525 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
9527 PR rtl-optimization/59896
9528 * lra-constraints.c (process_alt_operands): Check unused note for
9529 matched operands of insn with no output reloads.
9531 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
9533 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
9534 (mips_move_from_gpr_cost): Likewise.
9536 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
9538 PR rtl-optimization/59858
9539 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
9540 ira_class_hard_regs_num.
9541 (process_alt_operands): Increase reject for dying matched operand.
9543 2014-01-21 Jakub Jelinek <jakub@redhat.com>
9546 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
9547 smaller than size, perform several stores or loads and stores
9548 at dst + count - size to store or copy all of size bytes, rather
9549 than just last modesize bytes.
9551 2014-01-20 DJ Delorie <dj@redhat.com>
9553 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
9554 that CLOBBERs are REGs before propogating their values.
9556 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
9559 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
9560 (cgraph_inline_failed_type): New function.
9561 * cgraph.h (DEFCIFCODE): Add type.
9562 (cgraph_inline_failed_type_t): New enum.
9563 (cgraph_inline_failed_type): New prototype.
9564 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
9565 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
9566 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
9567 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
9568 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
9569 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
9570 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
9571 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
9572 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
9573 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
9574 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
9575 OPTIMIZATION_MISMATCH.
9576 * tree-inline.c (expand_call_inline): Emit errors during
9577 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
9579 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
9582 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
9583 mode attribute in insn output.
9585 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
9587 * output.h (output_constant): Delete.
9588 * varasm.c (output_constant): Make private.
9590 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
9592 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
9594 2014-01-20 Jakub Jelinek <jakub@redhat.com>
9597 * tree.h (fold_builtin_strcat): New prototype.
9598 * builtins.c (fold_builtin_strcat): No longer static. Add len
9599 argument, if non-NULL, don't call c_strlen. Optimize
9600 directly into __builtin_memcpy instead of __builtin_strcpy.
9601 (fold_builtin_2): Adjust fold_builtin_strcat caller.
9602 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
9604 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
9606 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
9607 for SImode_address_operand operands, having only a REG argument.
9609 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
9611 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
9612 loader name using mbig-endian.
9613 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
9615 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
9617 * doc/invoke.texi (-march): Clarify documentation for AArch64.
9621 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
9623 * config/aarch64/aarch64-protos.h
9624 (aarch64_cannot_change_mode_class_ptr): Declare.
9625 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
9626 aarch64_cannot_change_mode_class_ptr): New.
9627 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
9628 backend hook aarch64_cannot_change_mode_class.
9630 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
9632 * common/config/aarch64/aarch64-common.c
9633 (aarch64_handle_option): Don't handle any option order logic here.
9634 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
9635 selected_cpu, warn on architecture version mismatch.
9636 (aarch64_override_options): Fix parsing order for option strings.
9638 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
9639 Iain Sandoe <iain@codesourcery.com>
9642 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
9643 warning. Amend comment to reflect current functionality.
9645 2014-01-20 Richard Biener <rguenther@suse.de>
9648 * builtins.c (fold_builtin_strcat): Remove case better handled
9649 by tree-ssa-strlen.c.
9651 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
9653 * config/aarch64/aarch64.opt
9654 (mcpu, march, mtune): Make case-insensitive.
9656 2014-01-20 Jakub Jelinek <jakub@redhat.com>
9659 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
9660 if operands[1] is a REG or ZERO_EXTEND of a REG.
9662 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
9664 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
9666 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
9668 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
9669 long non-pic millicode calls.
9671 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
9673 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
9675 2014-01-19 Kito Cheng <kito@0xlab.org>
9677 * builtins.c (expand_movstr): Check movstr expand done or fail.
9679 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
9680 H.J. Lu <hongjiu.lu@intel.com>
9683 * config/i386/i386.md (*lea<mode>): Zero-extend return register
9684 to DImode for zero-extended addresses.
9686 2014-01-19 Jakub Jelinek <jakub@redhat.com>
9688 PR rtl-optimization/57763
9689 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
9690 on the new indirect jump_insn and increment LABEL_NUSES (label).
9692 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
9696 * config.gcc (x86_archs): New variable.
9697 (x86_64_archs): Likewise.
9698 (x86_cpus): Likewise.
9699 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
9700 --with-arch/--with-cpu= options.
9701 Support --with-arch=/--with-cpu={nehalem,westmere,
9702 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
9704 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
9706 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
9707 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
9709 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
9711 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
9713 2014-01-18 Jakub Jelinek <jakub@redhat.com>
9716 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
9717 clear cpp_get_options (parse_in)->warn_unused_macros for
9718 ix86_target_macros_internal with cpp_define.
9720 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
9722 * jump.c (delete_related_insns): Keep (use (insn))s.
9723 * reorg.c (redundant_insn): Check for barriers too.
9725 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
9727 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
9729 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
9731 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
9732 call to $$dyncall when TARGET_LONG_CALLS is true.
9734 2014-01-17 Jeff Law <law@redhat.com>
9736 * ree.c (combine_set_extension): Temporarily disable test for
9737 changing number of hard registers.
9739 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
9742 * ipa-inline-analysis.c (inline_free_summary):
9743 Do not free summary of aliases.
9745 2014-01-17 Jakub Jelinek <jakub@redhat.com>
9748 * gimplify.c (gimplify_expr): Use create_tmp_var
9749 instead of create_tmp_var_raw. If cond doesn't have
9750 integral type, don't add the IFN_ANNOTATE builtin at all.
9752 2014-01-17 Martin Jambor <mjambor@suse.cz>
9755 * ipa-cp.c (prev_edge_clone): New variable.
9756 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
9757 Also resize prev_edge_clone vector.
9758 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
9759 (ipcp_edge_removal_hook): New function.
9760 (ipcp_driver): Register ipcp_edge_removal_hook.
9762 2014-01-17 Andrew Pinski <apinski@cavium.com>
9763 Steve Ellcey <sellcey@mips.com>
9766 * config/mips/mips.c (mips_print_operand): Check operand mode instead
9769 2014-01-17 Jeff Law <law@redhat.com>
9772 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
9773 so that pass_ccp runs first.
9775 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
9777 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
9778 (ix86_adjust_cost): Use !TARGET_XXX.
9779 (do_reorder_for_imul): Likewise.
9780 (swap_top_of_ready_list): Likewise.
9781 (ix86_sched_reorder): Likewise.
9783 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
9785 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
9786 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
9787 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
9788 (intel_memset): New. Duplicate slm_memset.
9789 (intel_cost): New. Duplicate slm_cost.
9790 (m_INTEL): New macro.
9791 (processor_target_table): Add "intel".
9792 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
9793 with PROCESSOR_INTEL for "intel".
9794 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
9795 PROCESSOR_SILVERMONT.
9796 (ix86_issue_rate): Likewise.
9797 (ix86_adjust_cost): Likewise.
9798 (ia32_multipass_dfa_lookahead): Likewise.
9799 (swap_top_of_ready_list): Likewise.
9800 (ix86_sched_reorder): Likewise.
9801 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
9802 instead of TARGET_OPT_AGU.
9803 * config/i386/i386.h (TARGET_INTEL): New.
9804 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
9805 (processor_type): Add PROCESSOR_INTEL.
9806 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
9807 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
9809 2014-01-17 Marek Polacek <polacek@redhat.com>
9812 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
9815 2014-01-17 Richard Biener <rguenther@suse.de>
9817 PR tree-optimization/46590
9818 * opts.c (default_options_table): Add entries for
9819 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
9820 all enabled at -O1 but not for -Og.
9821 * common.opt (fbranch-count-reg): Remove Init(1).
9822 (fmove-loop-invariants): Likewise.
9823 (ftree-pta): Likewise.
9825 2014-01-17 Jakub Jelinek <jakub@redhat.com>
9827 * config/i386/i386.c (ix86_data_alignment): For compatibility with
9828 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
9829 decls to at least the GCC 4.8 used alignments.
9832 * tree-nested.c (convert_nonlocal_reference_stmt,
9833 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
9834 of GIMPLE_BIND stmts, adjust associated decls.
9836 2014-01-17 Richard Biener <rguenther@suse.de>
9838 PR tree-optimization/46590
9839 * vec.h (vec<>::bseach): New member function implementing
9840 binary search according to C89 bsearch.
9841 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
9842 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
9843 bitmap pointer again. Make accesses_in_loop a flat array.
9844 (mem_ref_obstack): New global.
9845 (outermost_indep_loop): Adjust for mem_ref->stored changes.
9846 (mark_ref_stored): Likewise.
9847 (ref_indep_loop_p_2): Likewise.
9848 (set_ref_stored_in_loop): New helper function.
9849 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
9850 (memref_free): Adjust.
9851 (record_mem_ref_loc): Simplify.
9852 (gather_mem_refs_stmt): Adjust.
9853 (sort_locs_in_loop_postorder_cmp): New function.
9854 (analyze_memory_references): Sort accesses_in_loop after
9855 loop postorder number.
9856 (find_ref_loc_in_loop_cmp): New function.
9857 (for_all_locs_in_loop): Find relevant cluster of locs in
9858 accesses_in_loop and iterate without recursion.
9859 (execute_sm): Avoid uninit warning.
9860 (struct ref_always_accessed): Simplify.
9861 (ref_always_accessed::operator ()): Likewise.
9862 (ref_always_accessed_p): Likewise.
9863 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
9864 loop postorder numbers here.
9865 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
9868 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
9871 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
9872 on decls for which assemble_alias has been called.
9874 2014-01-17 Nick Clifton <nickc@redhat.com>
9876 * config/msp430/msp430.opt: (mcpu): New option.
9877 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
9878 (msp430_option_override): Parse target_cpu. If the MCU name
9879 matches a generic string, clear target_mcu.
9880 (msp430_attr): Allow numeric interrupt values up to 63.
9881 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
9882 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
9884 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
9886 * config/msp430/msp430.md (popm): Use %J rather than %I.
9887 (addsi3): Use msp430_nonimmediate_operand for operand 2.
9888 (addhi_cy_i): Use immediate_operand for operand 2.
9889 * doc/invoke.texi: Document -mcpu option.
9891 2014-01-17 Richard Biener <rguenther@suse.de>
9893 PR rtl-optimization/38518
9894 * df.h (df_analyze_loop): Declare.
9895 * df-core.c: Include cfgloop.h.
9896 (df_analyze_1): Split out main part of df_analyze.
9897 (df_analyze): Adjust.
9898 (loop_inverted_post_order_compute): New function.
9899 (loop_post_order_compute): Likewise.
9900 (df_analyze_loop): New function avoiding whole-function
9902 * loop-invariant.c (find_defs): Use df_analyze_loop.
9903 (find_invariants): Adjust.
9904 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
9906 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
9908 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
9909 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
9911 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
9913 * ipa-ref.c (ipa_remove_stmt_references): Fix references
9914 traversal when removing references.
9916 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
9919 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
9921 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
9924 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
9925 pushing a reload for an autoinc when we had previously reloaded an
9926 inner part of the address.
9928 2014-01-16 Jakub Jelinek <jakub@redhat.com>
9930 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
9932 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
9933 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
9934 when not giving up or versioning for alias only because of
9936 (vect_analyze_data_ref_dependences): Set to true.
9937 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
9939 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
9940 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
9944 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
9947 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
9948 operand 0 predicate for gathers, use a new pseudo as subtarget.
9950 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
9953 * lra-constraints.c (process_alt_operands): Add printing debug info.
9954 Check absence of input/output reloads for matched operands too.
9956 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
9958 PR rtl-optimization/59835
9959 * ira.c (ira_init_register_move_cost): Increase cost for
9962 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
9964 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
9966 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
9969 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
9970 non-register objects. Use gen_(high/low)part more consistently.
9973 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
9976 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
9977 endian support, remove tests for WORDS_BIG_ENDIAN.
9978 (p8_mfvsrd_3_<mode>): Likewise.
9979 (reload_gpr_from_vsx<mode>): Likewise.
9980 (reload_gpr_from_vsxsf): Likewise.
9981 (p8_mfvsrd_4_disf): Likewise.
9983 2014-01-16 Richard Biener <rguenther@suse.de>
9985 PR rtl-optimization/46590
9986 * lcm.c (compute_antinout_edge): Use postorder iteration.
9987 (compute_laterin): Use inverted postorder iteration.
9989 2014-01-16 Nick Clifton <nickc@redhat.com>
9992 * varasm.c (output_constant): Return the number of bytes actually
9994 (output_constructor_array_range): Update the field size with the
9995 number of bytes emitted by output_constant.
9996 (output_constructor_regular_field): Likewise. Also do not
9997 complain if the total number of bytes emitted is now greater
9998 than the expected fieldpos.
9999 * output.h (output_constant): Update prototype and descriptive comment.
10001 2014-01-16 Marek Polacek <polacek@redhat.com>
10003 PR middle-end/59827
10004 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
10005 it is error_mark_node.
10007 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
10009 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
10010 VALID_AVX256_REG_OR_OI_MODE.
10012 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
10014 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
10015 current procedure should be profiled.
10017 2014-01-15 Andrew Pinski <apinski@cavium.com>
10019 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
10020 of moving from/to the STACK_REG register class.
10022 2014-01-15 Richard Henderson <rth@redhat.com>
10025 * reginfo.c (global_regs_decl): Globalize.
10026 * rtl.h (global_regs_decl): Declare.
10027 * ira.c (do_reload): Diagnose frame_pointer_needed and it
10028 reserved via global_regs.
10030 2014-01-15 Teresa Johnson <tejohnson@google.com>
10032 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
10034 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
10036 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
10037 and vmulosh rather than call gen_vec_widen_smult_*.
10038 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
10039 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
10040 (vec_widen_smult_even_v16qi): Likewise.
10041 (vec_widen_umult_even_v8hi): Likewise.
10042 (vec_widen_smult_even_v8hi): Likewise.
10043 (vec_widen_umult_odd_v16qi): Likewise.
10044 (vec_widen_smult_odd_v16qi): Likewise.
10045 (vec_widen_umult_odd_v8hi): Likewise.
10046 (vec_widen_smult_odd_v8hi): Likewise.
10047 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
10048 vmuloub rather than call gen_vec_widen_umult_*.
10049 (vec_widen_umult_lo_v16qi): Likewise.
10050 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
10051 vmulosb rather than call gen_vec_widen_smult_*.
10052 (vec_widen_smult_lo_v16qi): Likewise.
10053 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
10054 rather than call gen_vec_widen_umult_*.
10055 (vec_widen_umult_lo_v8hi): Likewise.
10056 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
10057 rather than call gen_vec_widen_smult_*.
10058 (vec_widen_smult_lo_v8hi): Likewise.
10060 2014-01-15 Jeff Law <law@redhat.com>
10062 PR tree-optimization/59747
10063 * ree.c (find_and_remove_re): Properly handle case where a second
10064 eliminated extension requires widening a copy created for elimination
10065 of a prior extension.
10066 (combine_set_extension): Ensure that the number of hard regs needed
10067 for a destination register does not change when we widen it.
10069 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
10071 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
10072 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
10073 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
10074 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
10075 (avr-*-rtems*): Likewise.
10076 (bfin*-rtems*): Likewise.
10077 (moxie-*-rtems*): Likewise.
10078 (h8300-*-rtems*): Likewise.
10079 (i[34567]86-*-rtems*): Likewise.
10080 (lm32-*-rtems*): Likewise.
10081 (m32r-*-rtems*): Likewise.
10082 (m68k-*-rtems*): Likewise.
10083 (microblaze*-*-rtems*): Likewise.
10084 (mips*-*-rtems*): Likewise.
10085 (powerpc-*-rtems*): Likewise.
10086 (sh-*-rtems*): Likewise.
10087 (sparc-*-rtems*): Likewise.
10088 (sparc64-*-rtems*): Likewise.
10089 (v850-*-rtems*): Likewise.
10090 (m32c-*-rtems*): Likewise.
10092 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
10094 PR rtl-optimization/59511
10095 * ira.c (ira_init_register_move_cost): Use memory costs for some
10096 cases of register move cost calculations.
10097 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
10098 instead of BB frequency.
10099 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
10100 * lra-assigns.c (find_hard_regno_for): Ditto.
10102 2014-01-15 Richard Biener <rguenther@suse.de>
10104 PR tree-optimization/59822
10105 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
10106 (vectorizable_load): Use it to hoist defs of uses of invariant
10107 loads out of the loop.
10109 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
10110 Kugan Vivekanandarajah <kuganv@linaro.org>
10113 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
10116 2014-01-15 Richard Biener <rguenther@suse.de>
10118 PR rtl-optimization/59802
10119 * lcm.c (compute_available): Use inverted postorder to seed
10120 the initial worklist.
10122 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10125 * config/s390/s390.c (s390_preferred_reload_class): Don't return
10126 ADDR_REGS for invalid symrefs in non-PIC code.
10128 2014-01-15 Jakub Jelinek <jakub@redhat.com>
10131 * builtins.c (determine_block_size): Initialize *probable_max_size
10132 even if len_rtx is CONST_INT.
10134 2014-01-14 Andrew Pinski <apinski@cavium.com>
10136 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
10137 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
10138 (cortexa53_tunings): Likewise.
10139 (aarch64_sched_issue_rate): New function.
10140 (TARGET_SCHED_ISSUE_RATE): Define.
10142 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
10144 * ira-costs.c (find_costs_and_classes): Add missed
10145 ira_init_register_move_cost_if_necessary.
10147 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
10150 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
10152 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
10155 * config/i386/i386.c (type_natural_mode): Add a bool parameter
10156 to indicate if type is used for function return value. Warn ABI
10157 change if the vector mode isn't available for function return value.
10158 (ix86_function_arg_advance): Pass false to type_natural_mode.
10159 (ix86_function_arg): Likewise.
10160 (ix86_gimplify_va_arg): Likewise.
10161 (function_arg_32): Don't warn ABI change.
10162 (ix86_function_value): Pass true to type_natural_mode.
10163 (ix86_return_in_memory): Likewise.
10164 (ix86_struct_value_rtx): Removed.
10165 (TARGET_STRUCT_VALUE_RTX): Likewise.
10167 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
10169 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
10170 converting a conditional jump into a conditional return.
10172 2014-01-14 Richard Biener <rguenther@suse.de>
10174 PR tree-optimization/58921
10175 PR tree-optimization/59006
10176 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
10177 hoisting invariant stmts.
10178 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
10179 invariant loads on the preheader edge if possible.
10181 2014-01-14 Joey Ye <joey.ye@arm.com>
10183 * doc/plugin.texi (Building GCC plugins): Update to C++.
10185 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
10187 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
10188 (_mm_rcp28_round_ss): Ditto.
10189 (_mm_rsqrt28_round_sd): Ditto.
10190 (_mm_rsqrt28_round_ss): Ditto.
10191 (_mm_rcp28_sd): Ditto.
10192 (_mm_rcp28_ss): Ditto.
10193 (_mm_rsqrt28_sd): Ditto.
10194 (_mm_rsqrt28_ss): Ditto.
10195 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
10196 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
10197 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
10198 (IX86_BUILTIN_RCP28SD): Ditto.
10199 (IX86_BUILTIN_RCP28SS): Ditto.
10200 (IX86_BUILTIN_RSQRT28SD): Ditto.
10201 (IX86_BUILTIN_RSQRT28SS): Ditto.
10202 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
10203 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
10204 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
10205 (ix86_expand_special_args_builtin): Expand new FTYPE.
10206 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
10207 (srcp14<mode>): Make insn unary.
10208 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
10209 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
10210 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
10211 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
10212 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
10213 Fix rounding: make it SAE only.
10214 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
10216 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
10218 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
10219 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
10220 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
10221 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
10222 (round_saeonly_mask_scalar_operand4): Ditto.
10223 (round_saeonly_mask_scalar_op3): Ditto.
10224 (round_saeonly_mask_scalar_op4): Ditto.
10226 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10228 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
10229 Implement -maltivec=be for vec_insert and vec_extract.
10231 2014-01-10 DJ Delorie <dj@redhat.com>
10233 * config/msp430/msp430.md (call_internal): Don't allow memory
10234 references with SP as the base register.
10235 (call_value_internal): Likewise.
10236 * config/msp430/constraints.md (Yc): New. For memory references
10237 that don't use SP as a base register.
10239 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
10240 "an integer without a # prefix"
10241 * config/msp430/msp430.md (epilogue_helper): Use it.
10243 2014-01-13 Jakub Jelinek <jakub@redhat.com>
10246 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
10247 AVX512F gather builtins.
10248 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
10249 on gather decls with INTEGER_TYPE masktype.
10250 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
10251 directly into the builtin rather than hoisting it before loop.
10253 PR tree-optimization/59387
10254 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
10255 (scev_const_prop): If folded_casts and type has undefined overflow,
10256 use force_gimple_operand instead of force_gimple_operand_gsi and
10257 for each added stmt if it is assign with
10258 arith_code_with_undefined_signed_overflow, call
10259 rewrite_to_defined_overflow.
10260 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
10261 gimple-fold.h instead.
10262 (arith_code_with_undefined_signed_overflow,
10263 rewrite_to_defined_overflow): Moved to ...
10264 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
10265 rewrite_to_defined_overflow): ... here. No longer static.
10266 Include gimplify-me.h.
10267 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
10268 rewrite_to_defined_overflow): New prototypes.
10270 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10272 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
10274 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
10276 * builtins.c (get_object_alignment_2): Minor tweak.
10277 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
10279 2014-01-13 Christian Bruel <christian.bruel@st.com>
10281 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
10282 optimized non constant lengths.
10284 2014-01-13 Jakub Jelinek <jakub@redhat.com>
10287 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
10288 load as __atomic_load_N if possible.
10290 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
10292 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
10294 (rs6000_expand_builtin): Adjust call.
10296 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
10299 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
10300 * config/rs6000/rs6000.c: Include target-globals.h.
10301 (rs6000_set_current_function): Instead of doing target_reinit
10302 unconditionally, use save_target_globals_default_opts and
10303 restore_target_globals.
10305 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
10307 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
10308 (rs6000_expand_builtin): Handle mffs and mtfsf.
10309 (rs6000_init_builtins): Define mffs and mtfsf.
10310 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
10311 (rs6000_mffs): New pattern.
10312 (rs6000_mtfsf): New pattern.
10314 2014-01-11 Bin Cheng <bin.cheng@arm.com>
10316 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
10317 Start narrowing with START. Apply candidate-use pair
10318 and check overall cost in narrowing.
10319 (iv_ca_prune): Pass new argument.
10321 2014-01-10 Jeff Law <law@redhat.com>
10323 PR middle-end/59743
10324 * ree.c (combine_reaching_defs): Ensure the defining statement
10325 occurs before the extension when optimizing extensions with
10326 different source and destination hard registers.
10328 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
10331 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
10332 vtables into the type inheritance graph.
10334 2014-01-10 Jakub Jelinek <jakub@redhat.com>
10336 PR rtl-optimization/59754
10337 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
10338 modes in the REGNO != REGNO case.
10340 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10342 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
10344 2014-01-10 Jakub Jelinek <jakub@redhat.com>
10346 PR tree-optimization/59745
10347 * tree-predcom.c (tree_predictive_commoning_loop): Call
10348 free_affine_expand_cache if giving up because components is NULL.
10350 * target-globals.c (save_target_globals): Allocate < 4KB structs using
10351 GC in payload of target_globals struct instead of allocating them on
10352 the heap and the larger structs separately using GC.
10353 * target-globals.h (struct target_globals): Make regs, hard_regs,
10354 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
10355 of GTY((skip)) and change type to void *.
10356 (reset_target_globals): Cast loads from those fields to corresponding
10359 2014-01-10 Steve Ellcey <sellcey@mips.com>
10362 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
10363 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
10364 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
10366 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
10369 * aarch64-modes.def (CC_Zmode): New flags mode.
10370 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
10371 represents an equality.
10372 (aarch64_get_condition_code): Handle CC_Zmode.
10373 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
10375 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10377 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
10378 extraction in good case.
10380 2014-01-10 Richard Biener <rguenther@suse.de>
10382 PR tree-optimization/59374
10383 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
10384 checking after SLP discovery. Mark stmts not participating
10385 in any SLP instance properly.
10387 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10389 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
10390 when handling a SET rtx.
10392 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10394 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
10395 (cortex-a57): Likewise.
10396 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
10398 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10400 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
10401 non-iwmmxt builtins.
10403 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
10407 * ipa-devirt.c record_target_from_binfo): Take as argument
10408 stack of binfos and lookup matching one for virtual inheritance.
10409 (possible_polymorphic_call_targets_1): Update.
10411 2014-01-10 Huacai Chen <chenhc@lemote.com>
10413 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
10414 kernel strings for Loongson-2E/2F/3A.
10416 2014-01-10 Jakub Jelinek <jakub@redhat.com>
10418 PR middle-end/59670
10419 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
10420 is_gimple_call before calling gimple_call_internal_p.
10422 2014-01-09 Steve Ellcey <sellcey@mips.com>
10424 * Makefile.in (TREE_FLOW_H): Remove.
10425 (TREE_SSA_H): Add file names from tree-flow.h.
10426 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
10427 * tree.h: Remove tree-flow.h reference.
10428 * hash-table.h: Remove tree-flow.h reference.
10429 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
10430 reference with tree-ssa-loop.h.
10432 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10434 * doc/invoke.texi: Add -maltivec={be,le} options, and document
10435 default element-order behavior for -maltivec.
10436 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
10437 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
10438 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
10439 when targeting big endian, at least for now.
10440 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
10442 2014-01-09 Jakub Jelinek <jakub@redhat.com>
10444 PR middle-end/47735
10445 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
10446 var satisfies use_register_for_decl, just take into account type
10447 alignment, rather than decl alignment.
10449 PR tree-optimization/59622
10450 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
10451 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
10452 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
10453 Don't devirtualize for inplace at all. For targets.length () == 1,
10454 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
10456 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
10458 * config/i386/i386.md (cpu): Remove the unused btver1.
10460 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
10462 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
10464 2014-01-09 Jakub Jelinek <jakub@redhat.com>
10467 * tree-core.h (struct target_globals): New forward declaration.
10468 (struct tree_target_option): Add globals field.
10469 * tree.h (TREE_TARGET_GLOBALS): Define.
10470 (prepare_target_option_nodes_for_pch): New prototype.
10471 * target-globals.h (struct target_globals): Define even if
10472 !SWITCHABLE_TARGET.
10473 * tree.c (prepare_target_option_node_for_pch,
10474 prepare_target_option_nodes_for_pch): New functions.
10475 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
10476 * config/i386/i386.c: Include target-globals.h.
10477 (ix86_set_current_function): Instead of doing target_reinit
10478 unconditionally, use save_target_globals_default_opts and
10479 restore_target_globals.
10481 2014-01-09 Richard Biener <rguenther@suse.de>
10483 PR tree-optimization/59715
10484 * tree-cfg.h (split_critical_edges): Declare.
10485 * tree-cfg.c (split_critical_edges): Export.
10486 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
10488 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
10490 * cfgexpand.c (expand_stack_vars): Optionally disable
10491 asan stack protection.
10492 (expand_used_vars): Likewise.
10493 (partition_stack_vars): Likewise.
10494 * asan.c (asan_emit_stack_protection): Optionally disable
10495 after return stack usage.
10496 (instrument_derefs): Optionally disable memory access instrumentation.
10497 (instrument_builtin_call): Likewise.
10498 (instrument_strlen_call): Likewise.
10499 (asan_protect_global): Optionally disable global variables protection.
10500 * doc/invoke.texi: Added doc for new options.
10501 * params.def: Added new options.
10502 * params.h: Likewise.
10504 2014-01-09 Jakub Jelinek <jakub@redhat.com>
10506 PR rtl-optimization/59724
10507 * ifcvt.c (cond_exec_process_if_block): Don't call
10508 flow_find_head_matching_sequence with 0 longest_match.
10509 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
10510 non-active insns if !stop_after.
10511 (try_head_merge_bb): Revert 2014-01-07 changes.
10513 2014-01-08 Jeff Law <law@redhat.com>
10515 * ree.c (get_sub_rtx): New function, extracted from...
10516 (merge_def_and_ext): Here.
10517 (combine_reaching_defs): Use get_sub_rtx.
10519 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
10521 * cgraph.h (varpool_variable_node): Do not choke on null node.
10523 2014-01-08 Catherine Moore <clm@codesourcery.com>
10525 * config/mips/mips.md (simple_return): Attempt to use JRC
10527 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
10529 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
10531 PR rtl-optimization/59137
10532 * reorg.c (steal_delay_list_from_target): Call update_block for
10534 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
10536 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10538 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
10539 two duplicate entries.
10541 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
10544 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
10546 * config/mips/mips.c (mips_truncated_op_cost): New function.
10547 (mips_rtx_costs): Adjust test for BADDU.
10548 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
10550 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
10552 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
10553 (*baddu_si): ...this new pattern.
10555 2014-01-08 Jakub Jelinek <jakub@redhat.com>
10558 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
10560 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
10562 PR middle-end/57748
10563 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
10565 (expand_expr, expand_normal): Adjust.
10566 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
10567 inner_reference_p. Use inner_reference_p to expand inner references.
10568 (store_expr): Adjust.
10569 * cfgexpand.c (expand_call_stmt): Adjust.
10571 2014-01-08 Rong Xu <xur@google.com>
10573 * gcov-io.c (gcov_var): Move from gcov-io.h.
10574 (gcov_position): Ditto.
10575 (gcov_is_error): Ditto.
10576 (gcov_rewrite): Ditto.
10577 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
10578 only part to libgcc/libgcov.h.
10580 2014-01-08 Marek Polacek <polacek@redhat.com>
10582 PR middle-end/59669
10583 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
10585 2014-01-08 Marek Polacek <polacek@redhat.com>
10588 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
10590 2014-01-08 Jakub Jelinek <jakub@redhat.com>
10592 PR rtl-optimization/59649
10593 * stor-layout.c (get_mode_bounds): For BImode return
10594 0 and STORE_FLAG_VALUE.
10596 2014-01-08 Richard Biener <rguenther@suse.de>
10598 PR middle-end/59630
10599 * gimple.h (is_gimple_builtin_call): Remove.
10600 (gimple_builtin_call_types_compatible_p): New.
10601 (gimple_call_builtin_p): New overload.
10602 * gimple.c (is_gimple_builtin_call): Remove.
10603 (validate_call): Rename to ...
10604 (gimple_builtin_call_types_compatible_p): ... this and export. Also
10605 check return types.
10606 (validate_type): New static function.
10607 (gimple_call_builtin_p): New overload and adjust.
10608 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
10609 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
10610 (gimple_fold_stmt_to_constant_1): Likewise.
10611 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
10613 2014-01-08 Richard Biener <rguenther@suse.de>
10615 PR middle-end/59471
10616 * gimplify.c (gimplify_expr): Gimplify register-register type
10617 VIEW_CONVERT_EXPRs to separate stmts.
10619 2014-01-07 Jeff Law <law@redhat.com>
10621 PR middle-end/53623
10622 * ree.c (combine_set_extension): Handle case where source
10623 and destination registers in an extension insn are different.
10624 (combine_reaching_defs): Allow source and destination registers
10625 in extension to be different under limited circumstances.
10626 (add_removable_extension): Remove restriction that the
10627 source and destination registers in the extension are the same.
10628 (find_and_remove_re): Emit a copy from the extension's
10629 destination to its source after the defining insn if
10630 the source and destination registers are different.
10632 PR middle-end/59285
10633 * ifcvt.c (merge_if_block): If we are merging a block with more than
10634 one successor with a block with no successors, remove any BARRIER
10635 after the second block.
10637 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
10639 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
10641 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
10644 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
10645 for 14-bit register offsets when INT14_OK_STRICT is false.
10647 2014-01-07 Roland Stigge <stigge@antcom.de>
10648 Michael Meissner <meissner@linux.vnet.ibm.com>
10651 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
10652 Only check TFmode for SPE constants. Don't check TImode or TDmode.
10654 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
10656 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
10659 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
10661 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
10662 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
10663 rtx is const0_rtx or not.
10665 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
10668 * target-globals.c (save_target_globals): Remove this_fn_optab
10670 * toplev.c: Include optabs.h.
10671 (target_reinit): Temporarily restore the global options if another
10672 set of options are in force.
10674 2014-01-07 Jakub Jelinek <jakub@redhat.com>
10676 PR rtl-optimization/58668
10677 * cfgcleanup.c (flow_find_cross_jump): Don't count
10678 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
10679 to determine what is counted.
10680 (flow_find_head_matching_sequence): Use active_insn_p to determine
10682 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
10684 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
10685 determine what is counted.
10687 PR tree-optimization/59643
10688 * tree-predcom.c (split_data_refs_to_components): If one dr is
10689 read and one write, determine_offset fails and the write isn't
10690 in the bad component, just put the read into the bad component.
10692 2014-01-07 Mike Stump <mikestump@comcast.net>
10693 Jakub Jelinek <jakub@redhat.com>
10696 * tree-core.h (struct tree_optimization_option): Change optabs
10697 type from unsigned char * to void *.
10698 * optabs.c (init_tree_optimization_optabs): Adjust
10699 TREE_OPTIMIZATION_OPTABS initialization.
10701 2014-01-06 Jakub Jelinek <jakub@redhat.com>
10704 * config/i386/i386.h (struct machine_function): Add
10705 no_drap_save_restore field.
10706 * config/i386/i386.c (ix86_save_reg): Use
10707 !cfun->machine->no_drap_save_restore instead of
10708 crtl->stack_realign_needed.
10709 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
10710 this function clears frame_pointer_needed. Set
10711 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
10712 and DRAP reg is needed.
10714 2014-01-06 Marek Polacek <polacek@redhat.com>
10717 * doc/implement-c.texi: Mention that other integer types are
10718 permitted as bit-field types in strictly conforming mode.
10720 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
10722 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
10723 is newly allocated.
10725 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
10727 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
10729 2014-01-06 Martin Jambor <mjambor@suse.cz>
10732 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
10734 * ipa-prop.c (ipa_print_node_params): Fix indentation.
10736 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
10740 * var-tracking.c (add_stores): Preserve the value of the source even if
10741 we don't record the store.
10743 2014-01-06 Terry Guo <terry.guo@arm.com>
10745 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
10747 2014-01-05 Iain Sandoe <iain@codesourcery.com>
10750 * config/darwin.c (darwin_function_section): Adjust return values to
10751 correspond to optimisation changes made in r206070.
10753 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
10755 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
10756 from prefetch_block tune setting.
10757 (nocona_cost): Correct size of prefetch block to 64.
10759 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
10761 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
10762 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
10763 used to save the static chain register in the computation of the offset
10764 from which the FP registers need to be restored.
10766 2014-01-04 Jakub Jelinek <jakub@redhat.com>
10768 PR tree-optimization/59519
10769 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
10770 ICE if get_current_def (current_new_name) is already non-NULL, as long
10771 as it is a phi result of some other phi in *new_exit_bb that has
10774 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
10775 or vmovdqu* for misaligned_operand.
10776 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
10777 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
10778 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
10779 aligned_mem for AVX512F masked aligned load and store builtins and for
10780 non-temporal moves.
10782 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
10784 PR tree-optimization/59651
10785 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
10786 Address range for negative step should be added by TYPE_SIZE_UNIT.
10788 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
10790 * config/m68k/m68k.c (handle_move_double): Handle pushes with
10791 overlapping registers also for registers other than the stack pointer.
10793 2014-01-03 Marek Polacek <polacek@redhat.com>
10796 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
10799 2014-01-03 Jakub Jelinek <jakub@redhat.com>
10802 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
10805 * config/i386/i386.md (MODE_SIZE): New mode attribute.
10806 (push splitter): Use <P:MODE_SIZE> instead of
10807 GET_MODE_SIZE (<P:MODE>mode).
10808 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
10809 (mov -1, reg peephole2): Likewise.
10810 * config/i386/sse.md (*mov<mode>_internal,
10811 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
10812 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
10813 *<code><mode>3, *andnot<mode>3<mask_name>,
10814 <mask_codefor><code><mode>3<mask_name>): Likewise.
10815 * config/i386/subst.md (mask_mode512bit_condition,
10816 sd_mask_mode512bit_condition): Likewise.
10818 2014-01-02 Xinliang David Li <davidxl@google.com>
10820 PR tree-optimization/59303
10821 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
10822 (dump_predicates): Better output format.
10823 (pred_equal_p): New function.
10824 (is_neq_relop_p): Ditto.
10825 (is_neq_zero_form_p): Ditto.
10826 (pred_expr_equal_p): Ditto.
10827 (pred_neg_p): Ditto.
10828 (simplify_pred): Ditto.
10829 (simplify_preds_2): Ditto.
10830 (simplify_preds_3): Ditto.
10831 (simplify_preds_4): Ditto.
10832 (simplify_preds): Ditto.
10833 (push_pred): Ditto.
10834 (push_to_worklist): Ditto.
10835 (get_pred_info_from_cmp): Ditto.
10836 (is_degenerated_phi): Ditto.
10837 (normalize_one_pred_1): Ditto.
10838 (normalize_one_pred): Ditto.
10839 (normalize_one_pred_chain): Ditto.
10840 (normalize_preds): Ditto.
10841 (normalize_cond_1): Remove function.
10842 (normalize_cond): Ditto.
10843 (is_gcond_subset_of): Ditto.
10844 (is_subset_of_any): Ditto.
10845 (is_or_set_subset_of): Ditto.
10846 (is_and_set_subset_of): Ditto.
10847 (is_norm_cond_subset_of): Ditto.
10848 (pred_chain_length_cmp): Ditto.
10849 (convert_control_dep_chain_into_preds): Type change.
10850 (find_predicates): Ditto.
10851 (find_def_preds): Ditto.
10852 (destroy_predicates_vecs): Ditto.
10853 (find_matching_predicates_in_rest_chains): Ditto.
10854 (use_pred_not_overlap_with_undef_path_pred): Ditto.
10855 (is_pred_expr_subset): Ditto.
10856 (is_pred_chain_subset_of): Ditto.
10857 (is_included_in): Ditto.
10858 (is_superset_of): Ditto.
10860 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
10862 Update copyright years.
10864 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
10866 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
10867 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
10868 config/arc/arc.md, config/arc/arc.opt,
10869 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
10870 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
10871 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
10872 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
10873 config/linux-protos.h, config/linux.c, config/winnt-c.c,
10874 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
10875 vtable-verify.c, vtable-verify.h: Use the standard form for the
10878 2014-01-02 Tobias Burnus <burnus@net-b.de>
10880 * gcc.c (process_command): Update copyright notice dates.
10881 * gcov-dump.c: Ditto.
10883 * doc/cpp.texi: Bump @copying's copyright year.
10884 * doc/cppinternals.texi: Ditto.
10885 * doc/gcc.texi: Ditto.
10886 * doc/gccint.texi: Ditto.
10887 * doc/gcov.texi: Ditto.
10888 * doc/install.texi: Ditto.
10889 * doc/invoke.texi: Ditto.
10891 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
10893 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
10895 2014-01-01 Jakub Jelinek <jakub@redhat.com>
10897 * config/i386/sse.md (*mov<mode>_internal): Guard
10898 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
10900 PR rtl-optimization/59647
10901 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
10902 new_rtx into UNSIGNED_FLOAT rtxes.
10904 Copyright (C) 2014 Free Software Foundation, Inc.
10906 Copying and distribution of this file, with or without modification,
10907 are permitted in any medium without royalty provided the copyright
10908 notice and this notice are preserved.