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): Improve cost for
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): Use address
11 costs when costing loads and stores to memory.
13 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
14 Philip Tomsich <philipp.tomsich@theobroma-systems.com>
16 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costing
19 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
21 * config/aarch64/aarch64.c (aarch64_rtx_costs): Set default costs.
23 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
24 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
26 * config/aarch64/aarch64.c (aarch64_strip_shift_or_extend): Rename
28 (aarch64_strip_extend): ...this, don't strip shifts, check RTX is
30 (aarch64_rtx_mult_cost): New.
31 (aarch64_rtx_costs): Use it, refactor as appropriate.
33 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
34 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
36 * config/aarch64/aarch64.c (aarch64_build_constant): Conditionally
37 emit instructions, return number of instructions which would
39 (aarch64_add_constant): Update call to aarch64_build_constant.
40 (aarch64_output_mi_thunk): Likewise.
41 (aarch64_rtx_costs): Estimate cost of a CONST_INT, cost
44 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
46 * config/aarch64/aarch64.c (aarch64_rtx_costs_wrapper): New.
47 (TARGET_RTX_COSTS): Call it.
49 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
51 * config/aarch64/aarch64.c (cortexa57_addrcost_table): New.
52 (cortexa57_vector_cost): Likewise.
53 (cortexa57_tunings): Use them.
55 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
57 * config/aarch64/aarch64-protos.h (scale_addr_mode_cost): New.
58 (cpu_addrcost_table): Use it.
59 * config/aarch64/aarch64.c (generic_addrcost_table): Initialize it.
60 (aarch64_address_cost): Rewrite using aarch64_classify_address,
63 2014-05-16 Richard Biener <rguenther@suse.de>
65 * tree-ssa-sccvn.c: Include tree-cfg.h and domwalk.h.
66 (set_ssa_val_to): Handle unexpected sets to VN_TOP.
67 (visit_phi): Ignore edges marked as not executable.
68 (class cond_dom_walker): New.
69 (cond_dom_walker::before_dom_children): Value-number
70 control statements and mark successor edges as not
71 executable if possible.
72 (run_scc_vn): First walk all control statements in
73 dominator order, marking edges as not executable.
74 * tree-inline.c (copy_edges_for_bb): Be not confused
75 about random edge flags.
77 2014-05-16 Richard Biener <rguenther@suse.de>
79 * tree-ssa-sccvn.c (visit_use): Also constant-fold calls.
81 2014-05-15 Peter Bergner <bergner@vnet.ibm.com>
84 * config/rs6000/htmxlintrin.h (_HTM_TBEGIN_STARTED): New define.
85 (__TM_simple_begin): Use it.
86 (__TM_begin): Likewise.
88 2014-05-15 Martin Jambor <mjambor@suse.cz>
91 * ipa-prop.c (update_indirect_edges_after_inlining): Check
92 type_preserved flag when the indirect edge is polymorphic.
94 2014-05-15 Martin Jambor <mjambor@suse.cz>
96 PR tree-optimization/61090
97 * tree-sra.c (sra_modify_expr): Pass the current gsi to
100 2014-05-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
102 * config/arm/arm.c (arm_option_override): Use the SCHED_PRESSURE_MODEL
103 enum name for PARAM_SCHED_PRESSURE_ALGORITHM.
105 2014-05-15 Jakub Jelinek <jakub@redhat.com>
107 PR tree-optimization/61158
108 * fold-const.c (fold_binary_loc): If X is zero-extended and
109 shiftc >= prec, make sure zerobits is all ones instead of
110 invoking undefined behavior.
112 2014-05-15 Zhenqiang Chen <zhenqiang.chen@linaro.org>
114 * regcprop.h: New file.
115 * regcprop.c (skip_debug_insn_p): New decl.
116 (replace_oldest_value_reg): Check skip_debug_insn_p.
117 (copyprop_hardreg_forward_bb_without_debug_insn): New function.
118 * shrink-wrap.c: Include regcprop.h.
119 (prepare_shrink_wrap): Call
120 copyprop_hardreg_forward_bb_without_debug_insn.
122 2014-05-15 Zhenqiang Chen <zhenqiang.chen@linaro.org>
124 * shrink-wrap.h: Update comment.
125 * shrink-wrap.c: Update comment.
126 (next_block_for_reg): Rename to live_edge_for_reg.
127 (live_edge_for_reg): Allow live_edge->dest has two predecessors.
128 (move_insn_for_shrink_wrap): Split live_edge.
129 (prepre_shrink_wrap): One more parameter for move_insn_for_shrink_wrap.
131 2014-05-14 Eric Botcazou <ebotcazou@adacore.com>
133 * config/sparc/sparc-protos.h (sparc_absnegfloat_split_legitimate):
135 * config/sparc/sparc.c (sparc_absnegfloat_split_legitimate): Likewise.
136 * config/sparc/sparc.md (fptype_ut699): New attribute.
137 (in_branch_delay): Return false if -mfix-ut699 is specified and
138 fptype_ut699 is set to single.
139 (truncdfsf2): Add fptype_ut699 attribute.
140 (fix_truncdfsi2): Likewise.
141 (floatsisf2): Change fptype attribute.
142 (fix_truncsfsi2): Likewise.
143 (negtf2_notv9): Delete.
144 (negtf2_v9): Likewise.
145 (negtf2_hq): New instruction.
146 (negtf2): New instruction and splitter.
147 (negdf2_notv9): Rewrite.
148 (abstf2_notv9): Delete.
149 (abstf2_hq_v9): Likewise.
150 (abstf2_v9): Likewise.
151 (abstf2_hq): New instruction.
152 (abstf2): New instruction and splitter.
153 (absdf2_notv9): Rewrite.
155 2014-05-14 Cary Coutant <ccoutant@google.com>
158 * opts.c (common_handle_option): Don't special-case "-g".
159 (set_debug_level): Default to at least level 2 with "-g".
161 2014-05-14 DJ Delorie <dj@redhat.com>
163 * config/msp430/msp430.c (msp430_builtin): Add
164 MSP430_BUILTIN_DELAY_CYCLES.
165 (msp430_init_builtins): Register void __delay_cycles(long long).
166 (msp430_builtin_decl): Add it.
167 (cg_magic_constant): New.
168 (msp430_expand_delay_cycles): New.
169 (msp430_expand_builtin): Call it.
170 (msp430_print_operand_raw): Change integer printing from "int" to
172 * config/msp430/msp430.md (define_constants): Add delay_cycles tags.
173 (delay_cycles_start): New.
174 (delay_cycles_end): New.
175 (delay_cycles_32): New.
176 (delay_cycles_32x): New.
177 (delay_cycles_16): New.
178 (delay_cycles_16x): New.
179 (delay_cycles_2): New.
180 (delay_cycles_1): New.
181 * doc/extend.texi: Document __delay_cycles().
183 2014-05-14 Sandra Loosemore <sandra@codesourcery.com>
185 * config/nios2/nios2.md (nios2_cbranch): Fix paste-o in
186 length attribute computation.
188 2014-05-14 Richard Sandiford <rdsandiford@googlemail.com>
191 * print-rtl.c (print_rtx): Suppress uids if flag_dump_unnumbered.
193 2014-05-14 Richard Sandiford <r.sandiford@uk.ibm.com>
196 * config/sparc/sparc.md: Fix types of low and high in DI constant
197 splitter. Use gen_int_mode in some other splitters.
199 2014-05-14 Martin Jambor <mjambor@suse.cz>
202 * ipa-prop.c (ipa_modify_formal_parameters): Reset DECL_LANG_SPECIFIC.
204 2014-05-14 James Norris <jnorris@codesourcery.com>
206 * omp-low.c (expand_parallel_call): Remove shadow variable.
207 (expand_omp_taskreg): Likewise.
209 2014-05-14 Ilya Tocar <ilya.tocar@intel.com>
211 * common/config/i386/i386-common.c
212 (OPTION_MASK_ISA_CLFLUSHOPT_SET): Define.
213 (OPTION_MASK_ISA_XSAVES_SET): Ditto.
214 (OPTION_MASK_ISA_XSAVEC_SET): Ditto.
215 (OPTION_MASK_ISA_CLFLUSHOPT_UNSET): Ditto.
216 (OPTION_MASK_ISA_XSAVES_UNSET): Ditto.
217 (OPTION_MASK_ISA_XSAVEC_UNSET): Ditto.
218 (ix86_handle_option): Handle OPT_mxsavec, OPT_mxsaves,
220 * config.gcc (i[34567]86-*-*): Add clflushoptintrin.h,
221 xsavecintrin.h, xsavesintrin.h.
223 * config/i386/clflushoptintrin.h: New.
224 * config/i386/xsavecintrin.h: Ditto.
225 * config/i386/xsavesintrin.h: Ditto.
226 * config/i386/cpuid.h (bit_CLFLUSHOPT): Define.
229 * config/i386/driver-i386.c (host_detect_local_cpu): Handle
230 -mclflushopt, -mxsavec, -mxsaves, -mno-xsaves, -mno-xsavec,
232 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
233 OPTION_MASK_ISA_CLFLUSHOPT, OPTION_MASK_ISA_XSAVEC,
234 OPTION_MASK_ISA_XSAVES.
235 * config/i386/i386.c (ix86_target_string): Handle -mclflushopt,
237 (PTA_CLFLUSHOPT) Define.
240 (ix86_option_override_internal): Handle new options.
241 (ix86_valid_target_attribute_inner_p): Ditto.
242 (ix86_builtins): Add IX86_BUILTIN_XSAVEC, IX86_BUILTIN_XSAVEC64,
243 IX86_BUILTIN_XSAVES, IX86_BUILTIN_XRSTORS, IX86_BUILTIN_XSAVES64,
244 IX86_BUILTIN_XRSTORS64, IX86_BUILTIN_CLFLUSHOPT.
245 (bdesc_special_args): Add __builtin_ia32_xsaves, __builtin_ia32_xrstors,
246 __builtin_ia32_xsavec, __builtin_ia32_xsaves64, __builtin_ia32_xrstors64,
247 __builtin_ia32_xsavec64.
248 (ix86_init_mmx_sse_builtins): Add __builtin_ia32_clflushopt.
249 (ix86_expand_builtin): Handle new builtins.
250 * config/i386/i386.h (TARGET_CLFLUSHOPT) Define.
251 (TARGET_CLFLUSHOPT_P): Ditto.
252 (TARGET_XSAVEC): Ditto.
253 (TARGET_XSAVEC_P): Ditto.
254 (TARGET_XSAVES): Ditto.
255 (TARGET_XSAVES_P): Ditto.
256 * config/i386/i386.md (ANY_XSAVE): Add UNSPECV_XSAVEC, UNSPECV_XSAVES.
257 (ANY_XSAVE64)" Add UNSPECV_XSAVEC64, UNSPECV_XSAVES64.
258 (attr xsave): Add xsavec, xsavec64, xsaves, xsaves64.
260 (ANY_XRSTOR64): Ditto.
262 (xrstor): Change into <xrstor>.
263 (xrstor_rex64): Change into <xrstor>_rex64.
264 (xrstor64): Change into <xrstor>64
266 * config/i386/i386.opt (mclflushopt): New.
269 * config/i386/x86intrin.h: Add clflushoptintrin.h, xsavesintrin.h,
271 * doc/invoke.texi: Document new options.
273 2014-05-14 Andrey Belevantsev <abel@ispras.ru>
275 PR rtl-optimization/60866
276 * sel-sched-ir (sel_init_new_insn): New parameter old_seqno.
277 Default it to -1. Pass it down to init_simplejump_data.
278 (init_simplejump_data): New parameter old_seqno. Pass it down
279 to get_seqno_for_a_jump.
280 (get_seqno_for_a_jump): New parameter old_seqno. Use it for
281 initializing new jump seqno as a last resort. Add comment.
282 (sel_redirect_edge_and_branch): Save old seqno of the conditional
283 jump and pass it down to sel_init_new_insn.
284 (sel_redirect_edge_and_branch_force): Likewise.
286 2014-05-14 Georg-Johann Lay <avr@gjlay.de>
288 * config/avr/avr.h (REG_CLASS_CONTENTS): Use unsigned suffix for
289 shifted values to avoid build warning.
291 2014-05-14 Eric Botcazou <ebotcazou@adacore.com>
293 * cfgcleanup.c (try_forward_edges): Use location_t for locations.
294 * cfgrtl.c (rtl_merge_blocks): Fix comment.
295 (cfg_layout_merge_blocks): Likewise.
296 * except.c (emit_to_new_bb_before): Remove prev_bb local variable.
298 2014-05-14 Andrey Belevantsev <abel@ispras.ru>
300 PR rtl-optimization/60901
301 * config/i386/i386.c (ix86_dependencies_evaluation_hook): Check that
302 bb predecessor belongs to the same scheduling region. Adjust comment.
304 2014-05-13 Peter Bergner <bergner@vnet.ibm.com>
306 * doc/sourcebuild.texi: (dfp_hw): Document.
307 (p8vector_hw): Likewise.
308 (powerpc_eabi_ok): Likewise.
309 (powerpc_elfv2): Likewise.
310 (powerpc_htm_ok): Likewise.
311 (ppc_recip_hw): Likewise.
314 2014-05-13 Cary Coutant <ccoutant@google.com>
316 * opts.c (finish_options): Use -ggnu-pubnames with -gsplit-dwarf.
318 2014-05-13 David Malcolm <dmalcolm@redhat.com>
320 * gengtype-parse.c (require3): Eliminate in favor of...
322 (require_template_declaration): Update to support optional single *
325 * gengtype.c (get_ultimate_base_class): Add a non-const overload.
326 (create_user_defined_type): Handle a single level of explicit
327 pointerness within template arguments.
328 (struct write_types_data): Add field "kind".
329 (filter_type_name): Handle "*" character.
330 (write_user_func_for_structure_ptr): Require a write_types_data
331 rather than just a prefix string, so that we can look up the kind
332 of the wtd and use it as an index into wrote_user_func_for_ptr,
333 ensuring that such functions are written at most once. Support
334 subclasses by invoking the marking function of the ultimate base
336 (write_user_func_for_structure_body): Require a write_types_data
337 rather than just a prefix string, so that we can pass this to
338 write_user_func_for_structure_ptr.
339 (write_func_for_structure): Likewise.
340 (ggc_wtd): Add initializer of new "kind" field.
343 * gengtype.h (enum write_types_kinds): New.
344 (struct type): Add field wrote_user_func_for_ptr to the "s"
347 2014-05-13 Richard Sandiford <r.sandiford@uk.ibm.com>
349 * fold-const.c (optimize_bit_field_compare): Use wi:: operations
350 instead of const_binop.
351 (fold_binary_loc): Likewise.
353 2014-05-13 Richard Sandiford <r.sandiford@uk.ibm.com>
355 * tree-dfa.h (get_addr_base_and_unit_offset_1): Update array index
356 calculation to match get_ref_base_and_extent.
358 2014-05-13 Catherine Moore <clm@codesourcery.com>
359 Sandra Loosemore <sandra@codesourcery.com>
361 * configure.ac: Fix assembly for explicit JALR relocation check.
362 * configure: Regenerate.
364 2014-05-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
366 * config/arm/arm.c (neon_itype): Remove NEON_RESULTPAIR.
367 (arm_init_neon_builtins): Remove handling of NEON_RESULTPAIR.
368 Remove associated type declarations and initialisations.
369 (arm_expand_neon_builtin): Likewise.
370 (neon_emit_pair_result_insn): Delete.
371 * config/arm/arm_neon_builtins (vtrn, vzip, vuzp): Delete.
372 * config/arm/neon.md (neon_vtrn<mode>): Delete.
373 (neon_vzip<mode>): Likewise.
374 (neon_vuzp<mode>): Likewise.
376 2014-05-13 Richard Biener <rguenther@suse.de>
379 * tree-inline.c (remap_gimple_stmt): Clear tail call flag,
380 it needs revisiting whether the call still may be tail-called.
382 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
384 * rtl.def (SYMBOL_REF): Remove middle "0" field.
385 * rtl.h (block_symbol): Reduce number of fields to 2.
386 (rtx_def): Add u2.symbol_ref_flags.
387 (SYMBOL_REF_FLAGS): Use it.
388 (SYMBOL_REF_DATA, SET_SYMBOL_REF_DECL, SYMBOL_REF_DECL)
389 (SET_SYMBOL_REF_CONSTANT, SYMBOL_REF_CONSTANT): Lower index.
390 * gengtype.c (adjust_field_rtx_def): Remove SYMBOL_REF_FLAGS handling.
391 Lower index of SYMBOL_REF_DATA.
392 * print-rtl.c (print_rtx): Lower index for SYMBOL_REF_DATA.
393 Print SYMBOL_REF_FLAGS at the same time.
394 * genattrtab.c (attr_rtx_1): Only initialize 1 "0" SYMBOL_REF field.
396 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
398 * rtl.def (VAR_LOCATION): Remove "i" field.
399 * rtl.h (rtx_def): Add u2.var_location_status.
400 (PAT_VAR_LOCATION_STATUS): Use it.
401 (gen_rtx_VAR_LOCATION): Declare.
402 * gengenrtl.c (excluded_rtx): Add VAR_LOCATION.
403 * emit-rtl.c (gen_rtx_VAR_LOCATION): New function.
404 * var-tracking.c (emit_note_insn_var_location): Remove casts.
406 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
408 * rtl.def (scratch): Fix outdated comment and remove "0" field.
409 * gengtype.c (adjust_field_rtx_def): Update accordingly.
411 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
413 * rtl.def (DEBUG_INSN, INSN, JUMP_INSN, CALL_INSN, JUMP_TABLE_DATA)
414 (BARRIER, CODE_LABEL, NOTE): Remove first "i" field.
415 * rtl.h (rtx_def): Add insn_uid to u2 field.
416 (RTX_FLAG_CHECK8): Delete in favor of...
417 (RTL_INSN_CHAIN_FLAG_CHECK): ...this new macro.
418 (INSN_DELETED_P): Update accordingly.
419 (INSN_UID): Use u2.insn_uid.
420 (INSN_CHAIN_CODE_P): Define.
421 (PREV_INSN, NEXT_INSN, BLOCK_FOR_INSN, PATTERN, INSN_LOCATION)
422 (INSN_CODE, REG_NOTES, CALL_INSN_FUNCTION_USAGE, CODE_LABEL_NUMBER)
423 (NOTE_DATA, NOTE_DELETED_LABEL_NAME, NOTE_BLOCK, NOTE_EH_HANDLER)
424 (NOTE_BASIC_BLOCK, NOTE_VAR_LOCATION, NOTE_CFI, NOTE_LABEL_NUMBER)
425 (NOTE_KIND, LABEL_NAME, LABEL_NUSES, JUMP_LABEL, LABEL_REFS): Lower
427 * print-rtl.c (print_rtx): Print INSN_UIDs before the main loop.
428 Update indices for insn-chain rtxes.
429 * gengtype.c (gen_rtx_next): Adjust test for insn-chain rtxes.
430 (adjust_field_rtx_def): Lower '0' indices for all insn-chain rtxes.
431 * emit-rtl.c (gen_label_rtx): Update gen_rtx_LABEL call.
432 * caller-save.c (init_caller_save): Update gen_rtx_INSN calls.
433 * combine.c (try_combine): Likewise.
434 * ira.c (setup_prohibited_mode_move_regs): Likewise.
436 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
438 * rtl.def (REG): Remove middle field.
439 * rtl.h (rtx_def): Add orignal_regno to u2.
440 (ORIGINAL_REGNO): Use it instead of field 1.
441 (REG_ATTRS): Lower field index accordingly.
442 * gengtype.c (adjust_field_rtx_def): Remove handling of
443 ORIGINAL_REGNO. Move REG_ATTRS index down.
444 * print-rtl.c (print_rtx): Move ORIGINAL_REGNO handling to the
445 code that prints the REGNO.
447 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
449 * print-rtl.c (print_rtx): Guard whole '0' block with ifndef
452 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
454 * rtl.h (rtx_def): Mark u2 as GTY ((skip)).
456 2014-05-13 Bin Cheng <bin.cheng@arm.com>
458 * tree-ssa-loop-ivopts.c (contain_complex_addr_expr): New.
459 (alloc_iv): Lower base expressions containing ADDR_EXPR.
461 2014-05-13 Ian Bolton <ian.bolton@arm.com>
463 * config/aarch64/aarch64-protos.h
464 (aarch64_hard_regno_caller_save_mode): New prototype.
465 * config/aarch64/aarch64.c (aarch64_hard_regno_caller_save_mode):
467 * config/aarch64/aarch64.h (HARD_REGNO_CALLER_SAVE_MODE): New macro.
469 2014-05-13 Christian Bruel <christian.bruel@st.com>
471 * target.def (mode_switching): New hook vector.
472 (mode_emit, mode_needed, mode_after, mode_entry): New hooks.
473 (mode_exit, modepriority_to_mode): Likewise.
474 * mode-switching.c (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Hookify.
475 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
476 * target.h: Include tm.h and hard-reg-set.h.
477 * doc/tm.texi.in (EMIT_MODE_SET, MODE_NEEDED, MODE_AFTER, MODE_ENTRY)
478 (MODE_EXIT, MODE_PRIORITY_TO_MODE): Delete and hookify.
479 * doc/tm.texi Regenerate.
480 * config/sh/sh.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
481 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
482 * config/sh/sh.c (sh_emit_mode_set, sh_mode_priority): Hookify.
483 (sh_mode_needed, sh_mode_after, sh_mode_entry, sh_mode_exit): Likewise.
484 * config/i386/i386.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
485 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
486 * config/i386/i386-protos.h (ix86_mode_needed, ix86_mode_after)
487 (ix86_mode_entrym, ix86_emit_mode_set): Remove external declaration.
488 * config/i386/i386.c (ix86_mode_needed, ix86_mode_after, ix86_mode_exit,
489 (ix86_mode_entry, ix86_mode_priority, ix86_emit_mode_set): Hookify.
490 * config/epiphany/epiphany.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY):
492 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
493 * config/epiphany/epiphany-protos.h (epiphany_mode_needed)
494 (emit_set_fp_mode, epiphany_mode_entry_exit, epiphany_mode_after)
495 (epiphany_mode_priority_to_mode): Remove declaration.
496 * config/epiphany/epiphany.c (emit_set_fp_mode): Hookify.
497 (epiphany_mode_needed, epiphany_mode_priority_to_mode): Likewise.
498 (epiphany_mode_entry, epiphany_mode_exit, epiphany_mode_after):
500 (epiphany_mode_priority_to_mode): Change priority type. Hookify.
501 (epiphany_mode_needed, epiphany_mode_entry_exit): Hookify.
502 (epiphany_mode_after, epiphany_mode_entry, emit_set_fp_mode): Hookify.
504 2014-05-13 Jakub Jelinek <jakub@redhat.com>
507 * config/i386/i386.c (ix86_expand_set_or_movmem): If count_exp
508 is const0_rtx, return immediately. Don't test count == 0 when
511 2014-05-13 Zhenqiang Chen <zhenqiang.chen@linaro.org>
513 * Makefile.in: add shrink-wrap.o.
514 * config/i386/i386.c: include "shrink-wrap.h"
515 * function.c: Likewise.
516 (requires_stack_frame_p, next_block_for_reg,
517 move_insn_for_shrink_wrap, prepare_shrink_wrap,
518 dup_block_and_redirect): Move to shrink-wrap.c
519 (thread_prologue_and_epilogue_insns): Extract three code segments
520 as functions in shrink-wrap.c
521 * function.h: Move #ifdef HAVE_simple_return ... #endif block to
523 * shrink-wrap.c: New file.
524 * shrink-wrap.h: New file.
526 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
528 * doc/extend.texi: Reflect current numbers of pragmas. Remove
529 reference to Solaris.
531 2014-05-12 Mike Stump <mikestump@comcast.net>
534 * genattrtab.c (filename): Add.
535 (convert_set_attr_alternative): Improve error message.
536 (check_defs): Restore read_md_filename for error messages.
537 (gen_insn): Save filename.
539 2014-05-12 Dimitris Papavasiliou <dpapavas@gmail.com>
541 * doc/invoke.texi: Document new switches -Wno-shadow-ivar,
542 -fno-local-ivars and -fivar-visibility.
543 * c-family/c.opt: Make -Wshadow also implicitly enable
546 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
548 * doc/tm.texi: Remove reference to deleted macro.
549 * doc/tm.texi.in: Likewise.
551 2014-05-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
554 * config/avr/avr.c (avr_out_store_psi): Use correct constant
557 2014-05-12 Georg-Johann Lay <avr@gjlay.de>
559 * config/arm/arm.h (License): Add GCC Runtime Library Exception.
560 * config/arm/aout.h (License): Same.
561 * config/arm/bpabi.h (License): Same.
562 * config/arm/elf.h (License): Same.
563 * config/arm/linux-elf.h (License): Same.
564 * config/arm/linux-gas.h (License): Same.
565 * config/arm/netbsd-elf.h (License): Same.
566 * config/arm/uclinux-eabi.h (License): Same.
567 * config/arm/uclinux-elf.h (License): Same.
568 * config/arm/vxworks.h (License): Same.
570 2014-05-11 Jakub Jelinek <jakub@redhat.com>
572 * tree.h (OMP_CLAUSE_LINEAR_STMT): Define.
573 * tree.c (omp_clause_num_ops): Increase OMP_CLAUSE_LINEAR
574 number of operands to 3.
575 (walk_tree_1): Walk all operands of OMP_CLAUSE_LINEAR.
576 * tree-nested.c (convert_nonlocal_omp_clauses,
577 convert_local_omp_clauses): Handle OMP_CLAUSE_DEPEND.
578 * gimplify.c (gimplify_scan_omp_clauses): Handle
579 OMP_CLAUSE_LINEAR_STMT.
580 * omp-low.c (lower_rec_input_clauses): Fix typo.
581 (maybe_add_implicit_barrier_cancel, lower_omp_1): Add
582 cast between Fortran boolean_type_node and C _Bool if
585 2014-05-11 Richard Sandiford <rdsandiford@googlemail.com>
587 PR tree-optimization/61136
588 * wide-int.h (multiple_of_p): Define a version that doesn't return
590 * fold-const.c (extract_muldiv_1): Use wi::multiple_of_p instead of an
591 integer_zerop/const_binop pair.
592 (multiple_of_p): Likewise, converting both operands to widest_int
595 2014-05-09 Teresa Johnson <tejohnson@google.com>
597 * cgraphunit.c (analyze_functions): Use correct dump file.
599 2014-05-09 Florian Weimer <fweimer@redhat.com>
601 * cfgexpand.c (stack_protect_decl_p): New function, extracted from
603 (stack_protect_return_slot_p): New function.
604 (expand_used_vars): Call stack_protect_decl_p and
605 stack_protect_return_slot_p for -fstack-protector-strong.
607 2014-05-09 Kenneth Zadeck <zadeck@naturalbridge.com>
610 * fold-const.c (fold_binary_loc): Changed width of mask.
612 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
614 * config/avr/avr-fixed.md (round<mode>3): Use -1U instead of -1 in
615 unsigned int initializers for regno_in, regno_out.
617 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
620 * config/avr/avr.md (cc): Add new attribute set_vzn.
621 (addqi3, addqq3, adduqq3, subqi3, subqq3, subuqq3, negqi2) [cc]:
622 Set cc insn attribute to set_vzn instead of set_zn for alternatives
623 with INC, DEC or NEG.
624 * config/avr/avr.c (avr_notice_update_cc): Handle SET_VZN.
625 (avr_out_plus_1): ADIW sets cc0 to CC_SET_CZN.
626 INC, DEC and ADD+ADC set cc0 to CC_CLOBBER.
628 2014-05-09 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
631 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
633 * wide-int.cc (UTItype): Define.
634 (UDWtype): Define for appropriate W_TYPE_SIZE.
636 2014-05-09 Richard Biener <rguenther@suse.de>
638 * Makefile.in (GTFILES): Remove tree-ssa-propagate.c.
639 * tree-ssa-propagate.c: Do not include gt-tree-ssa-propagate.h.
640 (interesting_ssa_edges, varying_ssa_edges): Move out of GC space.
641 (add_ssa_edge, process_ssa_edge_worklist, ssa_prop_init,
642 ssa_propagate): Adjust.
644 2014-05-08 Jeff Law <law@redhat.com>
646 PR tree-optimization/61009
647 * tree-ssa-threadedge.c (thread_through_normal_block): Return a
648 tri-state rather than a boolean. When a block is too big to
649 thread through, inform caller via negative return value.
650 (thread_across_edge): If a block was too big for normal threading,
651 then it's too big for a joiner too, so remove temporary equivalences
652 and return immediately.
654 2014-05-08 Manuel López-Ibáñez <manu@gcc.gnu.org>
655 Matthias Klose <doko@ubuntu.com>
658 * optc-gen.awk: Fix option handling for -Wunused-parameter.
660 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
663 * config/i386/i386.c (PTA_HASWELL): Remove PTA_RTM.
665 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
668 * config/alpha/alpha.c: Include gimple-iterator.h.
669 (alpha_gimple_fold_builtin): New function. Move
670 ALPHA_BUILTIN_UMULH folding from ...
671 (alpha_fold_builtin): ... here.
672 (TARGET_GIMPLE_FOLD_BUILTIN): New define.
674 2014-05-08 Wei Mi <wmi@google.com>
677 * config/i386/i386.c (ix86_compute_frame_layout): Update
678 preferred_stack_boundary for call, expanded from tls descriptor.
679 * config/i386/i386.md (*tls_global_dynamic_32_gnu): Update RTX
680 to depend on SP register.
681 (*tls_local_dynamic_base_32_gnu): Ditto.
682 (*tls_local_dynamic_32_once): Ditto.
683 (tls_global_dynamic_64_<mode>): Set
684 ix86_tls_descriptor_calls_expanded_in_cfun.
685 (tls_local_dynamic_base_64_<mode>): Ditto.
686 (tls_global_dynamic_32): Set
687 ix86_tls_descriptor_calls_expanded_in_cfun. Update RTX
688 to depend on SP register.
689 (tls_local_dynamic_base_32): Ditto.
691 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
693 * config/arm/arm_neon.h: Update comment.
694 * config/arm/neon-docgen.ml: Delete.
695 * config/arm/neon-gen.ml: Delete.
696 * doc/arm-neon-intrinsics.texi: Update comment.
698 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
700 * config/arm/arm_neon_builtins.def (vadd, vsub): Only define the v2sf
702 (vand, vorr, veor, vorn, vbic): Remove.
703 * config/arm/neon.md (neon_vadd, neon_vsub, neon_vadd_unspec): Adjust
705 (neon_vsub_unspec): Likewise.
706 (neon_vorr, neon_vand, neon_vbic, neon_veor, neon_vorn): Remove.
708 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
710 * config/arm/arm_neon.h (vadd_s8): GNU C implementation
711 (vadd_s16): Likewise.
712 (vadd_s32): Likewise.
713 (vadd_f32): Likewise.
715 (vadd_u16): Likewise.
716 (vadd_u32): Likewise.
717 (vadd_s64): Likewise.
718 (vadd_u64): Likewise.
719 (vaddq_s8): Likewise.
720 (vaddq_s16): Likewise.
721 (vaddq_s32): Likewise.
722 (vaddq_s64): Likewise.
723 (vaddq_f32): Likewise.
724 (vaddq_u8): Likewise.
725 (vaddq_u16): Likewise.
726 (vaddq_u32): Likewise.
727 (vaddq_u64): Likewise.
729 (vmul_s16): Likewise.
730 (vmul_s32): Likewise.
731 (vmul_f32): Likewise.
733 (vmul_u16): Likewise.
734 (vmul_u32): Likewise.
736 (vmulq_s8): Likewise.
737 (vmulq_s16): Likewise.
738 (vmulq_s32): Likewise.
739 (vmulq_f32): Likewise.
740 (vmulq_u8): Likewise.
741 (vmulq_u16): Likewise.
742 (vmulq_u32): Likewise.
744 (vsub_s16): Likewise.
745 (vsub_s32): Likewise.
746 (vsub_f32): Likewise.
748 (vsub_u16): Likewise.
749 (vsub_u32): Likewise.
750 (vsub_s64): Likewise.
751 (vsub_u64): Likewise.
752 (vsubq_s8): Likewise.
753 (vsubq_s16): Likewise.
754 (vsubq_s32): Likewise.
755 (vsubq_s64): Likewise.
756 (vsubq_f32): Likewise.
757 (vsubq_u8): Likewise.
758 (vsubq_u16): Likewise.
759 (vsubq_u32): Likewise.
760 (vsubq_u64): Likewise.
762 (vand_s16): Likewise.
763 (vand_s32): Likewise.
765 (vand_u16): Likewise.
766 (vand_u32): Likewise.
767 (vand_s64): Likewise.
768 (vand_u64): Likewise.
769 (vandq_s8): Likewise.
770 (vandq_s16): Likewise.
771 (vandq_s32): Likewise.
772 (vandq_s64): Likewise.
773 (vandq_u8): Likewise.
774 (vandq_u16): Likewise.
775 (vandq_u32): Likewise.
776 (vandq_u64): Likewise.
778 (vorr_s16): Likewise.
779 (vorr_s32): Likewise.
781 (vorr_u16): Likewise.
782 (vorr_u32): Likewise.
783 (vorr_s64): Likewise.
784 (vorr_u64): Likewise.
785 (vorrq_s8): Likewise.
786 (vorrq_s16): Likewise.
787 (vorrq_s32): Likewise.
788 (vorrq_s64): Likewise.
789 (vorrq_u8): Likewise.
790 (vorrq_u16): Likewise.
791 (vorrq_u32): Likewise.
792 (vorrq_u64): Likewise.
794 (veor_s16): Likewise.
795 (veor_s32): Likewise.
797 (veor_u16): Likewise.
798 (veor_u32): Likewise.
799 (veor_s64): Likewise.
800 (veor_u64): Likewise.
801 (veorq_s8): Likewise.
802 (veorq_s16): Likewise.
803 (veorq_s32): Likewise.
804 (veorq_s64): Likewise.
805 (veorq_u8): Likewise.
806 (veorq_u16): Likewise.
807 (veorq_u32): Likewise.
808 (veorq_u64): Likewise.
810 (vbic_s16): Likewise.
811 (vbic_s32): Likewise.
813 (vbic_u16): Likewise.
814 (vbic_u32): Likewise.
815 (vbic_s64): Likewise.
816 (vbic_u64): Likewise.
817 (vbicq_s8): Likewise.
818 (vbicq_s16): Likewise.
819 (vbicq_s32): Likewise.
820 (vbicq_s64): Likewise.
821 (vbicq_u8): Likewise.
822 (vbicq_u16): Likewise.
823 (vbicq_u32): Likewise.
824 (vbicq_u64): Likewise.
826 (vorn_s16): Likewise.
827 (vorn_s32): Likewise.
829 (vorn_u16): Likewise.
830 (vorn_u32): Likewise.
831 (vorn_s64): Likewise.
832 (vorn_u64): Likewise.
833 (vornq_s8): Likewise.
834 (vornq_s16): Likewise.
835 (vornq_s32): Likewise.
836 (vornq_s64): Likewise.
837 (vornq_u8): Likewise.
838 (vornq_u16): Likewise.
839 (vornq_u32): Likewise.
840 (vornq_u64): Likewise.
842 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
844 * wide-int.cc (UTItype): Define.
845 (UDWtype): Define for appropriate W_TYPE_SIZE.
847 2014-05-08 Marc Glisse <marc.glisse@inria.fr>
849 PR tree-optimization/59100
850 * tree-ssa-phiopt.c: Include tree-inline.h.
851 (neutral_element_p, absorbing_element_p): New functions.
852 (value_replacement): Handle conditional binary operations with a
853 neutral or absorbing element.
855 2014-05-08 Richard Biener <rguenther@suse.de>
857 * tree-ssa-sccvn.c (vn_get_expr_for): Valueize operands before
858 folding the expression.
859 (valueize_expr): Remove.
860 (visit_reference_op_load): Do not valueize the result of
862 (simplify_binary_expression): Likewise.
863 (simplify_unary_expression): Likewise.
865 2014-05-08 Richard Biener <rguenther@suse.de>
867 * gimplify.c (gimplify_call_expr): Use saved fnptrtype for
868 looking at TYPE_ARG_TYPES.
870 2014-05-08 Richard Biener <rguenther@suse.de>
872 * gimple-fold.c (gimple_fold_stmt_to_constant_1): Remove
873 pointer propagation special-case.
875 2014-05-08 Bin Cheng <bin.cheng@arm.com>
877 * tree-affine.c (tree_to_aff_combination): Handle MEM_REF for
878 core part of address expressions.
880 2014-05-08 Alan Modra <amodra@gmail.com>
883 * config/rs6000/rs6000.c (expand_block_move): Allow 64-bit
884 loads and stores when -mno-strict-align at any alignment.
885 (expand_block_clear): Similarly. Also correct calculation of
888 2014-05-07 Thomas Preud'homme <thomas.preudhomme@arm.com>
891 * tree-complex.c (expand_complex_move): Keep line info when expanding
893 * tree-ssa-uninit.c (warn_uninit): New argument. Ignore assignment
894 of complex expression. Use new argument to display correct location
895 for values coming from phi statement.
896 (warn_uninitialized_vars): Adapt to new signature of warn_uninit.
897 (warn_uninitialized_phi): Pass location of phi argument to
899 * tree-ssa.c (ssa_undefined_value_p): For SSA_NAME initialized by a
900 COMPLEX_EXPR, recurse on each part of the COMPLEX_EXPR.
902 2014-05-07 Segher Boessenkool <segher@kernel.crashing.org>
904 * config/rs6000/predicates.md (indexed_address_mem): New.
905 * config/rs6000/rs6000.md (type): Remove load_ext, load_ext_u,
906 load_ext_ux, load_ux, load_u, store_ux, store_u, fpload_ux, fpload_u,
907 fpstore_ux, fpstore_u.
908 (sign_extend, indexed, update): New.
909 (cell_micro): Adjust.
910 (*zero_extend<mode>di2_internal1, *zero_extendsidi2_lfiwzx,
911 *extendsidi2_lfiwax, *extendsidi2_nocell, *extendsfdf2_fpr,
912 *movsi_internal1, *movsi_internal1_single, *movhi_internal,
913 *movqi_internal, *movcc_internal1, mov<mode>_hardfloat,
914 *mov<mode>_softfloat, *mov<mode>_hardfloat32, *mov<mode>_hardfloat64,
915 *mov<mode>_softfloat64, *movdi_internal32, *movdi_internal64,
916 *mov<mode>_string, *ldmsi8, *ldmsi7, *ldmsi6, *ldmsi5, *ldmsi4,
917 *ldmsi3, *stmsi8, *stmsi7, *stmsi6, *stmsi5, *stmsi4, *stmsi3,
918 *movdi_update1, movdi_<mode>_update, movdi_<mode>_update_stack,
919 *movsi_update1, *movsi_update2, movsi_update, movsi_update_stack,
920 *movhi_update1, *movhi_update2, *movhi_update3, *movhi_update4,
921 *movqi_update1, *movqi_update2, *movqi_update3, *movsf_update1,
922 *movsf_update2, *movsf_update3, *movsf_update4, *movdf_update1,
923 *movdf_update2, load_toc_aix_si, load_toc_aix_di, probe_stack_<mode>,
924 *stmw, *lmw, as well as 10 anonymous patterns): Adjust.
926 * config/rs6000/dfp.md (movsd_store, movsd_load): Adjust.
927 * config/rs6000/vsx.md (*vsx_movti_32bit, *vsx_extract_<mode>_load,
928 *vsx_extract_<mode>_store): Adjust.
929 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_microcoded_insn,
930 is_cracked_insn, insn_must_be_first_in_group,
931 insn_must_be_last_in_group): Adjust.
933 * config/rs6000/40x.md (ppc403-load, ppc403-store, ppc405-float):
935 * config/rs6000/440.md (ppc440-load, ppc440-store, ppc440-fpload,
936 ppc440-fpstore): Adjust.
937 * config/rs6000/476.md (ppc476-load, ppc476-store, ppc476-fpload,
938 ppc476-fpstore): Adjust.
939 * config/rs6000/601.md (ppc601-load, ppc601-store, ppc601-fpload,
940 ppc601-fpstore): Adjust.
941 * config/rs6000/603.md (ppc603-load, ppc603-store, ppc603-fpload):
943 * config/rs6000/6xx.md (ppc604-load, ppc604-store, ppc604-fpload):
945 * config/rs6000/7450.md (ppc7450-load, ppc7450-store, ppc7450-fpload,
946 ppc7450-fpstore): Adjust.
947 * config/rs6000/7xx.md (ppc750-load, ppc750-store): Adjust.
948 * config/rs6000/8540.md (ppc8540_load, ppc8540_store): Adjust.
949 * config/rs6000/a2.md (ppca2-load, ppca2-fp-load, ppca2-fp-store):
951 * config/rs6000/cell.md (cell-load, cell-load-ux, cell-load-ext,
952 cell-fpload, cell-fpload-update, cell-store, cell-store-update,
953 cell-fpstore, cell-fpstore-update): Adjust.
954 * config/rs6000/e300c2c3.md (ppce300c3_load, ppce300c3_fpload,
955 ppce300c3_store, ppce300c3_fpstore): Adjust.
956 * config/rs6000/e500mc.md (e500mc_load, e500mc_fpload, e500mc_store,
957 e500mc_fpstore): Adjust.
958 * config/rs6000/e500mc64.md (e500mc64_load, e500mc64_fpload,
959 e500mc64_store, e500mc64_fpstore): Adjust.
960 * config/rs6000/e5500.md (e5500_load, e5500_fpload, e5500_store,
961 e5500_fpstore): Adjust.
962 * config/rs6000/e6500.md (e6500_load, e6500_fpload, e6500_store,
963 e6500_fpstore): Adjust.
964 * config/rs6000/mpc.md (mpccore-load, mpccore-store, mpccore-fpload):
966 * config/rs6000/power4.md (power4-load, power4-load-ext,
967 power4-load-ext-update, power4-load-ext-update-indexed,
968 power4-load-update-indexed, power4-load-update, power4-fpload,
969 power4-fpload-update, power4-store, power4-store-update,
970 power4-store-update-indexed, power4-fpstore, power4-fpstore-update):
972 * config/rs6000/power5.md (power5-load, power5-load-ext,
973 power5-load-ext-update, power5-load-ext-update-indexed,
974 power5-load-update-indexed, power5-load-update, power5-fpload,
975 power5-fpload-update, power5-store, power5-store-update,
976 power5-store-update-indexed, power5-fpstore, power5-fpstore-update):
978 * config/rs6000/power6.md (power6-load, power6-load-ext,
979 power6-load-update, power6-load-update-indexed,
980 power6-load-ext-update, power6-load-ext-update-indexed, power6-fpload,
981 power6-fpload-update, power6-store, power6-store-update,
982 power6-store-update-indexed, power6-fpstore, power6-fpstore-update):
984 * config/rs6000/power7.md (power7-load, power7-load-ext,
985 power7-load-update, power7-load-update-indexed,
986 power7-load-ext-update, power7-load-ext-update-indexed, power7-fpload,
987 power7-fpload-update, power7-store, power7-store-update,
988 power7-store-update-indexed, power7-fpstore, power7-fpstore-update):
990 * config/rs6000/power8.md (power8-load, power8-load-update,
991 power8-load-ext, power8-load-ext-update, power8-fpload,
992 power8-fpload-update, power8-store, power8-store-update-indexed,
993 power8-fpstore, power8-fpstore-update): Adjust.
994 * config/rs6000/rs64.md (rs64a-load, rs64a-store, rs64a-fpload):
996 * config/rs6000/titan.md (titan_lsu_load, titan_lsu_fpload,
997 titan_lsu_store, titan_lsu_fpstore): Adjust.
998 * config/rs6000/xfpu.md (fp-load, fp-store): Adjust.
1000 2014-05-07 Oleg Endo <olegendo@gcc.gnu.org>
1003 * config/sh/sh-mem.cc (sh_expand_strlen): Use loop when emitting
1004 unrolled byte insns. Emit address increments after move insns.
1006 2014-05-07 David Malcolm <dmalcolm@redhat.com>
1008 * gimple.h (gimple_builtin_call_types_compatible_p): Accept a
1009 const_gimple, rather than a gimple.
1010 (gimple_call_builtin_p): Likewise, for the three variants.
1012 * gimple.c (gimple_builtin_call_types_compatible_p): Likewise.
1013 (gimple_call_builtin_p): Likewise, for the three variants.
1015 2014-05-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
1017 PR tree-optimization/61095
1018 * tree-ssanames.c (get_nonzero_bits): Fix type extension in wi::shwi.
1020 2014-05-07 Richard Biener <rguenther@suse.de>
1022 PR tree-optimization/61034
1023 * tree-ssa-alias.c (call_may_clobber_ref_p_1): Export.
1024 (maybe_skip_until): Use translate to take into account
1025 lattices when trying to do disambiguations.
1026 (get_continuation_for_phi_1): Likewise.
1027 (get_continuation_for_phi): Adjust for added translate arguments.
1028 (walk_non_aliased_vuses): Likewise.
1029 * tree-ssa-alias.h (get_continuation_for_phi): Adjust prototype.
1030 (walk_non_aliased_vuses): Likewise.
1031 (call_may_clobber_ref_p_1): Declare.
1032 * tree-ssa-sccvn.c (vn_reference_lookup_3): Also disambiguate against
1033 calls. Stop early if we are only supposed to disambiguate.
1034 * tree-ssa-pre.c (translate_vuse_through_block): Adjust.
1036 2014-05-07 Joern Rennecke <joern.rennecke@embecosm.com>
1038 * config/epiphany/epiphany.c (epiphany_handle_interrupt_attribute):
1039 Emit an error when the function has arguments.
1041 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
1043 * cfgloop.h (unswitch_loops): Remove.
1044 * doc/passes.texi: Remove references to loop-unswitch.c
1045 * timevar.def (TV_LOOP_UNSWITCH): Remove.
1047 2014-05-07 Evgeny Stupachenko <evstupac@gmail.com>
1049 * tree-vect-data-refs.c (vect_grouped_load_supported): New
1050 check for loads group of length 3.
1051 (vect_permute_load_chain): New permutations for loads group of
1053 * tree-vect-stmts.c (vect_model_load_cost): Change cost
1054 of vec_perm_shuffle for the new permutations.
1056 2014-05-07 Alan Lawrence <alan.lawrence@arm.com>
1058 * config/aarch64/arm_neon.h (vtrn1_f32, vtrn1_p8, vtrn1_p16, vtrn1_s8,
1059 vtrn1_s16, vtrn1_s32, vtrn1_u8, vtrn1_u16, vtrn1_u32, vtrn1q_f32,
1060 vtrn1q_f64, vtrn1q_p8, vtrn1q_p16, vtrn1q_s8, vtrn1q_s16, vtrn1q_s32,
1061 vtrn1q_s64, vtrn1q_u8, vtrn1q_u16, vtrn1q_u32, vtrn1q_u64, vtrn2_f32,
1062 vtrn2_p8, vtrn2_p16, vtrn2_s8, vtrn2_s16, vtrn2_s32, vtrn2_u8,
1063 vtrn2_u16, vtrn2_u32, vtrn2q_f32, vtrn2q_f64, vtrn2q_p8, vtrn2q_p16,
1064 vtrn2q_s8, vtrn2q_s16, vtrn2q_s32, vtrn2q_s64, vtrn2q_u8, vtrn2q_u16,
1065 vtrn2q_u32, vtrn2q_u64): Replace temporary asm with __builtin_shuffle.
1067 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
1069 * loop-unswitch.c: Delete.
1071 2014-05-07 Richard Biener <rguenther@suse.de>
1073 * config.gcc: Always set need_64bit_hwint to yes.
1075 2014-05-07 Chung-Ju Wu <jasonwucj@gmail.com>
1077 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Have it in favor
1078 of using optimize_size.
1080 2014-05-06 Mike Stump <mikestump@comcast.net>
1082 * wide-int.h (wi::int_traits <HOST_WIDE_INT>): Always define.
1084 2014-05-06 Joseph Myers <joseph@codesourcery.com>
1086 * config/i386/sse.md (*mov<mode>_internal)
1087 (*<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>)
1088 (*<sse2_avx_avx512f>_loaddqu<mode><mask_name>)
1089 (<sse>_andnot<mode>3, <code><mode>3, *andnot<mode>3)
1090 (*<code><mode>3, *andnot<mode>3<mask_name>)
1091 (<mask_codefor><code><mode>3<mask_name>): Only consider
1092 TARGET_SSE_PACKED_SINGLE_INSN_OPTIMAL for modes of size 16.
1094 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
1097 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
1099 * lra-constraints.c (valid_address_p): Move earlier in file.
1100 Add a constraint argument to the address_info version.
1101 (satisfies_memory_constraint_p): New function.
1102 (satisfies_address_constraint_p): Likewise.
1103 (process_alt_operands, curr_insn_transform): Use them.
1104 (process_address): Pass the constraint to valid_address_p when
1105 checking address operands.
1107 2014-05-06 Richard Sandiford <r.sandiford@uk.ibm.com>
1109 * lto-cgraph.c (compute_ltrans_boundary): Make node variables local
1110 to their respective blocks. Fix inadvertent use of "node".
1112 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
1114 * emit-rtl.c (init_derived_machine_modes): New functionm, split
1116 (init_emit_once): ...here.
1117 * rtl.h (init_derived_machine_modes): Declare.
1118 * toplev.c (do_compile): Call it even if no_backend.
1120 2014-05-06 Kenneth Zadeck <zadeck@naturalbridge.com>
1121 Mike Stump <mikestump@comcast.net>
1122 Richard Sandiford <rdsandiford@googlemail.com>
1123 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1125 * alias.c (ao_ref_from_mem): Use wide-int interfaces.
1126 (rtx_equal_for_memref_p): Update comment.
1127 (adjust_offset_for_component_ref): Use wide-int interfaces.
1128 * builtins.c (get_object_alignment_2): Likewise.
1129 (c_readstr): Likewise.
1130 (target_char_cast): Add comment.
1131 (determine_block_size): Use wide-int interfaces.
1132 (expand_builtin_signbit): Likewise.
1133 (fold_builtin_int_roundingfn): Likewise.
1134 (fold_builtin_bitop): Likewise.
1135 (fold_builtin_bswap): Likewise.
1136 (fold_builtin_logarithm): Use signop.
1137 (fold_builtin_pow): Likewise.
1138 (fold_builtin_memory_op): Use wide-int interfaces.
1139 (fold_builtin_object_size): Likewise.
1140 * cfgloop.c (alloc_loop): Initialize nb_iterations_upper_bound and
1141 nb_iterations_estimate.
1142 (record_niter_bound): Use wide-int interfaces.
1143 (get_estimated_loop_iterations_int): Likewise.
1144 (get_estimated_loop_iterations): Likewise.
1145 (get_max_loop_iterations): Likewise.
1146 * cfgloop.h: Include wide-int.h.
1147 (struct nb_iter_bound): Change bound to widest_int.
1148 (struct loop): Change nb_iterations_upper_bound and
1149 nb_iterations_estimate to widest_int.
1150 (record_niter_bound): Switch to use widest_int.
1151 (get_estimated_loop_iterations): Likewise.
1152 (get_max_loop_iterations): Likewise.
1153 (gcov_type_to_double_int): Rename to gcov_type_to_wide_int and
1154 update for wide-int.
1155 * cgraph.c (cgraph_add_thunk): Use wide-int interfaces.
1156 * combine.c (try_combine): Likewise.
1157 (subst): Use CONST_SCALAR_INT_P rather than CONST_INT_P.
1158 * config/aarch64/aarch64.c (aapcs_vfp_sub_candidate): Use wide-int
1160 (aarch64_float_const_representable_p): Likewise.
1161 * config/arc/arc.c: Include wide-int.h.
1162 (arc_can_use_doloop_p): Use wide-int interfaces.
1163 * config/arm/arm.c (aapcs_vfp_sub_candidate): Likewise.
1164 (vfp3_const_double_index): Likewise.
1165 * config/avr/avr.c (avr_out_round): Likewise.
1166 (avr_fold_builtin): Likewise.
1167 * config/bfin/bfin.c (bfin_local_alignment): Likewise.
1168 (bfin_can_use_doloop_p): Likewise.
1169 * config/darwin.c (darwin_mergeable_constant_section): Likewise.
1170 (machopic_select_rtx_section): Update to handle CONST_WIDE_INT.
1171 * config/i386/i386.c: Include wide-int.h.
1172 (ix86_data_alignment): Use wide-int interfaces.
1173 (ix86_local_alignment): Likewise.
1174 (ix86_emit_swsqrtsf): Update real_from_integer.
1175 * config/msp430/msp430.c (msp430_attr): Use wide-int interfaces.
1176 * config/nds32/nds32.c (nds32_insert_attributes): Likewise.
1177 * config/rs6000/predicates.md (any_operand): Add const_wide_int.
1178 (zero_constant): Likewise.
1179 (input_operand): Likewise.
1180 (splat_input_operand): Likewise.
1181 (non_logical_cint_operand): Change const_double to const_wide_int.
1182 * config/rs6000/rs6000.c (num_insns_constant): Handle CONST_WIDE_INT.
1183 (easy_altivec_constant): Remove comment.
1184 (paired_expand_vector_init): Use CONSTANT_P.
1185 (rs6000_legitimize_address): Handle CONST_WIDE_INT.
1186 (rs6000_emit_move): Update checks.
1187 (rs6000_aggregate_candidate): Use wide-int interfaces.
1188 (rs6000_expand_ternop_builtin): Likewise.
1189 (rs6000_output_move_128bit): Handle CONST_WIDE_INT.
1190 (rs6000_assemble_integer): Likewise.
1191 (rs6000_hash_constant): Likewise.
1192 (output_toc): Likewise.
1193 (rs6000_rtx_costs): Likewise.
1194 (rs6000_emit_swrsqrt); Update call to real_from_integer.
1195 * config/rs6000/rs6000-c.c: Include wide-int.h.
1196 (altivec_resolve_overloaded_builtin): Use wide-int interfaces.
1197 * config/rs6000/rs6000.h (TARGET_SUPPORTS_WIDE_INT): New.
1198 * config/rs6000/rs6000.md: Use const_scalar_int_operand.
1199 Handle CONST_WIDE_INT.
1200 * config/sol2-c.c (solaris_pragma_align): Change low to unsigned HWI.
1201 Use tree_fits_uhwi_p.
1202 * config/sparc/sparc.c: Include wide-int.h.
1203 (sparc_fold_builtin): Use wide-int interfaces.
1204 * config/vax/vax.c: Include wide-int.h.
1205 (vax_float_literal): Use real_from_integer.
1206 * coretypes.h (struct hwivec_def): New.
1208 (const_hwivec): New.
1209 * cse.c (hash_rtx_cb): Handle CONST_WIDE_INT.
1210 (equiv_constant): Handle CONST_WIDE_INT.
1211 * cselib.c (rtx_equal_for_cselib_1): Use CASE_CONST_UNIQUE.
1212 (cselib_hash_rtx): Handle CONST_WIDE_INT.
1213 * dbxout.c (stabstr_U): Use wide-int interfaces.
1214 (dbxout_type): Update to use cst_fits_shwi_p.
1215 * defaults.h (LOG2_BITS_PER_UNIT): Define.
1216 (TARGET_SUPPORTS_WIDE_INT): Add default.
1217 * dfp.c: Include wide-int.h.
1218 (decimal_real_to_integer2): Use wide-int interfaces and rename to
1219 decimal_real_to_integer.
1220 * dfp.h (decimal_real_to_integer2): Return a wide_int and rename to
1221 decimal_real_to_integer.
1222 * doc/generic.texi (Constant expressions): Update for wide_int.
1223 * doc/rtl.texi (const_double): Likewise.
1224 (const_wide_int, CONST_WIDE_INT, CONST_WIDE_INT_VEC): New.
1225 (CONST_WIDE_INT_NUNITS, CONST_WIDE_INT_ELT): New.
1226 * doc/tm.texi.in (REAL_VALUE_TO_INT): Remove.
1227 (REAL_VALUE_FROM_INT): Remove.
1228 (TARGET_SUPPORTS_WIDE_INT): New.
1229 * doc/tm.texi: Regenerate.
1230 * dojump.c (prefer_and_bit_test): Use wide-int interfaces.
1231 * double-int.h: Include wide-int.h.
1232 (struct wi::int_traits): New.
1233 * dwarf2out.c (get_full_len): New.
1234 (dw_val_equal_p): Add case dw_val_class_wide_int.
1235 (size_of_loc_descr): Likewise.
1236 (output_loc_operands): Likewise.
1237 (insert_double): Remove.
1238 (insert_wide_int): New.
1240 (print_die): Add case dw_val_class_wide_int.
1241 (attr_checksum): Likewise.
1242 (attr_checksum_ordered): Likewise.
1243 (same_dw_val_p): Likewise.
1244 (size_of_die): Likewise.
1245 (value_format): Likewise.
1246 (output_die): Likewise.
1247 (double_int_type_size_in_bits): Rename to offset_int_type_size_in_bits.
1249 (clz_loc_descriptor): Use wide-int interfaces.
1250 (mem_loc_descriptor): Likewise. Handle CONST_WIDE_INT.
1251 (loc_descriptor): Use wide-int interfaces. Handle CONST_WIDE_INT.
1252 (round_up_to_align): Use wide-int interfaces.
1253 (field_byte_offset): Likewise.
1254 (insert_double): Rename to insert_wide_int. Use wide-int interfaces.
1255 (add_const_value_attribute): Handle CONST_WIDE_INT. Update
1256 CONST_DOUBLE handling. Use wide-int interfaces.
1257 (add_bound_info): Use tree_fits_uhwi_p. Use wide-int interfaces.
1258 (gen_enumeration_type_die): Use add_AT_wide.
1259 (hash_loc_operands): Add case dw_val_class_wide_int.
1260 (compare_loc_operands): Likewise.
1261 * dwarf2out.h: Include wide-int.h.
1262 (wide_int_ptr): New.
1263 (enum dw_val_class): Add dw_val_class_wide_int.
1264 (struct dw_val_struct): Add val_wide.
1265 * emit-rtl.c (const_wide_int_htab): New.
1266 (const_wide_int_htab_hash): New.
1267 (const_wide_int_htab_eq): New.
1268 (lookup_const_wide_int): New.
1269 (const_double_htab_hash): Use wide-int interfaces.
1270 (const_double_htab_eq): Likewise.
1271 (rtx_to_double_int): Conditionally compile for wide-int.
1272 (immed_double_int_const): Rename to immed_wide_int_const and
1273 update for wide-int.
1274 (immed_double_const): Conditionally compile for wide-int.
1275 (init_emit_once): Use wide-int interfaces.
1276 * explow.c (plus_constant): Likewise.
1277 * expmed.c (mask_rtx): Move further up file. Use wide-int interfaces.
1278 (lshift_value): Use wide-int interfaces.
1279 (expand_mult): Likewise.
1280 (choose_multiplier): Likewise.
1281 (expand_smod_pow2): Likewise.
1282 (make_tree): Likewise.
1283 * expr.c (convert_modes): Consolidate handling of constants.
1284 Use wide-int interfaces.
1285 (emit_group_load_1): Add note.
1286 (store_expr): Update comment.
1287 (get_inner_reference): Use wide-int interfaces.
1288 (expand_constructor): Update comment.
1289 (expand_expr_real_2): Use wide-int interfaces.
1290 (expand_expr_real_1): Likewise.
1291 (reduce_to_bit_field_precision): Likewise.
1292 (const_vector_from_tree): Likewise.
1293 * final.c: Include wide-int-print.h.
1294 (output_addr_const): Handle CONST_WIDE_INT. Use CONST_DOUBLE_AS_INT_P.
1295 * fixed-value.c: Include wide-int.h.
1296 (fixed_from_string): Use wide-int interfaces.
1297 (fixed_to_decimal): Likewise.
1298 (fixed_convert_from_real): Likewise.
1299 (real_convert_from_fixed): Likewise.
1300 * fold-const.h (mem_ref_offset): Return an offset_int.
1301 (div_if_zero_remainder): Remove code parameter.
1302 * fold-const.c (div_if_zero_remainder): Remove code parameter.
1303 Use wide-int interfaces.
1304 (may_negate_without_overflow_p): Use wide-int interfaces.
1305 (negate_expr_p): Likewise.
1306 (fold_negate_expr): Likewise.
1307 (int_const_binop_1): Likewise.
1308 (const_binop): Likewise.
1309 (fold_convert_const_int_from_int): Likewise.
1310 (fold_convert_const_int_from_real): Likewise.
1311 (fold_convert_const_int_from_fixed): Likewise.
1312 (fold_convert_const_fixed_from_int): Likewise.
1313 (all_ones_mask_p): Take an unsigned size. Use wide-int interfaces.
1314 (sign_bit_p): Use wide-int interfaces.
1315 (make_range_step): Likewise.
1316 (build_range_check): Likewise. Pass an integer of the correct type
1317 instead of using integer_one_node.
1318 (range_predecessor): Pass an integer of the correct type instead
1319 of using integer_one_node.
1320 (range_successor): Likewise.
1321 (merge_ranges): Likewise.
1322 (unextend): Use wide-int interfaces.
1323 (extract_muldiv_1): Likewise.
1324 (fold_div_compare): Likewise.
1325 (fold_single_bit_test): Likewise.
1326 (fold_sign_changed_comparison): Likewise.
1327 (try_move_mult_to_index): Update calls to div_if_zero_remainder.
1328 (fold_plusminus_mult_expr): Use wide-int interfaces.
1329 (native_encode_int): Likewise.
1330 (native_interpret_int): Likewise.
1331 (fold_unary_loc): Likewise.
1332 (pointer_may_wrap_p): Likewise.
1333 (size_low_cst): Likewise.
1334 (mask_with_tz): Likewise.
1335 (fold_binary_loc): Likewise.
1336 (fold_ternary_loc): Likewise.
1337 (multiple_of_p): Likewise.
1338 (tree_call_nonnegative_warnv_p): Update calls to
1339 tree_int_cst_min_precision and real_from_integer.
1340 (fold_negate_const): Use wide-int interfaces.
1341 (fold_abs_const): Likewise.
1342 (fold_relational_const): Use tree_int_cst_lt.
1343 (round_up_loc): Use wide-int interfaces.
1344 * genemit.c (gen_exp): Add CONST_WIDE_INT case.
1345 * gengenrtl.c (excluded_rtx): Add CONST_WIDE_INT case.
1346 * gengtype.c: Remove include of double-int.h.
1347 (do_typedef): Use wide-int interfaces.
1348 (open_base_files): Add wide-int.h.
1349 (main): Add offset_int and widest_int typedefs.
1350 * gengtype-lex.l: Handle "^".
1351 (CXX_KEYWORD): Add "static".
1352 * gengtype-parse.c (require3): New.
1353 (require_template_declaration): Handle constant template arguments
1354 and nested templates.
1355 * gengtype-state.c: Don't include "double-int.h".
1356 * genpreds.c (write_one_predicate_function): Update comment.
1357 (write_tm_constrs_h): Add check for hval and lval use in
1359 * genrecog.c (validate_pattern): Add CONST_WIDE_INT case.
1360 (add_to_sequence): Likewise.
1361 * gensupport.c (struct std_pred_table): Add const_scalar_int_operand
1362 and const_double_operand.
1363 * gimple.c (preprocess_case_label_vec_for_gimple): Use wide-int
1365 * gimple-fold.c (get_base_constructor): Likewise.
1366 (fold_array_ctor_reference): Likewise.
1367 (fold_nonarray_ctor_reference): Likewise.
1368 (fold_const_aggregate_ref_1): Likewise.
1369 (gimple_val_nonnegative_real_p): Likewise.
1370 (gimple_fold_indirect_ref): Likewise.
1371 * gimple-pretty-print.c (dump_ssaname_info): Likewise.
1372 * gimple-ssa-strength-reduction.c: Include wide-int-print.h.
1373 (struct slsr_cand_d): Change index to be widest_int.
1374 (struct incr_info_d): Change incr to be widest_int.
1375 (alloc_cand_and_find_basis): Use wide-int interfaces.
1376 (slsr_process_phi): Likewise.
1377 (backtrace_base_for_ref): Likewise. Return a widest_int.
1378 (restructure_reference): Take a widest_int instead of a double_int.
1379 (slsr_process_ref): Use wide-int interfaces.
1380 (create_mul_ssa_cand): Likewise.
1381 (create_mul_imm_cand): Likewise.
1382 (create_add_ssa_cand): Likewise.
1383 (create_add_imm_cand): Take a widest_int instead of a double_int.
1384 (slsr_process_add): Use wide-int interfaces.
1385 (slsr_process_cast): Likewise.
1386 (slsr_process_copy): Likewise.
1387 (dump_candidate): Likewise.
1388 (dump_incr_vec): Likewise.
1389 (replace_ref): Likewise.
1390 (cand_increment): Likewise. Return a widest_int.
1391 (cand_abs_increment): Likewise.
1392 (replace_mult_candidate): Take a widest_int instead of a double_int.
1393 (replace_unconditional_candidate): Use wide-int interfaces.
1394 (incr_vec_index): Take a widest_int instead of a double_int.
1395 (create_add_on_incoming_edge): Likewise.
1396 (create_phi_basis): Use wide-int interfaces.
1397 (replace_conditional_candidate): Likewise.
1398 (record_increment): Take a widest_int instead of a double_int.
1399 (record_phi_increments): Use wide-int interfaces.
1400 (phi_incr_cost): Take a widest_int instead of a double_int.
1401 (lowest_cost_path): Likewise.
1402 (total_savings): Likewise.
1403 (analyze_increments): Use wide-int interfaces.
1404 (ncd_with_phi): Take a widest_int instead of a double_int.
1405 (ncd_of_cand_and_phis): Likewise.
1406 (nearest_common_dominator_for_cands): Likewise.
1407 (insert_initializers): Use wide-int interfaces.
1408 (all_phi_incrs_profitable): Likewise.
1409 (replace_one_candidate): Likewise.
1410 (replace_profitable_candidates): Likewise.
1411 * godump.c: Include wide-int-print.h.
1412 (go_output_typedef): Use wide-int interfaces.
1413 * graphite-clast-to-gimple.c (gmp_cst_to_tree): Likewise.
1414 * graphite-sese-to-poly.c (tree_int_to_gmp): Likewise.
1415 (build_loop_iteration_domains): Likewise.
1416 * hooks.h: Include wide-int.h rather than double-int.h.
1417 (hook_bool_dint_dint_uint_bool_true): Delete.
1418 (hook_bool_wint_wint_uint_bool_true): Declare.
1419 * hooks.c (hook_bool_dint_dint_uint_bool_true): Removed.
1420 (hook_bool_wint_wint_uint_bool_true): New.
1421 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): Use wide-int
1423 (ubsan_expand_si_overflow_mul_check): Likewise.
1424 * ipa-devirt.c (get_polymorphic_call_info): Likewise.
1425 * ipa-prop.c (compute_complex_assign_jump_func): Likewise.
1426 (get_ancestor_addr_info): Likewise.
1427 (ipa_modify_call_arguments): Likewise.
1428 * loop-doloop.c (doloop_modify): Likewise.
1429 (doloop_optimize): Likewise.
1430 * loop-iv.c (iv_number_of_iterations): Likewise.
1431 * loop-unroll.c (decide_unroll_constant_iterations): Likewise.
1432 (unroll_loop_constant_iterations): Likewise.
1433 (decide_unroll_runtime_iterations): Likewise.
1434 (unroll_loop_runtime_iterations): Likewise.
1435 (decide_peel_simple): Likewise.
1436 (decide_unroll_stupid): Likewise.
1437 * lto-streamer-in.c (streamer_read_wi): Add.
1438 (input_cfg): Use wide-int interfaces.
1439 (lto_input_tree_1): Likewise.
1440 * lto-streamer-out.c (streamer_write_wi): Add.
1441 (hash_tree): Use wide-int interfaces.
1442 (output_cfg): Likewise.
1443 * Makefile.in (OBJS): Add wide-int.o and wide-int-print.o.
1444 (GTFILES): Add wide-int.h and signop.h.
1445 (TAGS): Look for .cc files too.
1446 * omp-low.c (scan_omp_1_op): Use wide-int interfaces.
1447 * optabs.c (expand_subword_shift): Likewise.
1448 (expand_doubleword_shift): Likewise.
1449 (expand_absneg_bit): Likewise.
1450 (expand_copysign_absneg): Likewise.
1451 (expand_copysign_bit): Likewise.
1452 * postreload.c (reload_cse_simplify_set): Likewise.
1453 * predict.c (predict_iv_comparison): Likewise.
1454 * pretty-print.h: Include wide-int-print.h.
1456 * print-rtl.c (print_rtx): Add CONST_WIDE_INT case.
1457 * print-tree.c: Include wide-int-print.h.
1458 (print_node_brief): Use wide-int interfaces.
1459 (print_node): Likewise.
1460 * read-rtl.c (validate_const_wide_int): New.
1461 (read_rtx_code): Add CONST_WIDE_INT case.
1462 * real.c: Include wide-int.h.
1463 (real_to_integer2): Delete.
1464 (real_to_integer): New function, returning a wide_int.
1465 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
1466 (ten_to_ptwo): Update call to real_from_integer.
1467 (real_digit): Likewise.
1468 * real.h: Include signop.h, wide-int.h and insn-modes.h.
1469 (real_to_integer2, REAL_VALUE_FROM_INT, REAL_VALUE_FROM_UNSIGNED_INT)
1470 (REAL_VALUE_TO_INT): Delete.
1471 (real_to_integer): Declare a wide-int form.
1472 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
1473 * recog.c (const_int_operand): Improve comment.
1474 (const_scalar_int_operand): New.
1475 (const_double_operand): Add a separate definition for CONST_WIDE_INT.
1476 * rtlanal.c (commutative_operand_precedence): Handle CONST_WIDE_INT.
1477 (split_double): Likewise.
1478 * rtl.c (DEF_RTL_EXPR): Handle CONST_WIDE_INT.
1479 (rtx_size): Likewise.
1480 (rtx_alloc_stat_v): New.
1481 (rtx_alloc_stat): Now calls rtx_alloc_stat_v.
1482 (cwi_output_hex): New.
1483 (iterative_hash_rtx): Handle CONST_WIDE_INT.
1484 (cwi_check_failed_bounds): New.
1485 * rtl.def (CONST_WIDE_INT): New.
1486 * rtl.h: Include <utility> and wide-int.h.
1487 (struct hwivec_def): New.
1488 (CWI_GET_NUM_ELEM): New.
1489 (CWI_PUT_NUM_ELEM): New.
1490 (struct rtx_def): Add num_elem and hwiv.
1491 (CASE_CONST_SCALAR_INT): Modify for TARGET_SUPPORTS_WIDE_INT.
1492 (CASE_CONST_UNIQUE): Likewise.
1493 (CASE_CONST_ANY): Likewise.
1494 (CONST_SCALAR_INT_P): Likewise.
1495 (CONST_WIDE_INT_P): New.
1497 (HWIVEC_CHECK): New.
1498 (cwi_check_failed_bounds): New.
1500 (HWIVEC_CHECK): New.
1501 (CONST_WIDE_INT_VEC) New.
1502 (CONST_WIDE_INT_NUNITS) New.
1503 (CONST_WIDE_INT_ELT) New.
1504 (rtx_mode_t): New type.
1505 (wi::int_traits <rtx_mode_t>): New.
1507 (wi::min_value): New.
1508 (wi::max_value): New.
1510 (const_wide_int_alloc): New.
1511 (immed_wide_int_const): New.
1512 * sched-vis.c (print_value): Handle CONST_WIDE_INT.
1513 * sel-sched-ir.c (lhs_and_rhs_separable_p): Update comment.
1514 * signop.h: New file.
1515 * simplify-rtx.c (mode_signbit_p): Handle CONST_WIDE_INT.
1516 (simplify_const_unary_operation): Use wide-int interfaces.
1517 (simplify_binary_operation_1): Likewise.
1518 (simplify_const_binary_operation): Likewise.
1519 (simplify_const_relational_operation): Likewise.
1520 (simplify_immed_subreg): Likewise.
1521 * stmt.c (expand_case): Likewise.
1522 * stor-layout.h (set_min_and_max_values_for_integral_type): Take a
1523 signop rather than a bool.
1524 * stor-layout.c (layout_type): Use wide-int interfaces.
1525 (initialize_sizetypes): Update calls to
1526 set_min_and_max_values_for_integral_type.
1527 (set_min_and_max_values_for_integral_type): Take a signop rather
1528 than a bool. Use wide-int interfaces.
1529 (fixup_signed_type): Update accordingly. Remove
1530 HOST_BITS_PER_DOUBLE_INT limit.
1531 (fixup_unsigned_type): Likewise.
1532 * system.h (STATIC_CONSTANT_P): New.
1533 (STATIC_ASSERT): New.
1534 * target.def (can_use_doloop_p): Take widest_ints rather than
1536 * target.h: Include wide-int.h rather than double-int.h.
1537 * targhooks.h (can_use_doloop_if_innermost): Take widest_ints rather
1539 * targhooks.c (default_cxx_get_cookie_size): Use tree_int_cst_lt
1540 rather than INT_CST_LT_UNSIGNED.
1541 (can_use_doloop_if_innermost): Take widest_ints rather than
1543 * tree-affine.c: Include wide-int-print.h.
1544 (double_int_ext_for_comb): Delete.
1545 (wide_int_ext_for_comb): New.
1546 (aff_combination_zero): Use wide-int interfaces.
1547 (aff_combination_const): Take a widest_int instead of a double_int.
1548 (aff_combination_elt): Use wide-int interfaces.
1549 (aff_combination_scale): Take a widest_int instead of a double_int.
1550 (aff_combination_add_elt): Likewise.
1551 (aff_combination_add_cst): Likewise.
1552 (aff_combination_add): Use wide-int interfaces.
1553 (aff_combination_convert): Likewise.
1554 (tree_to_aff_combination): Likewise.
1555 (add_elt_to_tree): Take a widest_int instead of a double_int.
1556 (aff_combination_to_tree): Use wide-int interfaces.
1557 (aff_combination_remove_elt): Likewise.
1558 (aff_combination_add_product): Take a widest_int instead of
1560 (aff_combination_mult): Use wide-int interfaces.
1561 (aff_combination_expand): Likewise.
1562 (double_int_constant_multiple_p): Delete.
1563 (wide_int_constant_multiple_p): New.
1564 (aff_combination_constant_multiple_p): Take a widest_int pointer
1565 instead of a double_int pointer.
1566 (print_aff): Use wide-int interfaces.
1567 (get_inner_reference_aff): Take a widest_int pointer
1568 instead of a double_int pointer.
1569 (aff_comb_cannot_overlap_p): Take widest_ints instead of double_ints.
1570 * tree-affine.h: Include wide-int.h.
1571 (struct aff_comb_elt): Change type of coef to widest_int.
1572 (struct affine_tree_combination): Change type of offset to widest_int.
1573 (double_int_ext_for_comb): Delete.
1574 (wide_int_ext_for_comb): New.
1575 (aff_combination_const): Use widest_int instead of double_int.
1576 (aff_combination_scale): Likewise.
1577 (aff_combination_add_elt): Likewise.
1578 (aff_combination_constant_multiple_p): Likewise.
1579 (get_inner_reference_aff): Likewise.
1580 (aff_comb_cannot_overlap_p): Likewise.
1581 (aff_combination_zero_p): Use wide-int interfaces.
1582 * tree.c: Include tree.h.
1583 (init_ttree): Use make_int_cst.
1584 (tree_code_size): Removed code for INTEGER_CST case.
1585 (tree_size): Add INTEGER_CST case.
1586 (make_node_stat): Update comment.
1587 (get_int_cst_ext_nunits, build_new_int_cst, build_int_cstu): New.
1588 (build_int_cst_type): Use wide-int interfaces.
1589 (double_int_to_tree): Likewise.
1590 (double_int_fits_to_tree_p): Delete.
1591 (force_fit_type_double): Delete.
1592 (force_fit_type): New.
1593 (int_cst_hash_hash): Use wide-int interfaces.
1594 (int_cst_hash_eq): Likewise.
1595 (build_int_cst_wide): Delete.
1596 (wide_int_to_tree): New.
1597 (cache_integer_cst): Use wide-int interfaces.
1598 (build_low_bits_mask): Likewise.
1599 (cst_and_fits_in_hwi): Likewise.
1600 (real_value_from_int_cst): Likewise.
1601 (make_int_cst_stat): New.
1602 (integer_zerop): Use wide_int interfaces.
1603 (integer_onep): Likewise.
1604 (integer_all_onesp): Likewise.
1605 (integer_pow2p): Likewise.
1606 (integer_nonzerop): Likewise.
1607 (tree_log2): Likewise.
1608 (tree_floor_log2): Likewise.
1609 (tree_ctz): Likewise.
1610 (int_size_in_bytes): Likewise.
1611 (mem_ref_offset): Return an offset_int rather than a double_int.
1612 (build_type_attribute_qual_variant): Use wide_int interfaces.
1613 (type_hash_eq): Likewise
1614 (tree_int_cst_equal): Likewise.
1615 (tree_int_cst_lt): Delete.
1616 (tree_int_cst_compare): Likewise.
1617 (tree_fits_shwi_p): Use wide_int interfaces.
1618 (tree_fits_uhwi_p): Likewise.
1619 (tree_int_cst_sign_bit): Likewise.
1620 (tree_int_cst_sgn): Likewise.
1621 (tree_int_cst_min_precision): Take a signop rather than a bool.
1622 (simple_cst_equal): Use wide_int interfaces.
1623 (compare_tree_int): Likewise.
1624 (iterative_hash_expr): Likewise.
1625 (int_fits_type_p): Likewise. Use tree_int_cst_lt rather than
1627 (get_type_static_bounds): Use wide_int interfaces.
1628 (tree_int_cst_elt_check_failed): New.
1629 (build_common_tree_nodes): Reordered to set prec before filling in
1631 (int_cst_value): Check cst_and_fits_in_hwi.
1632 (widest_int_cst_value): Use wide_int interfaces.
1633 (upper_bound_in_type): Likewise.
1634 (lower_bound_in_type): Likewise.
1635 (num_ending_zeros): Likewise.
1636 (drop_tree_overflow): Likewise.
1637 * tree-call-cdce.c (check_pow): Update call to real_from_integer.
1638 (gen_conditions_for_pow_cst_base): Likewise.
1639 * tree-cfg.c: Include wide-int.h and wide-int-print.h.
1640 (group_case_labels_stmt): Use wide-int interfaces.
1641 (verify_gimple_assign_binary): Likewise.
1642 (print_loop): Likewise.
1643 * tree-chrec.c (tree_fold_binomial): Likewise.
1644 * tree-core.h (struct tree_base): Add int_length.
1645 (struct tree_int_cst): Change rep of value.
1646 * tree-data-ref.c (dr_analyze_innermost): Use wide-int interfaces.
1647 (dr_may_alias_p): Likewise.
1648 (max_stmt_executions_tree): Likewise.
1649 * tree.def (INTEGER_CST): Update comment.
1650 * tree-dfa.c (get_ref_base_and_extent): Use wide-int interfaces.
1651 * tree-dfa.h (get_addr_base_and_unit_offset_1): Likewise.
1652 * tree-dump.c: Include wide-int.h and wide-int-print.h.
1653 (dequeue_and_dump): Use wide-int interfaces.
1654 * tree.h: Include wide-int.h.
1655 (NULL_TREE): Moved to earlier loc in file.
1656 (TREE_INT_CST_ELT_CHECK): New.
1657 (tree_int_cst_elt_check_failed): New.
1659 (TREE_INT_CST): Delete.
1660 (TREE_INT_CST_LOW): Use wide-int interfaces.
1661 (TREE_INT_CST_HIGH): Delete.
1662 (TREE_INT_CST_NUNITS): New.
1663 (TREE_INT_CST_EXT_NUNITS): Likewise.
1664 (TREE_INT_CST_OFFSET_NUNITS): Likewise.
1665 (TREE_INT_CST_ELT): Likewise.
1666 (INT_CST_LT): Delete.
1667 (tree_int_cst_elt_check): New (two forms).
1668 (type_code_size): Update comment.
1669 (make_int_cst_stat, make_int_cst): New.
1670 (tree_to_double_int): Delete.
1671 (double_int_fits_to_tree_p): Delete.
1672 (force_fit_type_double): Delete.
1673 (build_int_cstu): Replace with out-of-line function.
1674 (build_int_cst_wide): Delete.
1675 (tree_int_cst_lt): Define inline.
1676 (tree_int_cst_le): New.
1677 (tree_int_cst_compare): Define inline.
1678 (tree_int_cst_min_precision): Take a signop rather than a bool.
1679 (wi::int_traits <const_tree>): New.
1680 (wi::int_traits <tree>): New.
1681 (wi::extended_tree): New.
1682 (wi::int_traits <wi::extended_tree>): New.
1683 (wi::to_widest): New.
1684 (wi::to_offset): New.
1685 (wi::fits_to_tree_p): New.
1686 (wi::min_value): New.
1687 (wi::max_value): New.
1688 * tree-inline.c (remap_gimple_op_r): Use wide-int interfaces.
1689 (copy_tree_body_r): Likewise.
1690 * tree-object-size.c (compute_object_offset): Likewise.
1691 (addr_object_size): Likewise.
1692 * tree-predcom.c: Include wide-int-print.h.
1693 (struct dref_d): Change type of offset to widest_int.
1694 (dump_dref): Call wide-int printer.
1695 (aff_combination_dr_offset): Use wide-int interfaces.
1696 (determine_offset): Take a widest_int pointer rather than a
1698 (split_data_refs_to_components): Use wide-int interfaces.
1699 (suitable_component_p): Likewise.
1700 (order_drefs): Likewise.
1701 (add_ref_to_chain): Likewise.
1702 (valid_initializer_p): Likewise.
1703 (determine_roots_comp): Likewise.
1704 * tree-pretty-print.c: Include wide-int-print.h.
1705 (dump_generic_node): Use wide-int interfaces.
1706 * tree-sra.c (sra_ipa_modify_expr): Likewise.
1707 * tree-ssa-address.c (addr_for_mem_ref): Likewise.
1708 (move_fixed_address_to_symbol): Likewise.
1709 (move_hint_to_base): Likewise.
1710 (move_pointer_to_base): Likewise.
1711 (move_variant_to_index): Likewise.
1712 (most_expensive_mult_to_index): Likewise.
1713 (addr_to_parts): Likewise.
1714 (copy_ref_info): Likewise.
1715 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
1716 (indirect_refs_may_alias_p): Likewise.
1717 (stmt_kills_ref_p_1): Likewise.
1718 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
1719 * tree-ssa-ccp.c: Update comment at top of file. Include
1721 (struct prop_value_d): Change type of mask to widest_int.
1722 (extend_mask): New function.
1723 (dump_lattice_value): Use wide-int interfaces.
1724 (get_default_value): Likewise.
1725 (set_constant_value): Likewise.
1726 (set_value_varying): Likewise.
1727 (valid_lattice_transition): Likewise.
1728 (set_lattice_value): Likewise.
1729 (value_to_double_int): Delete.
1730 (value_to_wide_int): New.
1731 (get_value_from_alignment): Use wide-int interfaces.
1732 (get_value_for_expr): Likewise.
1733 (do_dbg_cnt): Likewise.
1734 (ccp_finalize): Likewise.
1735 (ccp_lattice_meet): Likewise.
1736 (bit_value_unop_1): Use widest_ints rather than double_ints.
1737 (bit_value_binop_1): Likewise.
1738 (bit_value_unop): Use wide-int interfaces.
1739 (bit_value_binop): Likewise.
1740 (bit_value_assume_aligned): Likewise.
1741 (evaluate_stmt): Likewise.
1742 (ccp_fold_stmt): Likewise.
1743 (visit_cond_stmt): Likewise.
1744 (ccp_visit_stmt): Likewise.
1745 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Likewise.
1746 (constant_pointer_difference): Likewise.
1747 (associate_pointerplus): Likewise.
1748 (combine_conversions): Likewise.
1749 * tree-ssa-loop.h: Include wide-int.h.
1750 (struct tree_niter_desc): Change type of max to widest_int.
1751 * tree-ssa-loop-im.c (mem_refs_may_alias_p): Use wide-int interfaces.
1752 * tree-ssa-loop-ivcanon.c (remove_exits_and_undefined_stmts): Likewise.
1753 (remove_redundant_iv_tests): Likewise.
1754 (canonicalize_loop_induction_variables): Likewise.
1755 * tree-ssa-loop-ivopts.c (alloc_iv): Likewise.
1756 (constant_multiple_of): Take a widest_int pointer instead of
1757 a double_int pointer.
1758 (get_computation_aff): Use wide-int interfaces.
1759 (ptr_difference_cost): Likewise.
1760 (difference_cost): Likewise.
1761 (get_loop_invariant_expr_id): Likewise.
1762 (get_computation_cost_at): Likewise.
1763 (iv_elimination_compare_lt): Likewise.
1764 (may_eliminate_iv): Likewise.
1765 * tree-ssa-loop-niter.h (estimated_loop_iterations): Use widest_int
1766 instead of double_int.
1767 (max_loop_iterations): Likewise.
1768 (max_stmt_executions): Likewise.
1769 (estimated_stmt_executions): Likewise.
1770 * tree-ssa-loop-niter.c: Include wide-int-print.h.
1771 (split_to_var_and_offset): Use wide-int interfaces.
1772 (determine_value_range): Likewise.
1773 (bound_difference_of_offsetted_base): Likewise.
1774 (bounds_add): Take a widest_int instead of a double_int.
1775 (number_of_iterations_ne_max): Use wide-int interfaces.
1776 (number_of_iterations_ne): Likewise.
1777 (number_of_iterations_lt_to_ne): Likewise.
1778 (assert_loop_rolls_lt): Likewise.
1779 (number_of_iterations_lt): Likewise.
1780 (number_of_iterations_le): Likewise.
1781 (number_of_iterations_cond): Likewise.
1782 (number_of_iterations_exit): Likewise.
1783 (finite_loop_p): Likewise.
1784 (derive_constant_upper_bound_assign): Likewise.
1785 (derive_constant_upper_bound): Return a widest_int.
1786 (derive_constant_upper_bound_ops): Likewise.
1787 (do_warn_aggressive_loop_optimizations): Use wide-int interfaces.
1788 (record_estimate): Take a widest_int rather than a double_int.
1789 (record_nonwrapping_iv): Use wide-int interfaces.
1790 (double_int_cmp): Delete.
1791 (wide_int_cmp): New.
1792 (bound_index): Take a widest_int rather than a double_int.
1793 (discover_iteration_bound_by_body_walk): Use wide-int interfaces.
1794 (maybe_lower_iteration_bound): Likewise.
1795 (estimate_numbers_of_iterations_loop): Likewise.
1796 (estimated_loop_iterations): Take a widest_int pointer than than
1797 a double_int pointer.
1798 (estimated_loop_iterations_int): Use wide-int interfaces.
1799 (max_loop_iterations): Take a widest_int pointer than than
1800 a double_int pointer.
1801 (max_loop_iterations_int): Use wide-int interfaces.
1802 (max_stmt_executions): Take a widest_int pointer than than
1803 a double_int pointer.
1804 (estimated_stmt_executions): Likewise.
1805 (n_of_executions_at_most): Use wide-int interfaces.
1806 (scev_probably_wraps_p): Likewise.
1807 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Update calls
1809 * tree-scalar-evolution.c (simplify_peeled_chrec): Use wide-int
1811 * tree-ssanames.c (set_range_info): Use wide_int_refs rather than
1812 double_ints. Adjust for trailing_wide_ints <3> representation.
1813 (set_nonzero_bits): Likewise.
1814 (get_range_info): Return wide_ints rather than double_ints.
1815 Adjust for trailing_wide_ints <3> representation.
1816 (get_nonzero_bits): Likewise.
1817 (duplicate_ssa_name_range_info): Adjust for trailing_wide_ints <3>
1819 * tree-ssanames.h (struct range_info_def): Replace min, max and
1820 nonzero_bits with a trailing_wide_ints <3>.
1821 (set_range_info): Use wide_int_refs rather than double_ints.
1822 (set_nonzero_bits): Likewise.
1823 (get_range_info): Return wide_ints rather than double_ints.
1824 (get_nonzero_bits): Likewise.
1825 * tree-ssa-phiopt.c (jump_function_from_stmt): Use wide-int interfaces.
1826 * tree-ssa-pre.c (phi_translate_1): Likewise.
1827 * tree-ssa-reassoc.c (decrement_power): Use calls to real_from_integer.
1828 (acceptable_pow_call): Likewise.
1829 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Use wide-int
1831 (vn_reference_fold_indirect): Likewise.
1832 (vn_reference_maybe_forwprop_address): Likewise.
1833 (valueize_refs_1): Likewise.
1834 * tree-ssa-structalias.c (get_constraint_for_ptr_offset): Likewise.
1835 * tree-ssa-uninit.c (is_value_included_in): Use wide-int interfaces,
1836 tree_int_cst_lt and tree_int_cst_le.
1837 * tree-streamer-in.c (unpack_ts_base_value_fields): Use wide-int
1839 (streamer_alloc_tree): Likewise.
1840 * tree-streamer-out.c (pack_ts_int_cst_value_fields): Likewise.
1841 (streamer_write_tree_header): Likewise.
1842 (streamer_write_integer_cst): Likewise.
1843 * tree-switch-conversion.c (emit_case_bit_tests): Likewise.
1844 (build_constructors): Likewise.
1845 (array_value_type): Likewise.
1846 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list): Likewise.
1847 (vect_check_gather): Likewise.
1848 * tree-vect-generic.c (build_replicated_const): Likewise.
1849 (expand_vector_divmod): Likewise.
1850 * tree-vect-loop.c (vect_transform_loop): Likewise.
1851 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
1852 (vect_do_peeling_for_alignment): Likewise.
1853 * tree-vect-patterns.c (vect_recog_divmod_pattern): Likewise.
1854 * tree-vrp.c: Include wide-int.h.
1855 (operand_less_p): Use wide-int interfaces and tree_int_cst_lt.
1856 (extract_range_from_assert): Use wide-int interfaces.
1857 (vrp_int_const_binop): Likewise.
1858 (zero_nonzero_bits_from_vr): Take wide_int pointers rather than
1859 double_int pointers.
1860 (ranges_from_anti_range): Use wide-int interfaces.
1861 (quad_int_cmp): Delete.
1862 (quad_int_pair_sort): Likewise.
1863 (extract_range_from_binary_expr_1): Use wide-int interfaces.
1864 (extract_range_from_unary_expr_1): Likewise.
1865 (adjust_range_with_scev): Likewise.
1866 (masked_increment): Take and return wide_ints rather than double_ints.
1867 (register_edge_assert_for_2): Use wide-int interfaces.
1868 (check_array_ref): Likewise.
1869 (search_for_addr_array): Likewise.
1870 (maybe_set_nonzero_bits): Likewise.
1871 (union_ranges): Pass an integer of the correct type instead of
1872 using integer_one_node.
1873 (intersect_ranges): Likewise.
1874 (simplify_truth_ops_using_ranges): Likewise.
1875 (simplify_bit_ops_using_ranges): Use wide-int interfaces.
1876 (range_fits_type_p): Likewise.
1877 (simplify_cond_using_ranges): Likewise. Take a signop rather than
1879 (simplify_conversion_using_ranges): Use wide-int interfaces.
1880 (simplify_float_conversion_using_ranges): Likewise.
1881 (vrp_finalize): Likewise.
1882 * value-prof.c (gimple_divmod_fixed_value_transform): Likewise.
1883 (gimple_stringops_transform): Likewise.
1884 * varasm.c (decode_addr_const): Likewise.
1885 (const_hash_1): Likewise.
1886 (const_rtx_hash_1): Likewise
1887 (output_constant): Likewise.
1888 (array_size_for_constructor): Likewise.
1889 (output_constructor_regular_field): Likewise.
1890 (output_constructor_bitfield): Likewise.
1891 * var-tracking.c (loc_cmp): Handle CONST_WIDE_INT.
1892 * mkconfig.sh: Include machmode.h to pick up BITS_PER_UNIT for
1894 * gencheck.c: Define BITS_PER_UNIT.
1897 * wide-int-print.cc: New.
1898 * wide-int-print.h: New.
1900 2014-05-06 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1902 * config/avr/avr.c (avr_can_eliminate): Mark unused argument.
1904 2014-05-06 Richard Biener <rguenther@suse.de>
1906 * tree-pass.h (TODO_verify_ssa, TODO_verify_flow,
1907 TODO_verify_stmts, TODO_verify_rtl_sharing): Remove.
1908 (TODO_verify_all): Adjust.
1909 * asan.c: Remove references to TODO_verify_ssa, TODO_verify_flow,
1910 TODO_verify_stmts and TODO_verify_rtl_sharing.
1911 * bb-reorder.c: Likewise.
1912 * cfgexpand.c: Likewise.
1913 * cprop.c: Likewise.
1915 * function.c: Likewise.
1916 * fwprop.c: Likewise.
1918 * gimple-ssa-isolate-paths.c: Likewise.
1919 * gimple-ssa-strength-reduction.c: Likewise.
1920 * ipa-split.c: Likewise.
1921 * loop-init.c: Likewise.
1922 * loop-unroll.c: Likewise.
1923 * lower-subreg.c: Likewise.
1924 * modulo-sched.c: Likewise.
1925 * postreload-gcse.c: Likewise.
1926 * predict.c: Likewise.
1927 * recog.c: Likewise.
1928 * sched-rgn.c: Likewise.
1929 * store-motion.c: Likewise.
1930 * tracer.c: Likewise.
1931 * trans-mem.c: Likewise.
1932 * tree-call-cdce.c: Likewise.
1933 * tree-cfg.c: Likewise.
1934 * tree-cfgcleanup.c: Likewise.
1935 * tree-complex.c: Likewise.
1936 * tree-eh.c: Likewise.
1937 * tree-emutls.c: Likewise.
1938 * tree-if-conv.c: Likewise.
1939 * tree-into-ssa.c: Likewise.
1940 * tree-loop-distribution.c: Likewise.
1941 * tree-object-size.c: Likewise.
1942 * tree-parloops.c: Likewise.
1943 * tree-pass.h: Likewise.
1944 * tree-sra.c: Likewise.
1945 * tree-ssa-ccp.c: Likewise.
1946 * tree-ssa-copy.c: Likewise.
1947 * tree-ssa-copyrename.c: Likewise.
1948 * tree-ssa-dce.c: Likewise.
1949 * tree-ssa-dom.c: Likewise.
1950 * tree-ssa-dse.c: Likewise.
1951 * tree-ssa-forwprop.c: Likewise.
1952 * tree-ssa-ifcombine.c: Likewise.
1953 * tree-ssa-loop-ch.c: Likewise.
1954 * tree-ssa-loop-ivcanon.c: Likewise.
1955 * tree-ssa-loop.c: Likewise.
1956 * tree-ssa-math-opts.c: Likewise.
1957 * tree-ssa-phiopt.c: Likewise.
1958 * tree-ssa-phiprop.c: Likewise.
1959 * tree-ssa-pre.c: Likewise.
1960 * tree-ssa-reassoc.c: Likewise.
1961 * tree-ssa-sink.c: Likewise.
1962 * tree-ssa-strlen.c: Likewise.
1963 * tree-ssa-tail-merge.c: Likewise.
1964 * tree-ssa-uncprop.c: Likewise.
1965 * tree-switch-conversion.c: Likewise.
1966 * tree-tailcall.c: Likewise.
1967 * tree-vect-generic.c: Likewise.
1968 * tree-vectorizer.c: Likewise.
1969 * tree-vrp.c: Likewise.
1971 * var-tracking.c: Likewise.
1972 * bt-load.c: Likewise.
1973 * cfgcleanup.c: Likewise.
1974 * combine-stack-adj.c: Likewise.
1975 * combine.c: Likewise.
1976 * compare-elim.c: Likewise.
1977 * config/epiphany/resolve-sw-modes.c: Likewise.
1978 * config/i386/i386.c: Likewise.
1979 * config/mips/mips.c: Likewise.
1980 * config/s390/s390.c: Likewise.
1981 * config/sh/sh_treg_combine.cc: Likewise.
1982 * config/sparc/sparc.c: Likewise.
1985 * final.c: Likewise.
1986 * ifcvt.c: Likewise.
1987 * mode-switching.c: Likewise.
1988 * passes.c: Likewise.
1989 * postreload.c: Likewise.
1991 * reg-stack.c: Likewise.
1992 * regcprop.c: Likewise.
1993 * regrename.c: Likewise.
1996 2014-05-06 Richard Biener <rguenther@suse.de>
1999 * bitmap.c (debug_bitmap): Dump to stderr, not stdout.
2000 * tree-ssa-structalias.c (dump_solution_for_var): Likewise.
2002 2014-05-05 Jan Hubicka <hubicka@ucw.cz>
2005 * ipa-devirt.c (get_class_context): Allow POD to change to non-POD.
2007 2014-05-05 Radovan Obradovic <robradovic@mips.com>
2008 Tom de Vries <tom@codesourcery.com>
2010 * target.def (call_fusage_contains_non_callee_clobbers): New
2012 * doc/tm.texi.in (@node Stack and Calling): Add Miscellaneous Register
2014 (@node Miscellaneous Register Hooks): New node.
2015 (@hook TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS): New hook.
2016 * doc/tm.texi: Regenerate.
2018 2014-05-05 Marek Polacek <polacek@redhat.com>
2021 * opts.c (common_handle_option): Call error_at instead of warning_at.
2023 2014-05-05 Richard Biener <rguenther@suse.de>
2025 * passes.c (execute_function_todo): Don't reset TODO_verify_ssa
2026 from last_verified if update_ssa ran. Move TODO_verify_rtl_sharing
2027 under the TODO_verify_il umbrella.
2029 2014-05-05 Richard Biener <rguenther@suse.de>
2031 * passes.c (execute_function_todo): Move TODO_verify_flow under
2032 the TODO_verify_ul umbrella.
2034 2014-05-05 Richard Biener <rguenther@suse.de>
2037 * fold-const.c (fold_binary_loc): Consistently avoid canonicalizing
2038 X & CST away from a CST that is the mask of a mode.
2040 2014-05-05 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2042 * config/picochip/picochip-protos.h (picochip_regno_nregs): Change
2043 int argument to enum machine_mode.
2044 (picochip_class_max_nregs): Ditto.
2045 * config/picochip/picochip.c (picochip_regno_nregs): Ditto.
2046 (picochip_class_max_nregs): Ditto.
2048 2014-05-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2050 * target.def: Add new target hook.
2051 * doc/tm.texi: Regenerate.
2052 * targhooks.h (default_keep_leaf_when_profiled): Add prototype.
2053 * targhooks.c (default_keep_leaf_when_profiled): New function.
2055 * config/s390/s390.c (s390_keep_leaf_when_profiled): New function.
2056 (TARGET_KEEP_LEAF_WHEN_PROFILED): Define.
2058 2014-05-05 Bin Cheng <bin.cheng@arm.com>
2060 PR tree-optimization/60363
2061 * gcc/tree-ssa-threadupdate.c (get_value_locus_in_path): New.
2062 (copy_phi_args): New parameters. Call get_value_locus_in_path.
2063 (update_destination_phis): New parameter.
2064 (create_edge_and_update_destination_phis): Ditto.
2065 (ssa_fix_duplicate_block_edges): Pass new arguments.
2066 (thread_single_edge): Ditto.
2068 2014-05-04 Peter Bergner <bergner@vnet.ibm.com>
2070 * config/rs6000/rs6000.h (RS6000_BTM_HARD_FLOAT): New define.
2071 (RS6000_BTM_COMMON): Add RS6000_BTM_HARD_FLOAT.
2072 (TARGET_EXTRA_BUILTINS): Add TARGET_HARD_FLOAT.
2073 * config/rs6000/rs6000-builtin.def (BU_MISC_1):
2074 Use RS6000_BTM_HARD_FLOAT.
2075 (BU_MISC_2): Likewise.
2076 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Handle
2077 RS6000_BTM_HARD_FLOAT.
2078 (rs6000_option_override_internal): Enforce -mhard-float if -mhard-dfp
2080 (rs6000_invalid_builtin): Add hard floating builtin support.
2081 (rs6000_expand_builtin): Relax the gcc_assert to allow the new
2082 hard float builtins.
2083 (rs6000_builtin_mask_names): Add RS6000_BTM_HARD_FLOAT.
2085 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
2087 * config/sh/sh_optimize_sett_clrt.cc (sh_optimize_sett_clrt::execute):
2088 Add missing function* argument.
2090 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
2092 * lra-constraints.c (valid_address_p): Move earlier in file.
2093 Add a constraint argument to the address_info version.
2094 (satisfies_memory_constraint_p): New function.
2095 (satisfies_address_constraint_p): Likewise.
2096 (process_alt_operands, curr_insn_transform): Use them.
2097 (process_address): Pass the constraint to valid_address_p when
2098 checking address operands.
2100 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
2102 * config/mips/mips.c (mips_isa_rev): New variable.
2103 (mips_set_architecture): Set it.
2104 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Set __mips_isa_rev
2106 (ISA_HAS_MUL3, ISA_HAS_FP_CONDMOVE, ISA_HAS_8CC, ISA_HAS_FP4)
2107 (ISA_HAS_PAIRED_SINGLE, ISA_HAS_MADD_MSUB, ISA_HAS_FP_RECIP_RSQRT)
2108 (ISA_HAS_CLZ_CLO, ISA_HAS_ROR, ISA_HAS_WSBH, ISA_HAS_PREFETCH)
2109 (ISA_HAS_SEB_SEH, ISA_HAS_EXT_INS, ISA_HAS_MXHC1)
2110 (ISA_HAS_HILO_INTERLOCKS, ISA_HAS_SYNCI, MIN_FPRS_PER_FMT): Reexpress
2111 conditions in terms of mips_isa_rev.
2112 (mips_isa_rev): Declare.
2114 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
2116 * config/sh/sh-mem.cc: Use tabs instead of spaces.
2117 (prob_unlikely, prob_likely): Make variables const.
2119 2014-05-03 Denis Chertykov <chertykov@gmail.com>
2121 * config/avr/avr.c (avr_adjust_insn_length): Handle JUMP_TABLE_DATA.
2123 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
2125 * config/sh/sh.h (SH_ASM_SPEC): Handle m1, m2*, m3* and m4* cases.
2127 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
2129 * config/sh/sh.h (ROUND_ADVANCE): Delete macro.
2130 (ROUND_REG, PASS_IN_REG_P): Move and rename macros to ...
2131 * config/sh/sh.c (sh_round_reg, sh_pass_in_reg_p): ... these new
2133 (sh_arg_partial_bytes, sh_function_arg, sh_function_arg_advance,
2134 sh_setup_incoming_varargs): Replace usage of PASS_IN_REG_P with
2136 Replace usage of ROUND_REG with sh_round_reg.
2137 Use CEIL instead of ROUND_ADVANCE.
2139 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
2142 * config/sh/sh.c: Include stdlib headers before everything else.
2144 2014-05-02 Jakub Jelinek <jakub@redhat.com>
2146 * gimplify.c (gimplify_adjust_omp_clauses_1): Handle
2147 GOVD_FIRSTPRIVATE | GOVD_LASTPRIVATE.
2148 (gimplify_adjust_omp_clauses): Simd region is never
2149 directly nested in combined parallel. Instead, for linear
2150 with copyin/copyout, if in combined for simd loop, make decl
2151 firstprivate/lastprivate on OMP_FOR.
2152 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
2153 expand_omp_for_static_chunk): When setting endvar, also set
2154 fd->loop.v to the same value.
2156 2014-05-02 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
2158 * hwint.h (zext_hwi): Fix signed overflow for prec == 63.
2160 2014-05-02 Alan Lawrence <alan.lawrence@arm.com>
2162 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_1): Tidy bit-flip
2165 2014-05-02 Marek Polacek <polacek@redhat.com>
2167 * doc/invoke.texi: Describe -fsanitize=float-divide-by-zero.
2169 2014-05-02 Kito Cheng <kito@0xlab.org>
2171 * defaults.h (HONOR_REG_ALLOC_ORDER): Change HONOR_REG_ALLOC_ORDER
2172 to a C expression marco.
2173 * ira-color.c (HONOR_REG_ALLOC_ORDER) : Ditto.
2174 * config/arm/arm.h (HONOR_REG_ALLOC_ORDER): Ditto.
2175 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Ditto.
2176 * doc/tm.texi (HONOR_REG_ALLOC_ORDER): Update document for
2177 HONOR_REG_ALLOC_ORDER.
2178 * doc/tm.texi.in (HONOR_REG_ALLOC_ORDER): Ditto.
2180 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2182 * config/arc/arc.c (TARGET_LRA_P): Undef before redefine.
2184 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2186 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
2188 2014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
2190 * tree-if-conv.c (is_cond_scalar_reduction): New function.
2191 (convert_scalar_cond_reduction): Likewise.
2192 (predicate_scalar_phi): Add recognition and transformation
2193 of simple conditioanl reduction to be vectorizable.
2195 2014-05-01 Marek Polacek <polacek@redhat.com>
2198 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
2200 2014-04-30 Alan Lawrence <alan.lawrence@arm.com>
2202 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
2203 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
2204 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
2205 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
2206 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
2207 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
2208 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
2209 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
2211 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
2213 * config/arc/arc.opt (mlra): Move comment above option name
2214 to avoid mis-parsing as language options.
2216 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2218 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
2219 * config/sol2.h: ... here.
2220 * config/sol2-10.h: Remove.
2222 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
2223 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
2224 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
2225 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
2226 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
2227 * config/sol2.h: ... here.
2228 (SECTION_NAME_FORMAT): Don't redefine.
2229 (STARTFILE_ARCH32_SPEC): Rename to ...
2230 (STARTFILE_ARCH_SPEC): ... this.
2231 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
2232 * config/sparc/sol2.h: ... here.
2233 (SECTION_NAME_FORMAT): Don't undef.
2234 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
2235 (SUBTARGET_EXTRA_SPECS): Remove.
2236 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
2238 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
2239 (MD_STARTFILE_PREFIX): Remove.
2240 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
2241 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
2242 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
2243 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
2244 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
2245 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
2246 * config/i386/sol2.h: ... here.
2247 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
2248 * config/i386/sol2-bi.h: Remove.
2249 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
2250 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
2252 * config/i386/t-sol2-64: Rename to ...
2253 * config/i386/t-sol2: ... this.
2254 * config/sparc/t-sol2-64: Rename to ...
2255 * config/sparc/t-sol2: ... this.
2257 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
2258 sol2_tm_file_head, sol2_tm_file_tail.
2259 Include ${cpu_type}/sol2.h before sol2.h.
2261 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
2262 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
2263 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
2264 Reflect i386/t-sol2-64 renaming.
2265 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
2266 Reflect sparc/t-sol2-64 renaming.
2268 2014-04-30 Richard Biener <rguenther@suse.de>
2270 * passes.c (execute_function_todo): Move TODO_verify_stmts
2271 and TODO_verify_ssa under the TODO_verify_il umbrella.
2272 * tree-ssa.h (verify_ssa): Adjust prototype.
2273 * tree-ssa.c (verify_ssa): Add parameter to tell whether
2274 we should verify SSA operands.
2275 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
2276 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
2277 whether we should verify whether not throwing stmts have EH info.
2278 * graphite-scop-detection.c (create_sese_edges): Adjust.
2279 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
2280 * tree-eh.c (lower_try_finally_switch): Do not add the
2281 default case label twice.
2283 2014-04-30 Marek Polacek <polacek@redhat.com>
2285 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
2286 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
2287 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
2288 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
2290 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
2292 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
2293 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
2294 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
2295 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
2296 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
2297 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
2298 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
2299 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
2301 2014-04-29 David Malcolm <dmalcolm@redhat.com>
2303 * tree-cfg.c (dump_function_to_file): Dump the return type of
2304 functions, in a line to itself before the function body, mimicking
2305 the layout of a C function.
2307 2014-04-29 Jakub Jelinek <jakub@redhat.com>
2309 PR tree-optimization/60971
2310 * tree-tailcall.c (process_assignment): Reject conversions which
2313 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
2315 * calls.c (initialize_argument_information): Always treat
2316 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
2317 (expand_call): Likewise.
2318 (emit_library_call_calue_1): Likewise.
2319 * expr.c (PUSH_ARGS_REVERSED): Do not define.
2320 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
2323 2014-04-29 Nick Clifton <nickc@redhat.com>
2325 * config/msp430/msp430.md (umulsidi): Fix typo.
2326 (mulhisi3): Enable even inside interrupt handlers.
2327 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
2328 bigger return address pushed in large mode.
2330 2014-04-29 Nick Clifton <nickc@redhat.com>
2332 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
2333 (arc_init_reg_tables): Use a machine_mode enum to iterate over
2335 * config/m32r/m32r.c (init_reg_tables): Likewise.
2336 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
2337 enum to hold the modes.
2339 2014-04-29 Richard Biener <rguenther@suse.de>
2341 * dominance.c (free_dominance_info): Add overload with
2343 (dom_info_state): Likewise.
2344 (dom_info_available_p): Likewise.
2345 * basic-block.h (free_dominance_info, dom_info_state,
2346 dom_info_available_p): Declare overloads.
2347 * passes.c (execute_function_todo): Verify that verifiers
2348 don't change dominator info state. Drop dominator info
2349 for IPA pass invocations.
2350 * cgraph.c (release_function_body): Restore asserts that
2351 dominator information is released.
2353 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
2355 * doc/invoke.texi: Fix typo.
2356 * tree-vrp.c: Fix typos.
2357 * gimple.c (infer_nonnull_range): Reorder operands of an && condition.
2359 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2361 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
2363 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
2365 * config/aarch64/aarch64-builtins.c
2366 (aarch64_types_storestruct_lane_qualifiers): New.
2367 (TYPES_STORESTRUCT_LANE): Likewise.
2368 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
2369 (st3_lane): Likewise.
2370 (st4_lane): Likewise.
2371 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
2372 (vec_store_lanesci_lane<mode>): Likewise.
2373 (vec_store_lanesxi_lane<mode>): Likewise.
2374 (aarch64_st2_lane<VQ:mode>): Likewise.
2375 (aarch64_st3_lane<VQ:mode>): Likewise.
2376 (aarch64_st4_lane<VQ:mode>): Likewise.
2377 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
2378 * config/aarch64/arm_neon.h
2379 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
2380 use new macro arguments.
2381 (__ST3_LANE_FUNC): Likewise.
2382 (__ST4_LANE_FUNC): Likewise.
2383 * config/aarch64/iterators.md (V_TWO_ELEM): New.
2384 (V_THREE_ELEM): Likewise.
2385 (V_FOUR_ELEM): Likewise.
2387 2014-04-28 David Malcolm <dmalcolm@redhat.com>
2389 * doc/gimple.texi: Replace the description of the now-defunct
2390 union gimple_statement_d with a diagram showing the
2391 gimple_statement_base class hierarchy and its relationships to
2392 the GSS_ and GIMPLE_ enums.
2394 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
2396 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
2397 * config/aarch64/aarch64.c
2398 (aarch64_cannot_change_mode_class): Weaken conditions.
2399 (aarch64_modes_tieable_p): New.
2400 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
2402 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
2404 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2405 (loadsync_<mode>): Change mode.
2406 (load_quadpti, store_quadpti): New.
2407 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2408 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2410 2014-04-28 Martin Jambor <mjambor@suse.cz>
2412 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
2413 same alias type as the original statement.
2414 (subreplacement_assignment_data): New type.
2415 (handle_unscalarized_data_in_subtree): New type of parameter,
2416 generate new memory accesses with same alias type as the original
2418 (load_assign_lhs_subreplacements): Likewise.
2419 (sra_modify_constructor_assign): Generate new memory accesses with
2420 same alias type as the original statement.
2422 2014-04-28 Richard Biener <rguenther@suse.de>
2424 * tree-pass.h (TODO_verify_il): Define.
2425 (TODO_verify_all): Complete properly.
2426 * passes.c (execute_function_todo): Move existing loop-closed
2427 SSA verification under TODO_verify_il.
2428 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
2429 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
2430 Fix tree sharing issue.
2432 2014-04-28 Richard Biener <rguenther@suse.de>
2435 * builtins.def (DEF_C11_BUILTIN): Add.
2436 (BUILT_IN_ALIGNED_ALLOC): Likewise.
2437 * coretypes.h (enum function_class): Add function_c11_misc.
2438 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
2439 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
2440 (call_may_clobber_ref_p_1): Likewise.
2441 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
2442 (mark_all_reaching_defs_necessary_1): Likewise.
2443 (propagate_necessity): Likewise.
2444 (eliminate_unnecessary_stmts): Likewise.
2445 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
2447 2014-04-28 Richard Biener <rguenther@suse.de>
2449 * tree-vrp.c (vrp_var_may_overflow): Remove.
2450 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
2451 with overflow immediately bump to one before that value and
2452 let iteration figure out overflow status.
2454 2014-04-28 Richard Biener <rguenther@suse.de>
2456 * configure.ac: Do valgrind header checks unconditionally.
2457 Add --enable-valgrind-annotations.
2458 * system.h: Guard valgrind header inclusion with
2459 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
2460 * alloc-pool.c (pool_alloc, pool_free): Use
2461 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
2462 to guard possibly dead code.
2463 * config.in: Regenerated.
2464 * configure: Likewise.
2466 2014-04-28 Jeff Law <law@redhat.com>
2468 PR tree-optimization/60902
2469 * tree-ssa-threadedge.c
2470 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
2471 over real defs when invalidating outputs from statements that do not
2472 produce useful outputs for threading.
2474 2014-04-28 Richard Biener <rguenther@suse.de>
2476 PR tree-optimization/60979
2477 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
2478 SCOPs that end in a block with a successor with abnormal
2481 2014-04-28 Richard Biener <rguenther@suse.de>
2483 * tree-pass.h (execute_pass_list): Adjust prototype.
2484 * passes.c (pass_manager::execute_early_local_passes): Adjust.
2485 (do_per_function): Change callback signature, push all actual
2486 work to the callbals.
2487 (do_per_function_toporder): Likewise.
2488 (execute_function_dump): Adjust.
2489 (execute_function_todo): Likewise.
2490 (clear_last_verified): Likewise.
2491 (verify_curr_properties): Likewise.
2492 (update_properties_after_pass): Likewise.
2493 (execute_pass_list_1): Split out from ...
2494 (execute_pass_list): ... here. Adjust.
2495 (execute_ipa_pass_list): Likewise.
2496 * cgraphunit.c (cgraph_add_new_function): Adjust.
2497 (analyze_function): Likewise.
2498 (expand_function): Likewise.
2499 * cgraph.c (release_function_body): Free dominance info
2500 here instead of asserting it was magically freed elsewhere.
2502 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
2504 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
2505 * configure: Regenerate.
2506 * config/sparc/sparc.opt (muser-mode): New option.
2507 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
2509 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
2510 * doc/invoke.texi (SPARC options): Document -muser-mode.
2512 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
2514 * cselib.c (find_slot_memmode): Delete.
2515 (cselib_hasher): Change compare_type to a struct.
2516 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
2518 (preserve_constants_and_equivs): Adjust for new compare_type.
2519 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
2520 (wrap_constant): Delete.
2521 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
2523 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
2525 * doc/install.texi (Building with profile feedback): Remove
2528 2014-04-26 Tom de Vries <tom@codesourcery.com>
2530 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
2533 2014-04-25 Cary Coutant <ccoutant@google.com>
2536 * dwarf2out.c (should_move_die_to_comdat): A type definition
2537 can contain a subprogram definition, but don't move it to a
2539 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
2540 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
2542 (clone_tree_hash): Rename to...
2543 (clone_tree_partial): ...this; change callers. Copy
2544 DW_TAG_subprogram DIEs as declarations.
2545 (copy_decls_walk): Don't copy children of a declaration into a
2548 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
2551 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
2554 2014-04-25 Jiong Wang <jiong.wang@arm.com>
2556 * config/arm/predicates.md (call_insn_operand): Add long_call check.
2557 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
2559 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
2562 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2564 * config/arm/arm.c (arm_cortex_a8_tune): Initialise T16-related fields.
2566 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2568 PR tree-optimization/60930
2569 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
2570 creating a multiply candidate by folding two constant
2571 multiplicands when the result overflows.
2573 2014-04-25 Jakub Jelinek <jakub@redhat.com>
2575 PR tree-optimization/60960
2576 * tree-vect-generic.c (expand_vector_operation): Only call
2577 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
2579 2014-04-25 Tom de Vries <tom@codesourcery.com>
2581 * expr.c (clobber_reg_mode): New function.
2582 * expr.h (clobber_reg): New function.
2584 2014-04-25 Tom de Vries <tom@codesourcery.com>
2586 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
2589 2014-04-25 Radovan Obradovic <robradovic@mips.com>
2590 Tom de Vries <tom@codesourcery.com>
2592 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
2594 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
2595 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
2596 new argument to find_all_hard_reg_sets call.
2598 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2600 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
2601 Use HOST_WIDE_INT_C for mask literal.
2602 (aarch_rev16_shleft_mask_imm_p): Likewise.
2604 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
2607 * config/sparc/sparc.md (ashlsi3_extend): Delete.
2609 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
2611 PR preprocessor/56540
2612 * config/i386/i386-c.c (ix86_target_macros): Define
2613 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
2615 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2617 * configure.ac (tga_func): Remove.
2618 (LIB_TLS_SPEC): Remove.
2619 * configure: Regenerate.
2620 * config.in: Regenerate.
2621 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
2623 2014-04-25 Richard Biener <rguenther@suse.de>
2626 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
2627 call stmt use/clobber sets during stmt walk instead of
2628 walking the possibly incomplete set of caller edges.
2630 2014-04-25 Richard Biener <rguenther@suse.de>
2633 * passes.c (apply_ipa_transforms): Inline into only caller ...
2634 (execute_one_pass): ... here. Properly bring in function
2635 bodies for nodes we want to apply IPA transforms to.
2637 2014-04-24 Cong Hou <congh@google.com>
2639 PR tree-optimization/60896
2640 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
2641 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
2642 (vect_mark_pattern_stmts): Set the def type of all statements in
2643 PATTERN_DEF_SEQ as vect_internal_def.
2645 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
2647 * doc/extend.texi (PowerPC Built-in Functions): Document new
2648 powerpc extended divide, bcd, pack/unpack 128-bit, builtin functions.
2649 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
2651 * config/rs6000/predicates.md (const_0_to_3_operand): New
2652 predicate to match 0..3 integer constants.
2654 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
2655 to support adding miscellaneous builtin functions.
2656 (BU_DFP_MISC_2): Likewise.
2657 (BU_P7_MISC_1): Likewise.
2658 (BU_P7_MISC_2): Likewise.
2659 (BU_P8V_MISC_3): Likewise.
2660 (BU_MISC_1): Likewise.
2661 (BU_MISC_2): Likewise.
2662 (DIVWE): Add extended divide builtin functions.
2665 (DIVWEUO): Likewise.
2669 (DIVDEUO): Likewise.
2670 (DXEX): Add decimal floating-point builtin functions.
2673 (DDEDPDQ): Likewise.
2675 (DENBCDQ): Likewise.
2682 (CDTBCD): Add new BCD builtin functions.
2686 (BCDADD_LT): Likewise.
2687 (BCDADD_EQ): Likewise.
2688 (BCDADD_GT): Likewise.
2689 (BCDADD_OV): Likewise.
2691 (BCDSUB_LT): Likewise.
2692 (BCDSUB_EQ): Likewise.
2693 (BCDSUB_GT): Likewise.
2694 (BCDSUB_OV): Likewise.
2695 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
2696 (UNPACK_TD): Likewise.
2697 (PACK_TF): Likewise.
2698 (UNPACK_TF): Likewise.
2699 (UNPACK_TF_0): Likewise.
2700 (UNPACK_TF_1): Likewise.
2701 (PACK_V1TI): Likewise.
2702 (UNPACK_V1TI): Likewise.
2704 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
2705 support for decimal floating point builtin functions.
2706 (rs6000_expand_ternop_builtin): Add checks for the new builtin
2707 functions that take constant arguments.
2708 (rs6000_invalid_builtin): Add decimal floating point builtin support.
2709 (rs6000_init_builtins): Setup long double, _Decimal64, and
2710 _Decimal128 types for new builtin functions.
2711 (builtin_function_type): Set the unsigned flags appropriately for
2712 the new builtin functions.
2713 (rs6000_opt_masks): Add support for decimal floating point builtin
2716 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
2717 floating point builtin functions.
2718 (RS6000_BTM_COMMON): Likewise.
2719 (RS6000_BTI_long_double): Likewise.
2720 (RS6000_BTI_dfloat64): Likewise.
2721 (RS6000_BTI_dfloat128): Likewise.
2722 (long_double_type_internal_node): Likewise.
2723 (dfloat64_type_internal_node): Likewise.
2724 (dfloat128_type_internal_node): Likewise.
2726 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
2727 2.07 bcd arithmetic instructions.
2728 (UNSPEC_BCDSUB): Likewise.
2729 (UNSPEC_BCD_OVERFLOW): Likewise.
2730 (UNSPEC_BCD_ADD_SUB): Likewise.
2731 (bcd_add_sub): Likewise.
2732 (BCD_TEST): Likewise.
2733 (bcd<bcd_add_sub>): Likewise.
2734 (bcd<bcd_add_sub>_test): Likewise.
2735 (bcd<bcd_add_sub>_test2): Likewise.
2736 (bcd<bcd_add_sub>_<code>): Likewise.
2737 (peephole2 for combined bcd ops): Likewise.
2739 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
2740 decimal floating point builtin functions.
2741 (UNSPEC_DENBCD): Likewise.
2742 (UNSPEC_DXEX): Likewise.
2743 (UNSPEC_DIEX): Likewise.
2744 (UNSPEC_DSCLI): Likewise.
2745 (UNSPEC_DSCRI): Likewise.
2746 (D64_D128): Likewise.
2747 (dfp_suffix): Likewise.
2748 (dfp_ddedpd_<mode>): Likewise.
2749 (dfp_denbcd_<mode>): Likewise.
2750 (dfp_dxex_<mode>): Likewise.
2751 (dfp_diex_<mode>): Likewise.
2752 (dfp_dscli_<mode>): Likewise.
2753 (dfp_dscri_<mode>): Likewise.
2755 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
2757 (UNSPEC_CDTBCD): Likewise.
2758 (UNSPEC_CBCDTD): Likewise.
2759 (UNSPEC_DIVE): Add support for new extended divide builtin functions.
2760 (UNSPEC_DIVEO): Likewise.
2761 (UNSPEC_DIVEU): Likewise.
2762 (UNSPEC_DIVEUO): Likewise.
2763 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
2764 pack/unpack 128-bit types.
2765 (UNSPEC_PACK_128BIT): Likewise.
2766 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
2767 (udiv<mode>3): Use idiv_ldiv mode attribute.
2768 (div<mode>3): Likewise.
2769 (addg6s): Add new BCD builtin functions.
2772 (UNSPEC_DIV_EXTEND): Add support for new extended divide instructions.
2773 (div_extend): Likewise.
2774 (div<div_extend>_<mode>"): Likewise.
2775 (FP128_64): Add support for new builtin functions to pack/unpack
2777 (unpack<mode>): Likewise.
2778 (unpacktf_0): Likewise.
2779 (unpacktf_1): Likewise.
2780 (unpack<mode>_dm): Likewise.
2781 (unpack<mode>_nodm): Likewise.
2782 (pack<mode>): Likewise.
2783 (unpackv1ti): Likewise.
2784 (packv1ti): Likewise.
2786 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
2788 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
2791 2014-04-24 Jakub Jelinek <jakub@redhat.com>
2793 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
2794 * gimplify.c (omp_is_private): Change last argument's type to int.
2795 Only diagnose lastprivate if the simd argument is 1, only diagnose
2796 linear if the simd argument is 2.
2797 (gimplify_omp_for): Adjust omp_is_private callers. When adding
2798 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
2799 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
2800 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
2801 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
2802 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
2803 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
2804 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
2805 * tree-nested.c (convert_nonlocal_omp_clauses,
2806 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
2808 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
2811 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
2814 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
2816 * flag-types.h (enum ivar_visibility): Add.
2818 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
2820 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
2821 function * argument.
2823 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
2825 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
2827 2014-04-24 Radovan Obradovic <robradovic@mips.com>
2828 Tom de Vries <tom@codesourcery.com>
2830 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
2831 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
2833 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
2834 * emit-rtl.c (try_split): Same.
2836 2014-04-24 Radovan Obradovic <robradovic@mips.com>
2837 Tom de Vries <tom@codesourcery.com>
2839 * common.opt (fuse-caller-save): New option.
2841 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
2843 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of
2844 elements for big-endian.
2846 2014-04-24 Richard Biener <rguenther@suse.de>
2848 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
2849 during TER and instead use the sepops interface for expanding
2850 non-GIMPLE_SINGLE_RHS.
2852 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2854 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
2855 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
2857 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2859 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
2860 assembler 64-bit option.
2861 * configure: Regenerate.
2863 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2865 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
2866 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
2867 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
2868 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
2869 (TARGET_CRYPTO): Take TARGET_SIMD into account.
2871 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2873 * config/aarch64/aarch64-builtins.c
2874 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
2875 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
2876 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
2877 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
2879 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
2880 (Vrevsuff): New mode attribute.
2882 2014-04-24 Terry Guo <terry.guo@arm.com>
2884 * config/arm/arm.h (machine_function): Define variable
2885 after_arm_reorg here.
2886 * config/arm/arm.c (after_arm_reorg): Remove the definition.
2887 (arm_split_constant): Update the way to access variable
2890 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
2892 2014-04-23 Tom de Vries <tom@codesourcery.com>
2894 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
2896 2014-04-23 David Malcolm <dmalcolm@redhat.com>
2898 * is-a.h: Update comments to reflect the following changes to the
2899 "pointerness" of the API, making the template parameter match the
2900 return type, allowing use of is-a.h with typedefs of pointers.
2901 (is_a_helper::cast): Return a T rather then a pointer to a T, so
2902 that the return type matches the parameter to the is_a_helper.
2904 (dyn_cast): Likewise.
2906 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
2907 pointer from the is-a.h API.
2909 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
2910 (is_a_helper <cgraph_node *>::test): ...this, matching change to
2912 (is_a_helper <varpool_node>::test): Likewise, convert to...
2913 (is_a_helper <varpool_node *>::test): ...this.
2915 (varpool_first_variable): Update for removal of implicit pointer
2916 from the is-a.h API.
2917 (varpool_next_variable): Likewise.
2918 (varpool_first_static_initializer): Likewise.
2919 (varpool_next_static_initializer): Likewise.
2920 (varpool_first_defined_variable): Likewise.
2921 (varpool_next_defined_variable): Likewise.
2922 (cgraph_first_defined_function): Likewise.
2923 (cgraph_next_defined_function): Likewise.
2924 (cgraph_first_function): Likewise.
2925 (cgraph_next_function): Likewise.
2926 (cgraph_first_function_with_gimple_body): Likewise.
2927 (cgraph_next_function_with_gimple_body): Likewise.
2928 (cgraph_alias_target): Likewise.
2929 (varpool_alias_target): Likewise.
2930 (cgraph_function_or_thunk_node): Likewise.
2931 (varpool_variable_node): Likewise.
2932 (symtab_real_symbol_p): Likewise.
2933 * cgraphunit.c (referred_to_p): Likewise.
2934 (analyze_functions): Likewise.
2935 (handle_alias_pairs): Likewise.
2936 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
2937 * gimple-ssa.h (gimple_vuse_op): Likewise.
2938 (gimple_vdef_op): Likewise.
2939 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
2940 * gimple.c (gimple_build_asm_1): Likewise.
2941 (gimple_build_try): Likewise.
2942 (gimple_build_resx): Likewise.
2943 (gimple_build_eh_dispatch): Likewise.
2944 (gimple_build_omp_for): Likewise.
2945 (gimple_omp_for_set_clauses): Likewise.
2947 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
2948 (is_a_helper <gimple_statement_asm *>::test): ...this.
2949 (is_a_helper <gimple_statement_bind>::test): Convert to...
2950 (is_a_helper <gimple_statement_bind *>::test): ...this.
2951 (is_a_helper <gimple_statement_call>::test): Convert to...
2952 (is_a_helper <gimple_statement_call *>::test): ...this.
2953 (is_a_helper <gimple_statement_catch>::test): Convert to...
2954 (is_a_helper <gimple_statement_catch *>::test): ...this.
2955 (is_a_helper <gimple_statement_resx>::test): Convert to...
2956 (is_a_helper <gimple_statement_resx *>::test): ...this.
2957 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
2958 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
2959 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
2960 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
2961 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
2962 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
2963 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
2964 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
2965 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
2966 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
2967 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
2968 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
2969 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
2970 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
2971 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
2972 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
2973 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
2974 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
2975 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
2976 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
2977 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
2978 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
2979 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
2980 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
2981 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
2982 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
2983 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
2984 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
2985 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
2986 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
2987 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
2988 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
2989 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
2990 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
2991 (is_a_helper <gimple_statement_phi>::test): Convert to...
2992 (is_a_helper <gimple_statement_phi *>::test): ...this.
2993 (is_a_helper <gimple_statement_transaction>::test): Convert to...
2994 (is_a_helper <gimple_statement_transaction *>::test): ...this.
2995 (is_a_helper <gimple_statement_try>::test): Convert to...
2996 (is_a_helper <gimple_statement_try *>::test): ...this.
2997 (is_a_helper <gimple_statement_wce>::test): Convert to...
2998 (is_a_helper <gimple_statement_wce *>::test): ...this.
2999 (is_a_helper <const gimple_statement_asm>::test): Convert to...
3000 (is_a_helper <const gimple_statement_asm *>::test): ...this.
3001 (is_a_helper <const gimple_statement_bind>::test): Convert to...
3002 (is_a_helper <const gimple_statement_bind *>::test): ...this.
3003 (is_a_helper <const gimple_statement_call>::test): Convert to...
3004 (is_a_helper <const gimple_statement_call *>::test): ...this.
3005 (is_a_helper <const gimple_statement_catch>::test): Convert to...
3006 (is_a_helper <const gimple_statement_catch *>::test): ...this.
3007 (is_a_helper <const gimple_statement_resx>::test): Convert to...
3008 (is_a_helper <const gimple_statement_resx *>::test): ...this.
3009 (is_a_helper <const gimple_statement_eh_dispatch>::test): Convert to...
3010 (is_a_helper <const gimple_statement_eh_dispatch *>::test): ...this.
3011 (is_a_helper <const gimple_statement_eh_filter>::test): Convert to...
3012 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
3013 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
3015 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
3017 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
3019 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
3021 (is_a_helper <const gimple_statement_omp_return>::test): Convert to...
3022 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
3023 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
3025 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
3026 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
3028 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
3029 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
3030 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
3031 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert to...
3032 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
3033 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
3035 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
3036 (is_a_helper <const gimple_statement_omp_target>::test): Convert to...
3037 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
3038 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
3040 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
3041 (is_a_helper <const gimple_statement_omp_single>::test): Convert to...
3042 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
3043 (is_a_helper <const gimple_statement_omp_teams>::test): Convert to...
3044 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
3045 (is_a_helper <const gimple_statement_omp_task>::test): Convert to...
3046 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
3047 (is_a_helper <const gimple_statement_phi>::test): Convert to...
3048 (is_a_helper <const gimple_statement_phi *>::test): ...this.
3049 (is_a_helper <const gimple_statement_transaction>::test): Convert to...
3050 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
3051 (is_a_helper <const gimple_statement_with_ops>::test): Convert to...
3052 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
3053 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
3054 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
3055 (is_a_helper <const gimple_statement_with_memory_ops>::test): Convert
3057 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
3059 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert to...
3060 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
3062 (gimple_use_ops): Update for removal of implicit pointer from the
3064 (gimple_set_use_ops): Likewise.
3065 (gimple_vuse): Likewise.
3066 (gimple_vdef): Likewise.
3067 (gimple_vuse_ptr): Likewise.
3068 (gimple_vdef_ptr): Likewise.
3069 (gimple_set_vuse): Likewise.
3070 (gimple_set_vdef): Likewise.
3071 (gimple_omp_return_set_lhs): Likewise.
3072 (gimple_omp_return_lhs): Likewise.
3073 (gimple_omp_return_lhs_ptr): Likewise.
3074 (gimple_call_fntype): Likewise.
3075 (gimple_call_set_fntype): Likewise.
3076 (gimple_call_set_internal_fn): Likewise.
3077 (gimple_call_use_set): Likewise.
3078 (gimple_call_clobber_set): Likewise.
3079 (gimple_bind_vars): Likewise.
3080 (gimple_bind_set_vars): Likewise.
3081 (gimple_bind_body_ptr): Likewise.
3082 (gimple_bind_set_body): Likewise.
3083 (gimple_bind_add_stmt): Likewise.
3084 (gimple_bind_block): Likewise.
3085 (gimple_bind_set_block): Likewise.
3086 (gimple_asm_ninputs): Likewise.
3087 (gimple_asm_noutputs): Likewise.
3088 (gimple_asm_nclobbers): Likewise.
3089 (gimple_asm_nlabels): Likewise.
3090 (gimple_asm_input_op): Likewise.
3091 (gimple_asm_input_op_ptr): Likewise.
3092 (gimple_asm_output_op): Likewise.
3093 (gimple_asm_output_op_ptr): Likewise.
3094 (gimple_asm_set_output_op): Likewise.
3095 (gimple_asm_clobber_op): Likewise.
3096 (gimple_asm_set_clobber_op): Likewise.
3097 (gimple_asm_label_op): Likewise.
3098 (gimple_asm_set_label_op): Likewise.
3099 (gimple_asm_string): Likewise.
3100 (gimple_catch_types): Likewise.
3101 (gimple_catch_types_ptr): Likewise.
3102 (gimple_catch_handler_ptr): Likewise.
3103 (gimple_catch_set_types): Likewise.
3104 (gimple_catch_set_handler): Likewise.
3105 (gimple_eh_filter_types): Likewise.
3106 (gimple_eh_filter_types_ptr): Likewise.
3107 (gimple_eh_filter_failure_ptr): Likewise.
3108 (gimple_eh_filter_set_types): Likewise.
3109 (gimple_eh_filter_set_failure): Likewise.
3110 (gimple_eh_must_not_throw_fndecl): Likewise.
3111 (gimple_eh_must_not_throw_set_fndecl): Likewise.
3112 (gimple_eh_else_n_body_ptr): Likewise.
3113 (gimple_eh_else_e_body_ptr): Likewise.
3114 (gimple_eh_else_set_n_body): Likewise.
3115 (gimple_eh_else_set_e_body): Likewise.
3116 (gimple_try_eval_ptr): Likewise.
3117 (gimple_try_cleanup_ptr): Likewise.
3118 (gimple_try_set_eval): Likewise.
3119 (gimple_try_set_cleanup): Likewise.
3120 (gimple_wce_cleanup_ptr): Likewise.
3121 (gimple_wce_set_cleanup): Likewise.
3122 (gimple_phi_capacity): Likewise.
3123 (gimple_phi_num_args): Likewise.
3124 (gimple_phi_result): Likewise.
3125 (gimple_phi_result_ptr): Likewise.
3126 (gimple_phi_set_result): Likewise.
3127 (gimple_phi_arg): Likewise.
3128 (gimple_phi_set_arg): Likewise.
3129 (gimple_resx_region): Likewise.
3130 (gimple_resx_set_region): Likewise.
3131 (gimple_eh_dispatch_region): Likewise.
3132 (gimple_eh_dispatch_set_region): Likewise.
3133 (gimple_omp_critical_name): Likewise.
3134 (gimple_omp_critical_name_ptr): Likewise.
3135 (gimple_omp_critical_set_name): Likewise.
3136 (gimple_omp_for_clauses): Likewise.
3137 (gimple_omp_for_clauses_ptr): Likewise.
3138 (gimple_omp_for_set_clauses): Likewise.
3139 (gimple_omp_for_collapse): Likewise.
3140 (gimple_omp_for_index): Likewise.
3141 (gimple_omp_for_index_ptr): Likewise.
3142 (gimple_omp_for_set_index): Likewise.
3143 (gimple_omp_for_initial): Likewise.
3144 (gimple_omp_for_initial_ptr): Likewise.
3145 (gimple_omp_for_set_initial): Likewise.
3146 (gimple_omp_for_final): Likewise.
3147 (gimple_omp_for_final_ptr): Likewise.
3148 (gimple_omp_for_set_final): Likewise.
3149 (gimple_omp_for_incr): Likewise.
3150 (gimple_omp_for_incr_ptr): Likewise.
3151 (gimple_omp_for_set_incr): Likewise.
3152 (gimple_omp_for_pre_body_ptr): Likewise.
3153 (gimple_omp_for_set_pre_body): Likewise.
3154 (gimple_omp_parallel_clauses): Likewise.
3155 (gimple_omp_parallel_clauses_ptr): Likewise.
3156 (gimple_omp_parallel_set_clauses): Likewise.
3157 (gimple_omp_parallel_child_fn): Likewise.
3158 (gimple_omp_parallel_child_fn_ptr): Likewise.
3159 (gimple_omp_parallel_set_child_fn): Likewise.
3160 (gimple_omp_parallel_data_arg): Likewise.
3161 (gimple_omp_parallel_data_arg_ptr): Likewise.
3162 (gimple_omp_parallel_set_data_arg): Likewise.
3163 (gimple_omp_task_clauses): Likewise.
3164 (gimple_omp_task_clauses_ptr): Likewise.
3165 (gimple_omp_task_set_clauses): Likewise.
3166 (gimple_omp_task_child_fn): Likewise.
3167 (gimple_omp_task_child_fn_ptr): Likewise.
3168 (gimple_omp_task_set_child_fn): Likewise.
3169 (gimple_omp_task_data_arg): Likewise.
3170 (gimple_omp_task_data_arg_ptr): Likewise.
3171 (gimple_omp_task_set_data_arg): Likewise.
3172 (gimple_omp_taskreg_clauses): Likewise.
3173 (gimple_omp_taskreg_clauses_ptr): Likewise.
3174 (gimple_omp_taskreg_set_clauses): Likewise.
3175 (gimple_omp_taskreg_child_fn): Likewise.
3176 (gimple_omp_taskreg_child_fn_ptr): Likewise.
3177 (gimple_omp_taskreg_set_child_fn): Likewise.
3178 (gimple_omp_taskreg_data_arg): Likewise.
3179 (gimple_omp_taskreg_data_arg_ptr): Likewise.
3180 (gimple_omp_taskreg_set_data_arg): Likewise.
3181 (gimple_omp_task_copy_fn): Likewise.
3182 (gimple_omp_task_copy_fn_ptr): Likewise.
3183 (gimple_omp_task_set_copy_fn): Likewise.
3184 (gimple_omp_task_arg_size): Likewise.
3185 (gimple_omp_task_arg_size_ptr): Likewise.
3186 (gimple_omp_task_set_arg_size): Likewise.
3187 (gimple_omp_task_arg_align): Likewise.
3188 (gimple_omp_task_arg_align_ptr): Likewise.
3189 (gimple_omp_task_set_arg_align): Likewise.
3190 (gimple_omp_single_clauses): Likewise.
3191 (gimple_omp_single_clauses_ptr): Likewise.
3192 (gimple_omp_single_set_clauses): Likewise.
3193 (gimple_omp_target_clauses): Likewise.
3194 (gimple_omp_target_clauses_ptr): Likewise.
3195 (gimple_omp_target_set_clauses): Likewise.
3196 (gimple_omp_target_child_fn): Likewise.
3197 (gimple_omp_target_child_fn_ptr): Likewise.
3198 (gimple_omp_target_set_child_fn): Likewise.
3199 (gimple_omp_target_data_arg): Likewise.
3200 (gimple_omp_target_data_arg_ptr): Likewise.
3201 (gimple_omp_target_set_data_arg): Likewise.
3202 (gimple_omp_teams_clauses): Likewise.
3203 (gimple_omp_teams_clauses_ptr): Likewise.
3204 (gimple_omp_teams_set_clauses): Likewise.
3205 (gimple_omp_sections_clauses): Likewise.
3206 (gimple_omp_sections_clauses_ptr): Likewise.
3207 (gimple_omp_sections_set_clauses): Likewise.
3208 (gimple_omp_sections_control): Likewise.
3209 (gimple_omp_sections_control_ptr): Likewise.
3210 (gimple_omp_sections_set_control): Likewise.
3211 (gimple_omp_for_set_cond): Likewise.
3212 (gimple_omp_for_cond): Likewise.
3213 (gimple_omp_atomic_store_set_val): Likewise.
3214 (gimple_omp_atomic_store_val): Likewise.
3215 (gimple_omp_atomic_store_val_ptr): Likewise.
3216 (gimple_omp_atomic_load_set_lhs): Likewise.
3217 (gimple_omp_atomic_load_lhs): Likewise.
3218 (gimple_omp_atomic_load_lhs_ptr): Likewise.
3219 (gimple_omp_atomic_load_set_rhs): Likewise.
3220 (gimple_omp_atomic_load_rhs): Likewise.
3221 (gimple_omp_atomic_load_rhs_ptr): Likewise.
3222 (gimple_omp_continue_control_def): Likewise.
3223 (gimple_omp_continue_control_def_ptr): Likewise.
3224 (gimple_omp_continue_set_control_def): Likewise.
3225 (gimple_omp_continue_control_use): Likewise.
3226 (gimple_omp_continue_control_use_ptr): Likewise.
3227 (gimple_omp_continue_set_control_use): Likewise.
3228 (gimple_transaction_body_ptr): Likewise.
3229 (gimple_transaction_label): Likewise.
3230 (gimple_transaction_label_ptr): Likewise.
3231 (gimple_transaction_set_body): Likewise.
3232 (gimple_transaction_set_label): Likewise.
3234 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
3235 * ipa-inline-analysis.c (inline_write_summary): Likewise.
3236 * ipa-ref.c (ipa_record_reference): Likewise.
3237 * ipa-reference.c (analyze_function): Likewise.
3238 (ipa_reference_write_optimization_summary): Likewise.
3239 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
3240 (address_taken_from_non_vtable_p): Likewise.
3241 (comdat_can_be_unshared_p_1): Likewise.
3242 * lto-cgraph.c (lto_output_ref): Likewise.
3243 (add_references): Likewise.
3244 (compute_ltrans_boundary): Likewise.
3245 (output_symtab): Likewise.
3246 (input_ref): Likewise.
3247 (input_cgraph_1): Likewise.
3248 (output_cgraph_opt_summary): Likewise.
3249 * lto-streamer-out.c (lto_output): Likewise.
3250 (output_symbol_p): Likewise.
3251 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
3252 (lsei_start_function_in_partition): Likewise.
3253 (lsei_next_variable_in_partition): Likewise.
3254 (lsei_start_variable_in_partition): Likewise.
3255 * symtab.c (insert_to_assembler_name_hash): Likewise.
3256 (unlink_from_assembler_name_hash): Likewise.
3257 (symtab_unregister_node): Likewise.
3258 (symtab_remove_node): Likewise.
3259 (dump_symtab_node): Likewise.
3260 (verify_symtab_base): Likewise.
3261 (verify_symtab_node): Likewise.
3262 (symtab_make_decl_local): Likewise.
3263 (symtab_alias_ultimate_target): Likewise.
3264 (symtab_resolve_alias): Likewise.
3265 (symtab_get_symbol_partitioning_class): Likewise.
3266 * tree-phinodes.c (allocate_phi_node): Likewise.
3267 (reserve_phi_args_for_new_edge): Likewise.
3268 (remove_phi_args): Likewise.
3269 * varpool.c (varpool_node_for_asm): Likewise.
3270 (varpool_remove_unreferenced_decls): Likewise.
3272 2014-04-23 Jeff Law <law@redhat.com>
3274 PR tree-optimization/60902
3275 * tree-ssa-threadedge.c
3276 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
3277 invalidate outputs from statements that do not produce useful
3278 outputs for threading.
3280 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
3282 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
3283 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
3284 machine descriptions for Stack Smashing Protector.
3286 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
3288 * aarch64.md (<optab>_rol<mode>3): New pattern.
3289 (<optab>_rolsi3_uxtw): Likewise.
3290 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
3292 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
3294 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
3295 (arm_cortex_a12_tune): Likewise.
3297 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3299 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
3301 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3303 * config/arm/arm.md (arm_rev16si2): New pattern.
3304 (arm_rev16si2_alt): Likewise.
3305 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
3307 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3309 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
3310 (rev16<mode>2_alt): Likewise.
3311 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
3312 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
3313 (aarch_rev16_shleft_mask_imm_p): Likewise.
3314 (aarch_rev16_p_1): Likewise.
3315 (aarch_rev16_p): Likewise.
3316 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
3317 (aarch_rev16_shright_mask_imm_p): Likewise.
3318 (aarch_rev16_shleft_mask_imm_p): Likewise.
3320 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3322 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
3323 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
3325 (cortex_a53_extra_costs): Likewise.
3326 (cortex_a57_extra_costs): Likewise.
3327 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
3328 (cortexa7_extra_costs): Likewise.
3329 (cortexa8_extra_costs): Likewise.
3330 (cortexa12_extra_costs): Likewise.
3331 (cortexa15_extra_costs): Likewise.
3332 (v7m_extra_costs): Likewise.
3333 (arm_new_rtx_costs): Handle BSWAP.
3335 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3337 * config/arm/arm.c (cortexa8_extra_costs): New table.
3338 (arm_cortex_a8_tune): New tuning struct.
3339 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
3341 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3343 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
3345 2014-04-23 Richard Biener <rguenther@suse.de>
3347 * Makefile.in (OBJS): Remove loop-unswitch.o.
3348 * tree-pass.h (make_pass_rtl_unswitch): Remove.
3349 * passes.def (pass_rtl_unswitch): Likewise.
3350 * loop-init.c (gate_rtl_unswitch): Likewise.
3351 (rtl_unswitch): Likewise.
3352 (pass_data_rtl_unswitch): Likewise.
3353 (pass_rtl_unswitch): Likewise.
3354 (make_pass_rtl_unswitch): Likewise.
3355 * rtl.h (reversed_condition): Likewise.
3356 (compare_and_jump_seq): Likewise.
3357 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
3359 * loop-unroll.c (compare_and_jump_seq): Likewise.
3361 2014-04-23 Richard Biener <rguenther@suse.de>
3363 PR tree-optimization/60903
3364 * tree-ssa-loop-im.c (analyze_memory_references): Remove
3365 commented code block.
3366 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
3367 loop flags to newly created BBs and edges.
3369 2014-04-23 Nick Clifton <nickc@redhat.com>
3371 * config/msp430/msp430.c (msp430_handle_option): Move function
3373 (msp430_option_override): Simplify mcu and mcpu option handling.
3374 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
3375 support for -mhwmult command line option.
3376 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
3377 -mhwmult command line option.
3378 (msp430_hwmult_enabled): Delete.
3379 (msp43o_output_labelref): Add support for -mhwmult command line option.
3380 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
3381 (umulsidi3): Likewise.
3382 * config/msp430/msp430.opt (mmcu): Add Report attribute.
3383 (mcpu, mlarge, msmall): Likewise.
3384 (mhwmult): New option.
3385 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
3387 (msp430_is_f5_mcu): Remove prototype.
3388 (msp430_use_f5_series_hwmult): Add prototype.
3389 * config/msp430/msp430-opts.h: New file.
3390 * common/config/msp430: New directory.
3391 * common/config/msp430/msp430-common.c: New file.
3392 * config.gcc (msp430): Remove target_has_targetm_common.
3393 * doc/invoke.texi: Document -mhwmult command line option.
3395 2014-04-23 Nick Clifton <nickc@redhat.com>
3397 * config/i386/cygwin.h (ENDFILE_SPEC): Include
3398 default-manifest.o if it can be found in the search path.
3399 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
3401 2014-04-23 Terry Guo <terry.guo@arm.com>
3403 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
3405 2014-04-23 Richard Biener <rguenther@suse.de>
3408 * tree-inline.c (declare_return_variable): Use mark_addressable.
3410 2014-04-23 Richard Biener <rguenther@suse.de>
3413 * loop-init.c (loop_optimizer_init): Make sure to apply
3414 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
3416 2014-04-22 Jakub Jelinek <jakub@redhat.com>
3419 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
3421 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
3422 if flag_sanitize_undefined_trap_on_error.
3423 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
3424 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
3425 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
3426 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
3427 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
3428 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
3429 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
3430 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
3431 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
3432 * ubsan.c (ubsan_instrument_unreachable): Return
3433 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
3434 (ubsan_expand_null_ifn): Emit __builtin_trap ()
3435 if flag_sanitize_undefined_trap_on_error and
3436 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
3437 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
3438 instrument_bool_enum_load): Emit __builtin_trap () if
3439 flag_sanitize_undefined_trap_on_error and
3440 __builtin_handle_*_abort () if !flag_sanitize_recover.
3441 * doc/invoke.texi (-fsanitize-recover,
3442 -fsanitize-undefined-trap-on-error): Document.
3444 2014-04-22 Christian Bruel <christian.bruel@st.com>
3446 * config/sh/sh.md (mov<mode>): Replace movQIHI.
3447 Force immediates to SImode.
3449 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
3451 * config/nios2/nios2.md (UNSPEC_ROUND): New.
3453 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
3454 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
3455 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
3456 (nios2_fpu_insn): Add entry for round.
3457 (N2FPU_NO_ERRNO_P): Define.
3458 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
3460 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
3462 2014-04-22 Richard Henderson <rth@redhat.com>
3464 * config/aarch64/aarch64 (addti3, subti3): New expanders.
3465 (add<GPI>3_compare0): Remove leading * from name.
3466 (add<GPI>3_carryin): Likewise.
3467 (sub<GPI>3_compare0): Likewise.
3468 (sub<GPI>3_carryin): Likewise.
3469 (<su_optab>mulditi3): New expander.
3470 (multi3): New expander.
3471 (madd<GPI>): Remove leading * from name.
3473 2014-04-22 Martin Jambor <mjambor@suse.cz>
3475 * cgraphclones.c (cgraph_function_versioning): Copy
3476 ipa_transforms_to_apply instead of asserting it is empty.
3478 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
3481 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
3482 on count_exp to get mode.
3484 2014-04-22 Andrew Pinski <apinski@cavium.com>
3486 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
3487 Handle TLS for ILP32.
3488 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
3489 (tlsie_small_<mode>): this and handle PTR.
3490 (tlsie_small_sidi): New pattern.
3491 (tlsle_small): Change to an expand to handle ILP32.
3492 (tlsle_small_<mode>): New pattern.
3493 (tlsdesc_small): Rename to ...
3494 (tlsdesc_small_<mode>): this and handle PTR.
3496 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3498 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
3500 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
3502 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
3503 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
3504 (aarch64_types_signed_poly_qualifiers): Likewise.
3505 (aarch64_types_unsigned_signed_qualifiers): Likewise.
3506 (aarch64_types_poly_signed_qualifiers): Likewise.
3507 (TYPES_REINTERP_SS): Type macro added.
3508 (TYPES_REINTERP_SU): Likewise.
3509 (TYPES_REINTERP_SP): Likewise.
3510 (TYPES_REINTERP_US): Likewise.
3511 (TYPES_REINTERP_PS): Likewise.
3512 (aarch64_fold_builtin): New expression folding added.
3513 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
3514 Declarations removed.
3515 (REINTERP_SS): Declarations added.
3516 (REINTERP_US): Likewise.
3517 (REINTERP_PS): Likewise.
3518 (REINTERP_SU): Likewise.
3519 (REINTERP_SP): Likewise.
3520 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
3521 (vreinterpretq_p8_f64): Likewise.
3522 (vreinterpret_p16_f64): Likewise.
3523 (vreinterpretq_p16_f64): Likewise.
3524 (vreinterpret_f32_f64): Likewise.
3525 (vreinterpretq_f32_f64): Likewise.
3526 (vreinterpret_f64_f32): Likewise.
3527 (vreinterpret_f64_p8): Likewise.
3528 (vreinterpret_f64_p16): Likewise.
3529 (vreinterpret_f64_s8): Likewise.
3530 (vreinterpret_f64_s16): Likewise.
3531 (vreinterpret_f64_s32): Likewise.
3532 (vreinterpret_f64_s64): Likewise.
3533 (vreinterpret_f64_u8): Likewise.
3534 (vreinterpret_f64_u16): Likewise.
3535 (vreinterpret_f64_u32): Likewise.
3536 (vreinterpret_f64_u64): Likewise.
3537 (vreinterpretq_f64_f32): Likewise.
3538 (vreinterpretq_f64_p8): Likewise.
3539 (vreinterpretq_f64_p16): Likewise.
3540 (vreinterpretq_f64_s8): Likewise.
3541 (vreinterpretq_f64_s16): Likewise.
3542 (vreinterpretq_f64_s32): Likewise.
3543 (vreinterpretq_f64_s64): Likewise.
3544 (vreinterpretq_f64_u8): Likewise.
3545 (vreinterpretq_f64_u16): Likewise.
3546 (vreinterpretq_f64_u32): Likewise.
3547 (vreinterpretq_f64_u64): Likewise.
3548 (vreinterpret_s64_f64): Likewise.
3549 (vreinterpretq_s64_f64): Likewise.
3550 (vreinterpret_u64_f64): Likewise.
3551 (vreinterpretq_u64_f64): Likewise.
3552 (vreinterpret_s8_f64): Likewise.
3553 (vreinterpretq_s8_f64): Likewise.
3554 (vreinterpret_s16_f64): Likewise.
3555 (vreinterpretq_s16_f64): Likewise.
3556 (vreinterpret_s32_f64): Likewise.
3557 (vreinterpretq_s32_f64): Likewise.
3558 (vreinterpret_u8_f64): Likewise.
3559 (vreinterpretq_u8_f64): Likewise.
3560 (vreinterpret_u16_f64): Likewise.
3561 (vreinterpretq_u16_f64): Likewise.
3562 (vreinterpret_u32_f64): Likewise.
3563 (vreinterpretq_u32_f64): Likewise.
3565 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
3567 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
3568 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
3569 (vreinterpret_p8_s8): Likewise.
3570 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
3571 (vreinterpret_p8_s16): Likewise.
3572 (vreinterpret_p8_s32): Likewise.
3573 (vreinterpret_p8_s64): Likewise.
3574 (vreinterpret_p8_f32): Likewise.
3575 (vreinterpret_p8_u8): Likewise.
3576 (vreinterpret_p8_u16): Likewise.
3577 (vreinterpret_p8_u32): Likewise.
3578 (vreinterpret_p8_u64): Likewise.
3579 (vreinterpret_p8_p16): Likewise.
3580 (vreinterpretq_p8_s8): Likewise.
3581 (vreinterpretq_p8_s16): Likewise.
3582 (vreinterpretq_p8_s32): Likewise.
3583 (vreinterpretq_p8_s64): Likewise.
3584 (vreinterpretq_p8_f32): Likewise.
3585 (vreinterpretq_p8_u8): Likewise.
3586 (vreinterpretq_p8_u16): Likewise.
3587 (vreinterpretq_p8_u32): Likewise.
3588 (vreinterpretq_p8_u64): Likewise.
3589 (vreinterpretq_p8_p16): Likewise.
3590 (vreinterpret_p16_s8): Likewise.
3591 (vreinterpret_p16_s16): Likewise.
3592 (vreinterpret_p16_s32): Likewise.
3593 (vreinterpret_p16_s64): Likewise.
3594 (vreinterpret_p16_f32): Likewise.
3595 (vreinterpret_p16_u8): Likewise.
3596 (vreinterpret_p16_u16): Likewise.
3597 (vreinterpret_p16_u32): Likewise.
3598 (vreinterpret_p16_u64): Likewise.
3599 (vreinterpret_p16_p8): Likewise.
3600 (vreinterpretq_p16_s8): Likewise.
3601 (vreinterpretq_p16_s16): Likewise.
3602 (vreinterpretq_p16_s32): Likewise.
3603 (vreinterpretq_p16_s64): Likewise.
3604 (vreinterpretq_p16_f32): Likewise.
3605 (vreinterpretq_p16_u8): Likewise.
3606 (vreinterpretq_p16_u16): Likewise.
3607 (vreinterpretq_p16_u32): Likewise.
3608 (vreinterpretq_p16_u64): Likewise.
3609 (vreinterpretq_p16_p8): Likewise.
3610 (vreinterpret_f32_s8): Likewise.
3611 (vreinterpret_f32_s16): Likewise.
3612 (vreinterpret_f32_s32): Likewise.
3613 (vreinterpret_f32_s64): Likewise.
3614 (vreinterpret_f32_u8): Likewise.
3615 (vreinterpret_f32_u16): Likewise.
3616 (vreinterpret_f32_u32): Likewise.
3617 (vreinterpret_f32_u64): Likewise.
3618 (vreinterpret_f32_p8): Likewise.
3619 (vreinterpret_f32_p16): Likewise.
3620 (vreinterpretq_f32_s8): Likewise.
3621 (vreinterpretq_f32_s16): Likewise.
3622 (vreinterpretq_f32_s32): Likewise.
3623 (vreinterpretq_f32_s64): Likewise.
3624 (vreinterpretq_f32_u8): Likewise.
3625 (vreinterpretq_f32_u16): Likewise.
3626 (vreinterpretq_f32_u32): Likewise.
3627 (vreinterpretq_f32_u64): Likewise.
3628 (vreinterpretq_f32_p8): Likewise.
3629 (vreinterpretq_f32_p16): Likewise.
3630 (vreinterpret_s64_s8): Likewise.
3631 (vreinterpret_s64_s16): Likewise.
3632 (vreinterpret_s64_s32): Likewise.
3633 (vreinterpret_s64_f32): Likewise.
3634 (vreinterpret_s64_u8): Likewise.
3635 (vreinterpret_s64_u16): Likewise.
3636 (vreinterpret_s64_u32): Likewise.
3637 (vreinterpret_s64_u64): Likewise.
3638 (vreinterpret_s64_p8): Likewise.
3639 (vreinterpret_s64_p16): Likewise.
3640 (vreinterpretq_s64_s8): Likewise.
3641 (vreinterpretq_s64_s16): Likewise.
3642 (vreinterpretq_s64_s32): Likewise.
3643 (vreinterpretq_s64_f32): Likewise.
3644 (vreinterpretq_s64_u8): Likewise.
3645 (vreinterpretq_s64_u16): Likewise.
3646 (vreinterpretq_s64_u32): Likewise.
3647 (vreinterpretq_s64_u64): Likewise.
3648 (vreinterpretq_s64_p8): Likewise.
3649 (vreinterpretq_s64_p16): Likewise.
3650 (vreinterpret_u64_s8): Likewise.
3651 (vreinterpret_u64_s16): Likewise.
3652 (vreinterpret_u64_s32): Likewise.
3653 (vreinterpret_u64_s64): Likewise.
3654 (vreinterpret_u64_f32): Likewise.
3655 (vreinterpret_u64_u8): Likewise.
3656 (vreinterpret_u64_u16): Likewise.
3657 (vreinterpret_u64_u32): Likewise.
3658 (vreinterpret_u64_p8): Likewise.
3659 (vreinterpret_u64_p16): Likewise.
3660 (vreinterpretq_u64_s8): Likewise.
3661 (vreinterpretq_u64_s16): Likewise.
3662 (vreinterpretq_u64_s32): Likewise.
3663 (vreinterpretq_u64_s64): Likewise.
3664 (vreinterpretq_u64_f32): Likewise.
3665 (vreinterpretq_u64_u8): Likewise.
3666 (vreinterpretq_u64_u16): Likewise.
3667 (vreinterpretq_u64_u32): Likewise.
3668 (vreinterpretq_u64_p8): Likewise.
3669 (vreinterpretq_u64_p16): Likewise.
3670 (vreinterpret_s8_s16): Likewise.
3671 (vreinterpret_s8_s32): Likewise.
3672 (vreinterpret_s8_s64): Likewise.
3673 (vreinterpret_s8_f32): Likewise.
3674 (vreinterpret_s8_u8): Likewise.
3675 (vreinterpret_s8_u16): Likewise.
3676 (vreinterpret_s8_u32): Likewise.
3677 (vreinterpret_s8_u64): Likewise.
3678 (vreinterpret_s8_p8): Likewise.
3679 (vreinterpret_s8_p16): Likewise.
3680 (vreinterpretq_s8_s16): Likewise.
3681 (vreinterpretq_s8_s32): Likewise.
3682 (vreinterpretq_s8_s64): Likewise.
3683 (vreinterpretq_s8_f32): Likewise.
3684 (vreinterpretq_s8_u8): Likewise.
3685 (vreinterpretq_s8_u16): Likewise.
3686 (vreinterpretq_s8_u32): Likewise.
3687 (vreinterpretq_s8_u64): Likewise.
3688 (vreinterpretq_s8_p8): Likewise.
3689 (vreinterpretq_s8_p16): Likewise.
3690 (vreinterpret_s16_s8): Likewise.
3691 (vreinterpret_s16_s32): Likewise.
3692 (vreinterpret_s16_s64): Likewise.
3693 (vreinterpret_s16_f32): Likewise.
3694 (vreinterpret_s16_u8): Likewise.
3695 (vreinterpret_s16_u16): Likewise.
3696 (vreinterpret_s16_u32): Likewise.
3697 (vreinterpret_s16_u64): Likewise.
3698 (vreinterpret_s16_p8): Likewise.
3699 (vreinterpret_s16_p16): Likewise.
3700 (vreinterpretq_s16_s8): Likewise.
3701 (vreinterpretq_s16_s32): Likewise.
3702 (vreinterpretq_s16_s64): Likewise.
3703 (vreinterpretq_s16_f32): Likewise.
3704 (vreinterpretq_s16_u8): Likewise.
3705 (vreinterpretq_s16_u16): Likewise.
3706 (vreinterpretq_s16_u32): Likewise.
3707 (vreinterpretq_s16_u64): Likewise.
3708 (vreinterpretq_s16_p8): Likewise.
3709 (vreinterpretq_s16_p16): Likewise.
3710 (vreinterpret_s32_s8): Likewise.
3711 (vreinterpret_s32_s16): Likewise.
3712 (vreinterpret_s32_s64): Likewise.
3713 (vreinterpret_s32_f32): Likewise.
3714 (vreinterpret_s32_u8): Likewise.
3715 (vreinterpret_s32_u16): Likewise.
3716 (vreinterpret_s32_u32): Likewise.
3717 (vreinterpret_s32_u64): Likewise.
3718 (vreinterpret_s32_p8): Likewise.
3719 (vreinterpret_s32_p16): Likewise.
3720 (vreinterpretq_s32_s8): Likewise.
3721 (vreinterpretq_s32_s16): Likewise.
3722 (vreinterpretq_s32_s64): Likewise.
3723 (vreinterpretq_s32_f32): Likewise.
3724 (vreinterpretq_s32_u8): Likewise.
3725 (vreinterpretq_s32_u16): Likewise.
3726 (vreinterpretq_s32_u32): Likewise.
3727 (vreinterpretq_s32_u64): Likewise.
3728 (vreinterpretq_s32_p8): Likewise.
3729 (vreinterpretq_s32_p16): Likewise.
3730 (vreinterpret_u8_s8): Likewise.
3731 (vreinterpret_u8_s16): Likewise.
3732 (vreinterpret_u8_s32): Likewise.
3733 (vreinterpret_u8_s64): Likewise.
3734 (vreinterpret_u8_f32): Likewise.
3735 (vreinterpret_u8_u16): Likewise.
3736 (vreinterpret_u8_u32): Likewise.
3737 (vreinterpret_u8_u64): Likewise.
3738 (vreinterpret_u8_p8): Likewise.
3739 (vreinterpret_u8_p16): Likewise.
3740 (vreinterpretq_u8_s8): Likewise.
3741 (vreinterpretq_u8_s16): Likewise.
3742 (vreinterpretq_u8_s32): Likewise.
3743 (vreinterpretq_u8_s64): Likewise.
3744 (vreinterpretq_u8_f32): Likewise.
3745 (vreinterpretq_u8_u16): Likewise.
3746 (vreinterpretq_u8_u32): Likewise.
3747 (vreinterpretq_u8_u64): Likewise.
3748 (vreinterpretq_u8_p8): Likewise.
3749 (vreinterpretq_u8_p16): Likewise.
3750 (vreinterpret_u16_s8): Likewise.
3751 (vreinterpret_u16_s16): Likewise.
3752 (vreinterpret_u16_s32): Likewise.
3753 (vreinterpret_u16_s64): Likewise.
3754 (vreinterpret_u16_f32): Likewise.
3755 (vreinterpret_u16_u8): Likewise.
3756 (vreinterpret_u16_u32): Likewise.
3757 (vreinterpret_u16_u64): Likewise.
3758 (vreinterpret_u16_p8): Likewise.
3759 (vreinterpret_u16_p16): Likewise.
3760 (vreinterpretq_u16_s8): Likewise.
3761 (vreinterpretq_u16_s16): Likewise.
3762 (vreinterpretq_u16_s32): Likewise.
3763 (vreinterpretq_u16_s64): Likewise.
3764 (vreinterpretq_u16_f32): Likewise.
3765 (vreinterpretq_u16_u8): Likewise.
3766 (vreinterpretq_u16_u32): Likewise.
3767 (vreinterpretq_u16_u64): Likewise.
3768 (vreinterpretq_u16_p8): Likewise.
3769 (vreinterpretq_u16_p16): Likewise.
3770 (vreinterpret_u32_s8): Likewise.
3771 (vreinterpret_u32_s16): Likewise.
3772 (vreinterpret_u32_s32): Likewise.
3773 (vreinterpret_u32_s64): Likewise.
3774 (vreinterpret_u32_f32): Likewise.
3775 (vreinterpret_u32_u8): Likewise.
3776 (vreinterpret_u32_u16): Likewise.
3777 (vreinterpret_u32_u64): Likewise.
3778 (vreinterpret_u32_p8): Likewise.
3779 (vreinterpret_u32_p16): Likewise.
3780 (vreinterpretq_u32_s8): Likewise.
3781 (vreinterpretq_u32_s16): Likewise.
3782 (vreinterpretq_u32_s32): Likewise.
3783 (vreinterpretq_u32_s64): Likewise.
3784 (vreinterpretq_u32_f32): Likewise.
3785 (vreinterpretq_u32_u8): Likewise.
3786 (vreinterpretq_u32_u16): Likewise.
3787 (vreinterpretq_u32_u64): Likewise.
3788 (vreinterpretq_u32_p8): Likewise.
3789 (vreinterpretq_u32_p16): Likewise.
3791 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
3793 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
3795 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator extended.
3797 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
3798 (vqnegd_s64): Likewise.
3799 (vqabs_s64): Likewise.
3800 (vqabsd_s64): Likewise.
3802 2014-04-22 Richard Henderson <rth@redhat.com>
3804 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
3805 computation to the top of the loop.
3807 2014-04-22 Renlin <renlin.li@arm.com>
3808 Jiong Wang <jiong.wang@arm.com>
3810 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
3811 * config/aarch64/aarch64.c (aarch64_layout_frame)
3812 (aarch64_initial_elimination_offset): Likewise.
3814 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
3816 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
3819 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
3821 * machmode.h (bitwise_mode_for_mode): Declare.
3822 * stor-layout.h (bitwise_type_for_mode): Likewise.
3823 * stor-layout.c (bitwise_mode_for_mode): New function.
3824 (bitwise_type_for_mode): Likewise.
3825 * builtins.c (fold_builtin_memory_op): Use it instead of
3826 int_mode_for_mode and build_nonstandard_integer_type.
3828 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3830 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
3831 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
3832 (*-*-solaris2*): Simplify.
3833 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
3834 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
3835 *-*-solaris2.9* handling.
3837 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
3839 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
3840 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
3842 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
3843 * configure: Regenerate.
3845 * config/i386/sol2-9.h: Remove.
3847 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
3848 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
3849 Remove Solaris 9 references.
3851 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
3853 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
3854 (floatuns<GPI:mode><GPF:mode>2): Remove.
3855 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
3856 and floatuns conversions.
3857 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
3858 and floatuns conversions.
3859 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
3860 (w1,w2): New mode attributes for inequal width conversions.
3862 2014-04-22 Renlin Li <Renlin.Li@arm.com>
3864 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
3865 the output asm format.
3867 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
3869 * config/aarch64/aarch64-simd.md
3870 (aarch64_cm<optab>di): Always split.
3871 (*aarch64_cm<optab>di): New.
3872 (aarch64_cmtstdi): Always split.
3873 (*aarch64_cmtstdi): New.
3875 2014-04-22 Jakub Jelinek <jakub@redhat.com>
3877 PR tree-optimization/60823
3878 * omp-low.c (ipa_simd_modify_function_body): Go through
3879 all SSA_NAMEs and for those refering to vector arguments
3880 which are going to be replaced adjust SSA_NAME_VAR and,
3881 if it is a default definition, change it into a non-default
3882 definition assigned at the beginning of function from new_decl.
3883 (ipa_simd_modify_stmt_ops): Rewritten.
3884 * tree-dfa.c (set_ssa_default_def): When removing default def,
3885 check for NULL loc instead of NULL *loc.
3887 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3889 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
3890 restrictions on core registers for DImode values in Thumb2.
3892 2014-04-22 Ian Bolton <ian.bolton@arm.com>
3894 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
3895 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
3897 2014-04-22 Ian Bolton <ian.bolton@arm.com>
3899 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
3900 (*iordi_notzesidi_di): Likewise.
3901 (*iordi_notsesidi_di): Likewise.
3903 2014-04-22 Ian Bolton <ian.bolton@arm.com>
3905 * config/arm/arm-protos.h (tune_params): New struct members.
3906 * config/arm/arm.c: Initialise tune_params per processor.
3907 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
3908 for speed, based on new tune_params.
3910 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
3912 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro added.
3913 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added macro.
3914 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment corrected.
3915 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
3916 * config/aarch64/arm_neon.h (vrnd_f64): Added.
3917 (vrnda_f64): Likewise.
3918 (vrndi_f64): Likewise.
3919 (vrndm_f64): Likewise.
3920 (vrndn_f64): Likewise.
3921 (vrndp_f64): Likewise.
3922 (vrndx_f64): Likewise.
3924 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3926 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
3927 GET_MODE_SIZE argument is enum machine_mode.
3929 2014-04-22 Jakub Jelinek <jakub@redhat.com>
3932 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
3933 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
3935 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
3938 * asan.c (asan_emit_stack_protection): Force the base to align to
3939 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
3940 appropriate bits if STRICT_ALIGNMENT.
3941 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
3943 (expand_used_vars): Leave a space in the stack frame for alignment
3944 if STRICT_ALIGNMENT.
3946 2014-04-21 David Malcolm <dmalcolm@redhat.com>
3948 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
3950 (gimple_store_p): Likewise.
3951 (gimple_assign_load_p): Likewise.
3952 (gimple_assign_cast_p): Likewise.
3953 (gimple_clobber_p): Likewise.
3955 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
3956 rather than a gimple.
3957 (gimple_assign_cast_p): Likewise.
3959 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
3962 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
3963 If mode is DDmode and TARGET_E500_DOUBLE allow move.
3965 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
3966 more debug information for E500 if -mdebug=reg.
3968 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
3971 * config/i386/i386.c (ix86_expand_builtin)
3972 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
3973 register for target RTX.
3974 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
3976 2014-04-18 Cong Hou <congh@google.com>
3978 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
3979 the widen-mult pattern by handling two operands with different sizes,
3980 and operands whose size is smaller than half of the result type.
3982 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
3984 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
3985 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
3986 (do_estimate_edge_time): Compute it.
3987 * ipa-inline.c (want_inline_small_function_p): Bypass
3988 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
3990 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
3992 * ipa-inline.c (spec_rem): New static variable.
3993 (dump_overall_stats): New function.
3994 (dump_inline_stats): New function.
3996 2014-04-18 Richard Henderson <rth@redhat.com>
3998 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
3999 to GET_MODE_SIZE, not a reg_class_t.
4001 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4003 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
4004 (vsx_xxmrglw_<mode>): Likewise.
4006 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
4009 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
4010 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
4011 (rs6000_init_hard_regno_mode_ok): Likewise.
4013 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
4015 * ipa-inline.c (inline_small_functions): Account only non-cold
4017 * doc/invoke.texi (inline-unit-growth): Update documentation.
4019 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
4021 * config/rs6000/rs6000.md (addti3, subti3): New.
4023 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
4026 * config/i386/i386.c (ix86_expand_clear): Remove outdated
4027 comment. Check optimize_insn_for_size_p instead of
4028 optimize_insn_for_speed_p.
4030 2014-04-17 Martin Jambor <mjambor@suse.cz>
4032 * gimple-iterator.c (gsi_start_edge): New function.
4033 * gimple-iterator.h (gsi_start_edge): Declare.
4034 * tree-sra.c (single_non_eh_succ): New function.
4035 (disqualify_ops_if_throwing_stmt): Renamed to
4036 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
4037 having one non-EH successor BB.
4038 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
4039 generate loads into replacements.
4040 (sra_modify_assign): Likewise and and also use the simple path for
4042 (sra_modify_function_body): Commit statements on edges.
4044 2014-04-17 Richard Biener <rguenther@suse.de>
4047 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
4048 comparison results and add clarifying comment.
4050 2014-04-17 Jakub Jelinek <jakub@redhat.com>
4052 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
4053 (blank_mode): Initialize it.
4054 (emit_mode_size_inline, emit_mode_nunits_inline,
4055 emit_mode_inner_inline): New functions.
4056 (emit_insn_modes_h): Call them and surround their output with
4057 #if GCC_VERSION >= 4001 ... #endif.
4058 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
4059 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
4060 mode_* arrays if the argument is __builtin_constant_p.
4061 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
4062 is enum machine_mode.
4064 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
4066 * passes.c (opt_pass::execute): Adjust.
4067 (pass_manager::execute_pass_mode_switching): Likewise.
4068 (early_local_passes::execute): Likewise.
4069 (execute_one_pass): Pass cfun to the pass's execute method.
4070 * tree-pass.h (opt_pass::execute): Add function * argument.
4071 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
4072 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
4073 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
4074 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
4075 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
4076 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
4077 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
4078 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
4079 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
4080 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
4081 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
4082 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
4083 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
4084 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
4085 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
4086 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
4087 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
4088 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
4089 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
4090 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
4091 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
4092 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
4093 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
4094 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
4095 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
4096 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
4097 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
4098 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
4099 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
4102 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
4104 * passes.c (opt_pass::gate): Take function * argument.
4105 (gate_all_early_local_passes): Merge into
4106 (early_local_passes::gate): this.
4107 (gate_all_early_optimizations): Merge into
4108 (all_early_optimizations::gate): this.
4109 (gate_all_optimizations): Mege into
4110 (all_optimizations::gate): this.
4111 (gate_all_optimizations_g): Merge into
4112 (all_optimizations_g::gate): this.
4113 (gate_rest_of_compilation): Mege into
4114 (rest_of_compilation::gate): this.
4115 (gate_postreload): Merge into
4116 (postreload::gate): this.
4117 (dump_one_pass): Pass cfun to the pass's gate method.
4118 (execute_ipa_summary_passes): Likewise.
4119 (execute_one_pass): Likewise.
4120 (ipa_write_summaries_2): Likewise.
4121 (ipa_write_optimization_summaries_1): Likewise.
4122 (ipa_read_summaries_1): Likewise.
4123 (ipa_read_optimization_summaries_1): Likewise.
4124 (execute_ipa_stmt_fixups): Likewise.
4125 * tree-pass.h (opt_pass::gate): Add function * argument.
4126 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
4127 combine-stack-adj.c, combine.c, compare-elim.c,
4128 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
4129 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
4130 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
4131 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
4132 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
4133 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
4134 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
4135 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
4136 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
4137 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
4138 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
4139 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
4140 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
4141 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
4142 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
4143 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
4144 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
4145 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
4146 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
4147 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
4148 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
4149 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
4150 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
4151 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
4152 var-tracking.c, vtable-verify.c, web.c: Adjust.
4154 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
4156 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
4157 * configure: Regenerate.
4159 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
4161 * passes.c (dump_one_pass): don't check pass->has_gate.
4162 (execute_ipa_summary_passes): Likewise.
4163 (execute_one_pass): Likewise.
4164 (ipa_write_summaries_2): Likewise.
4165 (ipa_write_optimization_summaries_1): Likewise.
4166 (ipa_read_optimization_summaries_1): Likewise.
4167 (execute_ipa_stmt_fixups): Likewise.
4168 * tree-pass.h (pass_data::has_gate): Remove.
4169 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
4170 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
4171 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
4172 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
4173 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
4174 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
4175 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
4176 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
4177 gimple-low.c, gimple-ssa-isolate-paths.c,
4178 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
4179 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
4180 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
4181 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
4182 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
4183 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
4184 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
4185 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
4186 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
4187 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
4188 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
4189 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
4190 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
4191 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
4192 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
4193 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
4194 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
4195 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
4196 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
4197 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
4198 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
4199 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
4202 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
4204 * pass_manager.h (pass_manager::register_dump_files_1): Remove
4206 * passes.c (pass_manager::register_dump_files_1): Merge into
4207 (pass_manager::register_dump_files): this, and remove its handling of
4208 properties since the pass always has the properties anyway.
4209 (pass_manager::pass_manager): Adjust.
4211 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
4213 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
4214 * passes.c (pass_manager::register_dump_files_1): Remove dead code
4215 dealing with properties.
4216 (pass_manager::register_dump_files): Adjust.
4218 2014-03-20 Mark Wielaard <mjw@redhat.com>
4220 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
4221 then represent the bound as normal constant value.
4223 2014-04-17 Jakub Jelinek <jakub@redhat.com>
4226 Forward port from 4.8 branch
4227 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
4229 * config/i386/bmiintrin.h (_blsi_u32): New.
4233 (_blsmsk_u32): Ditto.
4234 (_blsmsk_u64): Ditto.
4235 (_tzcnt_u32): Ditto.
4236 (_tzcnt_u64): Ditto.
4238 2014-04-17 Kito Cheng <kito@0xlab.org>
4240 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
4242 2014-04-17 Richard Biener <rguenther@suse.de>
4245 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
4246 boolean results for comparisons.
4248 2014-04-17 Richard Biener <rguenther@suse.de>
4250 PR tree-optimization/60836
4251 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
4252 initial PHI args to be gimple values.
4254 2014-04-17 Richard Biener <rguenther@suse.de>
4256 PR tree-optimization/60841
4257 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
4258 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
4259 of stmts to SLP build.
4260 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
4261 (vect_analyze_slp): Likewise.
4262 (vect_analyze_slp_instance): Likewise.
4263 (vect_build_slp_tree): Limit overall SLP tree growth.
4264 * tree-vectorizer.h (vect_analyze_data_refs,
4265 vect_analyze_slp): Adjust prototypes.
4267 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
4269 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
4272 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
4274 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
4275 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
4276 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
4277 for TARGET_SLOW_PSHUFB
4279 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
4281 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
4282 * config/i386/i386.c (intel_cost): Ditto.
4284 2014-04-17 Joey Ye <joey.ye@arm.com>
4286 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
4288 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
4290 * opts.c (common_handle_option): Disable -fipa-reference coorectly
4293 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
4295 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
4296 (type_all_derivations_known_p): New predicate.
4297 (type_all_ctors_visible_p): New predicate.
4298 (type_possibly_instantiated_p): New predicate.
4299 (get_odr_type): Compute all_derivations_known.
4300 (dump_odr_type): Dump the flag.
4301 (maybe_record_type): Cleanup.
4302 (record_target_from_binfo): Add bases_to_consider array;
4303 record bases for types w/o instances and skip CXX destructor.
4304 (possible_polymorphic_call_targets_1): Add bases_to_consider
4305 and consider_construction parameters; check if type may have instance.
4306 (get_polymorphic_call_info): Set maybe_in_construction to true
4307 when we know nothing.
4308 (record_targets_from_bases): Skip CXX destructors; they are
4309 never called for types in construction.
4310 (possible_polymorphic_call_targets): Do not record target when
4311 type may not have instance.
4313 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
4316 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
4317 external aliases alive, too.
4319 2014-04-16 Andrew Pinski <apinski@cavium.com>
4321 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
4324 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
4326 * final.c (compute_alignments): Do not apply loop alignment to a block
4327 falling through to the exit.
4329 2014-04-16 Catherine Moore <clm@codesourcery.com>
4331 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
4332 Adjust constraints for microMIPS store patterns.
4334 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
4336 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
4338 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
4340 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
4341 (append_use): Run at -O0.
4342 (append_vdef): Likewise.
4343 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
4344 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
4346 2014-04-16 Jakub Jelinek <jakub@redhat.com>
4348 PR tree-optimization/60844
4349 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
4350 (propagate_op_to_single_use, remove_visited_stmt_chain,
4351 linearize_expr, repropagate_negates, reassociate_bb): Use it
4352 instead of gsi_remove.
4354 2014-04-16 Martin Jambor <mjambor@suse.cz>
4356 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
4357 ipa_transforms_to_apply.
4358 (cgraph_function_versioning): Assert that old_node has empty
4359 ipa_transforms_to_apply.
4360 * trans-mem.c (ipa_tm_create_version): Likewise.
4361 * tree-inline.c (tree_function_versioning): Do not duplicate
4362 ipa_transforms_to_apply.
4364 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4367 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
4369 Pass necessary as flags on 64-bit Solaris/x86.
4370 Use lowercase relocs for x86_64-*-*.
4371 * configure: Regenerate.
4373 2014-04-15 Jan Hubicka <jh@suse.cz>
4375 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
4376 (maybe_record_node, likely_target_p): Use it.
4378 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4381 Revert following patch
4383 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
4386 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
4387 software floating point or no floating point registers, do not
4388 allow any type in the FPRs. Eliminate a test for SPE SIMD types
4389 in GPRs that occurs after we tested for GPRs that would never be
4392 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
4393 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
4394 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
4395 specifically allow DDmode, since that does not use the SPE SIMD
4398 2014-03-21 Mark Wielaard <mjw@redhat.com>
4400 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
4401 as unsigned or int depending on type and value used.
4403 2014-04-15 Richard Biener <rguenther@suse.de>
4405 PR rtl-optimization/56965
4406 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
4407 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
4409 * alias.c (true_dependence_1): Do not call
4410 nonoverlapping_component_refs_p.
4411 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
4412 nonoverlapping_component_refs_p.
4413 (indirect_refs_may_alias_p): Likewise.
4415 2014-04-15 Teresa Johnson <tejohnson@google.com>
4417 * cfg.c (dump_bb_info): Fix flags check.
4418 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
4420 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4422 PR rtl-optimization/60663
4423 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
4426 2014-04-15 Richard Biener <rguenther@suse.de>
4428 * lto-streamer.h (LTO_major_version): Bump to 4.
4430 2014-04-15 Richard Biener <rguenther@suse.de>
4432 * common.opt (lto_partition_model): New enum.
4433 (flto-partition=): Merge separate options with a single with argument,
4434 add -flto-partition=one support.
4435 * flag-types.h (enum lto_partition_model): Declare.
4436 * opts.c (finish_options): Remove duplicate -flto-partition=
4438 * lto-wrapper.c (run_gcc): Adjust.
4440 2014-04-15 Richard Biener <rguenther@suse.de>
4442 * alias.c (ncr_compar): New function.
4443 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
4445 2014-04-15 Richard Biener <rguenther@suse.de>
4447 * alias.c (record_component_aliases): Do not walk BINFOs.
4449 2014-04-15 Richard Biener <rguenther@suse.de>
4451 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
4452 Add struct function argument and adjust.
4453 (find_func_aliases_for_call): Likewise.
4454 (find_func_aliases): Likewise.
4455 (find_func_clobbers): Likewise.
4456 (intra_create_variable_infos): Likewise.
4457 (compute_points_to_sets): Likewise.
4458 (ipa_pta_execute): Adjust. Do not push/pop cfun.
4460 2014-04-15 Richard Biener <rguenther@suse.de>
4462 * tree.c (iterative_hash_expr): Use enum tree_code_class
4463 to store TREE_CODE_CLASS.
4464 (tree_block): Likewise.
4465 (tree_set_block): Likewise.
4466 * tree.h (fold_build_pointer_plus_loc): Use
4467 convert_to_ptrofftype_loc.
4469 2014-04-15 Jakub Jelinek <jakub@redhat.com>
4472 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
4475 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
4477 * cfgloop.h (struct loop): Move force_vectorize down.
4478 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
4479 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
4480 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
4481 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
4482 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
4483 * tree-core.h (enum annot_expr_kind): Add new kind values.
4484 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
4485 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
4487 * tree.def (ANNOTATE_EXPR): Tweak comment.
4489 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
4491 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
4494 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
4496 * tree.h (TYPE_IDENTIFIER): Declare.
4497 * tree.c (subrange_type_for_debug_p): Use it.
4498 * godump.c (go_format_type): Likewise.
4499 * dwarf2out.c (is_cxx_auto, modified_type_die,
4500 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
4501 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
4503 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
4506 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
4508 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
4510 * config/i386/i386.c (examine_argument): Return bool. Return true if
4511 parameter should be passed in memory.
4512 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
4513 (construct_container): Update calls to examine_argument.
4514 (function_arg_advance_64): Ditto.
4515 (return_in_memory_32): Merge with ix86_return_in_memory.
4516 (return_in_memory_64): Ditto.
4517 (return_in_memory_ms_64): Ditto.
4519 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
4521 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
4522 * coverage.c (coverage_compute_profile_id): Handle externally visible
4525 2014-04-14 Martin Jambor <mjambor@suse.cz>
4527 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
4528 DECL_DISREGARD_INLINE_LIMITS functions.
4530 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
4533 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
4535 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
4538 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
4539 optimize_insn_for_speed_p instead of
4540 optimize_function_for_speed_p.
4542 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
4544 * doc/invoke.texi (free): Document AArch64.
4546 2014-04-14 Richard Biener <rguenther@suse.de>
4548 PR tree-optimization/60042
4549 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
4550 (insert_into_preds_of_block): Do not prevent PHI insertion
4551 for REFERENCE exprs here ...
4552 (eliminate_dom_walker::before_dom_children): ... but prevent
4553 their use here under similar conditions when applied to the
4554 IL after PRE optimizations.
4556 2014-04-14 Richard Biener <rguenther@suse.de>
4558 * passes.def: Move early points-to after early SRA.
4560 2014-04-14 Richard Biener <rguenther@suse.de>
4562 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
4563 check for which sign-changes we allow when forwarding
4564 a converted value into a switch.
4566 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
4568 * stor-layout.c (place_field): Finalize non-constant offset for the
4571 2014-04-14 Richard Biener <rguenther@suse.de>
4573 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
4575 (expand_switch_using_bit_tests_p): Likewise.
4576 (process_switch): Compute and pass on speed_p based on the
4578 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
4579 optimize_bb_for_speed_p.
4581 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
4583 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
4584 * function.h (struct function): Rename has_force_vect_loops into
4585 has_force_vectorize_loops.
4586 * lto-streamer-in.c (input_cfg): Adjust for renaming.
4587 (input_struct_function_base): Likewise.
4588 * lto-streamer-out.c (output_cfg): Likewise.
4589 (output_struct_function_base): Likewise.
4590 * omp-low.c (expand_omp_simd): Likewise.
4591 * tree-cfg.c (move_sese_region_to_fn): Likewise.
4592 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
4593 (version_loop_for_if_conversion): Likewise.
4594 (tree_if_conversion): Likewise.
4595 (main_tree_if_conversion): Likewise.
4596 (gate_tree_if_conversion): Likewise.
4597 * tree-inline.c (copy_loops): Likewise.
4598 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
4599 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
4600 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
4601 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
4602 * tree-vectorizer.c (vectorize_loops): Likewise.
4603 * tree-vectorizer.h (unlimited_cost_model): Likewise.
4605 2014-04-14 Richard Biener <rguenther@suse.de>
4608 * lto-streamer-out.c (wrap_refs): New function.
4609 (lto_output): Wrap symbol references in global initializes in
4610 type-preserving MEM_REFs.
4612 2014-04-14 Christian Bruel <christian.bruel@st.com>
4614 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
4616 2014-04-14 Christian Bruel <christian.bruel@st.com>
4618 * config/sh/sh.md (setmemqi): New expand pattern.
4619 * config/sh/sh.h (CLEAR_RATIO): Define.
4620 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
4621 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
4623 2014-04-14 Richard Biener <rguenther@suse.de>
4626 * fold-const.c (negate_expr_p): Don't negate directional rounding
4628 (fold_negate_expr): Likewise.
4630 2014-04-14 Richard Biener <rguenther@suse.de>
4632 PR tree-optimization/59817
4633 PR tree-optimization/60453
4634 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
4635 recursion to catch all CHRECs in the scalar evolution and restrict
4636 the predicate for the remains appropriately.
4638 2014-04-12 Catherine Moore <clm@codesourcery.com>
4640 * config/mips/constraints.md: Add new register constraint "kb".
4641 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
4642 (*movhi_internal): Likewise.
4643 (*movqi_internal): Likewise.
4644 * config/mips/mips.h (M16_STORE_REGS): New register class.
4645 (REG_CLASS_NAMES): Add M16_STORE_REGS.
4646 (REG_CLASS_CONTENTS): Likewise.
4647 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
4649 2014-04-11 Tobias Burnus <burnus@net-b.de>
4652 * doc/invoke.texi (-Wformat-signedness): Document it.
4653 (Wformat=2): Mention that this enables -Wformat-signedness.
4655 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
4657 * common/config/epiphany/epiphany-common.c
4658 (epiphany_option_optimization_table): Enable section anchors by
4659 default at -O1 or higher.
4660 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
4661 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
4662 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
4663 carries no extra cost.
4664 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
4665 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
4666 * config/epiphany/predicates.md (memclob_operand): New predicate.
4667 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
4668 Use memclob_operand predicate and X constraint for operand 3.
4670 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
4672 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
4673 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
4676 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
4678 PR rtl-optimization/60651
4679 * mode-switching.c (optimize_mode_switching): Make sure to emit
4680 sets of a lower numbered entity before sets of a higher numbered
4681 entity to a mode of the same or lower priority.
4682 When creating a seginfo for a basic block that starts with a code
4683 label, move the insertion point past the code label.
4684 (new_seginfo): Document and enforce requirement that
4685 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
4686 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
4687 * doc/tm.texi: Regenerate.
4689 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
4692 * config/arc/arc.c (arc_save_restore): Fix assert typo.
4694 2013-04-11 Jakub Jelinek <jakub@redhat.com>
4696 * BASE-VER: Set to 4.10.0.
4698 2014-04-11 Tobias Burnus <burnus@net-b.de>
4701 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
4702 * doc/gcc.texi (Service): Update description in the @menu
4703 * doc/invoke.texi (Option Summary): Remove misplaced and
4706 2014-04-11 Steve Ellcey <sellcey@mips.com>
4707 Jakub Jelinek <jakub@redhat.com>
4710 * expr.c (convert_move): Use emit_store_flag_force instead of
4711 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
4714 2014-04-11 Richard Biener <rguenther@suse.de>
4717 * varasm.c (assemble_alias): Avoid endless error reporting
4718 recursion by setting TREE_ASM_WRITTEN.
4720 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4722 * config/s390/s390.md: Add a splitter for NOT rtx.
4724 2014-04-11 Jakub Jelinek <jakub@redhat.com>
4726 PR rtl-optimization/60663
4727 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
4729 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
4730 Jakub Jelinek <jakub@redhat.com>
4733 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
4734 flag from decl_node to node.
4736 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4739 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
4740 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
4741 ameliorating the cases where it can be.
4743 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
4746 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
4748 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
4749 (loadsync_<mode>): Change mode.
4750 (load_quadpti, store_quadpti): New.
4751 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
4752 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
4753 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
4755 2014-04-09 Cong Hou <congh@google.com>
4758 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
4761 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4763 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
4764 instead of vnor to exploit possible fusion opportunity in the
4766 (altivec_expand_vec_perm_const_le): Likewise.
4768 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
4770 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
4771 (loadsync_<mode>): Change mode.
4772 (load_quadpti, store_quadpti): New.
4773 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
4774 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
4776 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
4779 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
4780 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
4781 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
4783 2014-04-08 Richard Biener <rguenther@suse.de>
4786 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
4787 a 64bit widest int print double-int similar to on HWI64 hosts.
4789 2014-04-08 Richard Biener <rguenther@suse.de>
4791 PR tree-optimization/60785
4792 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
4793 default defs properly.
4795 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
4797 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
4798 (Weffc++): Likewise.
4800 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
4802 * ipa-devirt.c (maybe_record_node): When node is not recorded,
4803 set completep to false rather than true.
4805 2014-04-07 Douglas B Rupp <rupp@adacore.com>
4808 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
4809 ARM_TARGET2_DWARF_FORMAT.
4811 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
4814 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
4815 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
4818 2014-04-07 Richard Biener <rguenther@suse.de>
4820 PR tree-optimization/60766
4821 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
4822 (may_eliminate_iv): Convert cand_value_at result to desired type.
4824 2014-04-07 Jason Merrill <jason@redhat.com>
4827 * common.opt (-fno-gnu-unique): Add.
4828 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
4830 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4832 * haifa-sched.c: Fix outdated function reference and minor
4833 grammar errors in introductory comment.
4835 2014-04-07 Richard Biener <rguenther@suse.de>
4838 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
4840 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
4842 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
4845 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
4846 size accounting for thunks.
4847 (pa_asm_output_mi_thunk): Use final_start_function() and
4848 final_end_function() to output function start and end directives.
4850 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
4852 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
4853 device specific ISA/ feature information. Remove short_sp and
4854 errata_skip ds. Add avr_device_specific_features enum to have device
4856 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
4857 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
4858 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
4859 updated device specific info.
4860 * config/avr/avr-mcus.def: Merge device specific details to
4861 dev_attribute field.
4862 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
4864 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
4865 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
4866 assembler if RMW isa supported by current device.
4867 * config/avr/genmultilib.awk: Update as device info structure changed.
4868 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
4870 2014-04-04 Cong Hou <congh@google.com>
4872 PR tree-optimization/60656
4873 * tree-vect-stmts.c (supportable_widening_operation):
4874 Fix a bug that elements in a vector with vect_used_by_reduction
4875 property are incorrectly reordered when the operation on it is not
4876 consistant with the one in reduction operation.
4878 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
4880 PR rtl-optimization/60155
4881 * gcse.c (record_set_data): New function.
4882 (single_set_gcse): New function.
4883 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
4884 (hoist_code): Likewise.
4885 (get_pressure_class_and_nregs): Likewise.
4887 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
4889 * explow.c (probe_stack_range): Emit a final optimization blockage.
4891 2014-04-04 Anthony Green <green@moxielogic.com>
4893 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
4896 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
4899 * lto-cgraph.c (input_overwrite_node): Check that partitioning
4900 flags are set only during streaming.
4901 * ipa.c (process_references, walk_polymorphic_call_targets,
4902 symtab_remove_unreachable_nodes): Drop bodies of always inline
4903 after early inlining.
4904 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
4906 2014-04-04 Jakub Jelinek <jakub@redhat.com>
4907 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4910 * dwarf2out.c (const_ok_for_output_1): Reject expressions
4913 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4916 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
4918 (cortex_a53_fdivd): Likewise.
4920 2014-04-04 Martin Jambor <mjambor@suse.cz>
4923 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
4925 * cgraph.c (clone_of_p): Also return true if thunks match.
4926 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
4927 cgraph_function_or_thunk_node and an obsolete comment.
4928 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
4930 (build_function_decl_skip_args): Likewise.
4931 (set_new_clone_decl_and_node_flags): New function.
4932 (duplicate_thunk_for_node): Likewise.
4933 (redirect_edge_duplicating_thunks): Likewise.
4934 (cgraph_clone_node): New parameter args_to_skip, pass it to
4935 redirect_edge_duplicating_thunks which is called instead of
4936 cgraph_redirect_edge_callee.
4937 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
4938 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
4940 2014-04-04 Jeff Law <law@redhat.com>
4943 * config/arm/predicates.md (const_int_I_operand): New predicate.
4944 (const_int_M_operand): Similarly.
4945 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
4947 (insv_t2, extv_reg, extzv_t2): Likewise.
4948 (load_multiple_with_writeback): Similarly for const_int_I_operand.
4949 (pop_multiple_with_writeback_and_return): Likewise.
4950 (vfp_pop_multiple_with_writeback): Likewise
4952 2014-04-04 Richard Biener <rguenther@suse.de>
4955 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
4956 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
4958 * gimplify.h (gimple_add_tmp_var_fn): Declare.
4959 * gimplify.c (gimple_add_tmp_var_fn): New function.
4960 * gimple-expr.h (create_tmp_reg_fn): Declare.
4961 * gimple-expr.c (create_tmp_reg_fn): New function.
4962 * gimple-low.c (record_vars_into): Don't change cfun.
4963 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
4964 code generation without cfun.
4966 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
4969 * Makefile.in (install-driver): Fix shell scripting.
4971 2014-04-03 Cong Hou <congh@google.com>
4973 PR tree-optimization/60505
4974 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
4975 threshold of number of iterations below which no vectorization
4977 * tree-vect-loop.c (new_loop_vec_info):
4978 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
4979 * tree-vect-loop.c (vect_analyze_loop_operations):
4980 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
4981 * tree-vect-loop.c (vect_transform_loop):
4982 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
4983 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
4984 of iterations of the loop and see if we should build the epilogue.
4986 2014-04-03 Richard Biener <rguenther@suse.de>
4988 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
4989 (streamer_tree_cache_create): Adjust.
4990 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
4991 to allow optional nodes array.
4992 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
4993 (streamer_tree_cache_append): Likewise.
4994 (streamer_tree_cache_create): Create nodes array optionally
4995 as specified by parameter.
4996 * lto-streamer-out.c (create_output_block): Avoid maintaining
4997 the node array in the writer cache.
4998 (DFS_write_tree): Remove assertion.
4999 (produce_asm_for_decls): Free the out decl state hash table early.
5000 * lto-streamer-in.c (lto_data_in_create): Adjust for
5001 streamer_tree_cache_create prototype change.
5003 2014-04-03 Richard Biener <rguenther@suse.de>
5005 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
5006 set TREE_CHAIN to NULL_TREE.
5008 2014-04-03 Richard Biener <rguenther@suse.de>
5010 PR tree-optimization/60740
5011 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
5012 over all GIMPLE_COND operands.
5014 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
5016 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
5017 (Weffc++): Remove Scott's numbering, merge lists and reference
5020 2014-04-03 Nick Clifton <nickc@redhat.com>
5022 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
5025 2014-04-03 Martin Jambor <mjambor@suse.cz>
5027 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
5028 mention gcc_unreachable before failing.
5029 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
5032 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
5035 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
5036 inconsistent code and instead mark the context inconsistent.
5037 (possible_polymorphic_call_targets): For inconsistent contexts
5038 return empty complete list.
5040 2014-04-02 Anthony Green <green@moxielogic.com>
5042 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
5043 (extendqisi2, extendhisi2): Define.
5044 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
5045 (WCHAR_TYPE): Change to unsigned int.
5047 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5049 PR tree-optimization/60733
5050 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
5051 insertion point for PHI candidates to be the end of the feeding
5052 block for the PHI argument.
5054 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
5056 PR rtl-optimization/60650
5057 * lra-constraints.c (process_alt_operands): Decrease reject for
5058 earlyclobber matching.
5060 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5062 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
5064 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
5066 * config/spu/spu.c (pad_bb): Do not crash when the last
5067 insn is CODE_FOR_blockage.
5069 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
5071 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
5072 lies outside the target mode.
5074 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
5077 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
5078 software floating point or no floating point registers, do not
5079 allow any type in the FPRs. Eliminate a test for SPE SIMD types
5080 in GPRs that occurs after we tested for GPRs that would never be
5083 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
5084 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
5085 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
5086 specifically allow DDmode, since that does not use the SPE SIMD
5089 2014-04-02 Richard Biener <rguenther@suse.de>
5092 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
5093 MODE_INTs. Properly use negv_optab.
5094 (expand_abs): Likewise.
5096 2014-04-02 Richard Biener <rguenther@suse.de>
5099 * Makefile.in (install-driver): Guard extra installs with special
5102 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
5104 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
5107 2014-04-01 Richard Henderson <rth@redhat.com>
5110 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
5111 alternative enabled before register allocation.
5113 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
5115 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
5116 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
5118 (nios2_large_got_address): Remove unneeded 'sym' parameter.
5119 (nios2_got_address): Update nios2_large_got_address call site.
5120 (nios2_delegitimize_address): New function.
5121 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
5122 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
5123 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
5125 2014-04-01 Martin Husemann <martin@duskware.de>
5127 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
5130 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
5132 PR rtl-optimization/60604
5133 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
5134 check from register_operand.
5135 (register_operand): Redefine in terms of general_operand.
5136 (nonmemory_operand): Use register_operand for the non-constant cases.
5138 2014-04-01 Richard Biener <rguenther@suse.de>
5140 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
5142 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
5144 * doc/invoke.texi (mapp-regs): Clarify.
5146 2014-03-31 Ulrich Drepper <drepper@gmail.com>
5148 * config/i386/avx512fintrin.h (__v32hi): Define type.
5149 (__v64qi): Likewise.
5150 (_mm512_set1_epi8): Define.
5151 (_mm512_set1_epi16): Define.
5152 (_mm512_set4_epi32): Define.
5153 (_mm512_set4_epi64): Define.
5154 (_mm512_set4_pd): Define.
5155 (_mm512_set4_ps): Define.
5156 (_mm512_setr4_epi64): Define.
5157 (_mm512_setr4_epi32): Define.
5158 (_mm512_setr4_pd): Define.
5159 (_mm512_setr4_ps): Define.
5160 (_mm512_setzero_epi32): Define.
5162 2014-03-31 Martin Jambor <mjambor@suse.cz>
5165 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
5166 callsite_arguments_match_p. Updated all callers. Also check types of
5167 corresponding formal parameters and actual arguments.
5168 (not_all_callers_have_enough_arguments_p) Renamed to
5169 some_callers_have_mismatched_arguments_p.
5171 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
5173 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
5175 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
5178 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
5181 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
5183 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
5184 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
5186 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
5187 Use FMAMODE_NOVF512 mode iterator.
5188 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
5189 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
5190 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
5191 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
5193 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
5194 Use VF_128_256 mode iterator.
5195 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
5198 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
5200 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
5201 static chain if needed.
5203 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
5206 * lra-constraints.c (index_part_to_reg): New.
5207 (process_address): Use it.
5209 2014-03-27 Jeff Law <law@redhat.com>
5210 Jakub Jelinek <jakub@redhat.com>
5213 * expr.c (do_tablejump): Use simplify_gen_binary rather than
5214 gen_rtx_{PLUS,MULT} to build up the address expression.
5216 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
5217 creating non-canonical RTL.
5219 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
5222 * ipa-inline.c (want_inline_small_function_p): Short circuit large
5223 functions; reorganize to make cheap checks first.
5224 (inline_small_functions): Do not estimate growth when dumping;
5226 * ipa-inline.h (inline_summary): Add min_size.
5227 (growth_likely_positive): New function.
5228 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
5229 (set_cond_stmt_execution_predicate): Cleanup.
5230 (estimate_edge_size_and_time): Compute min_size.
5231 (estimate_calls_size_and_time): Likewise.
5232 (estimate_node_size_and_time): Likewise.
5233 (inline_update_overall_summary): Update min_size.
5234 (do_estimate_edge_time): Likewise.
5235 (do_estimate_edge_size): Update.
5236 (do_estimate_edge_hints): Update.
5237 (growth_likely_positive): New function.
5239 2014-03-28 Jakub Jelinek <jakub@redhat.com>
5242 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
5243 also if addr has VOIDmode.
5245 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5247 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
5248 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
5250 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
5251 instructions as well as AdvancedSIMD loads.
5253 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5255 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
5256 Use crypto_aese type.
5257 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
5258 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
5259 crypto_aese, crypto_aesmc. Move to types.md.
5260 * config/arm/types.md (crypto_aes): Split into crypto_aese,
5262 * config/arm/iterators.md (crypto_type): Likewise.
5264 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
5266 * cgraph.c: Include expr.h and tree-dfa.h.
5267 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
5270 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
5273 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
5274 regs from checking multi-reg pseudos.
5276 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5278 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
5280 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
5282 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
5283 if it would clobber the stack pointer, even temporarily.
5285 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
5287 * mode-switching.c: Make small adjustments to the top comment.
5289 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
5291 * config/rs6000/constraints.md (wD constraint): New constraint to
5292 match the constant integer to get the top DImode/DFmode out of a
5293 vector in a VSX register.
5295 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
5296 match the constant integer to get the top DImode/DFmode out of a
5297 vector in a VSX register.
5299 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
5302 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
5305 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
5306 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
5308 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
5309 Optimize vec_extract of 64-bit values, where the value being
5310 extracted is in the top word, where we can use scalar
5311 instructions. Add direct move and store support. Combine the big
5312 endian/little endian vector select load support into a single insn.
5313 (vsx_extract_<mode>_internal1): Likewise.
5314 (vsx_extract_<mode>_internal2): Likewise.
5315 (vsx_extract_<mode>_load): Likewise.
5316 (vsx_extract_<mode>_store): Likewise.
5317 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
5318 combined into vsx_extract_<mode>_load.
5319 (vsx_extract_<mode>_one_le): Likewise.
5321 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
5322 define the top 64-bit vector element.
5324 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
5327 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
5328 Document vec_vbpermq builtin.
5331 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
5332 enable use of xxsldwi and xxpermdi builtin functions.
5333 (vec_xxpermdi): Likewise.
5335 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
5336 Document use of vec_xxsldwi and vec_xxpermdi builtins.
5338 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
5340 PR rtl-optimization/60650
5341 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
5343 (find_spills_for): New.
5344 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
5345 Spill all pseudos on the second iteration.
5347 2014-03-27 Marek Polacek <polacek@redhat.com>
5350 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
5353 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5355 * config/s390/s390.c (s390_can_use_return_insn): Check for
5356 call-saved FPRs on 31 bit.
5358 2014-03-27 Jakub Jelinek <jakub@redhat.com>
5361 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
5362 if they need regimplification, just drop them instead of
5363 calling gimple_regimplify_operands on them.
5365 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
5368 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
5369 (aarch64_frame_pointer_required): Adjust logic.
5370 (aarch64_can_eliminate): Adjust logic.
5371 (aarch64_override_options_after_change): Adjust logic.
5373 2014-03-27 Dehao Chen <dehao@google.com>
5375 * ipa-inline.c (early_inliner): Update node's inline info.
5377 2014-03-26 Dehao Chen <dehao@google.com>
5379 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
5380 compiler inserted conditional jumps for NAN float check.
5382 2014-03-26 Jakub Jelinek <jakub@redhat.com>
5384 * ubsan.h (ubsan_create_data): Change second argument's type
5385 to const location_t *.
5386 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
5388 (ubsan_create_data): Change second argument to const location_t *PLOC.
5389 Create Loc field whenever PLOC is non-NULL.
5390 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
5391 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
5395 * real.c (real_to_integer2): Change type of low to UHWI.
5397 2014-03-26 Tobias Burnus <burnus@net-b.de>
5399 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
5400 (CILK_SELF_SPECS): New define.
5401 (driver_self_specs): Use it.
5403 2014-03-26 Richard Biener <rguenther@suse.de>
5405 * tree-pretty-print.c (percent_K_format): Implement special
5406 case for LTO and its stripped down BLOCK tree.
5408 2014-03-26 Jakub Jelinek <jakub@redhat.com>
5411 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
5413 * tree-vrp.c (simplify_internal_call_using_ranges): If only
5414 one range is range_int_cst_p, but not both, at least optimize
5415 addition/subtraction of 0 and multiplication by 0 or 1.
5416 * gimple-fold.c (gimple_fold_call): Fold
5417 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
5418 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
5419 INTEGER_CSTs, try to fold at least x * 0 and y - y.
5421 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
5423 PR rtl-optimization/60452
5424 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
5425 <case REG>: Return 1 for invalid offsets from the frame pointer.
5427 2014-03-26 Marek Polacek <polacek@redhat.com>
5430 * doc/extend.texi (C Extensions): Mention variable-length arrays in
5433 2014-03-26 Marek Polacek <polacek@redhat.com>
5436 * doc/extend.texi (Designated Inits): Describe what happens to omitted
5439 2014-03-26 Marek Polacek <polacek@redhat.com>
5442 * ira-color.c (update_conflict_hard_regno_costs): Perform the
5443 multiplication in unsigned type.
5445 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
5447 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
5449 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
5451 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
5453 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
5456 * cif-code.def (UNREACHABLE) New code.
5457 * ipa-inline.c (inline_small_functions): Skip edges to
5458 __builtlin_unreachable.
5459 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
5460 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
5461 predicate to __bulitin_unreachable.
5462 (set_cond_stmt_execution_predicate): Fix issue when
5463 invert_tree_comparison returns ERROR_MARK.
5464 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
5465 propagate to inline clones.
5466 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
5468 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
5469 * cgraphclones.c (cgraph_clone_node): If call destination is already
5470 ureachable, do not redirect it back.
5471 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
5474 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
5476 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
5477 Do not modify inline clones.
5479 2014-03-25 Jakub Jelinek <jakub@redhat.com>
5481 * config/i386/i386.md (general_sext_operand): New mode attr.
5482 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
5483 don't generate (sign_extend (const_int)).
5484 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
5485 operands[2]. Use We constraint instead of <i> and
5486 <general_sext_operand> predicate instead of <general_operand>.
5487 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
5488 * config/i386/constraints.md (We): New constraint.
5489 * config/i386/predicates.md (x86_64_sext_operand,
5490 sext_operand): New predicates.
5492 2014-03-25 Martin Jambor <mjambor@suse.cz>
5495 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
5496 inconsistent devirtualizations to __builtin_unreachable.
5498 2014-03-25 Marek Polacek <polacek@redhat.com>
5501 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
5503 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
5505 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
5506 order of elements for big-endian.
5508 2014-03-25 Richard Biener <rguenther@suse.de>
5511 * gimplify-me.c (gimple_regimplify_operands): Update the
5514 2014-03-25 Martin Jambor <mjambor@suse.cz>
5517 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
5518 (lto_output_varpool_node): Likewise.
5519 (input_overwrite_node): Likewise.
5520 (input_varpool_node): Likewise.
5522 2014-03-25 Richard Biener <rguenther@suse.de>
5524 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
5525 (run_gcc): Likewise.
5527 2014-03-25 Jakub Jelinek <jakub@redhat.com>
5529 * combine.c (simplify_compare_const): Add MODE argument.
5530 Handle mode_width 0 as very large mode_width.
5531 (try_combine, simplify_comparison): Adjust callers.
5533 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
5534 type to avoid signed integer overflow.
5535 * explow.c (plus_constant): Likewise.
5537 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
5539 * doc/generic.texi: Correct typos.
5541 2014-03-24 Tobias Burnus <burnus@net-b.de>
5543 * doc/invoke.texi (-flto): Expand section about
5544 using static libraries with LTO.
5546 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5548 PR rtl-optimization/60501
5549 * optabs.def (addptr3_optab): New optab.
5550 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
5551 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
5552 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
5554 * lra.c (emit_add3_insn): Use the addptr pattern if available.
5556 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
5558 2014-03-24 Ulrich Drepper <drepper@gmail.com>
5560 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
5563 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
5564 (_mm256_undefined_ps): Define.
5565 (_mm256_undefined_pd): Define.
5566 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
5567 (_mm_undefined_pd): Define.
5568 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
5569 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
5570 (_mm512_undefined_ps): Define.
5571 (_mm512_undefined_pd): Define.
5572 Use _mm*_undefined_*.
5573 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
5575 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
5577 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
5578 (lshr_simd): DI mode added.
5579 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
5580 (aarch64_ushr_simddi): Likewise.
5581 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
5582 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
5583 (vshrd_n_u64): Likewise.
5585 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5587 * Makefile.in (s-macro_list): Depend on cc1.
5589 2014-03-23 Teresa Johnson <tejohnson@google.com>
5591 * ipa-utils.c (ipa_print_order): Use specified dump file.
5593 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
5595 PR rtl-optimization/60601
5596 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
5598 * gcc.c (eval_spec_function): Initialize save_growing_value.
5600 2014-03-22 Jakub Jelinek <jakub@redhat.com>
5603 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
5604 code == MINUS_EXPR, never swap op0 with op1.
5606 * toplev.c (init_local_tick): Avoid signed integer multiplication
5608 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
5609 shift by first operand's bitsize.
5611 2014-03-21 Jakub Jelinek <jakub@redhat.com>
5614 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
5616 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
5617 TARGET_ISA_64BIT_P(x).
5619 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5621 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
5622 pattern for vector nor instead of subtract from splat(-1).
5623 (altivec_expand_vec_perm_const_le): Likewise.
5625 2014-03-21 Richard Henderson <rth@twiddle.net>
5628 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
5629 related insns after epilogue_completed.
5631 2014-03-21 Martin Jambor <mjambor@suse.cz>
5634 * cgraph.h (symtab_node): New flag body_removed.
5635 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
5636 when removing bodies.
5637 * symtab.c (dump_symtab_base): Dump body_removed flag.
5638 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
5639 had their bodies removed.
5641 2014-03-21 Martin Jambor <mjambor@suse.cz>
5644 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
5647 2014-03-21 Richard Biener <rguenther@suse.de>
5649 PR tree-optimization/60577
5650 * tree-core.h (struct tree_base): Document nothrow_flag use
5651 in VAR_DECL_NONALIASED.
5652 * tree.h (VAR_DECL_NONALIASED): New.
5653 (may_be_aliased): Adjust.
5654 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
5656 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
5658 * expr.c (expand_expr_real_1): Remove outdated comment.
5660 2014-03-20 Jakub Jelinek <jakub@redhat.com>
5663 * ira.c (adjust_cleared_regs): Call copy_rtx on
5664 *reg_equiv[REGNO (loc)].src_p before passing it to
5665 simplify_replace_fn_rtx.
5668 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
5669 into CONST, put pic register as first operand of PLUS. Use
5670 gen_const_mem for both 32-bit and 64-bit PIC got loads.
5672 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5674 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
5676 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
5678 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
5679 around for store forwarding issue in the FPU on the UT699.
5680 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
5681 loads and operations if -mfix-ut699 is specified.
5682 (divtf3_hq): Tweak attribute.
5683 (sqrttf2_hq): Likewise.
5685 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
5687 * calls.c (store_one_arg): Remove incorrect const qualification on the
5688 type of the temporary.
5689 * cfgexpand.c (expand_return): Likewise.
5690 * expr.c (expand_constructor): Likewise.
5691 (expand_expr_real_1): Likewise.
5693 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
5695 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
5698 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
5701 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
5703 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
5705 * config/arm/aarch-common-protos.h
5706 (alu_cost_table): Fix spelling of "extend".
5707 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
5709 2014-03-19 Richard Biener <rguenther@suse.de>
5712 * tree-core.h (tree_type_common): Re-order pointer members
5713 to reduce recursion depth during GC walks.
5715 2014-03-19 Marek Polacek <polacek@redhat.com>
5718 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
5719 before accessing it.
5721 2014-03-19 Richard Biener <rguenther@suse.de>
5724 * lto-streamer-in.c (input_function): In WPA stage do not drop
5727 2014-03-19 Jakub Jelinek <jakub@redhat.com>
5729 PR tree-optimization/60559
5730 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
5731 with build_zero_cst assignment.
5733 2014-03-18 Kai Tietz <ktietz@redhat.com>
5735 PR rtl-optimization/56356
5736 * sdbout.c (sdbout_parms): Verify that parms'
5737 incoming argument is valid.
5738 (sdbout_reg_parms): Likewise.
5740 2014-03-18 Richard Henderson <rth@redhat.com>
5743 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
5744 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
5745 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
5747 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
5749 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
5750 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
5751 Italicize plugin event names in description. Explain that
5752 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
5753 Remind that no GCC functions should be called after PLUGIN_FINISH.
5754 Explain what pragmas with expansion are.
5756 2014-03-18 Martin Liska <mliska@suse.cz>
5758 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
5759 gimple call statement is update.
5760 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
5761 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
5763 2014-03-18 Jakub Jelinek <jakub@redhat.com>
5766 * ubsan.c (ubsan_instrument_unreachable): Call
5767 initialize_sanitizer_builtins.
5768 (ubsan_pass): Likewise.
5771 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
5772 varpool_finalize_decl instead of rest_of_decl_compilation.
5774 2014-03-18 Richard Biener <rguenther@suse.de>
5776 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
5777 by using bitmap_and_compl instead of bitmap_and_compl_into.
5778 (df_rd_transfer_function): Likewise.
5780 2014-03-18 Richard Biener <rguenther@suse.de>
5782 * doc/lto.texi (fresolution): Fix typo.
5784 2014-03-18 Richard Biener <rguenther@suse.de>
5786 * doc/invoke.texi (flto): Update for changes in 4.9.
5788 2014-03-18 Richard Biener <rguenther@suse.de>
5790 * doc/loop.texi: Remove section on the removed lambda framework.
5791 Update loop docs with recent changes in preserving loop structure.
5793 2014-03-18 Richard Biener <rguenther@suse.de>
5795 * doc/lto.texi (-fresolution): Document.
5797 2014-03-18 Richard Biener <rguenther@suse.de>
5799 * doc/contrib.texi: Adjust my name.
5801 2014-03-18 Jakub Jelinek <jakub@redhat.com>
5804 * internal-fn.c: Include diagnostic-core.h.
5805 (expand_BUILTIN_EXPECT): New function.
5806 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
5807 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
5808 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
5809 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
5811 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
5812 Revert 3 argument __builtin_expect code.
5813 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
5814 * gimple-fold.c (gimple_fold_call): Likewise.
5815 * tree.h (fold_builtin_expect): New prototype.
5816 * builtins.c (build_builtin_expect_predicate): Add predictor
5817 argument, if non-NULL, create 3 argument __builtin_expect.
5818 (fold_builtin_expect): No longer static. Add ARG2 argument,
5819 pass it through to build_builtin_expect_predicate.
5820 (fold_builtin_2): Adjust caller.
5821 (fold_builtin_3): Handle BUILT_IN_EXPECT.
5822 * internal-fn.def (BUILTIN_EXPECT): New.
5824 2014-03-18 Tobias Burnus <burnus@net-b.de>
5827 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
5828 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
5829 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
5831 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
5834 * predict.c (combine_predictions_for_bb): Fix up formatting.
5835 (expr_expected_value_1, expr_expected_value): Add predictor argument,
5836 fill what it points to if non-NULL.
5837 (tree_predict_by_opcode): Adjust caller, use the predictor.
5838 * predict.def (PRED_COMPARE_AND_SWAP): Add.
5840 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
5842 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
5843 proper constant for the store mode.
5845 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
5847 * symtab.c (change_decl_assembler_name): Fix transparent alias
5850 2014-03-16 Renlin Li <Renlin.Li@arm.com>
5852 * config/aarch64/aarch64.c: Correct the comments about the
5853 aarch64 stack layout.
5855 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
5857 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
5858 check for GF_OMP_FOR_KIND_FOR.
5860 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
5862 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
5863 ymm and zmm register names.
5865 2014-03-17 Jakub Jelinek <jakub@redhat.com>
5868 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
5869 note creation for the 2010-08-31 changes.
5871 2014-03-17 Marek Polacek <polacek@redhat.com>
5874 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
5875 as -fno-tree-loop-vectorize.
5876 (expand_omp_simd): Likewise.
5878 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
5880 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
5881 (eligible_for_call_delay): New prototype.
5882 * config/sparc/sparc.c (tls_call_delay): Rename into...
5883 (eligible_for_call_delay): ...this. Return false if the instruction
5884 cannot be put in the delay slot of a branch.
5885 (eligible_for_restore_insn): Simplify.
5886 (eligible_for_return_delay): Return false if the instruction cannot be
5887 put in the delay slot of a branch and simplify.
5888 (eligible_for_sibcall_delay): Return false if the instruction cannot be
5889 put in the delay slot of a branch.
5890 * config/sparc/sparc.md (fix_ut699): New attribute.
5891 (tls_call_delay): Delete.
5892 (in_call_delay): Reimplement.
5893 (eligible_for_sibcall_delay): Rename into...
5894 (in_sibcall_delay): ...this.
5895 (eligible_for_return_delay): Rename into...
5896 (in_return_delay): ...this.
5897 (in_branch_delay): Reimplement.
5898 (in_uncond_branch_delay): Delete.
5899 (in_annul_branch_delay): Delete.
5901 2014-03-14 Richard Henderson <rth@redhat.com>
5904 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
5905 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
5906 (*floathi<X87MODEF>2_i387_with_temp): Remove.
5907 (floathi splitters): Remove.
5908 (float<SWI48x>xf2): New pattern.
5909 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
5910 code that tried to handle DImode for 32-bit, but which was excluded
5911 by the pattern's condition. Drop allocation of stack temporary.
5912 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
5913 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
5914 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
5915 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
5916 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
5917 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
5918 (*float<SWI48><MODEF>2_sse_interunit): Remove.
5919 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
5920 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
5921 (*float<SWI48x><X87MODEF>2_i387): Remove.
5922 (all float _with_temp splitters): Remove.
5923 (*float<SWI48x><MODEF>2_i387): New pattern.
5924 (*float<SWI48><MODEF>2_sse): New pattern.
5925 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
5926 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
5928 2014-03-14 Jakub Jelinek <jakub@redhat.com>
5929 Marek Polacek <polacek@redhat.com>
5932 * common.opt (dump_base_name_prefixed): New Variable.
5933 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
5934 if x_dump_base_name_prefixed is already set, set it at the end.
5936 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
5938 PR rtl-optimization/60508
5939 * lra-constraints.c (get_reload_reg): Add new parameter
5941 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
5942 Pass the new parameter values.
5944 2014-03-14 Richard Biener <rguenther@suse.de>
5946 * common.opt: Revert unintented changes from r205065.
5949 2014-03-14 Richard Biener <rguenther@suse.de>
5952 * cfghooks.c (split_block): Properly adjust all loops the
5953 block was a latch of.
5955 2014-03-14 Martin Jambor <mjambor@suse.cz>
5958 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
5961 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
5964 * config/avr/avr.c (avr_set_current_function): Pass function name
5965 through default_strip_name_encoding before sanity checking instead
5966 of skipping the first char of the assembler name.
5968 2014-03-13 Richard Henderson <rth@redhat.com>
5971 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
5972 (ix86_force_to_memory, ix86_free_from_memory): Remove.
5973 * config/i386/i386-protos.h: Likewise.
5974 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
5975 in the expander instead of a splitter.
5976 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
5977 any possibility of requiring a memory.
5978 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
5979 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
5980 (fp branch splitters): Update for ix86_split_fp_branch.
5981 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
5982 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
5983 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
5984 (*fop_<MODEF>_2_i387): Remove f/r alternative.
5985 (*fop_<MODEF>_3_i387): Likewise.
5986 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
5987 (splitters for the fop_* register patterns): Remove.
5988 (fscalexf4_i387): Rename from *fscalexf4_i387.
5989 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
5991 2014-03-13 Jakub Jelinek <jakub@redhat.com>
5993 PR tree-optimization/59779
5994 * tree-dfa.c (get_ref_base_and_extent): Use double_int
5995 type for bitsize and maxsize instead of HOST_WIDE_INT.
5997 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
5999 PR rtl-optimization/57320
6000 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
6001 the CFG after thread_prologue_and_epilogue_insns.
6003 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
6005 PR rtl-optimization/57189
6006 * lra-constraints.c (process_alt_operands): Disfavor spilling
6009 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
6011 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
6013 2014-03-13 Jakub Jelinek <jakub@redhat.com>
6015 PR tree-optimization/59025
6017 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
6018 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
6020 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
6023 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
6024 calls of avr_out_plus_1.
6026 2014-03-13 Bin Cheng <bin.cheng@arm.com>
6028 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
6029 BB's single pred and update the father loop's latch info later.
6031 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
6033 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
6037 (VEC_base): Likewise.
6038 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
6039 registers, we need to swap double words in little endian mode.
6041 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
6042 to be a container mode for 128-bit integer operations added in ISA
6043 2.07. Unlike TImode and PTImode, the preferred register set is
6044 the Altivec/VMX registers for the 128-bit operations.
6046 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
6048 (rs6000_split_128bit_ok_p): Likewise.
6050 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
6051 macros for creating ISA 2.07 normal and overloaded builtin
6052 functions with 3 arguments.
6053 (BU_P8V_OVERLOAD_3): Likewise.
6054 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
6055 for use as overloaded functions.
6056 (VPERM_1TI_UNS): Likewise.
6057 (VSEL_1TI): Likewise.
6058 (VSEL_1TI_UNS): Likewise.
6059 (ST_INTERNAL_1ti): Likewise.
6060 (LD_INTERNAL_1ti): Likewise.
6061 (XXSEL_1TI): Likewise.
6062 (XXSEL_1TI_UNS): Likewise.
6063 (VPERM_1TI): Likewise.
6064 (VPERM_1TI_UNS): Likewise.
6065 (XXPERMDI_1TI): Likewise.
6066 (SET_1TI): Likewise.
6067 (LXVD2X_V1TI): Likewise.
6068 (STXVD2X_V1TI): Likewise.
6069 (VEC_INIT_V1TI): Likewise.
6070 (VEC_SET_V1TI): Likewise.
6071 (VEC_EXT_V1TI): Likewise.
6072 (EQV_V1TI): Likewise.
6073 (NAND_V1TI): Likewise.
6074 (ORC_V1TI): Likewise.
6075 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
6076 added in ISA 2.07. Add both normal 'altivec' builtins, and the
6078 (VADDUQM): Likewise.
6079 (VSUBCUQ): Likewise.
6080 (VADDEUQM): Likewise.
6081 (VADDECUQ): Likewise.
6082 (VSUBEUQM): Likewise.
6083 (VSUBECUQ): Likewise.
6085 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
6086 __int128_t and __uint128_t types.
6087 (__uint128_type): Likewise.
6088 (altivec_categorize_keyword): Add support for vector __int128_t,
6089 vector __uint128_t, vector __int128, and vector unsigned __int128
6090 as a container type for TImode operations that need to be done in
6091 VSX/Altivec registers.
6092 (rs6000_macro_to_expand): Likewise.
6093 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
6094 to support 128-bit integer instructions vaddcuq, vadduqm,
6095 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
6096 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
6098 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
6099 for V1TImode, and set up preferences to use VSX/Altivec registers.
6100 Setup VSX reload handlers.
6101 (rs6000_debug_reg_global): Likewise.
6102 (rs6000_init_hard_regno_mode_ok): Likewise.
6103 (rs6000_preferred_simd_mode): Likewise.
6104 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
6105 (easy_altivec_constant): Likewise.
6106 (output_vec_const_move): Likewise.
6107 (rs6000_expand_vector_set): Convert V1TImode set and extract to
6109 (rs6000_expand_vector_extract): Likewise.
6110 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
6112 (rs6000_const_vec): Add support for V1TImode.
6113 (rs6000_emit_le_vsx_load): Swap double words when loading or
6114 storing TImode/V1TImode.
6115 (rs6000_emit_le_vsx_store): Likewise.
6116 (rs6000_emit_le_vsx_move): Likewise.
6117 (rs6000_emit_move): Add support for V1TImode.
6118 (altivec_expand_ld_builtin): Likewise.
6119 (altivec_expand_st_builtin): Likewise.
6120 (altivec_expand_vec_init_builtin): Likewise.
6121 (altivec_expand_builtin): Likewise.
6122 (rs6000_init_builtins): Add support for V1TImode type. Add
6123 support for ISA 2.07 128-bit integer builtins. Define type names
6124 for the VSX/Altivec vector types.
6125 (altivec_init_builtins): Add support for overloaded vector
6126 functions with V1TImode type.
6127 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
6128 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
6130 (rs6000_split_128bit_ok_p): Likewise.
6131 (rs6000_handle_altivec_attribute): Create V1TImode from vector
6132 __int128_t and vector __uint128_t.
6134 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
6135 and mode attributes.
6142 (VS_scalar): Likewise.
6143 (VS_double): Likewise.
6144 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
6146 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
6147 we support the ISA 2.07 128-bit integer arithmetic instructions.
6148 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
6149 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
6150 and TImode types for use with the builtin functions.
6151 (V1TI_type_node): Likewise.
6152 (unsigned_V1TI_type_node): Likewise.
6153 (intTI_type_internal_node): Likewise.
6154 (uintTI_type_internal_node): Likewise.
6156 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
6157 128-bit builtin functions.
6158 (UNSPEC_VADDEUQM): Likewise.
6159 (UNSPEC_VADDECUQ): Likewise.
6160 (UNSPEC_VSUBCUQ): Likewise.
6161 (UNSPEC_VSUBEUQM): Likewise.
6162 (UNSPEC_VSUBECUQ): Likewise.
6163 (VM): Add V1TImode to vector mode iterators.
6165 (VI_unit): Likewise.
6166 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
6167 (altivec_vaddcuq): Likewise.
6168 (altivec_vsubuqm): Likewise.
6169 (altivec_vsubcuq): Likewise.
6170 (altivec_vaddeuqm): Likewise.
6171 (altivec_vaddecuq): Likewise.
6172 (altivec_vsubeuqm): Likewise.
6173 (altivec_vsubecuq): Likewise.
6175 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
6177 (BOOL_128): Likewise.
6178 (BOOL_REGS_OUTPUT): Likewise.
6179 (BOOL_REGS_OP1): Likewise.
6180 (BOOL_REGS_OP2): Likewise.
6181 (BOOL_REGS_UNARY): Likewise.
6182 (BOOL_REGS_AND_CR0): Likewise.
6184 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
6185 128-bit integer builtin support.
6186 (vec_vadduqm): Likewise.
6187 (vec_vaddecuq): Likewise.
6188 (vec_vaddeuqm): Likewise.
6189 (vec_vsubecuq): Likewise.
6190 (vec_vsubeuqm): Likewise.
6191 (vec_vsubcuq): Likewise.
6192 (vec_vsubuqm): Likewise.
6194 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
6195 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
6196 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
6197 128-bit integer add/subtract to ISA 2.07.
6199 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
6201 * config/arc/arc.c (arc_predicate_delay_insns):
6202 Fix third argument passed to conditionalize_nonjump.
6204 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
6206 * config/aarch64/aarch64-builtins.c
6207 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
6208 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
6209 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
6210 instead of __builtin_lfloor.
6211 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
6213 2014-03-12 Jakub Jelinek <jakub@redhat.com>
6215 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
6216 (tree_ssa_ifcombine_bb_1): New function.
6217 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
6218 is an empty forwarder block to then_bb or vice versa and then_bb
6219 and else_bb are effectively swapped.
6221 2014-03-12 Christian Bruel <christian.bruel@st.com>
6224 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
6225 REG_CFA_DEF_CFA note.
6226 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
6227 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
6229 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
6231 PR tree-optimization/60454
6232 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
6234 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6236 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
6237 Do not define target_cpu_default2 to generic.
6238 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
6239 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
6240 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
6242 2014-03-12 Jakub Jelinek <jakub@redhat.com>
6243 Marc Glisse <marc.glisse@inria.fr>
6245 PR tree-optimization/60502
6246 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
6247 instead of build_low_bits_mask.
6249 2014-03-12 Jakub Jelinek <jakub@redhat.com>
6252 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
6253 if there are multiple uses, but op doesn't live on E edge.
6254 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
6255 clobber stmts before __builtin_unreachable.
6257 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
6259 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
6260 hard_frame_pointer_rtx.
6261 * cse.c (cse_insn): Remove volatile check.
6262 * cselib.c (cselib_process_insn): Likewise.
6263 * dse.c (scan_insn): Likewise.
6265 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
6267 * config/arc/arc.c (conditionalize_nonjump): New function,
6269 (arc_ifcvt): ... this.
6270 (arc_predicate_delay_insns): Use it.
6272 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
6274 * config/arc/predicates.md (extend_operand): During/after reload,
6275 allow const_int_operand.
6276 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
6277 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
6278 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
6280 (umulsi3_highpart_i): Likewise.
6282 2014-03-11 Richard Biener <rguenther@suse.de>
6284 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
6285 Add asserts to guard possible wrong-code bugs.
6287 2014-03-11 Richard Biener <rguenther@suse.de>
6289 PR tree-optimization/60429
6290 PR tree-optimization/60485
6291 * tree-ssa-structalias.c (set_union_with_increment): Properly
6292 take into account all fields that overlap the shifted vars.
6293 (do_sd_constraint): Likewise.
6294 (do_ds_constraint): Likewise.
6295 (get_constraint_for_ptr_offset): Likewise.
6297 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
6299 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
6300 (nios2_compute_frame_layout):
6301 Add calculation of cfun->machine->fp_save_offset.
6302 (nios2_expand_prologue): Correct setting of frame pointer register
6304 (nios2_expand_epilogue): Update recovery of stack pointer from
6305 frame pointer accordingly.
6306 (nios2_initial_elimination_offset): Update calculation of offset
6307 for eliminating to HARD_FRAME_POINTER_REGNUM.
6309 2014-03-10 Jakub Jelinek <jakub@redhat.com>
6312 * ipa.c (symtab_remove_unreachable_nodes): Don't call
6313 cgraph_get_create_node on VAR_DECLs.
6315 2014-03-10 Richard Biener <rguenther@suse.de>
6318 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
6320 2014-03-08 Douglas B Rupp <rupp@gnat.com>
6322 * config/vms/vms.opt (vms_float_format): New variable.
6324 2014-03-08 Tobias Burnus <burnus@net-b.de>
6326 * doc/invoke.texi (-fcilkplus): Update implementation status.
6328 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
6329 Richard Biener <rguenther@suse.de>
6331 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
6332 consistently accross all TUs.
6333 (run_gcc): Enable -fshort-double automatically at link at link-time
6334 and disallow override.
6336 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
6339 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
6340 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
6341 if they can't be used.
6343 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6345 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
6346 for Solaris 11/x86 ld.
6347 * configure: Regenerate.
6349 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6351 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
6352 (LIB_TLS_SPEC): Save as ld_tls_libs.
6353 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
6354 (HAVE_AS_IX86_TLSLDM): New test.
6355 * configure, config.in: Regenerate.
6356 * config/i386/i386.c (legitimize_tls_address): Fall back to
6357 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
6358 cannot support TLS_MODEL_LOCAL_DYNAMIC.
6359 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
6360 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
6362 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
6364 * common.opt (fira-loop-pressure): Mark as optimization.
6366 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
6368 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
6369 an OpenMP mappable type.
6371 2014-03-06 Matthias Klose <doko@ubuntu.com>
6373 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
6374 MULTILIB_OSDIRNAMES is not defined.
6376 2014-03-06 Jakub Jelinek <jakub@redhat.com>
6377 Meador Inge <meadori@codesourcery.com>
6380 * config/arm/arm.c (arm_tls_symbol_p): Remove.
6381 (arm_legitimize_address): Call legitimize_tls_address for any
6382 arm_tls_referenced_p expression, handle constant addend. Call it
6383 before testing for !TARGET_ARM.
6384 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
6386 2014-03-06 Richard Biener <rguenther@suse.de>
6392 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
6394 * tree-streamer.c (record_common_node): Assert we don't record
6395 nodes with type double.
6396 (preload_common_node): Skip type double, complex double and double
6397 pointer since it is now frontend dependent due to fshort-double option.
6399 2014-03-06 Richard Biener <rguenther@suse.de>
6401 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
6402 or -fno-lto is specified and the linker has full plugin support.
6403 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
6404 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
6405 * lto-wrapper.c (merge_and_complain): Merge compile-time
6406 optimization levels.
6407 (run_gcc): And pass it through to the link options.
6409 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
6413 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
6415 * cselib.c (remove_useless_values): Skip to avoid quadratic
6416 behavior if the condition moved from...
6417 (cselib_process_insn): ... here holds.
6419 2014-03-05 Jakub Jelinek <jakub@redhat.com>
6422 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
6423 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
6426 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
6427 (TM_H): Add x86-tune.def.
6429 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6431 * config/aarch64/aarch64.c (generic_tunings):
6432 Use cortexa57_extra_costs.
6434 2014-03-05 Jakub Jelinek <jakub@redhat.com>
6437 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
6438 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
6439 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
6442 2014-03-04 Heiher <r@hev.cc>
6444 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
6445 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
6447 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
6449 * config/i386/predicates.md (const2356_operand): Change to ...
6450 (const2367_operand): ... this.
6451 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
6453 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6454 (*avx512pf_scatterpf<mode>sf): Ditto.
6455 (avx512pf_scatterpf<mode>df): Ditto.
6456 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6457 (*avx512pf_scatterpf<mode>df): Ditto.
6458 * config/i386/i386.c (ix86_expand_builtin): Update
6459 incorrect hint operand error message.
6461 2014-03-04 Richard Biener <rguenther@suse.de>
6463 * lto-section-in.c (lto_get_section_data): Fix const cast.
6465 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
6467 * tree-streamer.c (record_common_node): Assert we don't record
6468 nodes with type double.
6469 (preload_common_node): Skip type double, complex double and double
6470 pointer since it is now frontend dependent due to fshort-double option.
6472 2014-03-04 Richard Biener <rguenther@suse.de>
6475 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
6476 (lto_input_toplevel_asms): Likewise.
6477 * lto-section-in.c (lto_get_section_data): Instead do it here
6480 2014-03-04 Richard Biener <rguenther@suse.de>
6482 PR tree-optimization/60382
6483 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
6484 dead PHIs a reduction.
6486 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
6488 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
6490 (_mm_prefetch): Move out of GCC target("sse") pragma.
6491 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
6492 GCC target("prfchw") pragma.
6493 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
6495 * config/i386/i386.c (ix86_option_override_internal): Enable
6496 -mprfchw with -mprefetchwt1.
6498 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
6500 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
6503 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
6505 * opts.h (CL_PCH_IGNORE): Define.
6506 * targhooks.c (option_affects_pch_p):
6507 Return false for options that have CL_PCH_IGNORE set.
6508 * opt-functions.awk: Process PchIgnore.
6509 * doc/options.texi: Document PchIgnore.
6511 * config/arc/arc.opt (misize): Add PchIgnore property.
6513 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6515 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
6516 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
6517 constraint on constants to permit them being loaded into
6518 GENERAL_REGS or BASE_REGS.
6520 2014-03-03 Nick Clifton <nickc@redhat.com>
6522 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
6523 anti-cacnonical alternatives.
6524 (negandhi3_real): New pattern.
6525 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
6527 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
6529 * config/avr/avr-mcus.def: Remove atxmega16x1.
6530 * config/avr/avr-tables.opt: Regenerate.
6531 * config/avr/t-multilib: Regenerate.
6532 * doc/avr-mmcu.texi: Regenerate.
6534 2014-03-03 Tobias Grosser <tobias@grosser.es>
6535 Mircea Namolaru <mircea.namolaru@inria.fr>
6537 PR tree-optimization/58028
6538 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
6541 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6543 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
6544 not handled by recognizers.
6546 2014-03-03 Jakub Jelinek <jakub@redhat.com>
6549 * function.c (expand_function_end): Don't emit
6550 clobber_return_register sequence if clobber_after is a BARRIER.
6551 * cfgexpand.c (construct_exit_block): Append instructions before
6552 return_label to prev_bb.
6554 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6556 * config/rs6000/constraints.md: Document reserved use of "wc".
6558 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
6561 * ipa.c (function_and_variable_visibility): When dissolving comdat
6562 group, also set all symbols to local.
6564 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
6569 2013-12-14 Jan Hubicka <jh@suse.cz>
6571 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
6573 2014-03-02 Jon Beniston <jon@beniston.com>
6579 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
6580 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
6581 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
6582 (simple_return, *simple_return): New patterns
6583 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
6584 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
6586 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
6588 * dwarf2out.c (gen_subprogram_die): Tidy.
6590 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
6593 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
6594 (*mov_t_msb_neg_negc): ... this new insn.
6596 2014-02-28 Jason Merrill <jason@redhat.com>
6599 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
6602 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
6605 * dwarf2out.c (decltype_auto_die): New static.
6606 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
6607 (gen_type_die_with_usage): Handle 'decltype(auto)'.
6608 (is_cxx_auto): Likewise.
6610 2014-02-28 Ian Bolton <ian.bolton@arm.com>
6612 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
6613 we are not using general regs only.
6615 2014-02-28 Richard Biener <rguenther@suse.de>
6618 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
6619 previous fix and only allow to remove trivial pre-headers
6620 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
6621 (remove_forwarder_block): Properly update the latch of a loop.
6623 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
6626 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
6627 (cselib_preserved_hash_table): New.
6628 (preserve_constants_and_equivs): Move preserved vals to it.
6629 (cselib_find_slot): Look it up first.
6630 (cselib_init): Initialize it.
6631 (cselib_finish): Release it.
6632 (dump_cselib_table): Dump it.
6634 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
6637 * cselib.c (remove_useless_values): Skip to avoid quadratic
6638 behavior if the condition moved from...
6639 (cselib_process_insn): ... here holds.
6641 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
6644 * var-tracking.c (vt_initialize): Apply the same condition to
6645 preserve the CFA base value.
6647 2014-02-28 Joey Ye <joey.ye@arm.com>
6650 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
6651 if reload in progress or completed.
6653 2014-02-28 Tobias Burnus <burnus@net-b.de>
6656 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
6659 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
6661 * doc/tm.texi.in (Condition Code Status): Update documention for
6662 relative locations of cc0-setter and cc0-user.
6664 2014-02-27 Jeff Law <law@redhat.com>
6666 PR rtl-optimization/52714
6667 * combine.c (try_combine): When splitting an unrecognized PARALLEL
6668 into two independent simple sets, if I3 is a jump, ensure the
6669 pattern we place into I3 is a (set (pc) ...).
6671 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
6672 Jeff Law <law@redhat.com>
6674 PR rtl-optimization/49847
6675 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
6676 are in different blocks.
6677 * doc/tm.texi (Condition Code Status): Update documention for
6678 relative locations of cc0-setter and cc0-user.
6680 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
6683 * lra.c (lra_emit_add): Check SUBREG too.
6685 2014-02-27 Andreas Schwab <schwab@suse.de>
6687 * config/m68k/m68k.c (m68k_option_override): Disable
6688 -flive-range-shrinkage for classic m68k.
6689 (m68k_override_options_after_change): Likewise.
6691 2014-02-27 Marek Polacek <polacek@redhat.com>
6694 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
6695 -Wmaybe-uninitialized.
6697 2014-02-27 Alan Modra <amodra@gmail.com>
6700 * reload1.c (emit_input_reload_insns): When reload_override_in,
6701 set old to rl->in_reg when rl->in_reg is a subreg.
6703 2014-02-26 Richard Biener <rguenther@suse.de>
6706 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
6708 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
6710 * common/config/i386/predicates.md (const1256_operand): Remove.
6711 (const2356_operand): New.
6712 (const_1_to_2_operand): Remove.
6713 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
6714 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
6715 (*avx512pf_gatherpf<mode>sf): Ditto.
6716 (avx512pf_gatherpf<mode>df): Ditto.
6717 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6718 (*avx512pf_gatherpf<mode>df): Ditto.
6719 (avx512pf_scatterpf<mode>sf): Ditto.
6720 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6721 (*avx512pf_scatterpf<mode>sf): Ditto.
6722 (avx512pf_scatterpf<mode>df): Ditto.
6723 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6724 (*avx512pf_scatterpf<mode>df): Ditto.
6725 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
6727 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
6729 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
6730 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
6731 (_mm512_mask_testn_epi64_mask): Move to ...
6732 * config/i386/avx512cdintrin.h: Here.
6733 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
6734 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
6735 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
6736 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
6737 TARGET_AVX512F from TARGET_AVX512CD.
6739 2014-02-26 Richard Biener <rguenther@suse.de>
6742 * ipa.c (walk_polymorphic_call_targets): Properly guard
6743 call to inline_update_overall_summary.
6745 2014-02-26 Bin Cheng <bin.cheng@arm.com>
6748 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
6749 and latches only if requested. Fix latch if it is removed.
6750 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
6751 LOOPS_HAVE_PREHEADERS.
6753 2014-02-25 Andrew Pinski <apinski@cavium.com>
6755 * builtins.c (expand_builtin_thread_pointer): Create a new target
6756 when the target is NULL.
6758 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
6760 PR rtl-optimization/60317
6761 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
6762 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
6763 * lra-assigns.c: Include params.h.
6764 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
6765 other reload pseudos considerations.
6767 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6769 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
6770 to use canonical form for nor<mode>3.
6772 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6775 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
6778 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
6780 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
6781 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
6782 (ix86_handle_option): Handle OPT_mprefetchwt1.
6783 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
6784 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
6786 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
6787 OPTION_MASK_ISA_PREFETCHWT1.
6788 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
6789 (PTA_PREFETCHWT1): New.
6790 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
6791 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
6792 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
6793 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
6794 (*prefetch_avx512pf_<mode>_: Change into ...
6795 (*prefetch_prefetchwt1_<mode>: This.
6796 * config/i386/i386.opt (mprefetchwt1): New.
6797 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
6798 (_mm_prefetch): Handle intent to write.
6799 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
6801 2014-02-25 Richard Biener <rguenther@suse.de>
6804 * emit-rtl.c (mem_attrs_htab): Remove.
6805 (mem_attrs_htab_hash): Likewise.
6806 (mem_attrs_htab_eq): Likewise.
6807 (set_mem_attrs): Always allocate new mem-attrs when something changed.
6808 (init_emit_once): Do not allocate mem_attrs_htab.
6810 2014-02-25 Richard Biener <rguenther@suse.de>
6813 * lto-opts.c (lto_write_options): Output non-explicit conservative
6814 -fwrapv, -fno-trapv and -fno-strict-overflow.
6815 * lto-wrapper.c (merge_and_complain): Handle merging those options.
6816 (run_gcc): And pass them through.
6818 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
6820 * sel-sched.c (calculate_new_fences): New parameter ptime.
6821 Calculate it as a maximum over all fence cycles.
6822 (sel_sched_region_2): Adjust the call to calculate_new_fences.
6823 Print the final schedule timing when sched_verbose.
6825 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
6827 PR rtl-optimization/60292
6828 * sel-sched.c (fill_vec_av_set): Do not reset target availability
6829 bit fot the fence instruction.
6831 2014-02-24 Alangi Derick <alangiderick@gmail.com>
6833 * calls.h: Fix typo in comment.
6835 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
6837 * config/pa/pa.c (pa_output_move_double): Don't valididate when
6838 adjusting offsetable addresses.
6840 2014-02-24 Guozhi Wei <carrot@google.com>
6842 * sparseset.h (sparseset_pop): Fix the wrong index.
6844 2014-02-24 Walter Lee <walt@tilera.com>
6846 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
6847 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
6849 * common/config/tilegx/tilegx-common.c
6850 (TARGET_DEFAULT_TARGET_FLAGS): Define.
6851 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
6853 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
6854 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
6855 (tilegx_gimplify_va_arg_expr): Handle big endian.
6856 (tilegx_expand_unaligned_load): Ditto.
6857 (tilegx_expand_unaligned_store): Ditto.
6858 (TARGET_RETURN_IN_MSB): New.
6859 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
6860 (TARGET_ENDIAN_DEFAULT): New.
6861 (TARGET_BIG_ENDIAN): Handle big endian.
6862 (BYTES_BIG_ENDIAN): Ditto.
6863 (WORDS_BIG_ENDIAN): Ditto.
6864 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
6867 * config/tilegx/tilegx.md (extv): Handle big endian.
6869 (insn_st<n>): Ditto.
6870 (insn_st<n>_add<bitsuffix>): Ditto.
6871 (insn_stnt<n>): Ditto.
6872 (insn_stnt<n>_add<bitsuffix>):Ditto.
6873 (vec_interleave_highv8qi): Handle big endian.
6874 (vec_interleave_highv8qi_be): New.
6875 (vec_interleave_highv8qi_le): New.
6876 (insn_v1int_h): Handle big endian.
6877 (vec_interleave_lowv8qi): Handle big endian.
6878 (vec_interleave_lowv8qi_be): New.
6879 (vec_interleave_lowv8qi_le): New.
6880 (insn_v1int_l): Handle big endian.
6881 (vec_interleave_highv4hi): Handle big endian.
6882 (vec_interleave_highv4hi_be): New.
6883 (vec_interleave_highv4hi_le): New.
6884 (insn_v2int_h): Handle big endian.
6885 (vec_interleave_lowv4hi): Handle big endian.
6886 (vec_interleave_lowv4hi_be): New.
6887 (vec_interleave_lowv4hi_le): New.
6888 (insn_v2int_l): Handle big endian.
6889 (vec_interleave_highv2si): Handle big endian.
6890 (vec_interleave_highv2si_be): New.
6891 (vec_interleave_highv2si_le): New.
6892 (insn_v4int_h): Handle big endian.
6893 (vec_interleave_lowv2si): Handle big endian.
6894 (vec_interleave_lowv2si_be): New.
6895 (vec_interleave_lowv2si_le): New.
6896 (insn_v4int_l): Handle big endian.
6897 * config/tilegx/tilegx.opt (mbig-endian): New option.
6898 (mlittle-endian): New option.
6899 * doc/install.texi: Document tilegxbe-linux.
6900 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
6902 2014-02-24 Martin Jambor <mjambor@suse.cz>
6905 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
6906 there are no parameter descriptors.
6908 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
6910 PR rtl-optimization/60268
6911 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
6912 initialization to ...
6913 (sched_rgn_init): ... here.
6914 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
6916 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
6918 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
6921 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
6923 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
6926 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
6928 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
6929 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
6931 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
6933 * config/microblaze/predicates.md: Add cmp_op predicate.
6934 * config/microblaze/microblaze.md: Add branch_compare instruction
6935 which uses cmp_op predicate and emits cmp insn before branch.
6936 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
6937 to microblaze_expand_conditional_branch and consolidate logic.
6938 (microblaze_expand_conditional_branch): emit branch_compare
6939 insn instead of handling cmp op separate from branch insn.
6941 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6943 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
6946 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6948 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
6949 define_insn with define_expand and new define_insn
6950 *altivec_lve<VI_char>x_internal.
6951 (altivec_stve<VI_char>x): Replace define_insn with define_expand
6952 and new define_insn *altivec_stve<VI_char>x_internal.
6953 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
6955 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
6957 (altivec_expand_stvex_be): New function.
6959 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
6961 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
6962 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
6963 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
6964 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
6966 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
6969 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
6970 instead of emit_move_insn.
6972 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6974 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
6976 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
6977 gen_altivec_vsumsws.
6979 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6981 * config/rs6000/altivec.md (altivec_lvxl): Rename as
6982 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
6983 (altivec_lvxl_<mode>): New define_expand incorporating
6984 -maltivec=be semantics where needed.
6985 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
6986 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
6987 semantics where needed.
6988 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
6989 (altivec_stvx_<mode>): New define_expand incorporating
6990 -maltivec=be semantics where needed.
6991 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
6992 VM2 iterator instead of V4SI.
6993 (altivec_stvxl_<mode>): New define_expand incorporating
6994 -maltivec=be semantics where needed.
6995 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
6996 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
6997 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
6998 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
6999 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
7000 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
7001 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
7002 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
7003 ALTIVEC_BUILTIN_STVXL.
7004 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
7005 (altivec_expand_stvx_be): Likewise.
7006 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
7007 (altivec_expand_lvx_be): Likewise.
7008 (altivec_expand_stvx_be): Likewise.
7009 (altivec_expand_builtin): Add cases for
7010 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
7011 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
7012 (altivec_init_builtins): Add definitions for
7013 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
7014 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
7016 2014-02-21 Catherine Moore <clm@codesourcery.com>
7018 * doc/invoke.texi (mvirt, mno-virt): Document.
7019 * config/mips/mips.opt (mvirt): New option.
7020 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
7022 2014-02-21 Richard Biener <rguenther@suse.de>
7024 PR tree-optimization/60276
7025 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
7026 (STMT_VINFO_MIN_NEG_DIST): New macro.
7027 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
7028 STMT_VINFO_MIN_NEG_DIST.
7029 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
7030 made for negative dependence distances still hold.
7032 2014-02-21 Richard Biener <rguenther@suse.de>
7035 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
7036 DECL_INITIAL for globals not in the current function context.
7038 2014-02-21 Jakub Jelinek <jakub@redhat.com>
7040 PR tree-optimization/56490
7041 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
7042 * tree-ssa-uninit.c: Include params.h.
7043 (compute_control_dep_chain): Add num_calls argument, return false
7044 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
7045 num_calls to recursive call.
7046 (find_predicates): Change dep_chain into normal array,
7047 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
7048 variable and adjust compute_control_dep_chain caller.
7049 (find_def_preds): Likewise.
7051 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
7053 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
7054 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
7056 2014-02-21 Nick Clifton <nickc@redhat.com>
7058 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
7059 (pushhi1): Likewise.
7060 (popqi1): Add mode to pre_dec.
7063 2014-02-21 Jakub Jelinek <jakub@redhat.com>
7065 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
7066 mode for mask of V8SFmode permutation.
7068 2014-02-20 Richard Henderson <rth@redhat.com>
7071 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
7072 a new pseudo for OLDVAL.
7074 2014-02-20 Jakub Jelinek <jakub@redhat.com>
7077 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
7078 gen_reg_rtx if d->testing_p.
7079 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
7080 if d->testing_p and we will certainly return true.
7081 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
7084 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
7086 * emit-rtl.c (gen_reg_rtx): Assert that
7087 crtl->emit.regno_pointer_align_length is non-zero.
7089 2014-02-20 Richard Henderson <rth@redhat.com>
7092 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
7093 on failure the store back into EXPECT.
7095 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
7096 Sandra Loosemore <sandra@codesourcery.com>
7098 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
7099 * config/nios2/nios2.c (nios2_function_profiler): Add
7100 -fPIC (flag_pic == 2) support.
7101 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
7102 (nios2_large_offset_p): New function.
7103 (nios2_unspec_reloc_p): Move up position, update to use
7104 nios2_large_offset_p.
7105 (nios2_unspec_address): Remove function.
7106 (nios2_unspec_offset): New function.
7107 (nios2_large_got_address): New function.
7108 (nios2_got_address): Add large offset support.
7109 (nios2_legitimize_tls_address): Update usage of removed and new
7111 (nios2_symbol_binds_local_p): New function.
7112 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
7113 (nios2_legitimize_address): Update to use nios2_large_offset_p.
7114 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
7115 (nios2_print_operand): Merge H/L processing, add hiadj/lo
7116 processing for (const (unspec ...)).
7117 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
7119 2014-02-20 Richard Biener <rguenther@suse.de>
7121 * tree-cfg.c (replace_uses_by): Mark altered BBs before
7122 doing the substitution.
7123 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
7125 2014-02-20 Martin Jambor <mjambor@suse.cz>
7128 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
7129 info when checking whether lattices are bottom.
7131 2014-02-20 Richard Biener <rguenther@suse.de>
7134 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
7137 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
7140 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
7141 parameter specifying the scaling.
7142 (inline_call): Update.
7143 (want_inline_recursively): Guard division by zero.
7144 (recursive_inlining): Update.
7145 * ipa-inline.h (clone_inlined_nodes): Update.
7147 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
7150 * config/i386/i386.c (classify_argument): Pass structures of size
7151 64 bytes or less in register.
7153 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
7154 Kirill Yukhin <kirill.yukhin@intel.com>
7156 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
7157 (_mm_rcp28_round_ss): Ditto.
7158 (_mm_rsqrt28_round_sd): Ditto.
7159 (_mm_rsqrt28_round_ss): Ditto.
7160 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
7161 (_mm_rcp14_round_ss): Ditto.
7162 (_mm_rsqrt14_round_sd): Ditto.
7163 (_mm_rsqrt14_round_ss): Ditto.
7164 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
7165 the first input operand, get rid of match_dup.
7166 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
7168 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
7170 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
7171 operand as the first input operand, set type attribute.
7172 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
7174 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
7175 operand as the first input operand, set type attribute.
7177 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7179 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
7182 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
7185 * config/i386/i386.c (construct_container): Remove TFmode check
7186 for X86_64_INTEGER_CLASS.
7188 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
7191 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
7192 only when -Wpsabi is enabled.
7194 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
7197 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
7198 passing arrays in registers are the same as for structs, so remove the
7199 special case for them.
7201 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
7203 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
7204 destination type, extract only the valid bits if the source type is not
7205 integral and has a different mode.
7207 2014-02-19 Richard Biener <rguenther@suse.de>
7210 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
7213 2014-02-19 Richard Biener <rguenther@suse.de>
7216 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
7217 (ipa_modify_call_arguments): Emit an argument load explicitely and
7218 preserve virtual SSA form there and for the replacement call.
7219 Do not update SSA form nor free dominance info.
7221 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
7223 * ipa.c (function_and_variable_visibility): Also clear WEAK
7224 flag when disolving COMDAT_GROUP.
7226 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
7228 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
7229 * ipa-prop.c (ipa_set_jf_known_type): Return early when
7231 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
7232 do more sanity checks.
7233 (detect_type_change): Return true when giving up early.
7234 (compute_complex_assign_jump_func): Fix type parameter of
7235 ipa_set_ancestor_jf.
7236 (compute_complex_ancestor_jump_func): Likewise.
7237 (update_jump_functions_after_inlining): Fix updating of
7239 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
7241 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
7243 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
7244 inline clones when edge disappears.
7246 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
7249 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
7250 Split 64-bit moves into 2 patterns. Do not allow the use of
7251 direct move for TDmode in little endian, since the decimal value
7252 has little endian bytes within a word, but the 64-bit pieces are
7253 ordered in a big endian fashion, and normal subreg's of TDmode are
7255 (mov<mode>_64bit_dm): Likewise.
7256 (movtd_64bit_nodm): Likewise.
7258 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
7260 PR tree-optimization/60174
7261 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
7262 statement of an SSA_NAME that occurs in an abnormal PHI node.
7264 2014-02-18 Jakub Jelinek <jakub@redhat.com>
7267 * final.c (SEEN_BB): Remove.
7268 (SEEN_NOTE, SEEN_EMITTED): Renumber.
7269 (final_scan_insn): Don't force_source_line on second
7270 NOTE_INSN_BASIC_BLOCK.
7272 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
7275 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
7276 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
7277 (type_natural_mode): Warn ABI change when %zmm register is not
7278 available for AVX512F vector value passing.
7280 2014-02-18 Kai Tietz <ktietz@redhat.com>
7283 * config/i386/i386.c (ix86_expand_prologue): Use value in
7284 rax register as displacement when restoring %r10 or %rax.
7285 Fix wrong offset when restoring both registers.
7287 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
7289 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
7290 assertion with conditional return.
7292 2014-02-18 Jakub Jelinek <jakub@redhat.com>
7293 Uros Bizjak <ubizjak@gmail.com>
7296 * config/i386/driver-i386.c (host_detect_local_cpu): If
7297 YMM state is not saved by the OS, also clear has_f16c. Move
7298 CPUID 0x80000001 handling before YMM state saving checking.
7300 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
7302 PR rtl-optimization/58960
7303 * haifa-sched.c (alloc_global_sched_pressure_data): New,
7304 factored out from ...
7305 (sched_init): ... here.
7306 (free_global_sched_pressure_data): New, factored out from ...
7307 (sched_finish): ... here.
7308 * sched-int.h (free_global_sched_pressure_data): Declare.
7309 * sched-rgn.c (nr_regions_initial): New static global.
7310 (haifa_find_rgns): Initialize it.
7311 (schedule_region): Disable sched-pressure for the newly
7314 2014-02-17 Richard Biener <rguenther@suse.de>
7316 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
7317 release SSA defs of pattern stmts.
7319 2014-02-17 Richard Biener <rguenther@suse.de>
7321 * tree-inline.c (expand_call_inline): Release the virtual
7322 operand defined by the call we are about to inline.
7324 2014-02-17 Richard Biener <rguenther@suse.de>
7326 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
7328 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
7329 Ilya Tocar <ilya.tocar@intel.com>
7331 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
7332 arguments order in builtin.
7333 (_mm512_permutexvar_epi64): Ditto.
7334 (_mm512_mask_permutexvar_epi64): Ditto
7335 (_mm512_maskz_permutexvar_epi32): Ditto
7336 (_mm512_permutexvar_epi32): Ditto
7337 (_mm512_mask_permutexvar_epi32): Ditto
7339 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7341 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
7342 (p8_vmrgow): Likewise.
7344 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7346 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
7349 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
7352 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
7353 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
7354 into 64-bit and 32-bit moves. On 64-bit moves, add support for
7355 using direct move instructions on ISA 2.07. Also adjust
7356 instruction length for 64-bit.
7357 (mov<mode>_64bit, TFmode/TDmode): Likewise.
7358 (mov<mode>_32bit, TFmode/TDmode): Likewise.
7360 2014-02-15 Alan Modra <amodra@gmail.com>
7364 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
7365 find_replacement on parts of insn rtl that might be reloaded.
7367 2014-02-15 Richard Biener <rguenther@suse.de>
7369 PR tree-optimization/60183
7370 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
7371 (tree_ssa_phiprop): Calculate and free post-dominators.
7373 2014-02-14 Jeff Law <law@redhat.com>
7375 PR rtl-optimization/60131
7376 * ree.c (get_extended_src_reg): New function.
7377 (combine_reaching_defs): Use it rather than assuming location of REG.
7378 (find_and_remove_re): Verify first operand of extension is
7379 a REG before adding the insns to the copy list.
7381 2014-02-14 Roland McGrath <mcgrathr@google.com>
7383 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
7384 * configure: Regenerated.
7385 * config.in: Regenerated.
7386 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
7387 instead of ASM_SHORT.
7389 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
7390 Richard Earnshaw <rearnsha@arm.com>
7392 PR rtl-optimization/59535
7393 * lra-constraints.c (process_alt_operands): Encourage alternative
7394 when unassigned pseudo class is superset of the alternative class.
7395 (inherit_reload_reg): Don't inherit when optimizing for code size.
7396 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
7397 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
7398 modes not less than 4 for Thumb1.
7400 2014-02-14 Kyle McMartin <kyle@redhat.com>
7403 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
7405 2014-02-14 Richard Biener <rguenther@suse.de>
7407 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
7408 (get_frame_arg): Drop the assert with langhook types_compatible_p.
7409 Do not strip INDIRECT_REFs.
7411 2014-02-14 Richard Biener <rguenther@suse.de>
7414 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
7415 DECL_FUNCTION_SPECIFIC_TARGET.
7416 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
7417 * tree-streamer-out.c (pack_ts_target_option): Remove.
7418 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
7419 (write_ts_function_decl_tree_pointers): Do not stream
7420 DECL_FUNCTION_SPECIFIC_TARGET.
7421 * tree-streamer-in.c (unpack_ts_target_option): Remove.
7422 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
7423 (lto_input_ts_function_decl_tree_pointers): Do not stream
7424 DECL_FUNCTION_SPECIFIC_TARGET.
7426 2014-02-14 Jakub Jelinek <jakub@redhat.com>
7428 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
7429 (get_initial_def_for_induction, vectorizable_induction): Ignore
7430 debug stmts when looking for exit_phi.
7431 (vectorizable_live_operation): Fix up condition.
7433 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
7435 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
7436 nreverse() because it changes the content of original tree list.
7438 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
7440 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
7441 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
7443 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
7445 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
7446 GNU coding standards.
7448 2014-02-13 Jakub Jelinek <jakub@redhat.com>
7451 * dwarf2out.c (gen_subprogram_die): Don't call
7452 add_calling_convention_attribute if subr_die is old_die.
7454 2014-02-13 Sharad Singhai <singhai@google.com>
7456 * doc/optinfo.texi: Fix order of nodes.
7458 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
7460 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
7461 operands[2], not operands[3].
7463 2014-02-13 Richard Biener <rguenther@suse.de>
7466 * doc/install.texi (ISL): Update recommended version to 0.12.2,
7467 mention the possibility of an in-tree build.
7468 (CLooG): Update recommended version to 0.18.1, mention the
7469 possibility of an in-tree build and clarify that the ISL
7470 bundled with CLooG does not work.
7472 2014-02-13 Jakub Jelinek <jakub@redhat.com>
7475 * expr.c (compress_float_constant): If x is a hard register,
7476 extend into a pseudo and then move to x.
7478 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
7480 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
7481 caused by bad second argument to warning_at() with -mhotpatch and
7482 nested functions (e.g. with gfortran).
7484 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
7486 * opts.c (option_name): Remove "enabled by default" rider.
7488 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
7490 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
7492 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
7493 Uros Bizjak <ubizjak@gmail.com>
7496 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
7497 * configure: Regenerated.
7499 2014-02-12 Richard Biener <rguenther@suse.de>
7501 * vec.c (vec_prefix::calculate_allocation): Move as
7502 inline variant to vec.h.
7503 (vec_prefix::calculate_allocation_1): New out-of-line version.
7504 * vec.h (vec_prefix::calculate_allocation_1): Declare.
7505 (vec_prefix::m_has_auto_buf): Rename to ...
7506 (vec_prefix::m_using_auto_storage): ... this.
7507 (vec_prefix::calculate_allocation): Inline the easy cases
7508 and dispatch to calculate_allocation_1 which doesn't need the
7510 (va_heap::reserve): Use gcc_checking_assert.
7511 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
7512 m_using_auto_storage.
7513 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
7515 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
7516 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
7517 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
7519 2014-02-12 Richard Biener <rguenther@suse.de>
7521 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
7522 when we found a dependence.
7524 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
7526 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
7528 (maybe_fold_stmt): ... into this new function.
7529 * omp-low.c (lower_omp): Update comment.
7531 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
7534 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
7537 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
7539 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
7540 identifiers in comments.
7541 (cortexa53_extra_costs): Likewise.
7542 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
7543 (cortexa7_extra_costs): Likewise.
7544 (cortexa12_extra_costs): Likewise.
7545 (cortexa15_extra_costs): Likewise.
7546 (v7m_extra_costs): Likewise.
7548 2014-02-12 Richard Biener <rguenther@suse.de>
7551 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
7552 of posix_memalign being successful.
7553 (lower_stmt): Restrict lowering of posix_memalign to when
7554 -ftree-bit-ccp is enabled.
7556 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
7558 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
7560 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
7562 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
7564 PR rtl-optimization/60116
7565 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
7566 other_insn once the combination has been validated.
7568 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
7571 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
7573 * ipa-devirt.c: Include demangle.h
7574 (odr_violation_reported): New static variable.
7575 (add_type_duplicate): Update odr_violations.
7576 (maybe_record_node): Add completep parameter; update it.
7577 (record_target_from_binfo): Add COMPLETEP parameter;
7578 update it as needed.
7579 (possible_polymorphic_call_targets_1): Likewise.
7580 (struct polymorphic_call_target_d): Add nonconstruction_targets;
7581 rename FINAL to COMPLETE.
7582 (record_targets_from_bases): Sanity check we found the binfo;
7583 fix COMPLETEP updating.
7584 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
7585 parameter, fix computing of COMPLETEP.
7586 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
7587 at LTO time do demangling.
7588 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
7589 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
7591 (gimple_get_virt_method_for_binfo): Likewise.
7592 * gimple-fold.h (gimple_get_virt_method_for_binfo,
7593 gimple_get_virt_method_for_vtable): Update prototypes.
7595 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
7598 * genautomata.c (add_presence_absence): Fix typo with
7599 {final_}presence_list.
7601 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
7604 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
7605 for VSX/Altivec vectors that land in GPR registers.
7607 2014-02-11 Richard Henderson <rth@redhat.com>
7608 Jakub Jelinek <jakub@redhat.com>
7611 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
7612 around drhs if type conversion to lacc->type is not useless.
7614 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7616 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
7618 (cortex-a57.cortex-a53): Likewise.
7619 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
7621 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7623 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
7626 2014-02-11 Renlin Li <Renlin.Li@arm.com>
7628 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
7629 add_options_for_arm_vfp3.
7631 2014-02-11 Jeff Law <law@redhat.com>
7634 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
7635 object with an undesirable mode.
7637 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7640 * config/i386/sol2-9.h: New file.
7641 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
7642 *-*-solaris2.9*): Use it.
7644 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
7646 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
7647 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
7649 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
7651 * config/microblaze/microblaze.c: Extend mcpu version format
7653 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
7655 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
7657 2014-02-10 Richard Henderson <rth@redhat.com>
7660 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
7661 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
7662 ms-abi vs -mno-accumulate-outgoing-args.
7663 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
7664 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
7667 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
7670 * cfgexpand.c (expand_asm_operands): Attach source location to
7671 ASM_INPUT rtx objects.
7672 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
7674 2014-02-10 Nick Clifton <nickc@redhat.com>
7676 * config/mn10300/mn10300.c (popcount): New function.
7677 (mn10300_expand_prologue): Include saved registers in stack usage
7680 2014-02-10 Jeff Law <law@redhat.com>
7683 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
7684 when changing the SET_DEST of a prior insn to avoid an input reload.
7686 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
7688 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
7689 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
7690 -mcall-openbsd, or -mcall-linux.
7691 (CC1_ENDIAN_BIG_SPEC): Remove.
7692 (CC1_ENDIAN_LITTLE_SPEC): Remove.
7693 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
7694 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
7695 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
7696 and %cc1_endian_default.
7697 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
7699 2014-02-10 Richard Biener <rguenther@suse.de>
7701 PR tree-optimization/60115
7702 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
7703 MEM_REF handling. Properly verify that the accesses are not
7704 out of the objects bound.
7706 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7708 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
7711 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
7713 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
7714 proper constants and fix formatting.
7715 (possible_polymorphic_call_targets): Fix formatting.
7717 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
7718 Ilya Tocar <ilya.tocar@intel.com>
7720 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
7721 (_mm512_loadu_epi32): Renamed into...
7722 (_mm512_loadu_si512): This.
7723 (_mm512_storeu_epi32): Renamed into...
7724 (_mm512_storeu_si512): This.
7725 (_mm512_maskz_ceil_ps): Removed.
7726 (_mm512_maskz_ceil_pd): Ditto.
7727 (_mm512_maskz_floor_ps): Ditto.
7728 (_mm512_maskz_floor_pd): Ditto.
7729 (_mm512_floor_round_ps): Ditto.
7730 (_mm512_floor_round_pd): Ditto.
7731 (_mm512_ceil_round_ps): Ditto.
7732 (_mm512_ceil_round_pd): Ditto.
7733 (_mm512_mask_floor_round_ps): Ditto.
7734 (_mm512_mask_floor_round_pd): Ditto.
7735 (_mm512_mask_ceil_round_ps): Ditto.
7736 (_mm512_mask_ceil_round_pd): Ditto.
7737 (_mm512_maskz_floor_round_ps): Ditto.
7738 (_mm512_maskz_floor_round_pd): Ditto.
7739 (_mm512_maskz_ceil_round_ps): Ditto.
7740 (_mm512_maskz_ceil_round_pd): Ditto.
7741 (_mm512_expand_pd): Ditto.
7742 (_mm512_expand_ps): Ditto.
7743 * config/i386/i386.c (ix86_builtins): Remove
7744 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
7745 (bdesc_args): Ditto.
7746 * config/i386/predicates.md (const1256_operand): New.
7747 (const_1_to_2_operand): Ditto.
7748 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
7749 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
7750 (*avx512pf_gatherpf<mode>sf): Ditto.
7751 (avx512pf_gatherpf<mode>df): Ditto.
7752 (*avx512pf_gatherpf<mode>df_mask): Ditto.
7753 (*avx512pf_gatherpf<mode>df): Ditto.
7754 (avx512pf_scatterpf<mode>sf): Ditto.
7755 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
7756 (*avx512pf_scatterpf<mode>sf): Ditto.
7757 (avx512pf_scatterpf<mode>df): Ditto.
7758 (*avx512pf_scatterpf<mode>df_mask): Ditto.
7759 (*avx512pf_scatterpf<mode>df): Ditto.
7760 (avx512f_expand<mode>): Removed.
7761 (<shift_insn><mode>3<mask_name>): Change predicate type.
7763 2014-02-08 Jakub Jelinek <jakub@redhat.com>
7765 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
7766 not at the end of datarefs vector use ordered_remove to avoid
7767 reordering datarefs vector.
7770 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
7771 mark local addressable non-static vars as GOVD_PRIVATE
7772 instead of GOVD_LOCAL.
7773 * omp-low.c (lower_omp_for): Move gimple_bind_vars
7774 and BLOCK_VARS of gimple_bind_block to new_stmt rather
7778 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
7779 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
7780 assume_aligned or alloc_align attributes.
7781 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
7782 arguments. Handle also assume_aligned and alloc_align attributes.
7783 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
7784 calls to functions with assume_aligned or alloc_align attributes.
7785 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
7787 2014-02-08 Terry Guo <terry.guo@arm.com>
7789 * doc/invoke.texi: Document ARM -march=armv7e-m.
7791 2014-02-08 Jakub Jelinek <jakub@redhat.com>
7793 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
7794 flag on __cilkrts_rethrow builtin.
7797 * ipa-cp.c (determine_versionability): Fail at -O0
7798 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
7799 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
7802 2014-02-04 Jakub Jelinek <jakub@redhat.com>
7805 * tree-inline.c (copy_forbidden): Fail for
7806 __attribute__((optimize (0))) functions.
7808 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
7810 * varpool.c: Include pointer-set.h.
7811 (varpool_remove_unreferenced_decls): Variables in other partitions
7812 will not be output; be however careful to not lose information
7815 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
7817 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
7818 lookup in the vtable constructor.
7820 2014-02-07 Jeff Law <law@redhat.com>
7823 * config/m68k/m68k.md (ashldi_extsi): Turn into a
7824 define_insn_and_split.
7826 * ipa-inline.c (inline_small_functions): Fix typos.
7828 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7830 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
7831 (s390_can_use_return_insn): Declare.
7832 * config/s390/s390.h (EPILOGUE_USES): Define.
7833 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
7835 (s390_chunkify_start): Handle return JUMP_LABELs.
7836 (s390_early_mach): Emit a main_pool instruction on the entry edge.
7837 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
7838 (s390_can_use_return_insn): New functions.
7839 (s390_fix_long_loop_prediction): Handle conditional returns.
7840 (TARGET_SET_UP_BY_PROLOGUE): Define.
7841 * config/s390/s390.md (ANY_RETURN): New code iterator.
7842 (*creturn, *csimple_return, return, simple_return): New patterns.
7844 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7846 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
7847 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
7848 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
7849 REG_CFA_RESTORE list when deciding not to restore a register.
7851 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7853 * config/s390/s390.c: Include tree-pass.h and context.h.
7854 (s390_early_mach): New function, split out from...
7855 (s390_emit_prologue): ...here.
7856 (pass_data_s390_early_mach): New pass structure.
7857 (pass_s390_early_mach): New class.
7858 (s390_option_override): Create and register early_mach pass.
7859 Move to end of file.
7861 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7863 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
7864 to match for the exit block.
7866 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7868 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
7869 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
7870 Reject misaligned operands.
7872 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7874 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
7876 2014-02-07 Richard Biener <rguenther@suse.de>
7879 * gimple-low.c (lower_builtin_posix_memalign): New function.
7880 (lower_stmt): Call it to lower posix_memalign in a way
7881 to make alignment info accessible.
7883 2014-02-07 Jakub Jelinek <jakub@redhat.com>
7886 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
7887 __builtin_setjmp_receiver.
7889 2014-02-07 Richard Biener <rguenther@suse.de>
7892 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
7893 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
7894 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
7895 Handle BUILT_IN_POSIX_MEMALIGN.
7896 (find_func_clobbers): Likewise.
7897 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
7898 (call_may_clobber_ref_p_1): Likewise.
7900 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7903 * ipa-devirt.c (record_target_from_binfo): Remove overactive
7906 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7909 * lto-cgraph.c (lto_output_node): Use
7910 symtab_get_symbol_partitioning_class.
7911 (lto_output_varpool_node): likewise.
7912 (symtab_get_symbol_partitioning_class): Move here from
7914 * cgraph.h (symbol_partitioning_class): Likewise.
7915 (symtab_get_symbol_partitioning_class): Declare.
7917 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7919 * ggc.h (ggc_internal_cleared_alloc): New macro.
7920 * vec.h (vec_safe_copy): Handle memory stats.
7921 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
7922 * target-globals.c (save_target_globals): Likewise.
7924 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7927 * expr.c (emit_move_resolve_push): Export; be bit more selective
7928 on when to clear alias set.
7929 * expr.h (emit_move_resolve_push): Declare.
7930 * function.h (struct function): Add tail_call_marked.
7931 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
7932 * config/i386/i386-protos.h (ix86_expand_push): Remove.
7933 * config/i386/i386.md (TImode move expander): De not call
7935 (FP push expanders): Preserve memory attributes.
7936 * config/i386/sse.md (push<mode>1): Remove.
7937 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
7938 (ix86_expand_push): Remove.
7939 * config/i386/mmx.md (push<mode>1): Remove.
7941 2014-02-06 Jakub Jelinek <jakub@redhat.com>
7943 PR rtl-optimization/60030
7944 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
7945 lopart with paradoxical subreg before shifting it up by hprec.
7947 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7949 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
7950 Remove extra newline at end of file.
7951 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
7952 (arm_issue_rate): Handle cortexa57.
7953 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
7954 (cortex-a57.cortex-a53): Likewise.
7956 2014-02-06 Jakub Jelinek <jakub@redhat.com>
7959 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
7960 don't record in REG_FRAME_RELATED_EXPR registers not set in that
7962 (arm_expand_prologue): Adjust all callers.
7963 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
7964 info, registers also at the lowest numbered registers side. Use
7965 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
7969 * var-tracking.c (adjust_mems): Before adding a SET to
7970 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
7972 2014-02-06 Alan Modra <amodra@gmail.com>
7975 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
7976 change SDmode to DDmode when lra_in_progress.
7978 2014-02-06 Jakub Jelinek <jakub@redhat.com>
7981 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
7982 free_data_ref on the dr first, and before goto again also set dr
7983 to the next dr. For simd_lane_access, free old datarefs[i] before
7984 overwriting it. For get_vectype_for_scalar_type failure, don't
7985 free_data_ref if simd_lane_access.
7987 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
7990 * tree.h (opts_for_fn): New inline function.
7991 (opt_for_fn): Define.
7992 * config/i386/i386.c (ix86_function_regparm): Use
7993 opt_for_fn (decl, optimize) instead of optimize.
7995 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
7997 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
7998 for SYMBOL_REF in large memory model.
8000 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8002 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
8004 (cortex-a57): Likewise.
8005 (cortex-a57.cortex-a53): Likewise.
8007 2014-02-06 Yury Gribov <y.gribov@samsung.com>
8008 Kugan Vivekanandarajah <kuganv@linaro.org>
8010 * config/arm/arm.c (arm_vector_alignment_reachable): Check
8012 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
8014 2014-02-06 Richard Biener <rguenther@suse.de>
8016 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
8017 set_loop_copy and initialize_original_copy_tables.
8019 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
8021 * config/aarch64/aarch64-simd.md
8022 (aarch64_ashr_simddi): Change QI to SI.
8024 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
8025 Jakub Jelinek <jakub@redhat.com>
8028 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
8031 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8033 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
8034 CODE_FOR_altivec_vpku[hw]um to
8035 CODE_FOR_altivec_vpku[hw]um_direct.
8036 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
8037 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
8038 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
8039 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
8041 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8043 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
8044 generation for -maltivec=be.
8045 (altivec_vsumsws): Simplify redundant test.
8047 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8049 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
8050 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
8051 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
8052 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
8053 gen_altivec_vpkuwum.
8054 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
8056 (altivec_vpks<VI_char>ss): Likewise.
8057 (altivec_vpks<VI_char>us): Likewise.
8058 (altivec_vpku<VI_char>us): Likewise.
8059 (altivec_vpku<VI_char>um): Likewise.
8060 (altivec_vpku<VI_char>um_direct): New (copy of
8061 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
8063 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
8064 target is little endian and -maltivec=be is not specified.
8065 (*altivec_vupkhs<VU_char>_direct): New (copy of
8066 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
8067 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
8068 target is little endian and -maltivec=be is not specified.
8069 (*altivec_vupkls<VU_char>_direct): New (copy of
8070 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
8071 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
8072 little endian and -maltivec=be is not specified.
8073 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
8074 little endian and -maltivec=be is not specified.
8076 2014-02-05 Richard Henderson <rth@redhat.com>
8079 * combine-stack-adj.c: Revert r206943.
8080 * sched-int.h (struct deps_desc): Add last_args_size.
8081 * sched-deps.c (init_deps): Initialize it.
8082 (sched_analyze_insn): Add OUTPUT dependencies between insns that
8083 contain REG_ARGS_SIZE notes.
8085 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
8087 * lto-cgraph.c (asm_nodes_output): Make global.
8088 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
8089 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
8090 (driver_handle_option): Handle OPT_fwpa.
8092 2014-02-05 Jakub Jelinek <jakub@redhat.com>
8095 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
8096 a comment typo and formatting issue. If odr_hash hasn't been
8097 created, return vNULL and set *completep to false.
8100 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
8101 bb with no successors.
8103 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
8106 * doc/invoke.texi (-march): Clarify documentation for ARM.
8110 2014-02-05 Richard Biener <rguenther@suse.de>
8112 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
8113 when not vectorizing because of too many alias checks.
8114 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
8115 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
8117 2014-02-05 Nick Clifton <nickc@redhat.com>
8119 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
8120 accept extended registers in any mode when compiling for the MN10300.
8122 2014-02-05 Yury Gribov <y.gribov@samsung.com>
8124 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
8125 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
8126 sanitization attributes.
8127 (can_inline_edge_p): Likewise.
8128 (sanitize_attrs_match_for_inline_p): New function.
8130 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
8132 * ipa-prop.c (detect_type_change): Shor circuit testing of
8133 type changes on THIS pointer.
8135 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
8138 * config/pa/pa.c (legitimize_tls_address): Return original address
8139 if not passed a SYMBOL_REF rtx.
8140 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
8142 (pa_emit_move_sequence): Simplify TLS source operands.
8143 (pa_legitimate_constant_p): Reject all TLS constants.
8144 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
8145 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
8147 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
8149 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
8150 groups when we know they are controlled by LTO.
8151 * varasm.c (default_binds_local_p_1): If object is in other partition,
8152 it will be resolved locally.
8154 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
8156 * config/host-linux.c (linux_gt_pch_use_address): Don't
8157 use SSIZE_MAX because it is not always defined.
8159 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
8162 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
8163 threshold for pseudo splitting.
8164 (update_ebb_live_info): Process call argument hard registers and
8165 hard registers from insn definition too.
8166 (max_small_class_regs_num): New constant.
8167 (inherit_in_ebb): Update live hard regs through EBBs. Update
8168 reloads_num only for small register classes. Don't split for
8171 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
8174 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
8177 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
8179 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
8182 2014-02-04 Marek Polacek <polacek@redhat.com>
8184 * gdbinit.in (pel): Define.
8186 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
8188 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
8191 2014-02-04 Richard Biener <rguenther@suse.de>
8194 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
8195 in function context local.
8196 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
8197 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
8198 similar to LTO_imported_decl_ref.
8200 2014-02-04 Jakub Jelinek <jakub@redhat.com>
8202 PR tree-optimization/60002
8203 * cgraphclones.c (build_function_decl_skip_args): Clear
8206 PR tree-optimization/60023
8207 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
8208 false to gsi_replace.
8209 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
8210 has been in some EH region and vec_stmt could throw, add
8211 vec_stmt into the same EH region.
8212 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
8213 has no lhs, ignore it.
8214 * internal-fn.c (expand_MASK_LOAD): Likewise.
8217 * tree-inline.c (copy_forbidden): Fail for
8218 __attribute__((optimize (0))) functions.
8221 * omp-low.c (simd_clone_struct_copy): If from->inbranch
8222 is set, copy one less argument.
8223 (expand_simd_clones): Don't subtract clone_info->inbranch
8224 from simd_clone_struct_alloc argument.
8226 PR rtl-optimization/57915
8227 * recog.c (simplify_while_replacing): If all unary/binary/relational
8228 operation arguments are constant, attempt to simplify those.
8231 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
8232 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
8234 2014-02-04 Richard Biener <rguenther@suse.de>
8236 PR tree-optimization/60012
8237 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
8238 TBAA disambiguation to all DDRs.
8240 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8243 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
8244 (LINK_SPEC): Use it for -shared, -shared-libgcc.
8246 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
8249 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
8251 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
8253 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
8254 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
8256 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
8259 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
8260 to figure out targets of polymorphic calls with known decl.
8261 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
8262 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
8263 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
8264 (get_polymorphic_call_info): ... here.
8265 (get_polymorphic_call_info_from_invariant): New function.
8267 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
8269 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
8270 lookup via vtable pointer; check for type consistency
8271 and turn inconsitent facts into UNREACHABLE.
8272 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
8273 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
8274 type inconsistent querries; return UNREACHABLE instead.
8276 2014-02-03 Richard Henderson <rth@twiddle.net>
8279 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
8280 already processed this node.
8281 (normalize_one_pred_1): Pass along mark_set.
8282 (normalize_one_pred): Create and destroy a pointer_set_t.
8283 (normalize_one_pred_chain): Likewise.
8285 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
8287 PR gcov-profile/58602
8288 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
8290 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
8293 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
8294 -fno-devirtualize; try to devirtualize by the knowledge of
8295 virtual table pointer given by aggregate propagation.
8296 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
8297 (ipa_print_node_jump_functions): Dump also offset that
8298 is relevant for polymorphic calls.
8299 (determine_known_aggregate_parts): Add arg_type parameter; use it
8300 instead of determining the type from pointer type.
8301 (ipa_compute_jump_functions_for_edge): Update call of
8302 determine_known_aggregate_parts.
8303 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
8304 (gimple_get_virt_method_for_binfo): ... here; simplify using
8305 vtable_pointer_value_to_vtable.
8306 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
8307 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
8308 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
8309 (vtable_pointer_value_to_vtable): Break out from ...; handle also
8311 (vtable_pointer_value_to_binfo): ... here.
8312 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
8314 2014-02-03 Teresa Johnson <tejohnson@google.com>
8316 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
8317 redef of outer loop index variable.
8319 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
8323 * doc/extend.texi (Function Attributes): Typo.
8325 2014-02-03 Cong Hou <congh@google.com>
8327 PR tree-optimization/60000
8328 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
8329 if the vectorized statement is a store. A store statement can only
8330 appear at the end of pattern statements.
8332 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
8334 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
8335 (ix86_option_override_internal): Default long double to 64-bit for
8336 32-bit Bionic and to 128-bit for 64-bit Bionic.
8338 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
8339 TARGET_LONG_DOUBLE_128 is true.
8340 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
8342 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
8343 (mlong-double-64): Negate -mlong-double-128.
8344 (mlong-double-128): New option.
8346 * config/i386/i386-c.c (ix86_target_macros): Define
8347 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
8349 * doc/invoke.texi: Document -mlong-double-128.
8351 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
8353 PR rtl-optimization/60024
8354 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
8356 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
8358 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
8360 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
8362 PR rtl-optimization/57662
8363 * sel-sched.c (code_motion_path_driver): Do not mark already not
8364 existing blocks in the visiting bitmap.
8366 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
8368 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
8369 on the insn being emitted.
8371 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
8372 Will Deacon <will.deacon@arm.com>
8374 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
8376 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8378 * config/arm/arm-tables.opt: Regenerate.
8380 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8382 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
8383 for vector types other than V16QImode.
8384 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
8385 define_expand, and call altivec_expand_vec_perm_le when producing
8386 code with little endian element order.
8387 (*altivec_vperm_<mode>_internal): New insn having previous
8388 behavior of altivec_vperm_<mode>.
8389 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
8390 altivec_expand_vec_perm_le when producing code with little endian
8392 (*altivec_vperm_<mode>_uns_internal): New insn having previous
8393 behavior of altivec_vperm_<mode>_uns.
8395 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8397 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
8398 (altivec_vsumsws): Add handling for -maltivec=be with a little
8400 (altivec_vsumsws_direct): New.
8401 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
8402 gen_altivec_vsumsws.
8404 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
8406 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
8407 vtable_pointer_value_to_binfo): New functions.
8408 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
8409 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
8411 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
8413 * config/nios2/nios2.md (load_got_register): Initialize GOT
8414 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
8415 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
8417 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
8419 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
8420 preserverd by passthrough, do not propagate the type.
8422 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
8424 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
8425 (mips_atomic_assign_expand_fenv): New function.
8426 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
8428 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
8430 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
8431 (__builtin_mips_set_fcsr): Likewise.
8432 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
8433 MIPS_USI_FTYPE_VOID.
8434 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
8435 (mips16_expand_set_fcsr): Likewise.
8436 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
8437 (mips16_set_fcsr_stub): Likewise.
8438 (mips16_get_fcsr_one_only_stub): New class.
8439 (mips16_set_fcsr_one_only_stub): Likewise.
8440 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
8441 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
8442 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
8443 (hard_float): New availability predicate.
8444 (mips_builtins): Add get_fcsr and set_fcsr.
8445 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
8446 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
8447 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
8448 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
8449 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
8452 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
8454 * config/mips/mips.c (mips_one_only_stub): New class.
8455 (mips_need_mips16_rdhwr_p): Replace with...
8456 (mips16_rdhwr_stub): ...this new variable.
8457 (mips16_stub_call_address): New function.
8458 (mips16_rdhwr_one_only_stub): New class.
8459 (mips_expand_thread_pointer): Use mips16_stub_call_address.
8460 (mips_output_mips16_rdhwr): Delete.
8461 (mips_finish_stub): New function.
8462 (mips_code_end): Use it to handle rdhwr stubs.
8464 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
8467 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
8468 when calculating size of integer atomic types.
8470 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
8472 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
8474 2014-02-01 Jakub Jelinek <jakub@redhat.com>
8476 PR tree-optimization/60003
8477 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
8478 * profile.c (branch_prob): Use gimple_call_builtin_p
8479 to check for BUILT_IN_SETJMP_RECEIVER.
8480 * tree-inline.c (copy_bb): Call notice_special_calls.
8482 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
8485 * lra-constraints.c (process_alt_operands): Update reload_sum only
8488 2014-01-31 Richard Henderson <rth@redhat.com>
8491 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
8492 until after else_eh is processed.
8494 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
8496 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
8497 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
8498 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
8499 in smmintrin.h, remove them.
8500 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
8501 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
8502 * config/i386/i386.md (ROUND_SAE): Fix value.
8503 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
8504 (const48_operand): New.
8505 * config/i386/subst.md (round), (round_expand): Use
8506 const_4_or_8_to_11_operand.
8507 (round_saeonly), (round_saeonly_expand): Use const48_operand.
8509 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
8511 * config/i386/constraints.md (Yk): Swap meaning with k.
8512 * config/i386/i386.md (movhi_internal): Change Yk to k.
8513 (movqi_internal): Ditto.
8514 (*k<logic><mode>): Ditto.
8517 (kandn<mode>): Ditto.
8518 (*<code>hi_1): Ditto.
8519 (*<code>qi_1): Ditto.
8520 (kxnor<mode>): Ditto.
8521 (kortestzhi): Ditto.
8522 (kortestchi): Ditto.
8524 (*one_cmplhi2_1): Ditto.
8525 (*one_cmplqi2_1): Ditto.
8526 * config/i386/sse.md (): Change k to Yk.
8527 (avx512f_load<mode>_mask): Ditto.
8528 (avx512f_blendm<mode>): Ditto.
8529 (avx512f_store<mode>_mask): Ditto.
8530 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
8531 (avx512f_storedqu<mode>_mask): Ditto.
8532 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
8534 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
8535 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
8536 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
8537 (avx512f_maskcmp<mode>3): Ditto.
8538 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
8539 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
8540 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
8541 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
8542 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
8543 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
8544 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
8545 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
8546 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
8547 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
8548 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
8549 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
8550 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
8551 (vec_extract_lo_<mode>_maskm): Ditto.
8552 (vec_extract_hi_<mode>_maskm): Ditto.
8553 (avx512f_vternlog<mode>_mask): Ditto.
8554 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
8555 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
8556 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
8557 (avx512f_<code>v8div16qi2_mask): Ditto.
8558 (avx512f_<code>v8div16qi2_mask_store): Ditto.
8559 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
8560 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
8561 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
8562 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
8563 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
8564 (*avx512pf_gatherpf<mode>df_mask): Ditto.
8565 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
8566 (*avx512pf_scatterpf<mode>df_mask): Ditto.
8567 (avx512cd_maskb_vec_dupv8di): Ditto.
8568 (avx512cd_maskw_vec_dupv16si): Ditto.
8569 (avx512f_vpermi2var<mode>3_maskz): Ditto.
8570 (avx512f_vpermi2var<mode>3_mask): Ditto.
8571 (avx512f_vpermi2var<mode>3_mask): Ditto.
8572 (avx512f_vpermt2var<mode>3_maskz): Ditto.
8573 (*avx512f_gathersi<mode>): Ditto.
8574 (*avx512f_gathersi<mode>_2): Ditto.
8575 (*avx512f_gatherdi<mode>): Ditto.
8576 (*avx512f_gatherdi<mode>_2): Ditto.
8577 (*avx512f_scattersi<mode>): Ditto.
8578 (*avx512f_scatterdi<mode>): Ditto.
8579 (avx512f_compress<mode>_mask): Ditto.
8580 (avx512f_compressstore<mode>_mask): Ditto.
8581 (avx512f_expand<mode>_mask): Ditto.
8582 * config/i386/subst.md (mask): Change k to Yk.
8583 (mask_scalar_merge): Ditto.
8586 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
8588 * doc/extend.texi (Vector Extensions): Document ?: in C++.
8590 2014-01-31 Richard Biener <rguenther@suse.de>
8593 * builtins.c (fold_builtin_memory_op): Make sure to not
8594 use a floating-point mode or a boolean or enumeral type for
8597 2014-01-30 DJ Delorie <dj@redhat.com>
8599 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
8600 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
8601 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
8602 whenever main() has an epilogue.
8604 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8606 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
8607 unused variable "field".
8608 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
8609 (vsx_mergeh_<mode>): Likewise.
8610 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
8611 (altivec_vmrghh): Likewise.
8612 (altivec_vmrghw): Likewise.
8613 (altivec_vmrglb): Likewise.
8614 (altivec_vmrglh): Likewise.
8615 (altivec_vmrglw): Likewise.
8616 (altivec_vspltb): Add missing uses.
8617 (altivec_vsplth): Likewise.
8618 (altivec_vspltw): Likewise.
8619 (altivec_vspltsf): Likewise.
8621 2014-01-30 Jakub Jelinek <jakub@redhat.com>
8624 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
8625 frame related instructions.
8627 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
8629 PR rtl-optimization/59959
8630 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
8631 any reload of register whose subreg is invalid.
8633 2014-01-30 Jakub Jelinek <jakub@redhat.com>
8635 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
8636 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
8637 Add missing return type - void.
8639 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8641 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
8642 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
8643 remove element index adjustment for endian (now handled in vsx.md
8645 (altivec_expand_vec_perm_const): Use
8646 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
8647 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
8648 (vsx_xxspltw_<mode>): Adjust element index for little endian.
8649 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
8650 define_expand and a new define_insn *altivec_vspltb_internal;
8651 adjust for -maltivec=be on a little endian target.
8652 (altivec_vspltb_direct): New.
8653 (altivec_vsplth): Divide into a define_expand and a new
8654 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
8655 little endian target.
8656 (altivec_vsplth_direct): New.
8657 (altivec_vspltw): Divide into a define_expand and a new
8658 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
8659 little endian target.
8660 (altivec_vspltw_direct): New.
8661 (altivec_vspltsf): Divide into a define_expand and a new
8662 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
8663 a little endian target.
8665 2014-01-30 Richard Biener <rguenther@suse.de>
8667 PR tree-optimization/59993
8668 * tree-ssa-forwprop.c (associate_pointerplus): Check we
8669 can propagate form the earlier stmt and avoid the transform
8670 when the intermediate result is needed.
8672 2014-01-30 Alangi Derick <alangiderick@gmail.com>
8674 * README.Portability: Fix typo.
8676 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
8678 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
8679 comparison_operator with ordered_comparison_operator.
8681 2014-01-30 Nick Clifton <nickc@redhat.com>
8683 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
8684 Rename to mn10300_store_multiple_regs.
8685 * config/mn10300/mn10300.c: Likewise.
8686 * config/mn10300/mn10300.md (store_movm): Fix typo: call
8687 store_multiple_regs.
8688 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
8689 Call mn10300_store_multiple_regs.
8691 2014-01-30 Nick Clifton <nickc@redhat.com>
8692 DJ Delorie <dj@redhat.com>
8694 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
8695 %fp 2 to keep registers after it properly word-aligned.
8696 (rl78_alloc_physical_registers_umul): Handle the case where both
8697 input operands are the same.
8699 2014-01-30 Richard Biener <rguenther@suse.de>
8701 PR tree-optimization/59903
8702 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
8705 2014-01-30 Jason Merrill <jason@redhat.com>
8708 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
8711 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
8713 2014-01-30 Richard Biener <rguenther@suse.de>
8715 PR tree-optimization/59951
8716 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
8718 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
8721 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
8722 SFmode to DFmode case.
8724 2014-01-29 DJ Delorie <dj@redhat.com>
8726 * config/msp430/msp430.opt (-minrt): New.
8727 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
8729 (ENDFILE_SPEC): Likewise.
8731 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
8733 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
8734 (estimate_function_body_sizes): Use it.
8736 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
8739 * dwarf2out.c (is_cxx_auto): New.
8740 (is_base_type): Use it.
8741 (gen_type_die_with_usage): Likewise.
8743 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8745 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
8746 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
8747 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
8748 -maltivec=be with LE targets.
8749 (vsx_mergeh_<mode>): Likewise.
8750 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
8751 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
8752 (altivec_vmrghb): Replace with define_expand and new
8753 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
8754 (altivec_vmrghb_direct): New define_insn.
8755 (altivec_vmrghh): Replace with define_expand and new
8756 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
8757 (altivec_vmrghh_direct): New define_insn.
8758 (altivec_vmrghw): Replace with define_expand and new
8759 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
8760 (altivec_vmrghw_direct): New define_insn.
8761 (*altivec_vmrghsf): Adjust for endianness.
8762 (altivec_vmrglb): Replace with define_expand and new
8763 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
8764 (altivec_vmrglb_direct): New define_insn.
8765 (altivec_vmrglh): Replace with define_expand and new
8766 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
8767 (altivec_vmrglh_direct): New define_insn.
8768 (altivec_vmrglw): Replace with define_expand and new
8769 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
8770 (altivec_vmrglw_direct): New define_insn.
8771 (*altivec_vmrglsf): Adjust for endianness.
8772 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
8773 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
8774 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
8775 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
8776 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
8777 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
8778 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
8779 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
8781 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
8783 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
8784 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
8787 2014-01-29 Richard Biener <rguenther@suse.de>
8789 PR tree-optimization/58742
8790 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
8791 associate_pointerplus_align.
8792 (associate_pointerplus_diff): New function.
8793 (associate_pointerplus): Likewise. Call associate_pointerplus_align
8794 and associate_pointerplus_diff.
8796 2014-01-29 Richard Biener <rguenther@suse.de>
8798 * lto-streamer.h (LTO_major_version): Bump to 3.
8799 (LTO_minor_version): Reset to 0.
8801 2014-01-29 Renlin Li <Renlin.Li@arm.com>
8803 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
8804 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
8805 (arm_file_start): Generate correct asm header for armv7ve.
8806 * config/arm/bpabi.h: Add multilib support for armv7ve.
8807 * config/arm/driver-arm.c: Change the architectures of cortex-a7
8808 and cortex-a15 to armv7ve.
8809 * config/arm/t-aprofile: Add multilib support for armv7ve.
8810 * doc/invoke.texi: Document -march=armv7ve.
8812 2014-01-29 Richard Biener <rguenther@suse.de>
8814 PR tree-optimization/58742
8815 * tree-ssa-forwprop.c (associate_plusminus): Return true
8816 if we changed sth, defer EH cleanup to ...
8817 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
8818 (simplify_mult): New function.
8820 2014-01-29 Jakub Jelinek <jakub@redhat.com>
8823 PR tree-optimization/59920
8824 * tree.c (build_common_builtin_nodes): Remove
8825 __builtin_setjmp_dispatcher initialization.
8826 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
8827 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
8828 instead of gsi_after_labels + manually skipping debug stmts.
8829 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
8830 ignore bbs with IFN_ABNORMAL_DISPATCHER.
8831 * tree-inline.c (copy_edges_for_bb): Remove
8832 can_make_abnormal_goto argument, instead add abnormal_goto_dest
8833 argument. Ignore computed_goto_p stmts. Don't call
8834 make_abnormal_goto_edges. If a call might need abnormal edges
8835 for non-local gotos, see if it already has an edge to
8836 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
8837 with true argument, don't do anything then, otherwise add
8838 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
8839 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
8841 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
8842 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
8843 (lower_stmt): Don't set data->calls_builtin_setjmp.
8844 (lower_builtin_setjmp): Adjust comment.
8845 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
8846 * tree-cfg.c (found_computed_goto): Remove.
8847 (factor_computed_gotos): Remove.
8848 (make_goto_expr_edges): Return bool, true for computed gotos.
8849 Don't call make_abnormal_goto_edges.
8850 (build_gimple_cfg): Don't set found_computed_goto, don't call
8851 factor_computed_gotos.
8852 (computed_goto_p): No longer static.
8853 (make_blocks): Don't set found_computed_goto.
8854 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
8855 (make_edges): If make_goto_expr_edges returns true, push bb
8856 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
8857 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
8858 vector. Record mapping between bbs and OpenMP regions if there
8859 are any, adjust make_gimple_omp_edges caller. Call
8860 handle_abnormal_edges.
8861 (make_abnormal_goto_edges): Remove.
8862 * tree-cfg.h (make_abnormal_goto_edges): Remove.
8863 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
8864 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
8865 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
8866 * internal-fn.def (ABNORMAL_DISPATCHER): New.
8867 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
8868 filling *region also set *region_idx to (*region)->entry->index.
8871 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
8872 For REGs set ORIGINAL_REGNO.
8874 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
8876 * doc/md.texi: Mention that a target shouldn't implement
8877 vec_widen_(s|u)mul_even/odd pair if it is less efficient
8880 2014-01-29 Jakub Jelinek <jakub@redhat.com>
8882 PR tree-optimization/59594
8883 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
8884 a copy of the datarefs vector rather than the vector itself.
8886 2014-01-28 Jason Merrill <jason@redhat.com>
8889 * dwarf2out.c (auto_die): New static.
8890 (gen_type_die_with_usage): Handle C++1y 'auto'.
8891 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
8894 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
8897 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
8898 (SPEC_X32): Likewise.
8899 (SPEC_64): Likewise.
8900 * config/i386/i386.c (ix86_option_override_internal): Turn off
8901 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
8903 (x86_file_start): Output .code16gcc for TARGET_16BIT.
8904 * config/i386/i386.h (TARGET_16BIT): New macro.
8905 (TARGET_16BIT_P): Likewise.
8906 * config/i386/i386.opt: Add m16.
8907 * doc/invoke.texi: Document -m16.
8909 2014-01-28 Jakub Jelinek <jakub@redhat.com>
8911 PR preprocessor/59935
8912 * input.c (location_get_source_line): Bail out on when line number
8913 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
8915 2014-01-28 Richard Biener <rguenther@suse.de>
8917 PR tree-optimization/58742
8918 * tree-ssa-forwprop.c (associate_plusminus): Handle
8919 pointer subtraction of the form (T)(P + A) - (T)P.
8921 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8923 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
8926 2014-01-28 Richard Biener <rguenther@suse.de>
8929 2014-01-28 Richard Biener <rguenther@suse.de>
8931 PR rtl-optimization/45364
8932 PR rtl-optimization/59890
8933 * var-tracking.c (local_get_addr_clear_given_value): Handle
8934 already cleared slot.
8935 (val_reset): Handle not allocated local_get_addr_cache.
8936 (vt_find_locations): Use post-order on the inverted CFG.
8938 2014-01-28 Richard Biener <rguenther@suse.de>
8940 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
8942 2014-01-28 Richard Biener <rguenther@suse.de>
8944 PR rtl-optimization/45364
8945 PR rtl-optimization/59890
8946 * var-tracking.c (local_get_addr_clear_given_value): Handle
8947 already cleared slot.
8948 (val_reset): Handle not allocated local_get_addr_cache.
8949 (vt_find_locations): Use post-order on the inverted CFG.
8951 2014-01-28 Alan Modra <amodra@gmail.com>
8953 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
8954 * configure.ac <recursive call for build != host>: Define
8955 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
8956 and LD_FOR_BUILD too.
8957 * configure: Regenerate.
8959 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
8961 * config/i386/i386.c (get_builtin_code_for_version): Separate
8962 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
8963 Broadwell from Haswell.
8965 2014-01-27 Steve Ellcey <sellcey@mips.com>
8967 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
8968 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
8969 * config/mips/mips.c (mips_option_override): Change setting
8971 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
8972 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
8973 Change from Mask to Var.
8975 2014-01-27 Jeff Law <law@redhat.com>
8977 * ipa-inline.c (inline_small_functions): Fix typo.
8979 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
8981 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
8982 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
8983 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
8984 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
8985 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
8986 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
8987 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
8988 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
8989 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
8990 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
8991 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
8992 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
8993 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
8994 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
8995 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
8996 (_mm512_storeu_epi64): Ditto.
8997 (_mm512_cmpge_epi32_mask): Ditto.
8998 (_mm512_cmpge_epu32_mask): Ditto.
8999 (_mm512_cmpge_epi64_mask): Ditto.
9000 (_mm512_cmpge_epu64_mask): Ditto.
9001 (_mm512_cmple_epi32_mask): Ditto.
9002 (_mm512_cmple_epu32_mask): Ditto.
9003 (_mm512_cmple_epi64_mask): Ditto.
9004 (_mm512_cmple_epu64_mask): Ditto.
9005 (_mm512_cmplt_epi32_mask): Ditto.
9006 (_mm512_cmplt_epu32_mask): Ditto.
9007 (_mm512_cmplt_epi64_mask): Ditto.
9008 (_mm512_cmplt_epu64_mask): Ditto.
9009 (_mm512_cmpneq_epi32_mask): Ditto.
9010 (_mm512_cmpneq_epu32_mask): Ditto.
9011 (_mm512_cmpneq_epi64_mask): Ditto.
9012 (_mm512_cmpneq_epu64_mask): Ditto.
9013 (_mm512_expand_pd): Ditto.
9014 (_mm512_expand_ps): Ditto.
9015 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
9016 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
9017 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
9018 * config/i386/i386.c (ix86_builtins): Add
9019 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
9020 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
9021 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
9022 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
9023 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
9024 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
9025 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
9026 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
9027 IX86_BUILTIN_PMOVUSQW512_MEM.
9028 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
9029 __builtin_ia32_pmovsqd512mem_mask,
9030 __builtin_ia32_pmovqd512mem_mask,
9031 __builtin_ia32_pmovusqw512mem_mask,
9032 __builtin_ia32_pmovsqw512mem_mask,
9033 __builtin_ia32_pmovqw512mem_mask,
9034 __builtin_ia32_pmovusdw512mem_mask,
9035 __builtin_ia32_pmovsdw512mem_mask,
9036 __builtin_ia32_pmovdw512mem_mask,
9037 __builtin_ia32_pmovqb512mem_mask,
9038 __builtin_ia32_pmovusqb512mem_mask,
9039 __builtin_ia32_pmovsqb512mem_mask,
9040 __builtin_ia32_pmovusdb512mem_mask,
9041 __builtin_ia32_pmovsdb512mem_mask,
9042 __builtin_ia32_pmovdb512mem_mask.
9043 (bdesc_args): Add __builtin_ia32_expanddf512,
9044 __builtin_ia32_expandsf512.
9045 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
9046 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
9047 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
9048 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
9049 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
9050 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
9051 (avx512f_<code>v8div16qi2_mask_store): This.
9052 (avx512f_expand<mode>): New.
9054 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
9056 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
9058 (_mm512_mask_prefetch_i64gather_pd): Ditto.
9059 (_mm512_prefetch_i32scatter_pd): Ditto.
9060 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
9061 (_mm512_prefetch_i64scatter_pd): Ditto.
9062 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
9063 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
9064 (_mm512_mask_prefetch_i64gather_ps): Ditto.
9065 (_mm512_prefetch_i32scatter_ps): Ditto.
9066 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
9067 (_mm512_prefetch_i64scatter_ps): Ditto.
9068 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
9069 * config/i386/i386-builtin-types.def: Define
9070 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
9071 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
9072 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
9073 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
9074 IX86_BUILTIN_SCATTERPFQPD.
9075 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
9076 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
9077 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
9078 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
9079 __builtin_ia32_scatterpfqps.
9080 (ix86_expand_builtin): Expand new built-ins.
9081 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
9082 fix memory access data type.
9083 (*avx512pf_gatherpf<mode>_mask): Ditto.
9084 (*avx512pf_gatherpf<mode>): Ditto.
9085 (avx512pf_scatterpf<mode>): Ditto.
9086 (*avx512pf_scatterpf<mode>_mask): Ditto.
9087 (*avx512pf_scatterpf<mode>): Ditto.
9088 (GATHER_SCATTER_SF_MEM_MODE): New.
9089 (avx512pf_gatherpf<mode>df): Ditto.
9090 (*avx512pf_gatherpf<mode>df_mask): Ditto.
9091 (*avx512pf_scatterpf<mode>df): Ditto.
9093 2014-01-27 Jakub Jelinek <jakub@redhat.com>
9096 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
9097 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
9100 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
9102 * common/config/arm/arm-common.c
9103 (arm_rewrite_mcpu): Handle multiple names.
9105 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
9107 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
9109 * gimple-builder.h (create_gimple_tmp): Delete.
9111 2014-01-27 Christian Bruel <christian.bruel@st.com>
9113 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
9116 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
9118 * config/pa/pa.md (call): Generate indirect long calls to non-local
9119 functions when outputing 32-bit code.
9120 (call_value): Likewise except for special call to buggy powf function.
9122 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
9123 portable runtime and PIC indirect calls.
9124 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
9125 and PIC call sequences. Use ldo instead of blr to set return register
9126 in PIC call sequence.
9128 2014-01-25 Walter Lee <walt@tilera.com>
9130 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
9131 avoid clobbering a live register.
9133 2014-01-25 Walter Lee <walt@tilera.com>
9135 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
9136 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
9137 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
9138 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
9140 2014-01-25 Walter Lee <walt@tilera.com>
9142 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
9143 arguments on even registers.
9144 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
9146 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
9147 (BIGGEST_ALIGNMENT): Ditto.
9148 (BIGGEST_FIELD_ALIGNMENT): Ditto.
9150 2014-01-25 Walter Lee <walt@tilera.com>
9152 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
9153 insns before bundling.
9154 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
9156 2014-01-25 Walter Lee <walt@tilera.com>
9158 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
9159 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
9160 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
9162 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
9164 * config/mips/constraints.md (kl): Delete.
9165 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
9166 define expands, using...
9167 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
9168 instructions for MIPS16.
9169 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
9170 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
9172 2014-01-25 Walter Lee <walt@tilera.com>
9174 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
9178 2014-01-25 Walter Lee <walt@tilera.com>
9180 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
9181 (TARGET_EXPAND_TO_RTL_HOOK): Define.
9183 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
9185 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
9188 2014-01-25 Jakub Jelinek <jakub@redhat.com>
9190 * print-rtl.c (in_call_function_usage): New var.
9191 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
9192 EXPR_LIST mode as mode and not as reg note name.
9195 * cfgloopmanip.c (copy_loop_info): If
9196 loop->warned_aggressive_loop_optimizations, make sure
9197 the flag is set in target loop too.
9199 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
9201 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
9203 * builtins.def: Likewise.
9204 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
9205 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
9206 * ira.c (ira_setup_eliminable_regset): Likewise.
9207 * omp-low.c (gate_expand_omp): Likewise.
9208 (execute_lower_omp): Likewise.
9209 (diagnose_sb_0): Likewise.
9210 (gate_diagnose_omp_blocks): Likewise.
9211 (simd_clone_clauses_extract): Likewise.
9214 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9216 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
9217 correction for little endian...
9218 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
9221 2014-01-24 Jeff Law <law@redhat.com>
9223 PR tree-optimization/59919
9224 * tree-vrp.c (find_assert_locations_1): Do not register asserts
9225 for non-returning calls.
9227 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
9229 * common/config/aarch64/aarch64-common.c
9230 (aarch64_rewrite_mcpu): Handle multiple names.
9231 * config/aarch64/aarch64.h
9232 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
9234 2014-01-24 Dodji Seketeli <dodji@redhat.com>
9236 * input.c (add_file_to_cache_tab): Handle the case where fopen
9239 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
9242 * config/i386/i386.md (pushsf splitter): Get stack adjustment
9243 from push operand if code of push isn't PRE_DEC.
9245 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
9248 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
9249 -mquad-memory-atomic. Update -mquad-memory documentation to say
9250 it is only used for non-atomic loads/stores.
9252 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
9253 -mquad-memory or -mquad-memory-atomic switches.
9255 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
9256 -mquad-memory-atomic to ISA 2.07 support.
9258 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
9259 to separate support of normal quad word memory operations (ldq, stq)
9260 from the atomic quad word memory operations.
9262 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
9263 support to separate non-atomic quad word operations from atomic
9264 quad word operations. Disable non-atomic quad word operations in
9265 little endian mode so that we don't have to swap words after the
9266 load and before the store.
9267 (quad_load_store_p): Add comment about atomic quad word support.
9268 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
9269 options printed with -mdebug=reg.
9271 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
9272 -mquad-memory-atomic as the test for whether we have quad word
9273 atomic instructions.
9274 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
9275 or -mp8-vector are used, allow byte/half-word atomic operations.
9277 * config/rs6000/sync.md (load_lockedti): Insure that the address
9278 is a proper indexed or indirect address for the lqarx instruction.
9279 On little endian systems, swap the hi/lo registers after the lqarx
9281 (load_lockedpti): Use indexed_or_indirect_operand predicate to
9282 insure the address is valid for the lqarx instruction.
9283 (store_conditionalti): Insure that the address is a proper indexed
9284 or indirect address for the stqcrx. instruction. On little endian
9285 systems, swap the hi/lo registers before doing the stqcrx.
9287 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
9288 insure the address is valid for the stqcrx. instruction.
9290 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
9291 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
9292 type of quad memory support is available.
9294 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
9297 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
9298 there is a danger of looping.
9300 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
9302 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
9303 force flag_ira_loop_pressure if set via command line.
9305 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
9307 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
9308 (ashr_simd): New builtin handling DI mode.
9309 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
9310 (aarch64_sshr_simddi): New match pattern.
9311 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
9312 (vshrd_n_s64): Likewise.
9313 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
9315 2014-01-23 Nick Clifton <nickc@redhat.com>
9317 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
9318 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
9319 favour of mcu specific scripts.
9320 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
9323 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
9324 Alex Velenko <Alex.Velenko@arm.com>
9326 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
9327 (vaddv_s16): Likewise.
9328 (vaddv_s32): Likewise.
9329 (vaddv_u8): Likewise.
9330 (vaddv_u16): Likewise.
9331 (vaddv_u32): Likewise.
9332 (vaddvq_s8): Likewise.
9333 (vaddvq_s16): Likewise.
9334 (vaddvq_s32): Likewise.
9335 (vaddvq_s64): Likewise.
9336 (vaddvq_u8): Likewise.
9337 (vaddvq_u16): Likewise.
9338 (vaddvq_u32): Likewise.
9339 (vaddvq_u64): Likewise.
9340 (vaddv_f32): Likewise.
9341 (vaddvq_f32): Likewise.
9342 (vaddvq_f64): Likewise.
9343 (vmaxv_f32): Likewise.
9344 (vmaxv_s8): Likewise.
9345 (vmaxv_s16): Likewise.
9346 (vmaxv_s32): Likewise.
9347 (vmaxv_u8): Likewise.
9348 (vmaxv_u16): Likewise.
9349 (vmaxv_u32): Likewise.
9350 (vmaxvq_f32): Likewise.
9351 (vmaxvq_f64): Likewise.
9352 (vmaxvq_s8): Likewise.
9353 (vmaxvq_s16): Likewise.
9354 (vmaxvq_s32): Likewise.
9355 (vmaxvq_u8): Likewise.
9356 (vmaxvq_u16): Likewise.
9357 (vmaxvq_u32): Likewise.
9358 (vmaxnmv_f32): Likewise.
9359 (vmaxnmvq_f32): Likewise.
9360 (vmaxnmvq_f64): Likewise.
9361 (vminv_f32): Likewise.
9362 (vminv_s8): Likewise.
9363 (vminv_s16): Likewise.
9364 (vminv_s32): Likewise.
9365 (vminv_u8): Likewise.
9366 (vminv_u16): Likewise.
9367 (vminv_u32): Likewise.
9368 (vminvq_f32): Likewise.
9369 (vminvq_f64): Likewise.
9370 (vminvq_s8): Likewise.
9371 (vminvq_s16): Likewise.
9372 (vminvq_s32): Likewise.
9373 (vminvq_u8): Likewise.
9374 (vminvq_u16): Likewise.
9375 (vminvq_u32): Likewise.
9376 (vminnmv_f32): Likewise.
9377 (vminnmvq_f32): Likewise.
9378 (vminnmvq_f64): Likewise.
9380 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
9382 * config/aarch64/aarch64-simd.md
9383 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
9384 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
9385 (*aarch64_mul3_elt<mode>): Likewise.
9386 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
9387 (*aarch64_mul3_elt_to_64v2df): Likewise.
9388 (*aarch64_mla_elt<mode>): Likewise.
9389 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
9390 (*aarch64_mls_elt<mode>): Likewise.
9391 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
9392 (*aarch64_fma4_elt<mode>): Likewise.
9393 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
9394 (*aarch64_fma4_elt_to_64v2df): Likewise.
9395 (*aarch64_fnma4_elt<mode>): Likewise.
9396 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
9397 (*aarch64_fnma4_elt_to_64v2df): Likewise.
9398 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
9399 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
9400 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
9401 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
9402 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
9403 (aarch64_sqdmull_lane<mode>_internal): Likewise.
9404 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
9406 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
9408 * config/aarch64/aarch64-simd.md
9409 (aarch64_be_checked_get_lane<mode>): New define_expand.
9410 * config/aarch64/aarch64-simd-builtins.def
9411 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
9412 New builtin definition.
9413 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
9414 Use new safe be builtin.
9416 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
9418 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
9420 (aarch64_be_st1<mode>): Likewise.
9421 (aarch_ld1<VALL:mode>): Define_expand modified.
9422 (aarch_st1<VALL:mode>): Likewise.
9423 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
9424 (UNSPEC_ST1): Likewise.
9426 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
9428 * config/microblaze/microblaze.md: Add trap insn and attribute
9430 2014-01-23 Dodji Seketeli <dodji@redhat.com>
9432 PR preprocessor/58580
9433 * input.h (location_get_source_line): Take an additional line_size
9435 (void diagnostics_file_cache_fini): Declare new function.
9436 * input.c (struct fcache): New type.
9437 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
9438 New static constants.
9439 (diagnostic_file_cache_init, total_lines_num)
9440 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
9441 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
9442 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
9443 (get_next_line, read_next_line, goto_next_line, read_line_num):
9444 New static function definitions.
9445 (diagnostic_file_cache_fini): New function.
9446 (location_get_source_line): Take an additional output line_len
9447 parameter. Re-write using lookup_or_add_file_to_cache_tab and
9449 * diagnostic.c (diagnostic_finish): Call
9450 diagnostic_file_cache_fini.
9451 (adjust_line): Take an additional input parameter for the length
9452 of the line, rather than calculating it with strlen.
9453 (diagnostic_show_locus): Adjust the use of
9454 location_get_source_line and adjust_line with respect to their new
9455 signature. While displaying a line now, do not stop at the first
9456 null byte. Rather, display the zero byte as a space and keep
9457 going until we reach the size of the line.
9458 * Makefile.in: Add vec.o to OBJS-libcommon
9460 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
9461 Ilya Tocar <ilya.tocar@intel.com>
9463 * config/i386/avx512fintrin.h (_mm512_kmov): New.
9464 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
9465 (__builtin_ia32_kmov16): Ditto.
9466 * config/i386/i386.md (UNSPEC_KMOV): New.
9469 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
9471 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
9472 (_mm512_storeu_si512): Ditto.
9474 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
9477 * rtl.h (get_referenced_operands): Declare.
9478 * recog.c (get_referenced_operands): New function.
9479 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
9480 operands have been referenced when recording LO_SUM references.
9482 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
9484 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
9486 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
9488 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
9489 Enable for generic and recent AMD targets.
9491 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
9493 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
9494 ARG_SIZE note when adjustment was eliminated.
9496 2014-01-22 Jeff Law <law@redhat.com>
9498 PR tree-optimization/59597
9499 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
9500 in file. Accept new argument REGISTERING and use it to modify
9501 dump output appropriately.
9502 (register_jump_thread): Corresponding changes.
9503 (mark_threaded_blocks): Reinstate code to cancel unprofitable
9504 thread paths involving joiner blocks. Add code to dump cancelled
9505 jump threading paths.
9507 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
9509 PR rtl-optimization/59477
9510 * lra-constraints.c (inherit_in_ebb): Process call for living hard
9511 regs. Update reloads_num and potential_reload_hard_regs for all insns.
9513 2014-01-22 Tom Tromey <tromey@redhat.com>
9515 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
9517 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
9519 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
9521 PR rtl-optimization/59896
9522 * lra-constraints.c (process_alt_operands): Check unused note for
9523 matched operands of insn with no output reloads.
9525 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
9527 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
9528 (mips_move_from_gpr_cost): Likewise.
9530 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
9532 PR rtl-optimization/59858
9533 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
9534 ira_class_hard_regs_num.
9535 (process_alt_operands): Increase reject for dying matched operand.
9537 2014-01-21 Jakub Jelinek <jakub@redhat.com>
9540 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
9541 smaller than size, perform several stores or loads and stores
9542 at dst + count - size to store or copy all of size bytes, rather
9543 than just last modesize bytes.
9545 2014-01-20 DJ Delorie <dj@redhat.com>
9547 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
9548 that CLOBBERs are REGs before propogating their values.
9550 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
9553 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
9554 (cgraph_inline_failed_type): New function.
9555 * cgraph.h (DEFCIFCODE): Add type.
9556 (cgraph_inline_failed_type_t): New enum.
9557 (cgraph_inline_failed_type): New prototype.
9558 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
9559 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
9560 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
9561 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
9562 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
9563 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
9564 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
9565 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
9566 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
9567 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
9568 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
9569 OPTIMIZATION_MISMATCH.
9570 * tree-inline.c (expand_call_inline): Emit errors during
9571 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
9573 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
9576 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
9577 mode attribute in insn output.
9579 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
9581 * output.h (output_constant): Delete.
9582 * varasm.c (output_constant): Make private.
9584 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
9586 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
9588 2014-01-20 Jakub Jelinek <jakub@redhat.com>
9591 * tree.h (fold_builtin_strcat): New prototype.
9592 * builtins.c (fold_builtin_strcat): No longer static. Add len
9593 argument, if non-NULL, don't call c_strlen. Optimize
9594 directly into __builtin_memcpy instead of __builtin_strcpy.
9595 (fold_builtin_2): Adjust fold_builtin_strcat caller.
9596 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
9598 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
9600 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
9601 for SImode_address_operand operands, having only a REG argument.
9603 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
9605 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
9606 loader name using mbig-endian.
9607 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
9609 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
9611 * doc/invoke.texi (-march): Clarify documentation for AArch64.
9615 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
9617 * config/aarch64/aarch64-protos.h
9618 (aarch64_cannot_change_mode_class_ptr): Declare.
9619 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
9620 aarch64_cannot_change_mode_class_ptr): New.
9621 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
9622 backend hook aarch64_cannot_change_mode_class.
9624 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
9626 * common/config/aarch64/aarch64-common.c
9627 (aarch64_handle_option): Don't handle any option order logic here.
9628 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
9629 selected_cpu, warn on architecture version mismatch.
9630 (aarch64_override_options): Fix parsing order for option strings.
9632 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
9633 Iain Sandoe <iain@codesourcery.com>
9636 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
9637 warning. Amend comment to reflect current functionality.
9639 2014-01-20 Richard Biener <rguenther@suse.de>
9642 * builtins.c (fold_builtin_strcat): Remove case better handled
9643 by tree-ssa-strlen.c.
9645 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
9647 * config/aarch64/aarch64.opt
9648 (mcpu, march, mtune): Make case-insensitive.
9650 2014-01-20 Jakub Jelinek <jakub@redhat.com>
9653 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
9654 if operands[1] is a REG or ZERO_EXTEND of a REG.
9656 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
9658 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
9660 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
9662 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
9663 long non-pic millicode calls.
9665 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
9667 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
9669 2014-01-19 Kito Cheng <kito@0xlab.org>
9671 * builtins.c (expand_movstr): Check movstr expand done or fail.
9673 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
9674 H.J. Lu <hongjiu.lu@intel.com>
9677 * config/i386/i386.md (*lea<mode>): Zero-extend return register
9678 to DImode for zero-extended addresses.
9680 2014-01-19 Jakub Jelinek <jakub@redhat.com>
9682 PR rtl-optimization/57763
9683 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
9684 on the new indirect jump_insn and increment LABEL_NUSES (label).
9686 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
9690 * config.gcc (x86_archs): New variable.
9691 (x86_64_archs): Likewise.
9692 (x86_cpus): Likewise.
9693 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
9694 --with-arch/--with-cpu= options.
9695 Support --with-arch=/--with-cpu={nehalem,westmere,
9696 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
9698 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
9700 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
9701 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
9703 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
9705 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
9707 2014-01-18 Jakub Jelinek <jakub@redhat.com>
9710 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
9711 clear cpp_get_options (parse_in)->warn_unused_macros for
9712 ix86_target_macros_internal with cpp_define.
9714 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
9716 * jump.c (delete_related_insns): Keep (use (insn))s.
9717 * reorg.c (redundant_insn): Check for barriers too.
9719 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
9721 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
9723 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
9725 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
9726 call to $$dyncall when TARGET_LONG_CALLS is true.
9728 2014-01-17 Jeff Law <law@redhat.com>
9730 * ree.c (combine_set_extension): Temporarily disable test for
9731 changing number of hard registers.
9733 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
9736 * ipa-inline-analysis.c (inline_free_summary):
9737 Do not free summary of aliases.
9739 2014-01-17 Jakub Jelinek <jakub@redhat.com>
9742 * gimplify.c (gimplify_expr): Use create_tmp_var
9743 instead of create_tmp_var_raw. If cond doesn't have
9744 integral type, don't add the IFN_ANNOTATE builtin at all.
9746 2014-01-17 Martin Jambor <mjambor@suse.cz>
9749 * ipa-cp.c (prev_edge_clone): New variable.
9750 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
9751 Also resize prev_edge_clone vector.
9752 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
9753 (ipcp_edge_removal_hook): New function.
9754 (ipcp_driver): Register ipcp_edge_removal_hook.
9756 2014-01-17 Andrew Pinski <apinski@cavium.com>
9757 Steve Ellcey <sellcey@mips.com>
9760 * config/mips/mips.c (mips_print_operand): Check operand mode instead
9763 2014-01-17 Jeff Law <law@redhat.com>
9766 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
9767 so that pass_ccp runs first.
9769 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
9771 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
9772 (ix86_adjust_cost): Use !TARGET_XXX.
9773 (do_reorder_for_imul): Likewise.
9774 (swap_top_of_ready_list): Likewise.
9775 (ix86_sched_reorder): Likewise.
9777 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
9779 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
9780 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
9781 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
9782 (intel_memset): New. Duplicate slm_memset.
9783 (intel_cost): New. Duplicate slm_cost.
9784 (m_INTEL): New macro.
9785 (processor_target_table): Add "intel".
9786 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
9787 with PROCESSOR_INTEL for "intel".
9788 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
9789 PROCESSOR_SILVERMONT.
9790 (ix86_issue_rate): Likewise.
9791 (ix86_adjust_cost): Likewise.
9792 (ia32_multipass_dfa_lookahead): Likewise.
9793 (swap_top_of_ready_list): Likewise.
9794 (ix86_sched_reorder): Likewise.
9795 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
9796 instead of TARGET_OPT_AGU.
9797 * config/i386/i386.h (TARGET_INTEL): New.
9798 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
9799 (processor_type): Add PROCESSOR_INTEL.
9800 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
9801 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
9803 2014-01-17 Marek Polacek <polacek@redhat.com>
9806 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
9809 2014-01-17 Richard Biener <rguenther@suse.de>
9811 PR tree-optimization/46590
9812 * opts.c (default_options_table): Add entries for
9813 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
9814 all enabled at -O1 but not for -Og.
9815 * common.opt (fbranch-count-reg): Remove Init(1).
9816 (fmove-loop-invariants): Likewise.
9817 (ftree-pta): Likewise.
9819 2014-01-17 Jakub Jelinek <jakub@redhat.com>
9821 * config/i386/i386.c (ix86_data_alignment): For compatibility with
9822 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
9823 decls to at least the GCC 4.8 used alignments.
9826 * tree-nested.c (convert_nonlocal_reference_stmt,
9827 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
9828 of GIMPLE_BIND stmts, adjust associated decls.
9830 2014-01-17 Richard Biener <rguenther@suse.de>
9832 PR tree-optimization/46590
9833 * vec.h (vec<>::bseach): New member function implementing
9834 binary search according to C89 bsearch.
9835 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
9836 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
9837 bitmap pointer again. Make accesses_in_loop a flat array.
9838 (mem_ref_obstack): New global.
9839 (outermost_indep_loop): Adjust for mem_ref->stored changes.
9840 (mark_ref_stored): Likewise.
9841 (ref_indep_loop_p_2): Likewise.
9842 (set_ref_stored_in_loop): New helper function.
9843 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
9844 (memref_free): Adjust.
9845 (record_mem_ref_loc): Simplify.
9846 (gather_mem_refs_stmt): Adjust.
9847 (sort_locs_in_loop_postorder_cmp): New function.
9848 (analyze_memory_references): Sort accesses_in_loop after
9849 loop postorder number.
9850 (find_ref_loc_in_loop_cmp): New function.
9851 (for_all_locs_in_loop): Find relevant cluster of locs in
9852 accesses_in_loop and iterate without recursion.
9853 (execute_sm): Avoid uninit warning.
9854 (struct ref_always_accessed): Simplify.
9855 (ref_always_accessed::operator ()): Likewise.
9856 (ref_always_accessed_p): Likewise.
9857 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
9858 loop postorder numbers here.
9859 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
9862 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
9865 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
9866 on decls for which assemble_alias has been called.
9868 2014-01-17 Nick Clifton <nickc@redhat.com>
9870 * config/msp430/msp430.opt: (mcpu): New option.
9871 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
9872 (msp430_option_override): Parse target_cpu. If the MCU name
9873 matches a generic string, clear target_mcu.
9874 (msp430_attr): Allow numeric interrupt values up to 63.
9875 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
9876 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
9878 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
9880 * config/msp430/msp430.md (popm): Use %J rather than %I.
9881 (addsi3): Use msp430_nonimmediate_operand for operand 2.
9882 (addhi_cy_i): Use immediate_operand for operand 2.
9883 * doc/invoke.texi: Document -mcpu option.
9885 2014-01-17 Richard Biener <rguenther@suse.de>
9887 PR rtl-optimization/38518
9888 * df.h (df_analyze_loop): Declare.
9889 * df-core.c: Include cfgloop.h.
9890 (df_analyze_1): Split out main part of df_analyze.
9891 (df_analyze): Adjust.
9892 (loop_inverted_post_order_compute): New function.
9893 (loop_post_order_compute): Likewise.
9894 (df_analyze_loop): New function avoiding whole-function
9896 * loop-invariant.c (find_defs): Use df_analyze_loop.
9897 (find_invariants): Adjust.
9898 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
9900 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
9902 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
9903 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
9905 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
9907 * ipa-ref.c (ipa_remove_stmt_references): Fix references
9908 traversal when removing references.
9910 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
9913 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
9915 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
9918 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
9919 pushing a reload for an autoinc when we had previously reloaded an
9920 inner part of the address.
9922 2014-01-16 Jakub Jelinek <jakub@redhat.com>
9924 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
9926 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
9927 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
9928 when not giving up or versioning for alias only because of
9930 (vect_analyze_data_ref_dependences): Set to true.
9931 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
9933 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
9934 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
9938 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
9941 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
9942 operand 0 predicate for gathers, use a new pseudo as subtarget.
9944 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
9947 * lra-constraints.c (process_alt_operands): Add printing debug info.
9948 Check absence of input/output reloads for matched operands too.
9950 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
9952 PR rtl-optimization/59835
9953 * ira.c (ira_init_register_move_cost): Increase cost for
9956 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
9958 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
9960 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
9963 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
9964 non-register objects. Use gen_(high/low)part more consistently.
9967 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
9970 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
9971 endian support, remove tests for WORDS_BIG_ENDIAN.
9972 (p8_mfvsrd_3_<mode>): Likewise.
9973 (reload_gpr_from_vsx<mode>): Likewise.
9974 (reload_gpr_from_vsxsf): Likewise.
9975 (p8_mfvsrd_4_disf): Likewise.
9977 2014-01-16 Richard Biener <rguenther@suse.de>
9979 PR rtl-optimization/46590
9980 * lcm.c (compute_antinout_edge): Use postorder iteration.
9981 (compute_laterin): Use inverted postorder iteration.
9983 2014-01-16 Nick Clifton <nickc@redhat.com>
9986 * varasm.c (output_constant): Return the number of bytes actually
9988 (output_constructor_array_range): Update the field size with the
9989 number of bytes emitted by output_constant.
9990 (output_constructor_regular_field): Likewise. Also do not
9991 complain if the total number of bytes emitted is now greater
9992 than the expected fieldpos.
9993 * output.h (output_constant): Update prototype and descriptive comment.
9995 2014-01-16 Marek Polacek <polacek@redhat.com>
9998 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
9999 it is error_mark_node.
10001 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
10003 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
10004 VALID_AVX256_REG_OR_OI_MODE.
10006 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
10008 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
10009 current procedure should be profiled.
10011 2014-01-15 Andrew Pinski <apinski@cavium.com>
10013 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
10014 of moving from/to the STACK_REG register class.
10016 2014-01-15 Richard Henderson <rth@redhat.com>
10019 * reginfo.c (global_regs_decl): Globalize.
10020 * rtl.h (global_regs_decl): Declare.
10021 * ira.c (do_reload): Diagnose frame_pointer_needed and it
10022 reserved via global_regs.
10024 2014-01-15 Teresa Johnson <tejohnson@google.com>
10026 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
10028 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
10030 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
10031 and vmulosh rather than call gen_vec_widen_smult_*.
10032 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
10033 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
10034 (vec_widen_smult_even_v16qi): Likewise.
10035 (vec_widen_umult_even_v8hi): Likewise.
10036 (vec_widen_smult_even_v8hi): Likewise.
10037 (vec_widen_umult_odd_v16qi): Likewise.
10038 (vec_widen_smult_odd_v16qi): Likewise.
10039 (vec_widen_umult_odd_v8hi): Likewise.
10040 (vec_widen_smult_odd_v8hi): Likewise.
10041 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
10042 vmuloub rather than call gen_vec_widen_umult_*.
10043 (vec_widen_umult_lo_v16qi): Likewise.
10044 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
10045 vmulosb rather than call gen_vec_widen_smult_*.
10046 (vec_widen_smult_lo_v16qi): Likewise.
10047 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
10048 rather than call gen_vec_widen_umult_*.
10049 (vec_widen_umult_lo_v8hi): Likewise.
10050 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
10051 rather than call gen_vec_widen_smult_*.
10052 (vec_widen_smult_lo_v8hi): Likewise.
10054 2014-01-15 Jeff Law <law@redhat.com>
10056 PR tree-optimization/59747
10057 * ree.c (find_and_remove_re): Properly handle case where a second
10058 eliminated extension requires widening a copy created for elimination
10059 of a prior extension.
10060 (combine_set_extension): Ensure that the number of hard regs needed
10061 for a destination register does not change when we widen it.
10063 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
10065 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
10066 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
10067 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
10068 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
10069 (avr-*-rtems*): Likewise.
10070 (bfin*-rtems*): Likewise.
10071 (moxie-*-rtems*): Likewise.
10072 (h8300-*-rtems*): Likewise.
10073 (i[34567]86-*-rtems*): Likewise.
10074 (lm32-*-rtems*): Likewise.
10075 (m32r-*-rtems*): Likewise.
10076 (m68k-*-rtems*): Likewise.
10077 (microblaze*-*-rtems*): Likewise.
10078 (mips*-*-rtems*): Likewise.
10079 (powerpc-*-rtems*): Likewise.
10080 (sh-*-rtems*): Likewise.
10081 (sparc-*-rtems*): Likewise.
10082 (sparc64-*-rtems*): Likewise.
10083 (v850-*-rtems*): Likewise.
10084 (m32c-*-rtems*): Likewise.
10086 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
10088 PR rtl-optimization/59511
10089 * ira.c (ira_init_register_move_cost): Use memory costs for some
10090 cases of register move cost calculations.
10091 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
10092 instead of BB frequency.
10093 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
10094 * lra-assigns.c (find_hard_regno_for): Ditto.
10096 2014-01-15 Richard Biener <rguenther@suse.de>
10098 PR tree-optimization/59822
10099 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
10100 (vectorizable_load): Use it to hoist defs of uses of invariant
10101 loads out of the loop.
10103 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
10104 Kugan Vivekanandarajah <kuganv@linaro.org>
10107 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
10110 2014-01-15 Richard Biener <rguenther@suse.de>
10112 PR rtl-optimization/59802
10113 * lcm.c (compute_available): Use inverted postorder to seed
10114 the initial worklist.
10116 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10119 * config/s390/s390.c (s390_preferred_reload_class): Don't return
10120 ADDR_REGS for invalid symrefs in non-PIC code.
10122 2014-01-15 Jakub Jelinek <jakub@redhat.com>
10125 * builtins.c (determine_block_size): Initialize *probable_max_size
10126 even if len_rtx is CONST_INT.
10128 2014-01-14 Andrew Pinski <apinski@cavium.com>
10130 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
10131 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
10132 (cortexa53_tunings): Likewise.
10133 (aarch64_sched_issue_rate): New function.
10134 (TARGET_SCHED_ISSUE_RATE): Define.
10136 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
10138 * ira-costs.c (find_costs_and_classes): Add missed
10139 ira_init_register_move_cost_if_necessary.
10141 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
10144 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
10146 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
10149 * config/i386/i386.c (type_natural_mode): Add a bool parameter
10150 to indicate if type is used for function return value. Warn ABI
10151 change if the vector mode isn't available for function return value.
10152 (ix86_function_arg_advance): Pass false to type_natural_mode.
10153 (ix86_function_arg): Likewise.
10154 (ix86_gimplify_va_arg): Likewise.
10155 (function_arg_32): Don't warn ABI change.
10156 (ix86_function_value): Pass true to type_natural_mode.
10157 (ix86_return_in_memory): Likewise.
10158 (ix86_struct_value_rtx): Removed.
10159 (TARGET_STRUCT_VALUE_RTX): Likewise.
10161 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
10163 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
10164 converting a conditional jump into a conditional return.
10166 2014-01-14 Richard Biener <rguenther@suse.de>
10168 PR tree-optimization/58921
10169 PR tree-optimization/59006
10170 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
10171 hoisting invariant stmts.
10172 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
10173 invariant loads on the preheader edge if possible.
10175 2014-01-14 Joey Ye <joey.ye@arm.com>
10177 * doc/plugin.texi (Building GCC plugins): Update to C++.
10179 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
10181 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
10182 (_mm_rcp28_round_ss): Ditto.
10183 (_mm_rsqrt28_round_sd): Ditto.
10184 (_mm_rsqrt28_round_ss): Ditto.
10185 (_mm_rcp28_sd): Ditto.
10186 (_mm_rcp28_ss): Ditto.
10187 (_mm_rsqrt28_sd): Ditto.
10188 (_mm_rsqrt28_ss): Ditto.
10189 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
10190 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
10191 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
10192 (IX86_BUILTIN_RCP28SD): Ditto.
10193 (IX86_BUILTIN_RCP28SS): Ditto.
10194 (IX86_BUILTIN_RSQRT28SD): Ditto.
10195 (IX86_BUILTIN_RSQRT28SS): Ditto.
10196 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
10197 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
10198 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
10199 (ix86_expand_special_args_builtin): Expand new FTYPE.
10200 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
10201 (srcp14<mode>): Make insn unary.
10202 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
10203 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
10204 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
10205 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
10206 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
10207 Fix rounding: make it SAE only.
10208 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
10210 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
10212 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
10213 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
10214 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
10215 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
10216 (round_saeonly_mask_scalar_operand4): Ditto.
10217 (round_saeonly_mask_scalar_op3): Ditto.
10218 (round_saeonly_mask_scalar_op4): Ditto.
10220 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10222 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
10223 Implement -maltivec=be for vec_insert and vec_extract.
10225 2014-01-10 DJ Delorie <dj@redhat.com>
10227 * config/msp430/msp430.md (call_internal): Don't allow memory
10228 references with SP as the base register.
10229 (call_value_internal): Likewise.
10230 * config/msp430/constraints.md (Yc): New. For memory references
10231 that don't use SP as a base register.
10233 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
10234 "an integer without a # prefix"
10235 * config/msp430/msp430.md (epilogue_helper): Use it.
10237 2014-01-13 Jakub Jelinek <jakub@redhat.com>
10240 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
10241 AVX512F gather builtins.
10242 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
10243 on gather decls with INTEGER_TYPE masktype.
10244 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
10245 directly into the builtin rather than hoisting it before loop.
10247 PR tree-optimization/59387
10248 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
10249 (scev_const_prop): If folded_casts and type has undefined overflow,
10250 use force_gimple_operand instead of force_gimple_operand_gsi and
10251 for each added stmt if it is assign with
10252 arith_code_with_undefined_signed_overflow, call
10253 rewrite_to_defined_overflow.
10254 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
10255 gimple-fold.h instead.
10256 (arith_code_with_undefined_signed_overflow,
10257 rewrite_to_defined_overflow): Moved to ...
10258 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
10259 rewrite_to_defined_overflow): ... here. No longer static.
10260 Include gimplify-me.h.
10261 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
10262 rewrite_to_defined_overflow): New prototypes.
10264 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10266 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
10268 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
10270 * builtins.c (get_object_alignment_2): Minor tweak.
10271 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
10273 2014-01-13 Christian Bruel <christian.bruel@st.com>
10275 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
10276 optimized non constant lengths.
10278 2014-01-13 Jakub Jelinek <jakub@redhat.com>
10281 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
10282 load as __atomic_load_N if possible.
10284 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
10286 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
10288 (rs6000_expand_builtin): Adjust call.
10290 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
10293 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
10294 * config/rs6000/rs6000.c: Include target-globals.h.
10295 (rs6000_set_current_function): Instead of doing target_reinit
10296 unconditionally, use save_target_globals_default_opts and
10297 restore_target_globals.
10299 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
10301 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
10302 (rs6000_expand_builtin): Handle mffs and mtfsf.
10303 (rs6000_init_builtins): Define mffs and mtfsf.
10304 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
10305 (rs6000_mffs): New pattern.
10306 (rs6000_mtfsf): New pattern.
10308 2014-01-11 Bin Cheng <bin.cheng@arm.com>
10310 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
10311 Start narrowing with START. Apply candidate-use pair
10312 and check overall cost in narrowing.
10313 (iv_ca_prune): Pass new argument.
10315 2014-01-10 Jeff Law <law@redhat.com>
10317 PR middle-end/59743
10318 * ree.c (combine_reaching_defs): Ensure the defining statement
10319 occurs before the extension when optimizing extensions with
10320 different source and destination hard registers.
10322 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
10325 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
10326 vtables into the type inheritance graph.
10328 2014-01-10 Jakub Jelinek <jakub@redhat.com>
10330 PR rtl-optimization/59754
10331 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
10332 modes in the REGNO != REGNO case.
10334 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10336 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
10338 2014-01-10 Jakub Jelinek <jakub@redhat.com>
10340 PR tree-optimization/59745
10341 * tree-predcom.c (tree_predictive_commoning_loop): Call
10342 free_affine_expand_cache if giving up because components is NULL.
10344 * target-globals.c (save_target_globals): Allocate < 4KB structs using
10345 GC in payload of target_globals struct instead of allocating them on
10346 the heap and the larger structs separately using GC.
10347 * target-globals.h (struct target_globals): Make regs, hard_regs,
10348 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
10349 of GTY((skip)) and change type to void *.
10350 (reset_target_globals): Cast loads from those fields to corresponding
10353 2014-01-10 Steve Ellcey <sellcey@mips.com>
10356 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
10357 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
10358 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
10360 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
10363 * aarch64-modes.def (CC_Zmode): New flags mode.
10364 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
10365 represents an equality.
10366 (aarch64_get_condition_code): Handle CC_Zmode.
10367 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
10369 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10371 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
10372 extraction in good case.
10374 2014-01-10 Richard Biener <rguenther@suse.de>
10376 PR tree-optimization/59374
10377 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
10378 checking after SLP discovery. Mark stmts not participating
10379 in any SLP instance properly.
10381 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10383 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
10384 when handling a SET rtx.
10386 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10388 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
10389 (cortex-a57): Likewise.
10390 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
10392 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10394 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
10395 non-iwmmxt builtins.
10397 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
10401 * ipa-devirt.c record_target_from_binfo): Take as argument
10402 stack of binfos and lookup matching one for virtual inheritance.
10403 (possible_polymorphic_call_targets_1): Update.
10405 2014-01-10 Huacai Chen <chenhc@lemote.com>
10407 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
10408 kernel strings for Loongson-2E/2F/3A.
10410 2014-01-10 Jakub Jelinek <jakub@redhat.com>
10412 PR middle-end/59670
10413 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
10414 is_gimple_call before calling gimple_call_internal_p.
10416 2014-01-09 Steve Ellcey <sellcey@mips.com>
10418 * Makefile.in (TREE_FLOW_H): Remove.
10419 (TREE_SSA_H): Add file names from tree-flow.h.
10420 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
10421 * tree.h: Remove tree-flow.h reference.
10422 * hash-table.h: Remove tree-flow.h reference.
10423 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
10424 reference with tree-ssa-loop.h.
10426 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10428 * doc/invoke.texi: Add -maltivec={be,le} options, and document
10429 default element-order behavior for -maltivec.
10430 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
10431 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
10432 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
10433 when targeting big endian, at least for now.
10434 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
10436 2014-01-09 Jakub Jelinek <jakub@redhat.com>
10438 PR middle-end/47735
10439 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
10440 var satisfies use_register_for_decl, just take into account type
10441 alignment, rather than decl alignment.
10443 PR tree-optimization/59622
10444 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
10445 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
10446 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
10447 Don't devirtualize for inplace at all. For targets.length () == 1,
10448 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
10450 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
10452 * config/i386/i386.md (cpu): Remove the unused btver1.
10454 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
10456 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
10458 2014-01-09 Jakub Jelinek <jakub@redhat.com>
10461 * tree-core.h (struct target_globals): New forward declaration.
10462 (struct tree_target_option): Add globals field.
10463 * tree.h (TREE_TARGET_GLOBALS): Define.
10464 (prepare_target_option_nodes_for_pch): New prototype.
10465 * target-globals.h (struct target_globals): Define even if
10466 !SWITCHABLE_TARGET.
10467 * tree.c (prepare_target_option_node_for_pch,
10468 prepare_target_option_nodes_for_pch): New functions.
10469 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
10470 * config/i386/i386.c: Include target-globals.h.
10471 (ix86_set_current_function): Instead of doing target_reinit
10472 unconditionally, use save_target_globals_default_opts and
10473 restore_target_globals.
10475 2014-01-09 Richard Biener <rguenther@suse.de>
10477 PR tree-optimization/59715
10478 * tree-cfg.h (split_critical_edges): Declare.
10479 * tree-cfg.c (split_critical_edges): Export.
10480 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
10482 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
10484 * cfgexpand.c (expand_stack_vars): Optionally disable
10485 asan stack protection.
10486 (expand_used_vars): Likewise.
10487 (partition_stack_vars): Likewise.
10488 * asan.c (asan_emit_stack_protection): Optionally disable
10489 after return stack usage.
10490 (instrument_derefs): Optionally disable memory access instrumentation.
10491 (instrument_builtin_call): Likewise.
10492 (instrument_strlen_call): Likewise.
10493 (asan_protect_global): Optionally disable global variables protection.
10494 * doc/invoke.texi: Added doc for new options.
10495 * params.def: Added new options.
10496 * params.h: Likewise.
10498 2014-01-09 Jakub Jelinek <jakub@redhat.com>
10500 PR rtl-optimization/59724
10501 * ifcvt.c (cond_exec_process_if_block): Don't call
10502 flow_find_head_matching_sequence with 0 longest_match.
10503 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
10504 non-active insns if !stop_after.
10505 (try_head_merge_bb): Revert 2014-01-07 changes.
10507 2014-01-08 Jeff Law <law@redhat.com>
10509 * ree.c (get_sub_rtx): New function, extracted from...
10510 (merge_def_and_ext): Here.
10511 (combine_reaching_defs): Use get_sub_rtx.
10513 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
10515 * cgraph.h (varpool_variable_node): Do not choke on null node.
10517 2014-01-08 Catherine Moore <clm@codesourcery.com>
10519 * config/mips/mips.md (simple_return): Attempt to use JRC
10521 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
10523 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
10525 PR rtl-optimization/59137
10526 * reorg.c (steal_delay_list_from_target): Call update_block for
10528 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
10530 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10532 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
10533 two duplicate entries.
10535 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
10538 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
10540 * config/mips/mips.c (mips_truncated_op_cost): New function.
10541 (mips_rtx_costs): Adjust test for BADDU.
10542 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
10544 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
10546 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
10547 (*baddu_si): ...this new pattern.
10549 2014-01-08 Jakub Jelinek <jakub@redhat.com>
10552 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
10554 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
10556 PR middle-end/57748
10557 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
10559 (expand_expr, expand_normal): Adjust.
10560 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
10561 inner_reference_p. Use inner_reference_p to expand inner references.
10562 (store_expr): Adjust.
10563 * cfgexpand.c (expand_call_stmt): Adjust.
10565 2014-01-08 Rong Xu <xur@google.com>
10567 * gcov-io.c (gcov_var): Move from gcov-io.h.
10568 (gcov_position): Ditto.
10569 (gcov_is_error): Ditto.
10570 (gcov_rewrite): Ditto.
10571 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
10572 only part to libgcc/libgcov.h.
10574 2014-01-08 Marek Polacek <polacek@redhat.com>
10576 PR middle-end/59669
10577 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
10579 2014-01-08 Marek Polacek <polacek@redhat.com>
10582 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
10584 2014-01-08 Jakub Jelinek <jakub@redhat.com>
10586 PR rtl-optimization/59649
10587 * stor-layout.c (get_mode_bounds): For BImode return
10588 0 and STORE_FLAG_VALUE.
10590 2014-01-08 Richard Biener <rguenther@suse.de>
10592 PR middle-end/59630
10593 * gimple.h (is_gimple_builtin_call): Remove.
10594 (gimple_builtin_call_types_compatible_p): New.
10595 (gimple_call_builtin_p): New overload.
10596 * gimple.c (is_gimple_builtin_call): Remove.
10597 (validate_call): Rename to ...
10598 (gimple_builtin_call_types_compatible_p): ... this and export. Also
10599 check return types.
10600 (validate_type): New static function.
10601 (gimple_call_builtin_p): New overload and adjust.
10602 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
10603 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
10604 (gimple_fold_stmt_to_constant_1): Likewise.
10605 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
10607 2014-01-08 Richard Biener <rguenther@suse.de>
10609 PR middle-end/59471
10610 * gimplify.c (gimplify_expr): Gimplify register-register type
10611 VIEW_CONVERT_EXPRs to separate stmts.
10613 2014-01-07 Jeff Law <law@redhat.com>
10615 PR middle-end/53623
10616 * ree.c (combine_set_extension): Handle case where source
10617 and destination registers in an extension insn are different.
10618 (combine_reaching_defs): Allow source and destination registers
10619 in extension to be different under limited circumstances.
10620 (add_removable_extension): Remove restriction that the
10621 source and destination registers in the extension are the same.
10622 (find_and_remove_re): Emit a copy from the extension's
10623 destination to its source after the defining insn if
10624 the source and destination registers are different.
10626 PR middle-end/59285
10627 * ifcvt.c (merge_if_block): If we are merging a block with more than
10628 one successor with a block with no successors, remove any BARRIER
10629 after the second block.
10631 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
10633 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
10635 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
10638 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
10639 for 14-bit register offsets when INT14_OK_STRICT is false.
10641 2014-01-07 Roland Stigge <stigge@antcom.de>
10642 Michael Meissner <meissner@linux.vnet.ibm.com>
10645 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
10646 Only check TFmode for SPE constants. Don't check TImode or TDmode.
10648 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
10650 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
10653 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
10655 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
10656 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
10657 rtx is const0_rtx or not.
10659 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
10662 * target-globals.c (save_target_globals): Remove this_fn_optab
10664 * toplev.c: Include optabs.h.
10665 (target_reinit): Temporarily restore the global options if another
10666 set of options are in force.
10668 2014-01-07 Jakub Jelinek <jakub@redhat.com>
10670 PR rtl-optimization/58668
10671 * cfgcleanup.c (flow_find_cross_jump): Don't count
10672 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
10673 to determine what is counted.
10674 (flow_find_head_matching_sequence): Use active_insn_p to determine
10676 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
10678 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
10679 determine what is counted.
10681 PR tree-optimization/59643
10682 * tree-predcom.c (split_data_refs_to_components): If one dr is
10683 read and one write, determine_offset fails and the write isn't
10684 in the bad component, just put the read into the bad component.
10686 2014-01-07 Mike Stump <mikestump@comcast.net>
10687 Jakub Jelinek <jakub@redhat.com>
10690 * tree-core.h (struct tree_optimization_option): Change optabs
10691 type from unsigned char * to void *.
10692 * optabs.c (init_tree_optimization_optabs): Adjust
10693 TREE_OPTIMIZATION_OPTABS initialization.
10695 2014-01-06 Jakub Jelinek <jakub@redhat.com>
10698 * config/i386/i386.h (struct machine_function): Add
10699 no_drap_save_restore field.
10700 * config/i386/i386.c (ix86_save_reg): Use
10701 !cfun->machine->no_drap_save_restore instead of
10702 crtl->stack_realign_needed.
10703 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
10704 this function clears frame_pointer_needed. Set
10705 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
10706 and DRAP reg is needed.
10708 2014-01-06 Marek Polacek <polacek@redhat.com>
10711 * doc/implement-c.texi: Mention that other integer types are
10712 permitted as bit-field types in strictly conforming mode.
10714 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
10716 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
10717 is newly allocated.
10719 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
10721 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
10723 2014-01-06 Martin Jambor <mjambor@suse.cz>
10726 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
10728 * ipa-prop.c (ipa_print_node_params): Fix indentation.
10730 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
10734 * var-tracking.c (add_stores): Preserve the value of the source even if
10735 we don't record the store.
10737 2014-01-06 Terry Guo <terry.guo@arm.com>
10739 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
10741 2014-01-05 Iain Sandoe <iain@codesourcery.com>
10744 * config/darwin.c (darwin_function_section): Adjust return values to
10745 correspond to optimisation changes made in r206070.
10747 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
10749 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
10750 from prefetch_block tune setting.
10751 (nocona_cost): Correct size of prefetch block to 64.
10753 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
10755 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
10756 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
10757 used to save the static chain register in the computation of the offset
10758 from which the FP registers need to be restored.
10760 2014-01-04 Jakub Jelinek <jakub@redhat.com>
10762 PR tree-optimization/59519
10763 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
10764 ICE if get_current_def (current_new_name) is already non-NULL, as long
10765 as it is a phi result of some other phi in *new_exit_bb that has
10768 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
10769 or vmovdqu* for misaligned_operand.
10770 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
10771 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
10772 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
10773 aligned_mem for AVX512F masked aligned load and store builtins and for
10774 non-temporal moves.
10776 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
10778 PR tree-optimization/59651
10779 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
10780 Address range for negative step should be added by TYPE_SIZE_UNIT.
10782 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
10784 * config/m68k/m68k.c (handle_move_double): Handle pushes with
10785 overlapping registers also for registers other than the stack pointer.
10787 2014-01-03 Marek Polacek <polacek@redhat.com>
10790 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
10793 2014-01-03 Jakub Jelinek <jakub@redhat.com>
10796 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
10799 * config/i386/i386.md (MODE_SIZE): New mode attribute.
10800 (push splitter): Use <P:MODE_SIZE> instead of
10801 GET_MODE_SIZE (<P:MODE>mode).
10802 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
10803 (mov -1, reg peephole2): Likewise.
10804 * config/i386/sse.md (*mov<mode>_internal,
10805 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
10806 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
10807 *<code><mode>3, *andnot<mode>3<mask_name>,
10808 <mask_codefor><code><mode>3<mask_name>): Likewise.
10809 * config/i386/subst.md (mask_mode512bit_condition,
10810 sd_mask_mode512bit_condition): Likewise.
10812 2014-01-02 Xinliang David Li <davidxl@google.com>
10814 PR tree-optimization/59303
10815 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
10816 (dump_predicates): Better output format.
10817 (pred_equal_p): New function.
10818 (is_neq_relop_p): Ditto.
10819 (is_neq_zero_form_p): Ditto.
10820 (pred_expr_equal_p): Ditto.
10821 (pred_neg_p): Ditto.
10822 (simplify_pred): Ditto.
10823 (simplify_preds_2): Ditto.
10824 (simplify_preds_3): Ditto.
10825 (simplify_preds_4): Ditto.
10826 (simplify_preds): Ditto.
10827 (push_pred): Ditto.
10828 (push_to_worklist): Ditto.
10829 (get_pred_info_from_cmp): Ditto.
10830 (is_degenerated_phi): Ditto.
10831 (normalize_one_pred_1): Ditto.
10832 (normalize_one_pred): Ditto.
10833 (normalize_one_pred_chain): Ditto.
10834 (normalize_preds): Ditto.
10835 (normalize_cond_1): Remove function.
10836 (normalize_cond): Ditto.
10837 (is_gcond_subset_of): Ditto.
10838 (is_subset_of_any): Ditto.
10839 (is_or_set_subset_of): Ditto.
10840 (is_and_set_subset_of): Ditto.
10841 (is_norm_cond_subset_of): Ditto.
10842 (pred_chain_length_cmp): Ditto.
10843 (convert_control_dep_chain_into_preds): Type change.
10844 (find_predicates): Ditto.
10845 (find_def_preds): Ditto.
10846 (destroy_predicates_vecs): Ditto.
10847 (find_matching_predicates_in_rest_chains): Ditto.
10848 (use_pred_not_overlap_with_undef_path_pred): Ditto.
10849 (is_pred_expr_subset): Ditto.
10850 (is_pred_chain_subset_of): Ditto.
10851 (is_included_in): Ditto.
10852 (is_superset_of): Ditto.
10854 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
10856 Update copyright years.
10858 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
10860 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
10861 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
10862 config/arc/arc.md, config/arc/arc.opt,
10863 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
10864 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
10865 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
10866 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
10867 config/linux-protos.h, config/linux.c, config/winnt-c.c,
10868 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
10869 vtable-verify.c, vtable-verify.h: Use the standard form for the
10872 2014-01-02 Tobias Burnus <burnus@net-b.de>
10874 * gcc.c (process_command): Update copyright notice dates.
10875 * gcov-dump.c: Ditto.
10877 * doc/cpp.texi: Bump @copying's copyright year.
10878 * doc/cppinternals.texi: Ditto.
10879 * doc/gcc.texi: Ditto.
10880 * doc/gccint.texi: Ditto.
10881 * doc/gcov.texi: Ditto.
10882 * doc/install.texi: Ditto.
10883 * doc/invoke.texi: Ditto.
10885 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
10887 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
10889 2014-01-01 Jakub Jelinek <jakub@redhat.com>
10891 * config/i386/sse.md (*mov<mode>_internal): Guard
10892 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
10894 PR rtl-optimization/59647
10895 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
10896 new_rtx into UNSIGNED_FLOAT rtxes.
10898 Copyright (C) 2014 Free Software Foundation, Inc.
10900 Copying and distribution of this file, with or without modification,
10901 are permitted in any medium without royalty provided the copyright
10902 notice and this notice are preserved.