1 2014-05-15 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3 * regcprop.h: New file.
4 * regcprop.c (skip_debug_insn_p): New decl.
5 (replace_oldest_value_reg): Check skip_debug_insn_p.
6 (copyprop_hardreg_forward_bb_without_debug_insn.): New function.
7 * shrink-wrap.c: include regcprop.h
9 Call copyprop_hardreg_forward_bb_without_debug_insn.
11 2014-05-15 Zhenqiang Chen <zhenqiang.chen@linaro.org>
13 * shrink-wrap.h: Update comment.
14 * shrink-wrap.c: Update comment.
15 (next_block_for_reg): Rename to live_edge_for_reg.
16 (live_edge_for_reg): Allow live_edge->dest has two predecessors.
17 (move_insn_for_shrink_wrap): Split live_edge.
18 (prepre_shrink_wrap): One more parameter for move_insn_for_shrink_wrap.
20 2014-05-14 Eric Botcazou <ebotcazou@adacore.com>
22 * config/sparc/sparc-protos.h (sparc_absnegfloat_split_legitimate):
24 * config/sparc/sparc.c (sparc_absnegfloat_split_legitimate): Likewise.
25 * config/sparc/sparc.md (fptype_ut699): New attribute.
26 (in_branch_delay): Return false if -mfix-ut699 is specified and
27 fptype_ut699 is set to single.
28 (truncdfsf2): Add fptype_ut699 attribute.
29 (fix_truncdfsi2): Likewise.
30 (floatsisf2): Change fptype attribute.
31 (fix_truncsfsi2): Likewise.
32 (negtf2_notv9): Delete.
33 (negtf2_v9): Likewise.
34 (negtf2_hq): New instruction.
35 (negtf2): New instruction and splitter.
36 (negdf2_notv9): Rewrite.
37 (abstf2_notv9): Delete.
38 (abstf2_hq_v9): Likewise.
39 (abstf2_v9): Likewise.
40 (abstf2_hq): New instruction.
41 (abstf2): New instruction and splitter.
42 (absdf2_notv9): Rewrite.
44 2014-05-14 Cary Coutant <ccoutant@google.com>
47 * opts.c (common_handle_option): Don't special-case "-g".
48 (set_debug_level): Default to at least level 2 with "-g".
50 2014-05-14 DJ Delorie <dj@redhat.com>
52 * config/msp430/msp430.c (msp430_builtin): Add
53 MSP430_BUILTIN_DELAY_CYCLES.
54 (msp430_init_builtins): Register void __delay_cycles(long long).
55 (msp430_builtin_decl): Add it.
56 (cg_magic_constant): New.
57 (msp430_expand_delay_cycles): New.
58 (msp430_expand_builtin): Call it.
59 (msp430_print_operand_raw): Change integer printing from "int" to
61 * config/msp430/msp430.md (define_constants): Add delay_cycles tags.
62 (delay_cycles_start): New.
63 (delay_cycles_end): New.
64 (delay_cycles_32): New.
65 (delay_cycles_32x): New.
66 (delay_cycles_16): New.
67 (delay_cycles_16x): New.
68 (delay_cycles_2): New.
69 (delay_cycles_1): New.
70 * doc/extend.texi: Document __delay_cycles().
72 2014-05-14 Sandra Loosemore <sandra@codesourcery.com>
74 * config/nios2/nios2.md (nios2_cbranch): Fix paste-o in
75 length attribute computation.
77 2014-05-14 Richard Sandiford <rdsandiford@googlemail.com>
80 * print-rtl.c (print_rtx): Suppress uids if flag_dump_unnumbered.
82 2014-05-14 Richard Sandiford <r.sandiford@uk.ibm.com>
85 * config/sparc/sparc.md: Fix types of low and high in DI constant
86 splitter. Use gen_int_mode in some other splitters.
88 2014-05-14 Martin Jambor <mjambor@suse.cz>
91 * ipa-prop.c (ipa_modify_formal_parameters): Reset DECL_LANG_SPECIFIC.
93 2014-05-14 James Norris <jnorris@codesourcery.com>
95 * omp-low.c (expand_parallel_call): Remove shadow variable.
96 (expand_omp_taskreg): Likewise.
98 2014-05-14 Ilya Tocar <ilya.tocar@intel.com>
100 * common/config/i386/i386-common.c
101 (OPTION_MASK_ISA_CLFLUSHOPT_SET): Define.
102 (OPTION_MASK_ISA_XSAVES_SET): Ditto.
103 (OPTION_MASK_ISA_XSAVEC_SET): Ditto.
104 (OPTION_MASK_ISA_CLFLUSHOPT_UNSET): Ditto.
105 (OPTION_MASK_ISA_XSAVES_UNSET): Ditto.
106 (OPTION_MASK_ISA_XSAVEC_UNSET): Ditto.
107 (ix86_handle_option): Handle OPT_mxsavec, OPT_mxsaves,
109 * config.gcc (i[34567]86-*-*): Add clflushoptintrin.h,
110 xsavecintrin.h, xsavesintrin.h.
112 * config/i386/clflushoptintrin.h: New.
113 * config/i386/xsavecintrin.h: Ditto.
114 * config/i386/xsavesintrin.h: Ditto.
115 * config/i386/cpuid.h (bit_CLFLUSHOPT): Define.
118 * config/i386/driver-i386.c (host_detect_local_cpu): Handle
119 -mclflushopt, -mxsavec, -mxsaves, -mno-xsaves, -mno-xsavec,
121 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
122 OPTION_MASK_ISA_CLFLUSHOPT, OPTION_MASK_ISA_XSAVEC,
123 OPTION_MASK_ISA_XSAVES.
124 * config/i386/i386.c (ix86_target_string): Handle -mclflushopt,
126 (PTA_CLFLUSHOPT) Define.
129 (ix86_option_override_internal): Handle new options.
130 (ix86_valid_target_attribute_inner_p): Ditto.
131 (ix86_builtins): Add IX86_BUILTIN_XSAVEC, IX86_BUILTIN_XSAVEC64,
132 IX86_BUILTIN_XSAVES, IX86_BUILTIN_XRSTORS, IX86_BUILTIN_XSAVES64,
133 IX86_BUILTIN_XRSTORS64, IX86_BUILTIN_CLFLUSHOPT.
134 (bdesc_special_args): Add __builtin_ia32_xsaves, __builtin_ia32_xrstors,
135 __builtin_ia32_xsavec, __builtin_ia32_xsaves64, __builtin_ia32_xrstors64,
136 __builtin_ia32_xsavec64.
137 (ix86_init_mmx_sse_builtins): Add __builtin_ia32_clflushopt.
138 (ix86_expand_builtin): Handle new builtins.
139 * config/i386/i386.h (TARGET_CLFLUSHOPT) Define.
140 (TARGET_CLFLUSHOPT_P): Ditto.
141 (TARGET_XSAVEC): Ditto.
142 (TARGET_XSAVEC_P): Ditto.
143 (TARGET_XSAVES): Ditto.
144 (TARGET_XSAVES_P): Ditto.
145 * config/i386/i386.md (ANY_XSAVE): Add UNSPECV_XSAVEC, UNSPECV_XSAVES.
146 (ANY_XSAVE64)" Add UNSPECV_XSAVEC64, UNSPECV_XSAVES64.
147 (attr xsave): Add xsavec, xsavec64, xsaves, xsaves64.
149 (ANY_XRSTOR64): Ditto.
151 (xrstor): Change into <xrstor>.
152 (xrstor_rex64): Change into <xrstor>_rex64.
153 (xrstor64): Change into <xrstor>64
155 * config/i386/i386.opt (mclflushopt): New.
158 * config/i386/x86intrin.h: Add clflushoptintrin.h, xsavesintrin.h,
160 * doc/invoke.texi: Document new options.
162 2014-05-14 Andrey Belevantsev <abel@ispras.ru>
164 PR rtl-optimization/60866
165 * sel-sched-ir (sel_init_new_insn): New parameter old_seqno.
166 Default it to -1. Pass it down to init_simplejump_data.
167 (init_simplejump_data): New parameter old_seqno. Pass it down
168 to get_seqno_for_a_jump.
169 (get_seqno_for_a_jump): New parameter old_seqno. Use it for
170 initializing new jump seqno as a last resort. Add comment.
171 (sel_redirect_edge_and_branch): Save old seqno of the conditional
172 jump and pass it down to sel_init_new_insn.
173 (sel_redirect_edge_and_branch_force): Likewise.
175 2014-05-14 Georg-Johann Lay <avr@gjlay.de>
177 * config/avr/avr.h (REG_CLASS_CONTENTS): Use unsigned suffix for
178 shifted values to avoid build warning.
180 2014-05-14 Eric Botcazou <ebotcazou@adacore.com>
182 * cfgcleanup.c (try_forward_edges): Use location_t for locations.
183 * cfgrtl.c (rtl_merge_blocks): Fix comment.
184 (cfg_layout_merge_blocks): Likewise.
185 * except.c (emit_to_new_bb_before): Remove prev_bb local variable.
187 2014-05-14 Andrey Belevantsev <abel@ispras.ru>
189 PR rtl-optimization/60901
190 * config/i386/i386.c (ix86_dependencies_evaluation_hook): Check that
191 bb predecessor belongs to the same scheduling region. Adjust comment.
193 2014-05-13 Peter Bergner <bergner@vnet.ibm.com>
195 * doc/sourcebuild.texi: (dfp_hw): Document.
196 (p8vector_hw): Likewise.
197 (powerpc_eabi_ok): Likewise.
198 (powerpc_elfv2): Likewise.
199 (powerpc_htm_ok): Likewise.
200 (ppc_recip_hw): Likewise.
203 2014-05-13 Cary Coutant <ccoutant@google.com>
205 * opts.c (finish_options): Use -ggnu-pubnames with -gsplit-dwarf.
207 2014-05-13 David Malcolm <dmalcolm@redhat.com>
209 * gengtype-parse.c (require3): Eliminate in favor of...
211 (require_template_declaration): Update to support optional single *
214 * gengtype.c (get_ultimate_base_class): Add a non-const overload.
215 (create_user_defined_type): Handle a single level of explicit
216 pointerness within template arguments.
217 (struct write_types_data): Add field "kind".
218 (filter_type_name): Handle "*" character.
219 (write_user_func_for_structure_ptr): Require a write_types_data
220 rather than just a prefix string, so that we can look up the kind
221 of the wtd and use it as an index into wrote_user_func_for_ptr,
222 ensuring that such functions are written at most once. Support
223 subclasses by invoking the marking function of the ultimate base
225 (write_user_func_for_structure_body): Require a write_types_data
226 rather than just a prefix string, so that we can pass this to
227 write_user_func_for_structure_ptr.
228 (write_func_for_structure): Likewise.
229 (ggc_wtd): Add initializer of new "kind" field.
232 * gengtype.h (enum write_types_kinds): New.
233 (struct type): Add field wrote_user_func_for_ptr to the "s"
236 2014-05-13 Richard Sandiford <r.sandiford@uk.ibm.com>
238 * fold-const.c (optimize_bit_field_compare): Use wi:: operations
239 instead of const_binop.
240 (fold_binary_loc): Likewise.
242 2014-05-13 Richard Sandiford <r.sandiford@uk.ibm.com>
244 * tree-dfa.h (get_addr_base_and_unit_offset_1): Update array index
245 calculation to match get_ref_base_and_extent.
247 2014-05-13 Catherine Moore <clm@codesourcery.com>
248 Sandra Loosemore <sandra@codesourcery.com>
250 * configure.ac: Fix assembly for explicit JALR relocation check.
251 * configure: Regenerate.
253 2014-05-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
255 * config/arm/arm.c (neon_itype): Remove NEON_RESULTPAIR.
256 (arm_init_neon_builtins): Remove handling of NEON_RESULTPAIR.
257 Remove associated type declarations and initialisations.
258 (arm_expand_neon_builtin): Likewise.
259 (neon_emit_pair_result_insn): Delete.
260 * config/arm/arm_neon_builtins (vtrn, vzip, vuzp): Delete.
261 * config/arm/neon.md (neon_vtrn<mode>): Delete.
262 (neon_vzip<mode>): Likewise.
263 (neon_vuzp<mode>): Likewise.
265 2014-05-13 Richard Biener <rguenther@suse.de>
268 * tree-inline.c (remap_gimple_stmt): Clear tail call flag,
269 it needs revisiting whether the call still may be tail-called.
271 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
273 * rtl.def (SYMBOL_REF): Remove middle "0" field.
274 * rtl.h (block_symbol): Reduce number of fields to 2.
275 (rtx_def): Add u2.symbol_ref_flags.
276 (SYMBOL_REF_FLAGS): Use it.
277 (SYMBOL_REF_DATA, SET_SYMBOL_REF_DECL, SYMBOL_REF_DECL)
278 (SET_SYMBOL_REF_CONSTANT, SYMBOL_REF_CONSTANT): Lower index.
279 * gengtype.c (adjust_field_rtx_def): Remove SYMBOL_REF_FLAGS handling.
280 Lower index of SYMBOL_REF_DATA.
281 * print-rtl.c (print_rtx): Lower index for SYMBOL_REF_DATA.
282 Print SYMBOL_REF_FLAGS at the same time.
283 * genattrtab.c (attr_rtx_1): Only initialize 1 "0" SYMBOL_REF field.
285 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
287 * rtl.def (VAR_LOCATION): Remove "i" field.
288 * rtl.h (rtx_def): Add u2.var_location_status.
289 (PAT_VAR_LOCATION_STATUS): Use it.
290 (gen_rtx_VAR_LOCATION): Declare.
291 * gengenrtl.c (excluded_rtx): Add VAR_LOCATION.
292 * emit-rtl.c (gen_rtx_VAR_LOCATION): New function.
293 * var-tracking.c (emit_note_insn_var_location): Remove casts.
295 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
297 * rtl.def (scratch): Fix outdated comment and remove "0" field.
298 * gengtype.c (adjust_field_rtx_def): Update accordingly.
300 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
302 * rtl.def (DEBUG_INSN, INSN, JUMP_INSN, CALL_INSN, JUMP_TABLE_DATA)
303 (BARRIER, CODE_LABEL, NOTE): Remove first "i" field.
304 * rtl.h (rtx_def): Add insn_uid to u2 field.
305 (RTX_FLAG_CHECK8): Delete in favor of...
306 (RTL_INSN_CHAIN_FLAG_CHECK): ...this new macro.
307 (INSN_DELETED_P): Update accordingly.
308 (INSN_UID): Use u2.insn_uid.
309 (INSN_CHAIN_CODE_P): Define.
310 (PREV_INSN, NEXT_INSN, BLOCK_FOR_INSN, PATTERN, INSN_LOCATION)
311 (INSN_CODE, REG_NOTES, CALL_INSN_FUNCTION_USAGE, CODE_LABEL_NUMBER)
312 (NOTE_DATA, NOTE_DELETED_LABEL_NAME, NOTE_BLOCK, NOTE_EH_HANDLER)
313 (NOTE_BASIC_BLOCK, NOTE_VAR_LOCATION, NOTE_CFI, NOTE_LABEL_NUMBER)
314 (NOTE_KIND, LABEL_NAME, LABEL_NUSES, JUMP_LABEL, LABEL_REFS): Lower
316 * print-rtl.c (print_rtx): Print INSN_UIDs before the main loop.
317 Update indices for insn-chain rtxes.
318 * gengtype.c (gen_rtx_next): Adjust test for insn-chain rtxes.
319 (adjust_field_rtx_def): Lower '0' indices for all insn-chain rtxes.
320 * emit-rtl.c (gen_label_rtx): Update gen_rtx_LABEL call.
321 * caller-save.c (init_caller_save): Update gen_rtx_INSN calls.
322 * combine.c (try_combine): Likewise.
323 * ira.c (setup_prohibited_mode_move_regs): Likewise.
325 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
327 * rtl.def (REG): Remove middle field.
328 * rtl.h (rtx_def): Add orignal_regno to u2.
329 (ORIGINAL_REGNO): Use it instead of field 1.
330 (REG_ATTRS): Lower field index accordingly.
331 * gengtype.c (adjust_field_rtx_def): Remove handling of
332 ORIGINAL_REGNO. Move REG_ATTRS index down.
333 * print-rtl.c (print_rtx): Move ORIGINAL_REGNO handling to the
334 code that prints the REGNO.
336 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
338 * print-rtl.c (print_rtx): Guard whole '0' block with ifndef
341 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
343 * rtl.h (rtx_def): Mark u2 as GTY ((skip)).
345 2014-05-13 Bin Cheng <bin.cheng@arm.com>
347 * tree-ssa-loop-ivopts.c (contain_complex_addr_expr): New.
348 (alloc_iv): Lower base expressions containing ADDR_EXPR.
350 2014-05-13 Ian Bolton <ian.bolton@arm.com>
352 * config/aarch64/aarch64-protos.h
353 (aarch64_hard_regno_caller_save_mode): New prototype.
354 * config/aarch64/aarch64.c (aarch64_hard_regno_caller_save_mode):
356 * config/aarch64/aarch64.h (HARD_REGNO_CALLER_SAVE_MODE): New macro.
358 2014-05-13 Christian Bruel <christian.bruel@st.com>
360 * target.def (mode_switching): New hook vector.
361 (mode_emit, mode_needed, mode_after, mode_entry): New hooks.
362 (mode_exit, modepriority_to_mode): Likewise.
363 * mode-switching.c (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Hookify.
364 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
365 * target.h: Include tm.h and hard-reg-set.h.
366 * doc/tm.texi.in (EMIT_MODE_SET, MODE_NEEDED, MODE_AFTER, MODE_ENTRY)
367 (MODE_EXIT, MODE_PRIORITY_TO_MODE): Delete and hookify.
368 * doc/tm.texi Regenerate.
369 * config/sh/sh.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
370 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
371 * config/sh/sh.c (sh_emit_mode_set, sh_mode_priority): Hookify.
372 (sh_mode_needed, sh_mode_after, sh_mode_entry, sh_mode_exit): Likewise.
373 * config/i386/i386.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
374 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
375 * config/i386/i386-protos.h (ix86_mode_needed, ix86_mode_after)
376 (ix86_mode_entrym, ix86_emit_mode_set): Remove external declaration.
377 * config/i386/i386.c (ix86_mode_needed, ix86_mode_after, ix86_mode_exit,
378 (ix86_mode_entry, ix86_mode_priority, ix86_emit_mode_set): Hookify.
379 * config/epiphany/epiphany.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY):
381 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
382 * config/epiphany/epiphany-protos.h (epiphany_mode_needed)
383 (emit_set_fp_mode, epiphany_mode_entry_exit, epiphany_mode_after)
384 (epiphany_mode_priority_to_mode): Remove declaration.
385 * config/epiphany/epiphany.c (emit_set_fp_mode): Hookify.
386 (epiphany_mode_needed, epiphany_mode_priority_to_mode): Likewise.
387 (epiphany_mode_entry, epiphany_mode_exit, epiphany_mode_after):
389 (epiphany_mode_priority_to_mode): Change priority type. Hookify.
390 (epiphany_mode_needed, epiphany_mode_entry_exit): Hookify.
391 (epiphany_mode_after, epiphany_mode_entry, emit_set_fp_mode): Hookify.
393 2014-05-13 Jakub Jelinek <jakub@redhat.com>
396 * config/i386/i386.c (ix86_expand_set_or_movmem): If count_exp
397 is const0_rtx, return immediately. Don't test count == 0 when
400 2014-05-13 Zhenqiang Chen <zhenqiang.chen@linaro.org>
402 * Makefile.in: add shrink-wrap.o.
403 * config/i386/i386.c: include "shrink-wrap.h"
404 * function.c: Likewise.
405 (requires_stack_frame_p, next_block_for_reg,
406 move_insn_for_shrink_wrap, prepare_shrink_wrap,
407 dup_block_and_redirect): Move to shrink-wrap.c
408 (thread_prologue_and_epilogue_insns): Extract three code segments
409 as functions in shrink-wrap.c
410 * function.h: Move #ifdef HAVE_simple_return ... #endif block to
412 * shrink-wrap.c: New file.
413 * shrink-wrap.h: New file.
415 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
417 * doc/extend.texi: Reflect current numbers of pragmas. Remove
418 reference to Solaris.
420 2014-05-12 Mike Stump <mikestump@comcast.net>
423 * genattrtab.c (filename): Add.
424 (convert_set_attr_alternative): Improve error message.
425 (check_defs): Restore read_md_filename for error messages.
426 (gen_insn): Save filename.
428 2014-05-12 Dimitris Papavasiliou <dpapavas@gmail.com>
430 * doc/invoke.texi: Document new switches -Wno-shadow-ivar,
431 -fno-local-ivars and -fivar-visibility.
432 * c-family/c.opt: Make -Wshadow also implicitly enable
435 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
437 * doc/tm.texi: Remove reference to deleted macro.
438 * doc/tm.texi.in: Likewise.
440 2014-05-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
443 * config/avr/avr.c (avr_out_store_psi): Use correct constant
446 2014-05-12 Georg-Johann Lay <avr@gjlay.de>
448 * config/arm/arm.h (License): Add GCC Runtime Library Exception.
449 * config/arm/aout.h (License): Same.
450 * config/arm/bpabi.h (License): Same.
451 * config/arm/elf.h (License): Same.
452 * config/arm/linux-elf.h (License): Same.
453 * config/arm/linux-gas.h (License): Same.
454 * config/arm/netbsd-elf.h (License): Same.
455 * config/arm/uclinux-eabi.h (License): Same.
456 * config/arm/uclinux-elf.h (License): Same.
457 * config/arm/vxworks.h (License): Same.
459 2014-05-11 Jakub Jelinek <jakub@redhat.com>
461 * tree.h (OMP_CLAUSE_LINEAR_STMT): Define.
462 * tree.c (omp_clause_num_ops): Increase OMP_CLAUSE_LINEAR
463 number of operands to 3.
464 (walk_tree_1): Walk all operands of OMP_CLAUSE_LINEAR.
465 * tree-nested.c (convert_nonlocal_omp_clauses,
466 convert_local_omp_clauses): Handle OMP_CLAUSE_DEPEND.
467 * gimplify.c (gimplify_scan_omp_clauses): Handle
468 OMP_CLAUSE_LINEAR_STMT.
469 * omp-low.c (lower_rec_input_clauses): Fix typo.
470 (maybe_add_implicit_barrier_cancel, lower_omp_1): Add
471 cast between Fortran boolean_type_node and C _Bool if
474 2014-05-11 Richard Sandiford <rdsandiford@googlemail.com>
476 PR tree-optimization/61136
477 * wide-int.h (multiple_of_p): Define a version that doesn't return
479 * fold-const.c (extract_muldiv_1): Use wi::multiple_of_p instead of an
480 integer_zerop/const_binop pair.
481 (multiple_of_p): Likewise, converting both operands to widest_int
484 2014-05-09 Teresa Johnson <tejohnson@google.com>
486 * cgraphunit.c (analyze_functions): Use correct dump file.
488 2014-05-09 Florian Weimer <fweimer@redhat.com>
490 * cfgexpand.c (stack_protect_decl_p): New function, extracted from
492 (stack_protect_return_slot_p): New function.
493 (expand_used_vars): Call stack_protect_decl_p and
494 stack_protect_return_slot_p for -fstack-protector-strong.
496 2014-05-09 Kenneth Zadeck <zadeck@naturalbridge.com>
499 * fold-const.c (fold_binary_loc): Changed width of mask.
501 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
503 * config/avr/avr-fixed.md (round<mode>3): Use -1U instead of -1 in
504 unsigned int initializers for regno_in, regno_out.
506 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
509 * config/avr/avr.md (cc): Add new attribute set_vzn.
510 (addqi3, addqq3, adduqq3, subqi3, subqq3, subuqq3, negqi2) [cc]:
511 Set cc insn attribute to set_vzn instead of set_zn for alternatives
512 with INC, DEC or NEG.
513 * config/avr/avr.c (avr_notice_update_cc): Handle SET_VZN.
514 (avr_out_plus_1): ADIW sets cc0 to CC_SET_CZN.
515 INC, DEC and ADD+ADC set cc0 to CC_CLOBBER.
517 2014-05-09 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
520 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
522 * wide-int.cc (UTItype): Define.
523 (UDWtype): Define for appropriate W_TYPE_SIZE.
525 2014-05-09 Richard Biener <rguenther@suse.de>
527 * Makefile.in (GTFILES): Remove tree-ssa-propagate.c.
528 * tree-ssa-propagate.c: Do not include gt-tree-ssa-propagate.h.
529 (interesting_ssa_edges, varying_ssa_edges): Move out of GC space.
530 (add_ssa_edge, process_ssa_edge_worklist, ssa_prop_init,
531 ssa_propagate): Adjust.
533 2014-05-08 Jeff Law <law@redhat.com>
535 PR tree-optimization/61009
536 * tree-ssa-threadedge.c (thread_through_normal_block): Return a
537 tri-state rather than a boolean. When a block is too big to
538 thread through, inform caller via negative return value.
539 (thread_across_edge): If a block was too big for normal threading,
540 then it's too big for a joiner too, so remove temporary equivalences
541 and return immediately.
543 2014-05-08 Manuel López-Ibáñez <manu@gcc.gnu.org>
544 Matthias Klose <doko@ubuntu.com>
547 * optc-gen.awk: Fix option handling for -Wunused-parameter.
549 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
552 * config/i386/i386.c (PTA_HASWELL): Remove PTA_RTM.
554 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
557 * config/alpha/alpha.c: Include gimple-iterator.h.
558 (alpha_gimple_fold_builtin): New function. Move
559 ALPHA_BUILTIN_UMULH folding from ...
560 (alpha_fold_builtin): ... here.
561 (TARGET_GIMPLE_FOLD_BUILTIN): New define.
563 2014-05-08 Wei Mi <wmi@google.com>
566 * config/i386/i386.c (ix86_compute_frame_layout): Update
567 preferred_stack_boundary for call, expanded from tls descriptor.
568 * config/i386/i386.md (*tls_global_dynamic_32_gnu): Update RTX
569 to depend on SP register.
570 (*tls_local_dynamic_base_32_gnu): Ditto.
571 (*tls_local_dynamic_32_once): Ditto.
572 (tls_global_dynamic_64_<mode>): Set
573 ix86_tls_descriptor_calls_expanded_in_cfun.
574 (tls_local_dynamic_base_64_<mode>): Ditto.
575 (tls_global_dynamic_32): Set
576 ix86_tls_descriptor_calls_expanded_in_cfun. Update RTX
577 to depend on SP register.
578 (tls_local_dynamic_base_32): Ditto.
580 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
582 * config/arm/arm_neon.h: Update comment.
583 * config/arm/neon-docgen.ml: Delete.
584 * config/arm/neon-gen.ml: Delete.
585 * doc/arm-neon-intrinsics.texi: Update comment.
587 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
589 * config/arm/arm_neon_builtins.def (vadd, vsub): Only define the v2sf
591 (vand, vorr, veor, vorn, vbic): Remove.
592 * config/arm/neon.md (neon_vadd, neon_vsub, neon_vadd_unspec): Adjust
594 (neon_vsub_unspec): Likewise.
595 (neon_vorr, neon_vand, neon_vbic, neon_veor, neon_vorn): Remove.
597 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
599 * config/arm/arm_neon.h (vadd_s8): GNU C implementation
600 (vadd_s16): Likewise.
601 (vadd_s32): Likewise.
602 (vadd_f32): Likewise.
604 (vadd_u16): Likewise.
605 (vadd_u32): Likewise.
606 (vadd_s64): Likewise.
607 (vadd_u64): Likewise.
608 (vaddq_s8): Likewise.
609 (vaddq_s16): Likewise.
610 (vaddq_s32): Likewise.
611 (vaddq_s64): Likewise.
612 (vaddq_f32): Likewise.
613 (vaddq_u8): Likewise.
614 (vaddq_u16): Likewise.
615 (vaddq_u32): Likewise.
616 (vaddq_u64): Likewise.
618 (vmul_s16): Likewise.
619 (vmul_s32): Likewise.
620 (vmul_f32): Likewise.
622 (vmul_u16): Likewise.
623 (vmul_u32): Likewise.
625 (vmulq_s8): Likewise.
626 (vmulq_s16): Likewise.
627 (vmulq_s32): Likewise.
628 (vmulq_f32): Likewise.
629 (vmulq_u8): Likewise.
630 (vmulq_u16): Likewise.
631 (vmulq_u32): Likewise.
633 (vsub_s16): Likewise.
634 (vsub_s32): Likewise.
635 (vsub_f32): Likewise.
637 (vsub_u16): Likewise.
638 (vsub_u32): Likewise.
639 (vsub_s64): Likewise.
640 (vsub_u64): Likewise.
641 (vsubq_s8): Likewise.
642 (vsubq_s16): Likewise.
643 (vsubq_s32): Likewise.
644 (vsubq_s64): Likewise.
645 (vsubq_f32): Likewise.
646 (vsubq_u8): Likewise.
647 (vsubq_u16): Likewise.
648 (vsubq_u32): Likewise.
649 (vsubq_u64): Likewise.
651 (vand_s16): Likewise.
652 (vand_s32): Likewise.
654 (vand_u16): Likewise.
655 (vand_u32): Likewise.
656 (vand_s64): Likewise.
657 (vand_u64): Likewise.
658 (vandq_s8): Likewise.
659 (vandq_s16): Likewise.
660 (vandq_s32): Likewise.
661 (vandq_s64): Likewise.
662 (vandq_u8): Likewise.
663 (vandq_u16): Likewise.
664 (vandq_u32): Likewise.
665 (vandq_u64): Likewise.
667 (vorr_s16): Likewise.
668 (vorr_s32): Likewise.
670 (vorr_u16): Likewise.
671 (vorr_u32): Likewise.
672 (vorr_s64): Likewise.
673 (vorr_u64): Likewise.
674 (vorrq_s8): Likewise.
675 (vorrq_s16): Likewise.
676 (vorrq_s32): Likewise.
677 (vorrq_s64): Likewise.
678 (vorrq_u8): Likewise.
679 (vorrq_u16): Likewise.
680 (vorrq_u32): Likewise.
681 (vorrq_u64): Likewise.
683 (veor_s16): Likewise.
684 (veor_s32): Likewise.
686 (veor_u16): Likewise.
687 (veor_u32): Likewise.
688 (veor_s64): Likewise.
689 (veor_u64): Likewise.
690 (veorq_s8): Likewise.
691 (veorq_s16): Likewise.
692 (veorq_s32): Likewise.
693 (veorq_s64): Likewise.
694 (veorq_u8): Likewise.
695 (veorq_u16): Likewise.
696 (veorq_u32): Likewise.
697 (veorq_u64): Likewise.
699 (vbic_s16): Likewise.
700 (vbic_s32): Likewise.
702 (vbic_u16): Likewise.
703 (vbic_u32): Likewise.
704 (vbic_s64): Likewise.
705 (vbic_u64): Likewise.
706 (vbicq_s8): Likewise.
707 (vbicq_s16): Likewise.
708 (vbicq_s32): Likewise.
709 (vbicq_s64): Likewise.
710 (vbicq_u8): Likewise.
711 (vbicq_u16): Likewise.
712 (vbicq_u32): Likewise.
713 (vbicq_u64): Likewise.
715 (vorn_s16): Likewise.
716 (vorn_s32): Likewise.
718 (vorn_u16): Likewise.
719 (vorn_u32): Likewise.
720 (vorn_s64): Likewise.
721 (vorn_u64): Likewise.
722 (vornq_s8): Likewise.
723 (vornq_s16): Likewise.
724 (vornq_s32): Likewise.
725 (vornq_s64): Likewise.
726 (vornq_u8): Likewise.
727 (vornq_u16): Likewise.
728 (vornq_u32): Likewise.
729 (vornq_u64): Likewise.
731 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
733 * wide-int.cc (UTItype): Define.
734 (UDWtype): Define for appropriate W_TYPE_SIZE.
736 2014-05-08 Marc Glisse <marc.glisse@inria.fr>
738 PR tree-optimization/59100
739 * tree-ssa-phiopt.c: Include tree-inline.h.
740 (neutral_element_p, absorbing_element_p): New functions.
741 (value_replacement): Handle conditional binary operations with a
742 neutral or absorbing element.
744 2014-05-08 Richard Biener <rguenther@suse.de>
746 * tree-ssa-sccvn.c (vn_get_expr_for): Valueize operands before
747 folding the expression.
748 (valueize_expr): Remove.
749 (visit_reference_op_load): Do not valueize the result of
751 (simplify_binary_expression): Likewise.
752 (simplify_unary_expression): Likewise.
754 2014-05-08 Richard Biener <rguenther@suse.de>
756 * gimplify.c (gimplify_call_expr): Use saved fnptrtype for
757 looking at TYPE_ARG_TYPES.
759 2014-05-08 Richard Biener <rguenther@suse.de>
761 * gimple-fold.c (gimple_fold_stmt_to_constant_1): Remove
762 pointer propagation special-case.
764 2014-05-08 Bin Cheng <bin.cheng@arm.com>
766 * tree-affine.c (tree_to_aff_combination): Handle MEM_REF for
767 core part of address expressions.
769 2014-05-08 Alan Modra <amodra@gmail.com>
772 * config/rs6000/rs6000.c (expand_block_move): Allow 64-bit
773 loads and stores when -mno-strict-align at any alignment.
774 (expand_block_clear): Similarly. Also correct calculation of
777 2014-05-07 Thomas Preud'homme <thomas.preudhomme@arm.com>
780 * tree-complex.c (expand_complex_move): Keep line info when expanding
782 * tree-ssa-uninit.c (warn_uninit): New argument. Ignore assignment
783 of complex expression. Use new argument to display correct location
784 for values coming from phi statement.
785 (warn_uninitialized_vars): Adapt to new signature of warn_uninit.
786 (warn_uninitialized_phi): Pass location of phi argument to
788 * tree-ssa.c (ssa_undefined_value_p): For SSA_NAME initialized by a
789 COMPLEX_EXPR, recurse on each part of the COMPLEX_EXPR.
791 2014-05-07 Segher Boessenkool <segher@kernel.crashing.org>
793 * config/rs6000/predicates.md (indexed_address_mem): New.
794 * config/rs6000/rs6000.md (type): Remove load_ext, load_ext_u,
795 load_ext_ux, load_ux, load_u, store_ux, store_u, fpload_ux, fpload_u,
796 fpstore_ux, fpstore_u.
797 (sign_extend, indexed, update): New.
798 (cell_micro): Adjust.
799 (*zero_extend<mode>di2_internal1, *zero_extendsidi2_lfiwzx,
800 *extendsidi2_lfiwax, *extendsidi2_nocell, *extendsfdf2_fpr,
801 *movsi_internal1, *movsi_internal1_single, *movhi_internal,
802 *movqi_internal, *movcc_internal1, mov<mode>_hardfloat,
803 *mov<mode>_softfloat, *mov<mode>_hardfloat32, *mov<mode>_hardfloat64,
804 *mov<mode>_softfloat64, *movdi_internal32, *movdi_internal64,
805 *mov<mode>_string, *ldmsi8, *ldmsi7, *ldmsi6, *ldmsi5, *ldmsi4,
806 *ldmsi3, *stmsi8, *stmsi7, *stmsi6, *stmsi5, *stmsi4, *stmsi3,
807 *movdi_update1, movdi_<mode>_update, movdi_<mode>_update_stack,
808 *movsi_update1, *movsi_update2, movsi_update, movsi_update_stack,
809 *movhi_update1, *movhi_update2, *movhi_update3, *movhi_update4,
810 *movqi_update1, *movqi_update2, *movqi_update3, *movsf_update1,
811 *movsf_update2, *movsf_update3, *movsf_update4, *movdf_update1,
812 *movdf_update2, load_toc_aix_si, load_toc_aix_di, probe_stack_<mode>,
813 *stmw, *lmw, as well as 10 anonymous patterns): Adjust.
815 * config/rs6000/dfp.md (movsd_store, movsd_load): Adjust.
816 * config/rs6000/vsx.md (*vsx_movti_32bit, *vsx_extract_<mode>_load,
817 *vsx_extract_<mode>_store): Adjust.
818 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_microcoded_insn,
819 is_cracked_insn, insn_must_be_first_in_group,
820 insn_must_be_last_in_group): Adjust.
822 * config/rs6000/40x.md (ppc403-load, ppc403-store, ppc405-float):
824 * config/rs6000/440.md (ppc440-load, ppc440-store, ppc440-fpload,
825 ppc440-fpstore): Adjust.
826 * config/rs6000/476.md (ppc476-load, ppc476-store, ppc476-fpload,
827 ppc476-fpstore): Adjust.
828 * config/rs6000/601.md (ppc601-load, ppc601-store, ppc601-fpload,
829 ppc601-fpstore): Adjust.
830 * config/rs6000/603.md (ppc603-load, ppc603-store, ppc603-fpload):
832 * config/rs6000/6xx.md (ppc604-load, ppc604-store, ppc604-fpload):
834 * config/rs6000/7450.md (ppc7450-load, ppc7450-store, ppc7450-fpload,
835 ppc7450-fpstore): Adjust.
836 * config/rs6000/7xx.md (ppc750-load, ppc750-store): Adjust.
837 * config/rs6000/8540.md (ppc8540_load, ppc8540_store): Adjust.
838 * config/rs6000/a2.md (ppca2-load, ppca2-fp-load, ppca2-fp-store):
840 * config/rs6000/cell.md (cell-load, cell-load-ux, cell-load-ext,
841 cell-fpload, cell-fpload-update, cell-store, cell-store-update,
842 cell-fpstore, cell-fpstore-update): Adjust.
843 * config/rs6000/e300c2c3.md (ppce300c3_load, ppce300c3_fpload,
844 ppce300c3_store, ppce300c3_fpstore): Adjust.
845 * config/rs6000/e500mc.md (e500mc_load, e500mc_fpload, e500mc_store,
846 e500mc_fpstore): Adjust.
847 * config/rs6000/e500mc64.md (e500mc64_load, e500mc64_fpload,
848 e500mc64_store, e500mc64_fpstore): Adjust.
849 * config/rs6000/e5500.md (e5500_load, e5500_fpload, e5500_store,
850 e5500_fpstore): Adjust.
851 * config/rs6000/e6500.md (e6500_load, e6500_fpload, e6500_store,
852 e6500_fpstore): Adjust.
853 * config/rs6000/mpc.md (mpccore-load, mpccore-store, mpccore-fpload):
855 * config/rs6000/power4.md (power4-load, power4-load-ext,
856 power4-load-ext-update, power4-load-ext-update-indexed,
857 power4-load-update-indexed, power4-load-update, power4-fpload,
858 power4-fpload-update, power4-store, power4-store-update,
859 power4-store-update-indexed, power4-fpstore, power4-fpstore-update):
861 * config/rs6000/power5.md (power5-load, power5-load-ext,
862 power5-load-ext-update, power5-load-ext-update-indexed,
863 power5-load-update-indexed, power5-load-update, power5-fpload,
864 power5-fpload-update, power5-store, power5-store-update,
865 power5-store-update-indexed, power5-fpstore, power5-fpstore-update):
867 * config/rs6000/power6.md (power6-load, power6-load-ext,
868 power6-load-update, power6-load-update-indexed,
869 power6-load-ext-update, power6-load-ext-update-indexed, power6-fpload,
870 power6-fpload-update, power6-store, power6-store-update,
871 power6-store-update-indexed, power6-fpstore, power6-fpstore-update):
873 * config/rs6000/power7.md (power7-load, power7-load-ext,
874 power7-load-update, power7-load-update-indexed,
875 power7-load-ext-update, power7-load-ext-update-indexed, power7-fpload,
876 power7-fpload-update, power7-store, power7-store-update,
877 power7-store-update-indexed, power7-fpstore, power7-fpstore-update):
879 * config/rs6000/power8.md (power8-load, power8-load-update,
880 power8-load-ext, power8-load-ext-update, power8-fpload,
881 power8-fpload-update, power8-store, power8-store-update-indexed,
882 power8-fpstore, power8-fpstore-update): Adjust.
883 * config/rs6000/rs64.md (rs64a-load, rs64a-store, rs64a-fpload):
885 * config/rs6000/titan.md (titan_lsu_load, titan_lsu_fpload,
886 titan_lsu_store, titan_lsu_fpstore): Adjust.
887 * config/rs6000/xfpu.md (fp-load, fp-store): Adjust.
889 2014-05-07 Oleg Endo <olegendo@gcc.gnu.org>
892 * config/sh/sh-mem.cc (sh_expand_strlen): Use loop when emitting
893 unrolled byte insns. Emit address increments after move insns.
895 2014-05-07 David Malcolm <dmalcolm@redhat.com>
897 * gimple.h (gimple_builtin_call_types_compatible_p): Accept a
898 const_gimple, rather than a gimple.
899 (gimple_call_builtin_p): Likewise, for the three variants.
901 * gimple.c (gimple_builtin_call_types_compatible_p): Likewise.
902 (gimple_call_builtin_p): Likewise, for the three variants.
904 2014-05-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
906 PR tree-optimization/61095
907 * tree-ssanames.c (get_nonzero_bits): Fix type extension in wi::shwi.
909 2014-05-07 Richard Biener <rguenther@suse.de>
911 PR tree-optimization/61034
912 * tree-ssa-alias.c (call_may_clobber_ref_p_1): Export.
913 (maybe_skip_until): Use translate to take into account
914 lattices when trying to do disambiguations.
915 (get_continuation_for_phi_1): Likewise.
916 (get_continuation_for_phi): Adjust for added translate arguments.
917 (walk_non_aliased_vuses): Likewise.
918 * tree-ssa-alias.h (get_continuation_for_phi): Adjust prototype.
919 (walk_non_aliased_vuses): Likewise.
920 (call_may_clobber_ref_p_1): Declare.
921 * tree-ssa-sccvn.c (vn_reference_lookup_3): Also disambiguate against
922 calls. Stop early if we are only supposed to disambiguate.
923 * tree-ssa-pre.c (translate_vuse_through_block): Adjust.
925 2014-05-07 Joern Rennecke <joern.rennecke@embecosm.com>
927 * config/epiphany/epiphany.c (epiphany_handle_interrupt_attribute):
928 Emit an error when the function has arguments.
930 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
932 * cfgloop.h (unswitch_loops): Remove.
933 * doc/passes.texi: Remove references to loop-unswitch.c
934 * timevar.def (TV_LOOP_UNSWITCH): Remove.
936 2014-05-07 Evgeny Stupachenko <evstupac@gmail.com>
938 * tree-vect-data-refs.c (vect_grouped_load_supported): New
939 check for loads group of length 3.
940 (vect_permute_load_chain): New permutations for loads group of
942 * tree-vect-stmts.c (vect_model_load_cost): Change cost
943 of vec_perm_shuffle for the new permutations.
945 2014-05-07 Alan Lawrence <alan.lawrence@arm.com>
947 * config/aarch64/arm_neon.h (vtrn1_f32, vtrn1_p8, vtrn1_p16, vtrn1_s8,
948 vtrn1_s16, vtrn1_s32, vtrn1_u8, vtrn1_u16, vtrn1_u32, vtrn1q_f32,
949 vtrn1q_f64, vtrn1q_p8, vtrn1q_p16, vtrn1q_s8, vtrn1q_s16, vtrn1q_s32,
950 vtrn1q_s64, vtrn1q_u8, vtrn1q_u16, vtrn1q_u32, vtrn1q_u64, vtrn2_f32,
951 vtrn2_p8, vtrn2_p16, vtrn2_s8, vtrn2_s16, vtrn2_s32, vtrn2_u8,
952 vtrn2_u16, vtrn2_u32, vtrn2q_f32, vtrn2q_f64, vtrn2q_p8, vtrn2q_p16,
953 vtrn2q_s8, vtrn2q_s16, vtrn2q_s32, vtrn2q_s64, vtrn2q_u8, vtrn2q_u16,
954 vtrn2q_u32, vtrn2q_u64): Replace temporary asm with __builtin_shuffle.
956 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
958 * loop-unswitch.c: Delete.
960 2014-05-07 Richard Biener <rguenther@suse.de>
962 * config.gcc: Always set need_64bit_hwint to yes.
964 2014-05-07 Chung-Ju Wu <jasonwucj@gmail.com>
966 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Have it in favor
967 of using optimize_size.
969 2014-05-06 Mike Stump <mikestump@comcast.net>
971 * wide-int.h (wi::int_traits <HOST_WIDE_INT>): Always define.
973 2014-05-06 Joseph Myers <joseph@codesourcery.com>
975 * config/i386/sse.md (*mov<mode>_internal)
976 (*<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>)
977 (*<sse2_avx_avx512f>_loaddqu<mode><mask_name>)
978 (<sse>_andnot<mode>3, <code><mode>3, *andnot<mode>3)
979 (*<code><mode>3, *andnot<mode>3<mask_name>)
980 (<mask_codefor><code><mode>3<mask_name>): Only consider
981 TARGET_SSE_PACKED_SINGLE_INSN_OPTIMAL for modes of size 16.
983 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
986 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
988 * lra-constraints.c (valid_address_p): Move earlier in file.
989 Add a constraint argument to the address_info version.
990 (satisfies_memory_constraint_p): New function.
991 (satisfies_address_constraint_p): Likewise.
992 (process_alt_operands, curr_insn_transform): Use them.
993 (process_address): Pass the constraint to valid_address_p when
994 checking address operands.
996 2014-05-06 Richard Sandiford <r.sandiford@uk.ibm.com>
998 * lto-cgraph.c (compute_ltrans_boundary): Make node variables local
999 to their respective blocks. Fix inadvertent use of "node".
1001 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
1003 * emit-rtl.c (init_derived_machine_modes): New functionm, split
1005 (init_emit_once): ...here.
1006 * rtl.h (init_derived_machine_modes): Declare.
1007 * toplev.c (do_compile): Call it even if no_backend.
1009 2014-05-06 Kenneth Zadeck <zadeck@naturalbridge.com>
1010 Mike Stump <mikestump@comcast.net>
1011 Richard Sandiford <rdsandiford@googlemail.com>
1012 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1014 * alias.c (ao_ref_from_mem): Use wide-int interfaces.
1015 (rtx_equal_for_memref_p): Update comment.
1016 (adjust_offset_for_component_ref): Use wide-int interfaces.
1017 * builtins.c (get_object_alignment_2): Likewise.
1018 (c_readstr): Likewise.
1019 (target_char_cast): Add comment.
1020 (determine_block_size): Use wide-int interfaces.
1021 (expand_builtin_signbit): Likewise.
1022 (fold_builtin_int_roundingfn): Likewise.
1023 (fold_builtin_bitop): Likewise.
1024 (fold_builtin_bswap): Likewise.
1025 (fold_builtin_logarithm): Use signop.
1026 (fold_builtin_pow): Likewise.
1027 (fold_builtin_memory_op): Use wide-int interfaces.
1028 (fold_builtin_object_size): Likewise.
1029 * cfgloop.c (alloc_loop): Initialize nb_iterations_upper_bound and
1030 nb_iterations_estimate.
1031 (record_niter_bound): Use wide-int interfaces.
1032 (get_estimated_loop_iterations_int): Likewise.
1033 (get_estimated_loop_iterations): Likewise.
1034 (get_max_loop_iterations): Likewise.
1035 * cfgloop.h: Include wide-int.h.
1036 (struct nb_iter_bound): Change bound to widest_int.
1037 (struct loop): Change nb_iterations_upper_bound and
1038 nb_iterations_estimate to widest_int.
1039 (record_niter_bound): Switch to use widest_int.
1040 (get_estimated_loop_iterations): Likewise.
1041 (get_max_loop_iterations): Likewise.
1042 (gcov_type_to_double_int): Rename to gcov_type_to_wide_int and
1043 update for wide-int.
1044 * cgraph.c (cgraph_add_thunk): Use wide-int interfaces.
1045 * combine.c (try_combine): Likewise.
1046 (subst): Use CONST_SCALAR_INT_P rather than CONST_INT_P.
1047 * config/aarch64/aarch64.c (aapcs_vfp_sub_candidate): Use wide-int
1049 (aarch64_float_const_representable_p): Likewise.
1050 * config/arc/arc.c: Include wide-int.h.
1051 (arc_can_use_doloop_p): Use wide-int interfaces.
1052 * config/arm/arm.c (aapcs_vfp_sub_candidate): Likewise.
1053 (vfp3_const_double_index): Likewise.
1054 * config/avr/avr.c (avr_out_round): Likewise.
1055 (avr_fold_builtin): Likewise.
1056 * config/bfin/bfin.c (bfin_local_alignment): Likewise.
1057 (bfin_can_use_doloop_p): Likewise.
1058 * config/darwin.c (darwin_mergeable_constant_section): Likewise.
1059 (machopic_select_rtx_section): Update to handle CONST_WIDE_INT.
1060 * config/i386/i386.c: Include wide-int.h.
1061 (ix86_data_alignment): Use wide-int interfaces.
1062 (ix86_local_alignment): Likewise.
1063 (ix86_emit_swsqrtsf): Update real_from_integer.
1064 * config/msp430/msp430.c (msp430_attr): Use wide-int interfaces.
1065 * config/nds32/nds32.c (nds32_insert_attributes): Likewise.
1066 * config/rs6000/predicates.md (any_operand): Add const_wide_int.
1067 (zero_constant): Likewise.
1068 (input_operand): Likewise.
1069 (splat_input_operand): Likewise.
1070 (non_logical_cint_operand): Change const_double to const_wide_int.
1071 * config/rs6000/rs6000.c (num_insns_constant): Handle CONST_WIDE_INT.
1072 (easy_altivec_constant): Remove comment.
1073 (paired_expand_vector_init): Use CONSTANT_P.
1074 (rs6000_legitimize_address): Handle CONST_WIDE_INT.
1075 (rs6000_emit_move): Update checks.
1076 (rs6000_aggregate_candidate): Use wide-int interfaces.
1077 (rs6000_expand_ternop_builtin): Likewise.
1078 (rs6000_output_move_128bit): Handle CONST_WIDE_INT.
1079 (rs6000_assemble_integer): Likewise.
1080 (rs6000_hash_constant): Likewise.
1081 (output_toc): Likewise.
1082 (rs6000_rtx_costs): Likewise.
1083 (rs6000_emit_swrsqrt); Update call to real_from_integer.
1084 * config/rs6000/rs6000-c.c: Include wide-int.h.
1085 (altivec_resolve_overloaded_builtin): Use wide-int interfaces.
1086 * config/rs6000/rs6000.h (TARGET_SUPPORTS_WIDE_INT): New.
1087 * config/rs6000/rs6000.md: Use const_scalar_int_operand.
1088 Handle CONST_WIDE_INT.
1089 * config/sol2-c.c (solaris_pragma_align): Change low to unsigned HWI.
1090 Use tree_fits_uhwi_p.
1091 * config/sparc/sparc.c: Include wide-int.h.
1092 (sparc_fold_builtin): Use wide-int interfaces.
1093 * config/vax/vax.c: Include wide-int.h.
1094 (vax_float_literal): Use real_from_integer.
1095 * coretypes.h (struct hwivec_def): New.
1097 (const_hwivec): New.
1098 * cse.c (hash_rtx_cb): Handle CONST_WIDE_INT.
1099 (equiv_constant): Handle CONST_WIDE_INT.
1100 * cselib.c (rtx_equal_for_cselib_1): Use CASE_CONST_UNIQUE.
1101 (cselib_hash_rtx): Handle CONST_WIDE_INT.
1102 * dbxout.c (stabstr_U): Use wide-int interfaces.
1103 (dbxout_type): Update to use cst_fits_shwi_p.
1104 * defaults.h (LOG2_BITS_PER_UNIT): Define.
1105 (TARGET_SUPPORTS_WIDE_INT): Add default.
1106 * dfp.c: Include wide-int.h.
1107 (decimal_real_to_integer2): Use wide-int interfaces and rename to
1108 decimal_real_to_integer.
1109 * dfp.h (decimal_real_to_integer2): Return a wide_int and rename to
1110 decimal_real_to_integer.
1111 * doc/generic.texi (Constant expressions): Update for wide_int.
1112 * doc/rtl.texi (const_double): Likewise.
1113 (const_wide_int, CONST_WIDE_INT, CONST_WIDE_INT_VEC): New.
1114 (CONST_WIDE_INT_NUNITS, CONST_WIDE_INT_ELT): New.
1115 * doc/tm.texi.in (REAL_VALUE_TO_INT): Remove.
1116 (REAL_VALUE_FROM_INT): Remove.
1117 (TARGET_SUPPORTS_WIDE_INT): New.
1118 * doc/tm.texi: Regenerate.
1119 * dojump.c (prefer_and_bit_test): Use wide-int interfaces.
1120 * double-int.h: Include wide-int.h.
1121 (struct wi::int_traits): New.
1122 * dwarf2out.c (get_full_len): New.
1123 (dw_val_equal_p): Add case dw_val_class_wide_int.
1124 (size_of_loc_descr): Likewise.
1125 (output_loc_operands): Likewise.
1126 (insert_double): Remove.
1127 (insert_wide_int): New.
1129 (print_die): Add case dw_val_class_wide_int.
1130 (attr_checksum): Likewise.
1131 (attr_checksum_ordered): Likewise.
1132 (same_dw_val_p): Likewise.
1133 (size_of_die): Likewise.
1134 (value_format): Likewise.
1135 (output_die): Likewise.
1136 (double_int_type_size_in_bits): Rename to offset_int_type_size_in_bits.
1138 (clz_loc_descriptor): Use wide-int interfaces.
1139 (mem_loc_descriptor): Likewise. Handle CONST_WIDE_INT.
1140 (loc_descriptor): Use wide-int interfaces. Handle CONST_WIDE_INT.
1141 (round_up_to_align): Use wide-int interfaces.
1142 (field_byte_offset): Likewise.
1143 (insert_double): Rename to insert_wide_int. Use wide-int interfaces.
1144 (add_const_value_attribute): Handle CONST_WIDE_INT. Update
1145 CONST_DOUBLE handling. Use wide-int interfaces.
1146 (add_bound_info): Use tree_fits_uhwi_p. Use wide-int interfaces.
1147 (gen_enumeration_type_die): Use add_AT_wide.
1148 (hash_loc_operands): Add case dw_val_class_wide_int.
1149 (compare_loc_operands): Likewise.
1150 * dwarf2out.h: Include wide-int.h.
1151 (wide_int_ptr): New.
1152 (enum dw_val_class): Add dw_val_class_wide_int.
1153 (struct dw_val_struct): Add val_wide.
1154 * emit-rtl.c (const_wide_int_htab): New.
1155 (const_wide_int_htab_hash): New.
1156 (const_wide_int_htab_eq): New.
1157 (lookup_const_wide_int): New.
1158 (const_double_htab_hash): Use wide-int interfaces.
1159 (const_double_htab_eq): Likewise.
1160 (rtx_to_double_int): Conditionally compile for wide-int.
1161 (immed_double_int_const): Rename to immed_wide_int_const and
1162 update for wide-int.
1163 (immed_double_const): Conditionally compile for wide-int.
1164 (init_emit_once): Use wide-int interfaces.
1165 * explow.c (plus_constant): Likewise.
1166 * expmed.c (mask_rtx): Move further up file. Use wide-int interfaces.
1167 (lshift_value): Use wide-int interfaces.
1168 (expand_mult): Likewise.
1169 (choose_multiplier): Likewise.
1170 (expand_smod_pow2): Likewise.
1171 (make_tree): Likewise.
1172 * expr.c (convert_modes): Consolidate handling of constants.
1173 Use wide-int interfaces.
1174 (emit_group_load_1): Add note.
1175 (store_expr): Update comment.
1176 (get_inner_reference): Use wide-int interfaces.
1177 (expand_constructor): Update comment.
1178 (expand_expr_real_2): Use wide-int interfaces.
1179 (expand_expr_real_1): Likewise.
1180 (reduce_to_bit_field_precision): Likewise.
1181 (const_vector_from_tree): Likewise.
1182 * final.c: Include wide-int-print.h.
1183 (output_addr_const): Handle CONST_WIDE_INT. Use CONST_DOUBLE_AS_INT_P.
1184 * fixed-value.c: Include wide-int.h.
1185 (fixed_from_string): Use wide-int interfaces.
1186 (fixed_to_decimal): Likewise.
1187 (fixed_convert_from_real): Likewise.
1188 (real_convert_from_fixed): Likewise.
1189 * fold-const.h (mem_ref_offset): Return an offset_int.
1190 (div_if_zero_remainder): Remove code parameter.
1191 * fold-const.c (div_if_zero_remainder): Remove code parameter.
1192 Use wide-int interfaces.
1193 (may_negate_without_overflow_p): Use wide-int interfaces.
1194 (negate_expr_p): Likewise.
1195 (fold_negate_expr): Likewise.
1196 (int_const_binop_1): Likewise.
1197 (const_binop): Likewise.
1198 (fold_convert_const_int_from_int): Likewise.
1199 (fold_convert_const_int_from_real): Likewise.
1200 (fold_convert_const_int_from_fixed): Likewise.
1201 (fold_convert_const_fixed_from_int): Likewise.
1202 (all_ones_mask_p): Take an unsigned size. Use wide-int interfaces.
1203 (sign_bit_p): Use wide-int interfaces.
1204 (make_range_step): Likewise.
1205 (build_range_check): Likewise. Pass an integer of the correct type
1206 instead of using integer_one_node.
1207 (range_predecessor): Pass an integer of the correct type instead
1208 of using integer_one_node.
1209 (range_successor): Likewise.
1210 (merge_ranges): Likewise.
1211 (unextend): Use wide-int interfaces.
1212 (extract_muldiv_1): Likewise.
1213 (fold_div_compare): Likewise.
1214 (fold_single_bit_test): Likewise.
1215 (fold_sign_changed_comparison): Likewise.
1216 (try_move_mult_to_index): Update calls to div_if_zero_remainder.
1217 (fold_plusminus_mult_expr): Use wide-int interfaces.
1218 (native_encode_int): Likewise.
1219 (native_interpret_int): Likewise.
1220 (fold_unary_loc): Likewise.
1221 (pointer_may_wrap_p): Likewise.
1222 (size_low_cst): Likewise.
1223 (mask_with_tz): Likewise.
1224 (fold_binary_loc): Likewise.
1225 (fold_ternary_loc): Likewise.
1226 (multiple_of_p): Likewise.
1227 (tree_call_nonnegative_warnv_p): Update calls to
1228 tree_int_cst_min_precision and real_from_integer.
1229 (fold_negate_const): Use wide-int interfaces.
1230 (fold_abs_const): Likewise.
1231 (fold_relational_const): Use tree_int_cst_lt.
1232 (round_up_loc): Use wide-int interfaces.
1233 * genemit.c (gen_exp): Add CONST_WIDE_INT case.
1234 * gengenrtl.c (excluded_rtx): Add CONST_WIDE_INT case.
1235 * gengtype.c: Remove include of double-int.h.
1236 (do_typedef): Use wide-int interfaces.
1237 (open_base_files): Add wide-int.h.
1238 (main): Add offset_int and widest_int typedefs.
1239 * gengtype-lex.l: Handle "^".
1240 (CXX_KEYWORD): Add "static".
1241 * gengtype-parse.c (require3): New.
1242 (require_template_declaration): Handle constant template arguments
1243 and nested templates.
1244 * gengtype-state.c: Don't include "double-int.h".
1245 * genpreds.c (write_one_predicate_function): Update comment.
1246 (write_tm_constrs_h): Add check for hval and lval use in
1248 * genrecog.c (validate_pattern): Add CONST_WIDE_INT case.
1249 (add_to_sequence): Likewise.
1250 * gensupport.c (struct std_pred_table): Add const_scalar_int_operand
1251 and const_double_operand.
1252 * gimple.c (preprocess_case_label_vec_for_gimple): Use wide-int
1254 * gimple-fold.c (get_base_constructor): Likewise.
1255 (fold_array_ctor_reference): Likewise.
1256 (fold_nonarray_ctor_reference): Likewise.
1257 (fold_const_aggregate_ref_1): Likewise.
1258 (gimple_val_nonnegative_real_p): Likewise.
1259 (gimple_fold_indirect_ref): Likewise.
1260 * gimple-pretty-print.c (dump_ssaname_info): Likewise.
1261 * gimple-ssa-strength-reduction.c: Include wide-int-print.h.
1262 (struct slsr_cand_d): Change index to be widest_int.
1263 (struct incr_info_d): Change incr to be widest_int.
1264 (alloc_cand_and_find_basis): Use wide-int interfaces.
1265 (slsr_process_phi): Likewise.
1266 (backtrace_base_for_ref): Likewise. Return a widest_int.
1267 (restructure_reference): Take a widest_int instead of a double_int.
1268 (slsr_process_ref): Use wide-int interfaces.
1269 (create_mul_ssa_cand): Likewise.
1270 (create_mul_imm_cand): Likewise.
1271 (create_add_ssa_cand): Likewise.
1272 (create_add_imm_cand): Take a widest_int instead of a double_int.
1273 (slsr_process_add): Use wide-int interfaces.
1274 (slsr_process_cast): Likewise.
1275 (slsr_process_copy): Likewise.
1276 (dump_candidate): Likewise.
1277 (dump_incr_vec): Likewise.
1278 (replace_ref): Likewise.
1279 (cand_increment): Likewise. Return a widest_int.
1280 (cand_abs_increment): Likewise.
1281 (replace_mult_candidate): Take a widest_int instead of a double_int.
1282 (replace_unconditional_candidate): Use wide-int interfaces.
1283 (incr_vec_index): Take a widest_int instead of a double_int.
1284 (create_add_on_incoming_edge): Likewise.
1285 (create_phi_basis): Use wide-int interfaces.
1286 (replace_conditional_candidate): Likewise.
1287 (record_increment): Take a widest_int instead of a double_int.
1288 (record_phi_increments): Use wide-int interfaces.
1289 (phi_incr_cost): Take a widest_int instead of a double_int.
1290 (lowest_cost_path): Likewise.
1291 (total_savings): Likewise.
1292 (analyze_increments): Use wide-int interfaces.
1293 (ncd_with_phi): Take a widest_int instead of a double_int.
1294 (ncd_of_cand_and_phis): Likewise.
1295 (nearest_common_dominator_for_cands): Likewise.
1296 (insert_initializers): Use wide-int interfaces.
1297 (all_phi_incrs_profitable): Likewise.
1298 (replace_one_candidate): Likewise.
1299 (replace_profitable_candidates): Likewise.
1300 * godump.c: Include wide-int-print.h.
1301 (go_output_typedef): Use wide-int interfaces.
1302 * graphite-clast-to-gimple.c (gmp_cst_to_tree): Likewise.
1303 * graphite-sese-to-poly.c (tree_int_to_gmp): Likewise.
1304 (build_loop_iteration_domains): Likewise.
1305 * hooks.h: Include wide-int.h rather than double-int.h.
1306 (hook_bool_dint_dint_uint_bool_true): Delete.
1307 (hook_bool_wint_wint_uint_bool_true): Declare.
1308 * hooks.c (hook_bool_dint_dint_uint_bool_true): Removed.
1309 (hook_bool_wint_wint_uint_bool_true): New.
1310 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): Use wide-int
1312 (ubsan_expand_si_overflow_mul_check): Likewise.
1313 * ipa-devirt.c (get_polymorphic_call_info): Likewise.
1314 * ipa-prop.c (compute_complex_assign_jump_func): Likewise.
1315 (get_ancestor_addr_info): Likewise.
1316 (ipa_modify_call_arguments): Likewise.
1317 * loop-doloop.c (doloop_modify): Likewise.
1318 (doloop_optimize): Likewise.
1319 * loop-iv.c (iv_number_of_iterations): Likewise.
1320 * loop-unroll.c (decide_unroll_constant_iterations): Likewise.
1321 (unroll_loop_constant_iterations): Likewise.
1322 (decide_unroll_runtime_iterations): Likewise.
1323 (unroll_loop_runtime_iterations): Likewise.
1324 (decide_peel_simple): Likewise.
1325 (decide_unroll_stupid): Likewise.
1326 * lto-streamer-in.c (streamer_read_wi): Add.
1327 (input_cfg): Use wide-int interfaces.
1328 (lto_input_tree_1): Likewise.
1329 * lto-streamer-out.c (streamer_write_wi): Add.
1330 (hash_tree): Use wide-int interfaces.
1331 (output_cfg): Likewise.
1332 * Makefile.in (OBJS): Add wide-int.o and wide-int-print.o.
1333 (GTFILES): Add wide-int.h and signop.h.
1334 (TAGS): Look for .cc files too.
1335 * omp-low.c (scan_omp_1_op): Use wide-int interfaces.
1336 * optabs.c (expand_subword_shift): Likewise.
1337 (expand_doubleword_shift): Likewise.
1338 (expand_absneg_bit): Likewise.
1339 (expand_copysign_absneg): Likewise.
1340 (expand_copysign_bit): Likewise.
1341 * postreload.c (reload_cse_simplify_set): Likewise.
1342 * predict.c (predict_iv_comparison): Likewise.
1343 * pretty-print.h: Include wide-int-print.h.
1345 * print-rtl.c (print_rtx): Add CONST_WIDE_INT case.
1346 * print-tree.c: Include wide-int-print.h.
1347 (print_node_brief): Use wide-int interfaces.
1348 (print_node): Likewise.
1349 * read-rtl.c (validate_const_wide_int): New.
1350 (read_rtx_code): Add CONST_WIDE_INT case.
1351 * real.c: Include wide-int.h.
1352 (real_to_integer2): Delete.
1353 (real_to_integer): New function, returning a wide_int.
1354 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
1355 (ten_to_ptwo): Update call to real_from_integer.
1356 (real_digit): Likewise.
1357 * real.h: Include signop.h, wide-int.h and insn-modes.h.
1358 (real_to_integer2, REAL_VALUE_FROM_INT, REAL_VALUE_FROM_UNSIGNED_INT)
1359 (REAL_VALUE_TO_INT): Delete.
1360 (real_to_integer): Declare a wide-int form.
1361 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
1362 * recog.c (const_int_operand): Improve comment.
1363 (const_scalar_int_operand): New.
1364 (const_double_operand): Add a separate definition for CONST_WIDE_INT.
1365 * rtlanal.c (commutative_operand_precedence): Handle CONST_WIDE_INT.
1366 (split_double): Likewise.
1367 * rtl.c (DEF_RTL_EXPR): Handle CONST_WIDE_INT.
1368 (rtx_size): Likewise.
1369 (rtx_alloc_stat_v): New.
1370 (rtx_alloc_stat): Now calls rtx_alloc_stat_v.
1371 (cwi_output_hex): New.
1372 (iterative_hash_rtx): Handle CONST_WIDE_INT.
1373 (cwi_check_failed_bounds): New.
1374 * rtl.def (CONST_WIDE_INT): New.
1375 * rtl.h: Include <utility> and wide-int.h.
1376 (struct hwivec_def): New.
1377 (CWI_GET_NUM_ELEM): New.
1378 (CWI_PUT_NUM_ELEM): New.
1379 (struct rtx_def): Add num_elem and hwiv.
1380 (CASE_CONST_SCALAR_INT): Modify for TARGET_SUPPORTS_WIDE_INT.
1381 (CASE_CONST_UNIQUE): Likewise.
1382 (CASE_CONST_ANY): Likewise.
1383 (CONST_SCALAR_INT_P): Likewise.
1384 (CONST_WIDE_INT_P): New.
1386 (HWIVEC_CHECK): New.
1387 (cwi_check_failed_bounds): New.
1389 (HWIVEC_CHECK): New.
1390 (CONST_WIDE_INT_VEC) New.
1391 (CONST_WIDE_INT_NUNITS) New.
1392 (CONST_WIDE_INT_ELT) New.
1393 (rtx_mode_t): New type.
1394 (wi::int_traits <rtx_mode_t>): New.
1396 (wi::min_value): New.
1397 (wi::max_value): New.
1399 (const_wide_int_alloc): New.
1400 (immed_wide_int_const): New.
1401 * sched-vis.c (print_value): Handle CONST_WIDE_INT.
1402 * sel-sched-ir.c (lhs_and_rhs_separable_p): Update comment.
1403 * signop.h: New file.
1404 * simplify-rtx.c (mode_signbit_p): Handle CONST_WIDE_INT.
1405 (simplify_const_unary_operation): Use wide-int interfaces.
1406 (simplify_binary_operation_1): Likewise.
1407 (simplify_const_binary_operation): Likewise.
1408 (simplify_const_relational_operation): Likewise.
1409 (simplify_immed_subreg): Likewise.
1410 * stmt.c (expand_case): Likewise.
1411 * stor-layout.h (set_min_and_max_values_for_integral_type): Take a
1412 signop rather than a bool.
1413 * stor-layout.c (layout_type): Use wide-int interfaces.
1414 (initialize_sizetypes): Update calls to
1415 set_min_and_max_values_for_integral_type.
1416 (set_min_and_max_values_for_integral_type): Take a signop rather
1417 than a bool. Use wide-int interfaces.
1418 (fixup_signed_type): Update accordingly. Remove
1419 HOST_BITS_PER_DOUBLE_INT limit.
1420 (fixup_unsigned_type): Likewise.
1421 * system.h (STATIC_CONSTANT_P): New.
1422 (STATIC_ASSERT): New.
1423 * target.def (can_use_doloop_p): Take widest_ints rather than
1425 * target.h: Include wide-int.h rather than double-int.h.
1426 * targhooks.h (can_use_doloop_if_innermost): Take widest_ints rather
1428 * targhooks.c (default_cxx_get_cookie_size): Use tree_int_cst_lt
1429 rather than INT_CST_LT_UNSIGNED.
1430 (can_use_doloop_if_innermost): Take widest_ints rather than
1432 * tree-affine.c: Include wide-int-print.h.
1433 (double_int_ext_for_comb): Delete.
1434 (wide_int_ext_for_comb): New.
1435 (aff_combination_zero): Use wide-int interfaces.
1436 (aff_combination_const): Take a widest_int instead of a double_int.
1437 (aff_combination_elt): Use wide-int interfaces.
1438 (aff_combination_scale): Take a widest_int instead of a double_int.
1439 (aff_combination_add_elt): Likewise.
1440 (aff_combination_add_cst): Likewise.
1441 (aff_combination_add): Use wide-int interfaces.
1442 (aff_combination_convert): Likewise.
1443 (tree_to_aff_combination): Likewise.
1444 (add_elt_to_tree): Take a widest_int instead of a double_int.
1445 (aff_combination_to_tree): Use wide-int interfaces.
1446 (aff_combination_remove_elt): Likewise.
1447 (aff_combination_add_product): Take a widest_int instead of
1449 (aff_combination_mult): Use wide-int interfaces.
1450 (aff_combination_expand): Likewise.
1451 (double_int_constant_multiple_p): Delete.
1452 (wide_int_constant_multiple_p): New.
1453 (aff_combination_constant_multiple_p): Take a widest_int pointer
1454 instead of a double_int pointer.
1455 (print_aff): Use wide-int interfaces.
1456 (get_inner_reference_aff): Take a widest_int pointer
1457 instead of a double_int pointer.
1458 (aff_comb_cannot_overlap_p): Take widest_ints instead of double_ints.
1459 * tree-affine.h: Include wide-int.h.
1460 (struct aff_comb_elt): Change type of coef to widest_int.
1461 (struct affine_tree_combination): Change type of offset to widest_int.
1462 (double_int_ext_for_comb): Delete.
1463 (wide_int_ext_for_comb): New.
1464 (aff_combination_const): Use widest_int instead of double_int.
1465 (aff_combination_scale): Likewise.
1466 (aff_combination_add_elt): Likewise.
1467 (aff_combination_constant_multiple_p): Likewise.
1468 (get_inner_reference_aff): Likewise.
1469 (aff_comb_cannot_overlap_p): Likewise.
1470 (aff_combination_zero_p): Use wide-int interfaces.
1471 * tree.c: Include tree.h.
1472 (init_ttree): Use make_int_cst.
1473 (tree_code_size): Removed code for INTEGER_CST case.
1474 (tree_size): Add INTEGER_CST case.
1475 (make_node_stat): Update comment.
1476 (get_int_cst_ext_nunits, build_new_int_cst, build_int_cstu): New.
1477 (build_int_cst_type): Use wide-int interfaces.
1478 (double_int_to_tree): Likewise.
1479 (double_int_fits_to_tree_p): Delete.
1480 (force_fit_type_double): Delete.
1481 (force_fit_type): New.
1482 (int_cst_hash_hash): Use wide-int interfaces.
1483 (int_cst_hash_eq): Likewise.
1484 (build_int_cst_wide): Delete.
1485 (wide_int_to_tree): New.
1486 (cache_integer_cst): Use wide-int interfaces.
1487 (build_low_bits_mask): Likewise.
1488 (cst_and_fits_in_hwi): Likewise.
1489 (real_value_from_int_cst): Likewise.
1490 (make_int_cst_stat): New.
1491 (integer_zerop): Use wide_int interfaces.
1492 (integer_onep): Likewise.
1493 (integer_all_onesp): Likewise.
1494 (integer_pow2p): Likewise.
1495 (integer_nonzerop): Likewise.
1496 (tree_log2): Likewise.
1497 (tree_floor_log2): Likewise.
1498 (tree_ctz): Likewise.
1499 (int_size_in_bytes): Likewise.
1500 (mem_ref_offset): Return an offset_int rather than a double_int.
1501 (build_type_attribute_qual_variant): Use wide_int interfaces.
1502 (type_hash_eq): Likewise
1503 (tree_int_cst_equal): Likewise.
1504 (tree_int_cst_lt): Delete.
1505 (tree_int_cst_compare): Likewise.
1506 (tree_fits_shwi_p): Use wide_int interfaces.
1507 (tree_fits_uhwi_p): Likewise.
1508 (tree_int_cst_sign_bit): Likewise.
1509 (tree_int_cst_sgn): Likewise.
1510 (tree_int_cst_min_precision): Take a signop rather than a bool.
1511 (simple_cst_equal): Use wide_int interfaces.
1512 (compare_tree_int): Likewise.
1513 (iterative_hash_expr): Likewise.
1514 (int_fits_type_p): Likewise. Use tree_int_cst_lt rather than
1516 (get_type_static_bounds): Use wide_int interfaces.
1517 (tree_int_cst_elt_check_failed): New.
1518 (build_common_tree_nodes): Reordered to set prec before filling in
1520 (int_cst_value): Check cst_and_fits_in_hwi.
1521 (widest_int_cst_value): Use wide_int interfaces.
1522 (upper_bound_in_type): Likewise.
1523 (lower_bound_in_type): Likewise.
1524 (num_ending_zeros): Likewise.
1525 (drop_tree_overflow): Likewise.
1526 * tree-call-cdce.c (check_pow): Update call to real_from_integer.
1527 (gen_conditions_for_pow_cst_base): Likewise.
1528 * tree-cfg.c: Include wide-int.h and wide-int-print.h.
1529 (group_case_labels_stmt): Use wide-int interfaces.
1530 (verify_gimple_assign_binary): Likewise.
1531 (print_loop): Likewise.
1532 * tree-chrec.c (tree_fold_binomial): Likewise.
1533 * tree-core.h (struct tree_base): Add int_length.
1534 (struct tree_int_cst): Change rep of value.
1535 * tree-data-ref.c (dr_analyze_innermost): Use wide-int interfaces.
1536 (dr_may_alias_p): Likewise.
1537 (max_stmt_executions_tree): Likewise.
1538 * tree.def (INTEGER_CST): Update comment.
1539 * tree-dfa.c (get_ref_base_and_extent): Use wide-int interfaces.
1540 * tree-dfa.h (get_addr_base_and_unit_offset_1): Likewise.
1541 * tree-dump.c: Include wide-int.h and wide-int-print.h.
1542 (dequeue_and_dump): Use wide-int interfaces.
1543 * tree.h: Include wide-int.h.
1544 (NULL_TREE): Moved to earlier loc in file.
1545 (TREE_INT_CST_ELT_CHECK): New.
1546 (tree_int_cst_elt_check_failed): New.
1548 (TREE_INT_CST): Delete.
1549 (TREE_INT_CST_LOW): Use wide-int interfaces.
1550 (TREE_INT_CST_HIGH): Delete.
1551 (TREE_INT_CST_NUNITS): New.
1552 (TREE_INT_CST_EXT_NUNITS): Likewise.
1553 (TREE_INT_CST_OFFSET_NUNITS): Likewise.
1554 (TREE_INT_CST_ELT): Likewise.
1555 (INT_CST_LT): Delete.
1556 (tree_int_cst_elt_check): New (two forms).
1557 (type_code_size): Update comment.
1558 (make_int_cst_stat, make_int_cst): New.
1559 (tree_to_double_int): Delete.
1560 (double_int_fits_to_tree_p): Delete.
1561 (force_fit_type_double): Delete.
1562 (build_int_cstu): Replace with out-of-line function.
1563 (build_int_cst_wide): Delete.
1564 (tree_int_cst_lt): Define inline.
1565 (tree_int_cst_le): New.
1566 (tree_int_cst_compare): Define inline.
1567 (tree_int_cst_min_precision): Take a signop rather than a bool.
1568 (wi::int_traits <const_tree>): New.
1569 (wi::int_traits <tree>): New.
1570 (wi::extended_tree): New.
1571 (wi::int_traits <wi::extended_tree>): New.
1572 (wi::to_widest): New.
1573 (wi::to_offset): New.
1574 (wi::fits_to_tree_p): New.
1575 (wi::min_value): New.
1576 (wi::max_value): New.
1577 * tree-inline.c (remap_gimple_op_r): Use wide-int interfaces.
1578 (copy_tree_body_r): Likewise.
1579 * tree-object-size.c (compute_object_offset): Likewise.
1580 (addr_object_size): Likewise.
1581 * tree-predcom.c: Include wide-int-print.h.
1582 (struct dref_d): Change type of offset to widest_int.
1583 (dump_dref): Call wide-int printer.
1584 (aff_combination_dr_offset): Use wide-int interfaces.
1585 (determine_offset): Take a widest_int pointer rather than a
1587 (split_data_refs_to_components): Use wide-int interfaces.
1588 (suitable_component_p): Likewise.
1589 (order_drefs): Likewise.
1590 (add_ref_to_chain): Likewise.
1591 (valid_initializer_p): Likewise.
1592 (determine_roots_comp): Likewise.
1593 * tree-pretty-print.c: Include wide-int-print.h.
1594 (dump_generic_node): Use wide-int interfaces.
1595 * tree-sra.c (sra_ipa_modify_expr): Likewise.
1596 * tree-ssa-address.c (addr_for_mem_ref): Likewise.
1597 (move_fixed_address_to_symbol): Likewise.
1598 (move_hint_to_base): Likewise.
1599 (move_pointer_to_base): Likewise.
1600 (move_variant_to_index): Likewise.
1601 (most_expensive_mult_to_index): Likewise.
1602 (addr_to_parts): Likewise.
1603 (copy_ref_info): Likewise.
1604 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
1605 (indirect_refs_may_alias_p): Likewise.
1606 (stmt_kills_ref_p_1): Likewise.
1607 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
1608 * tree-ssa-ccp.c: Update comment at top of file. Include
1610 (struct prop_value_d): Change type of mask to widest_int.
1611 (extend_mask): New function.
1612 (dump_lattice_value): Use wide-int interfaces.
1613 (get_default_value): Likewise.
1614 (set_constant_value): Likewise.
1615 (set_value_varying): Likewise.
1616 (valid_lattice_transition): Likewise.
1617 (set_lattice_value): Likewise.
1618 (value_to_double_int): Delete.
1619 (value_to_wide_int): New.
1620 (get_value_from_alignment): Use wide-int interfaces.
1621 (get_value_for_expr): Likewise.
1622 (do_dbg_cnt): Likewise.
1623 (ccp_finalize): Likewise.
1624 (ccp_lattice_meet): Likewise.
1625 (bit_value_unop_1): Use widest_ints rather than double_ints.
1626 (bit_value_binop_1): Likewise.
1627 (bit_value_unop): Use wide-int interfaces.
1628 (bit_value_binop): Likewise.
1629 (bit_value_assume_aligned): Likewise.
1630 (evaluate_stmt): Likewise.
1631 (ccp_fold_stmt): Likewise.
1632 (visit_cond_stmt): Likewise.
1633 (ccp_visit_stmt): Likewise.
1634 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Likewise.
1635 (constant_pointer_difference): Likewise.
1636 (associate_pointerplus): Likewise.
1637 (combine_conversions): Likewise.
1638 * tree-ssa-loop.h: Include wide-int.h.
1639 (struct tree_niter_desc): Change type of max to widest_int.
1640 * tree-ssa-loop-im.c (mem_refs_may_alias_p): Use wide-int interfaces.
1641 * tree-ssa-loop-ivcanon.c (remove_exits_and_undefined_stmts): Likewise.
1642 (remove_redundant_iv_tests): Likewise.
1643 (canonicalize_loop_induction_variables): Likewise.
1644 * tree-ssa-loop-ivopts.c (alloc_iv): Likewise.
1645 (constant_multiple_of): Take a widest_int pointer instead of
1646 a double_int pointer.
1647 (get_computation_aff): Use wide-int interfaces.
1648 (ptr_difference_cost): Likewise.
1649 (difference_cost): Likewise.
1650 (get_loop_invariant_expr_id): Likewise.
1651 (get_computation_cost_at): Likewise.
1652 (iv_elimination_compare_lt): Likewise.
1653 (may_eliminate_iv): Likewise.
1654 * tree-ssa-loop-niter.h (estimated_loop_iterations): Use widest_int
1655 instead of double_int.
1656 (max_loop_iterations): Likewise.
1657 (max_stmt_executions): Likewise.
1658 (estimated_stmt_executions): Likewise.
1659 * tree-ssa-loop-niter.c: Include wide-int-print.h.
1660 (split_to_var_and_offset): Use wide-int interfaces.
1661 (determine_value_range): Likewise.
1662 (bound_difference_of_offsetted_base): Likewise.
1663 (bounds_add): Take a widest_int instead of a double_int.
1664 (number_of_iterations_ne_max): Use wide-int interfaces.
1665 (number_of_iterations_ne): Likewise.
1666 (number_of_iterations_lt_to_ne): Likewise.
1667 (assert_loop_rolls_lt): Likewise.
1668 (number_of_iterations_lt): Likewise.
1669 (number_of_iterations_le): Likewise.
1670 (number_of_iterations_cond): Likewise.
1671 (number_of_iterations_exit): Likewise.
1672 (finite_loop_p): Likewise.
1673 (derive_constant_upper_bound_assign): Likewise.
1674 (derive_constant_upper_bound): Return a widest_int.
1675 (derive_constant_upper_bound_ops): Likewise.
1676 (do_warn_aggressive_loop_optimizations): Use wide-int interfaces.
1677 (record_estimate): Take a widest_int rather than a double_int.
1678 (record_nonwrapping_iv): Use wide-int interfaces.
1679 (double_int_cmp): Delete.
1680 (wide_int_cmp): New.
1681 (bound_index): Take a widest_int rather than a double_int.
1682 (discover_iteration_bound_by_body_walk): Use wide-int interfaces.
1683 (maybe_lower_iteration_bound): Likewise.
1684 (estimate_numbers_of_iterations_loop): Likewise.
1685 (estimated_loop_iterations): Take a widest_int pointer than than
1686 a double_int pointer.
1687 (estimated_loop_iterations_int): Use wide-int interfaces.
1688 (max_loop_iterations): Take a widest_int pointer than than
1689 a double_int pointer.
1690 (max_loop_iterations_int): Use wide-int interfaces.
1691 (max_stmt_executions): Take a widest_int pointer than than
1692 a double_int pointer.
1693 (estimated_stmt_executions): Likewise.
1694 (n_of_executions_at_most): Use wide-int interfaces.
1695 (scev_probably_wraps_p): Likewise.
1696 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Update calls
1698 * tree-scalar-evolution.c (simplify_peeled_chrec): Use wide-int
1700 * tree-ssanames.c (set_range_info): Use wide_int_refs rather than
1701 double_ints. Adjust for trailing_wide_ints <3> representation.
1702 (set_nonzero_bits): Likewise.
1703 (get_range_info): Return wide_ints rather than double_ints.
1704 Adjust for trailing_wide_ints <3> representation.
1705 (get_nonzero_bits): Likewise.
1706 (duplicate_ssa_name_range_info): Adjust for trailing_wide_ints <3>
1708 * tree-ssanames.h (struct range_info_def): Replace min, max and
1709 nonzero_bits with a trailing_wide_ints <3>.
1710 (set_range_info): Use wide_int_refs rather than double_ints.
1711 (set_nonzero_bits): Likewise.
1712 (get_range_info): Return wide_ints rather than double_ints.
1713 (get_nonzero_bits): Likewise.
1714 * tree-ssa-phiopt.c (jump_function_from_stmt): Use wide-int interfaces.
1715 * tree-ssa-pre.c (phi_translate_1): Likewise.
1716 * tree-ssa-reassoc.c (decrement_power): Use calls to real_from_integer.
1717 (acceptable_pow_call): Likewise.
1718 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Use wide-int
1720 (vn_reference_fold_indirect): Likewise.
1721 (vn_reference_maybe_forwprop_address): Likewise.
1722 (valueize_refs_1): Likewise.
1723 * tree-ssa-structalias.c (get_constraint_for_ptr_offset): Likewise.
1724 * tree-ssa-uninit.c (is_value_included_in): Use wide-int interfaces,
1725 tree_int_cst_lt and tree_int_cst_le.
1726 * tree-streamer-in.c (unpack_ts_base_value_fields): Use wide-int
1728 (streamer_alloc_tree): Likewise.
1729 * tree-streamer-out.c (pack_ts_int_cst_value_fields): Likewise.
1730 (streamer_write_tree_header): Likewise.
1731 (streamer_write_integer_cst): Likewise.
1732 * tree-switch-conversion.c (emit_case_bit_tests): Likewise.
1733 (build_constructors): Likewise.
1734 (array_value_type): Likewise.
1735 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list): Likewise.
1736 (vect_check_gather): Likewise.
1737 * tree-vect-generic.c (build_replicated_const): Likewise.
1738 (expand_vector_divmod): Likewise.
1739 * tree-vect-loop.c (vect_transform_loop): Likewise.
1740 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
1741 (vect_do_peeling_for_alignment): Likewise.
1742 * tree-vect-patterns.c (vect_recog_divmod_pattern): Likewise.
1743 * tree-vrp.c: Include wide-int.h.
1744 (operand_less_p): Use wide-int interfaces and tree_int_cst_lt.
1745 (extract_range_from_assert): Use wide-int interfaces.
1746 (vrp_int_const_binop): Likewise.
1747 (zero_nonzero_bits_from_vr): Take wide_int pointers rather than
1748 double_int pointers.
1749 (ranges_from_anti_range): Use wide-int interfaces.
1750 (quad_int_cmp): Delete.
1751 (quad_int_pair_sort): Likewise.
1752 (extract_range_from_binary_expr_1): Use wide-int interfaces.
1753 (extract_range_from_unary_expr_1): Likewise.
1754 (adjust_range_with_scev): Likewise.
1755 (masked_increment): Take and return wide_ints rather than double_ints.
1756 (register_edge_assert_for_2): Use wide-int interfaces.
1757 (check_array_ref): Likewise.
1758 (search_for_addr_array): Likewise.
1759 (maybe_set_nonzero_bits): Likewise.
1760 (union_ranges): Pass an integer of the correct type instead of
1761 using integer_one_node.
1762 (intersect_ranges): Likewise.
1763 (simplify_truth_ops_using_ranges): Likewise.
1764 (simplify_bit_ops_using_ranges): Use wide-int interfaces.
1765 (range_fits_type_p): Likewise.
1766 (simplify_cond_using_ranges): Likewise. Take a signop rather than
1768 (simplify_conversion_using_ranges): Use wide-int interfaces.
1769 (simplify_float_conversion_using_ranges): Likewise.
1770 (vrp_finalize): Likewise.
1771 * value-prof.c (gimple_divmod_fixed_value_transform): Likewise.
1772 (gimple_stringops_transform): Likewise.
1773 * varasm.c (decode_addr_const): Likewise.
1774 (const_hash_1): Likewise.
1775 (const_rtx_hash_1): Likewise
1776 (output_constant): Likewise.
1777 (array_size_for_constructor): Likewise.
1778 (output_constructor_regular_field): Likewise.
1779 (output_constructor_bitfield): Likewise.
1780 * var-tracking.c (loc_cmp): Handle CONST_WIDE_INT.
1781 * mkconfig.sh: Include machmode.h to pick up BITS_PER_UNIT for
1783 * gencheck.c: Define BITS_PER_UNIT.
1786 * wide-int-print.cc: New.
1787 * wide-int-print.h: New.
1789 2014-05-06 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1791 * config/avr/avr.c (avr_can_eliminate): Mark unused argument.
1793 2014-05-06 Richard Biener <rguenther@suse.de>
1795 * tree-pass.h (TODO_verify_ssa, TODO_verify_flow,
1796 TODO_verify_stmts, TODO_verify_rtl_sharing): Remove.
1797 (TODO_verify_all): Adjust.
1798 * asan.c: Remove references to TODO_verify_ssa, TODO_verify_flow,
1799 TODO_verify_stmts and TODO_verify_rtl_sharing.
1800 * bb-reorder.c: Likewise.
1801 * cfgexpand.c: Likewise.
1802 * cprop.c: Likewise.
1804 * function.c: Likewise.
1805 * fwprop.c: Likewise.
1807 * gimple-ssa-isolate-paths.c: Likewise.
1808 * gimple-ssa-strength-reduction.c: Likewise.
1809 * ipa-split.c: Likewise.
1810 * loop-init.c: Likewise.
1811 * loop-unroll.c: Likewise.
1812 * lower-subreg.c: Likewise.
1813 * modulo-sched.c: Likewise.
1814 * postreload-gcse.c: Likewise.
1815 * predict.c: Likewise.
1816 * recog.c: Likewise.
1817 * sched-rgn.c: Likewise.
1818 * store-motion.c: Likewise.
1819 * tracer.c: Likewise.
1820 * trans-mem.c: Likewise.
1821 * tree-call-cdce.c: Likewise.
1822 * tree-cfg.c: Likewise.
1823 * tree-cfgcleanup.c: Likewise.
1824 * tree-complex.c: Likewise.
1825 * tree-eh.c: Likewise.
1826 * tree-emutls.c: Likewise.
1827 * tree-if-conv.c: Likewise.
1828 * tree-into-ssa.c: Likewise.
1829 * tree-loop-distribution.c: Likewise.
1830 * tree-object-size.c: Likewise.
1831 * tree-parloops.c: Likewise.
1832 * tree-pass.h: Likewise.
1833 * tree-sra.c: Likewise.
1834 * tree-ssa-ccp.c: Likewise.
1835 * tree-ssa-copy.c: Likewise.
1836 * tree-ssa-copyrename.c: Likewise.
1837 * tree-ssa-dce.c: Likewise.
1838 * tree-ssa-dom.c: Likewise.
1839 * tree-ssa-dse.c: Likewise.
1840 * tree-ssa-forwprop.c: Likewise.
1841 * tree-ssa-ifcombine.c: Likewise.
1842 * tree-ssa-loop-ch.c: Likewise.
1843 * tree-ssa-loop-ivcanon.c: Likewise.
1844 * tree-ssa-loop.c: Likewise.
1845 * tree-ssa-math-opts.c: Likewise.
1846 * tree-ssa-phiopt.c: Likewise.
1847 * tree-ssa-phiprop.c: Likewise.
1848 * tree-ssa-pre.c: Likewise.
1849 * tree-ssa-reassoc.c: Likewise.
1850 * tree-ssa-sink.c: Likewise.
1851 * tree-ssa-strlen.c: Likewise.
1852 * tree-ssa-tail-merge.c: Likewise.
1853 * tree-ssa-uncprop.c: Likewise.
1854 * tree-switch-conversion.c: Likewise.
1855 * tree-tailcall.c: Likewise.
1856 * tree-vect-generic.c: Likewise.
1857 * tree-vectorizer.c: Likewise.
1858 * tree-vrp.c: Likewise.
1860 * var-tracking.c: Likewise.
1861 * bt-load.c: Likewise.
1862 * cfgcleanup.c: Likewise.
1863 * combine-stack-adj.c: Likewise.
1864 * combine.c: Likewise.
1865 * compare-elim.c: Likewise.
1866 * config/epiphany/resolve-sw-modes.c: Likewise.
1867 * config/i386/i386.c: Likewise.
1868 * config/mips/mips.c: Likewise.
1869 * config/s390/s390.c: Likewise.
1870 * config/sh/sh_treg_combine.cc: Likewise.
1871 * config/sparc/sparc.c: Likewise.
1874 * final.c: Likewise.
1875 * ifcvt.c: Likewise.
1876 * mode-switching.c: Likewise.
1877 * passes.c: Likewise.
1878 * postreload.c: Likewise.
1880 * reg-stack.c: Likewise.
1881 * regcprop.c: Likewise.
1882 * regrename.c: Likewise.
1885 2014-05-06 Richard Biener <rguenther@suse.de>
1888 * bitmap.c (debug_bitmap): Dump to stderr, not stdout.
1889 * tree-ssa-structalias.c (dump_solution_for_var): Likewise.
1891 2014-05-05 Jan Hubicka <hubicka@ucw.cz>
1894 * ipa-devirt.c (get_class_context): Allow POD to change to non-POD.
1896 2014-05-05 Radovan Obradovic <robradovic@mips.com>
1897 Tom de Vries <tom@codesourcery.com>
1899 * target.def (call_fusage_contains_non_callee_clobbers): New
1901 * doc/tm.texi.in (@node Stack and Calling): Add Miscellaneous Register
1903 (@node Miscellaneous Register Hooks): New node.
1904 (@hook TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS): New hook.
1905 * doc/tm.texi: Regenerate.
1907 2014-05-05 Marek Polacek <polacek@redhat.com>
1910 * opts.c (common_handle_option): Call error_at instead of warning_at.
1912 2014-05-05 Richard Biener <rguenther@suse.de>
1914 * passes.c (execute_function_todo): Don't reset TODO_verify_ssa
1915 from last_verified if update_ssa ran. Move TODO_verify_rtl_sharing
1916 under the TODO_verify_il umbrella.
1918 2014-05-05 Richard Biener <rguenther@suse.de>
1920 * passes.c (execute_function_todo): Move TODO_verify_flow under
1921 the TODO_verify_ul umbrella.
1923 2014-05-05 Richard Biener <rguenther@suse.de>
1926 * fold-const.c (fold_binary_loc): Consistently avoid canonicalizing
1927 X & CST away from a CST that is the mask of a mode.
1929 2014-05-05 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1931 * config/picochip/picochip-protos.h (picochip_regno_nregs): Change
1932 int argument to enum machine_mode.
1933 (picochip_class_max_nregs): Ditto.
1934 * config/picochip/picochip.c (picochip_regno_nregs): Ditto.
1935 (picochip_class_max_nregs): Ditto.
1937 2014-05-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1939 * target.def: Add new target hook.
1940 * doc/tm.texi: Regenerate.
1941 * targhooks.h (default_keep_leaf_when_profiled): Add prototype.
1942 * targhooks.c (default_keep_leaf_when_profiled): New function.
1944 * config/s390/s390.c (s390_keep_leaf_when_profiled): New function.
1945 (TARGET_KEEP_LEAF_WHEN_PROFILED): Define.
1947 2014-05-05 Bin Cheng <bin.cheng@arm.com>
1949 PR tree-optimization/60363
1950 * gcc/tree-ssa-threadupdate.c (get_value_locus_in_path): New.
1951 (copy_phi_args): New parameters. Call get_value_locus_in_path.
1952 (update_destination_phis): New parameter.
1953 (create_edge_and_update_destination_phis): Ditto.
1954 (ssa_fix_duplicate_block_edges): Pass new arguments.
1955 (thread_single_edge): Ditto.
1957 2014-05-04 Peter Bergner <bergner@vnet.ibm.com>
1959 * config/rs6000/rs6000.h (RS6000_BTM_HARD_FLOAT): New define.
1960 (RS6000_BTM_COMMON): Add RS6000_BTM_HARD_FLOAT.
1961 (TARGET_EXTRA_BUILTINS): Add TARGET_HARD_FLOAT.
1962 * config/rs6000/rs6000-builtin.def (BU_MISC_1):
1963 Use RS6000_BTM_HARD_FLOAT.
1964 (BU_MISC_2): Likewise.
1965 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Handle
1966 RS6000_BTM_HARD_FLOAT.
1967 (rs6000_option_override_internal): Enforce -mhard-float if -mhard-dfp
1969 (rs6000_invalid_builtin): Add hard floating builtin support.
1970 (rs6000_expand_builtin): Relax the gcc_assert to allow the new
1971 hard float builtins.
1972 (rs6000_builtin_mask_names): Add RS6000_BTM_HARD_FLOAT.
1974 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
1976 * config/sh/sh_optimize_sett_clrt.cc (sh_optimize_sett_clrt::execute):
1977 Add missing function* argument.
1979 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
1981 * lra-constraints.c (valid_address_p): Move earlier in file.
1982 Add a constraint argument to the address_info version.
1983 (satisfies_memory_constraint_p): New function.
1984 (satisfies_address_constraint_p): Likewise.
1985 (process_alt_operands, curr_insn_transform): Use them.
1986 (process_address): Pass the constraint to valid_address_p when
1987 checking address operands.
1989 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
1991 * config/mips/mips.c (mips_isa_rev): New variable.
1992 (mips_set_architecture): Set it.
1993 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Set __mips_isa_rev
1995 (ISA_HAS_MUL3, ISA_HAS_FP_CONDMOVE, ISA_HAS_8CC, ISA_HAS_FP4)
1996 (ISA_HAS_PAIRED_SINGLE, ISA_HAS_MADD_MSUB, ISA_HAS_FP_RECIP_RSQRT)
1997 (ISA_HAS_CLZ_CLO, ISA_HAS_ROR, ISA_HAS_WSBH, ISA_HAS_PREFETCH)
1998 (ISA_HAS_SEB_SEH, ISA_HAS_EXT_INS, ISA_HAS_MXHC1)
1999 (ISA_HAS_HILO_INTERLOCKS, ISA_HAS_SYNCI, MIN_FPRS_PER_FMT): Reexpress
2000 conditions in terms of mips_isa_rev.
2001 (mips_isa_rev): Declare.
2003 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
2005 * config/sh/sh-mem.cc: Use tabs instead of spaces.
2006 (prob_unlikely, prob_likely): Make variables const.
2008 2014-05-03 Denis Chertykov <chertykov@gmail.com>
2010 * config/avr/avr.c (avr_adjust_insn_length): Handle JUMP_TABLE_DATA.
2012 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
2014 * config/sh/sh.h (SH_ASM_SPEC): Handle m1, m2*, m3* and m4* cases.
2016 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
2018 * config/sh/sh.h (ROUND_ADVANCE): Delete macro.
2019 (ROUND_REG, PASS_IN_REG_P): Move and rename macros to ...
2020 * config/sh/sh.c (sh_round_reg, sh_pass_in_reg_p): ... these new
2022 (sh_arg_partial_bytes, sh_function_arg, sh_function_arg_advance,
2023 sh_setup_incoming_varargs): Replace usage of PASS_IN_REG_P with
2025 Replace usage of ROUND_REG with sh_round_reg.
2026 Use CEIL instead of ROUND_ADVANCE.
2028 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
2031 * config/sh/sh.c: Include stdlib headers before everything else.
2033 2014-05-02 Jakub Jelinek <jakub@redhat.com>
2035 * gimplify.c (gimplify_adjust_omp_clauses_1): Handle
2036 GOVD_FIRSTPRIVATE | GOVD_LASTPRIVATE.
2037 (gimplify_adjust_omp_clauses): Simd region is never
2038 directly nested in combined parallel. Instead, for linear
2039 with copyin/copyout, if in combined for simd loop, make decl
2040 firstprivate/lastprivate on OMP_FOR.
2041 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
2042 expand_omp_for_static_chunk): When setting endvar, also set
2043 fd->loop.v to the same value.
2045 2014-05-02 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
2047 * hwint.h (zext_hwi): Fix signed overflow for prec == 63.
2049 2014-05-02 Alan Lawrence <alan.lawrence@arm.com>
2051 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_1): Tidy bit-flip
2054 2014-05-02 Marek Polacek <polacek@redhat.com>
2056 * doc/invoke.texi: Describe -fsanitize=float-divide-by-zero.
2058 2014-05-02 Kito Cheng <kito@0xlab.org>
2060 * defaults.h (HONOR_REG_ALLOC_ORDER): Change HONOR_REG_ALLOC_ORDER
2061 to a C expression marco.
2062 * ira-color.c (HONOR_REG_ALLOC_ORDER) : Ditto.
2063 * config/arm/arm.h (HONOR_REG_ALLOC_ORDER): Ditto.
2064 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Ditto.
2065 * doc/tm.texi (HONOR_REG_ALLOC_ORDER): Update document for
2066 HONOR_REG_ALLOC_ORDER.
2067 * doc/tm.texi.in (HONOR_REG_ALLOC_ORDER): Ditto.
2069 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2071 * config/arc/arc.c (TARGET_LRA_P): Undef before redefine.
2073 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2075 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
2077 2014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
2079 * tree-if-conv.c (is_cond_scalar_reduction): New function.
2080 (convert_scalar_cond_reduction): Likewise.
2081 (predicate_scalar_phi): Add recognition and transformation
2082 of simple conditioanl reduction to be vectorizable.
2084 2014-05-01 Marek Polacek <polacek@redhat.com>
2087 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
2089 2014-04-30 Alan Lawrence <alan.lawrence@arm.com>
2091 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
2092 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
2093 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
2094 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
2095 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
2096 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
2097 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
2098 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
2100 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
2102 * config/arc/arc.opt (mlra): Move comment above option name
2103 to avoid mis-parsing as language options.
2105 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2107 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
2108 * config/sol2.h: ... here.
2109 * config/sol2-10.h: Remove.
2111 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
2112 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
2113 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
2114 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
2115 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
2116 * config/sol2.h: ... here.
2117 (SECTION_NAME_FORMAT): Don't redefine.
2118 (STARTFILE_ARCH32_SPEC): Rename to ...
2119 (STARTFILE_ARCH_SPEC): ... this.
2120 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
2121 * config/sparc/sol2.h: ... here.
2122 (SECTION_NAME_FORMAT): Don't undef.
2123 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
2124 (SUBTARGET_EXTRA_SPECS): Remove.
2125 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
2127 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
2128 (MD_STARTFILE_PREFIX): Remove.
2129 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
2130 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
2131 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
2132 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
2133 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
2134 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
2135 * config/i386/sol2.h: ... here.
2136 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
2137 * config/i386/sol2-bi.h: Remove.
2138 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
2139 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
2141 * config/i386/t-sol2-64: Rename to ...
2142 * config/i386/t-sol2: ... this.
2143 * config/sparc/t-sol2-64: Rename to ...
2144 * config/sparc/t-sol2: ... this.
2146 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
2147 sol2_tm_file_head, sol2_tm_file_tail.
2148 Include ${cpu_type}/sol2.h before sol2.h.
2150 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
2151 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
2152 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
2153 Reflect i386/t-sol2-64 renaming.
2154 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
2155 Reflect sparc/t-sol2-64 renaming.
2157 2014-04-30 Richard Biener <rguenther@suse.de>
2159 * passes.c (execute_function_todo): Move TODO_verify_stmts
2160 and TODO_verify_ssa under the TODO_verify_il umbrella.
2161 * tree-ssa.h (verify_ssa): Adjust prototype.
2162 * tree-ssa.c (verify_ssa): Add parameter to tell whether
2163 we should verify SSA operands.
2164 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
2165 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
2166 whether we should verify whether not throwing stmts have EH info.
2167 * graphite-scop-detection.c (create_sese_edges): Adjust.
2168 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
2169 * tree-eh.c (lower_try_finally_switch): Do not add the
2170 default case label twice.
2172 2014-04-30 Marek Polacek <polacek@redhat.com>
2174 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
2175 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
2176 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
2177 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
2179 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
2181 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
2182 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
2183 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
2184 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
2185 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
2186 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
2187 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
2188 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
2190 2014-04-29 David Malcolm <dmalcolm@redhat.com>
2192 * tree-cfg.c (dump_function_to_file): Dump the return type of
2193 functions, in a line to itself before the function body, mimicking
2194 the layout of a C function.
2196 2014-04-29 Jakub Jelinek <jakub@redhat.com>
2198 PR tree-optimization/60971
2199 * tree-tailcall.c (process_assignment): Reject conversions which
2202 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
2204 * calls.c (initialize_argument_information): Always treat
2205 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
2206 (expand_call): Likewise.
2207 (emit_library_call_calue_1): Likewise.
2208 * expr.c (PUSH_ARGS_REVERSED): Do not define.
2209 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
2212 2014-04-29 Nick Clifton <nickc@redhat.com>
2214 * config/msp430/msp430.md (umulsidi): Fix typo.
2215 (mulhisi3): Enable even inside interrupt handlers.
2216 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
2217 bigger return address pushed in large mode.
2219 2014-04-29 Nick Clifton <nickc@redhat.com>
2221 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
2222 (arc_init_reg_tables): Use a machine_mode enum to iterate over
2224 * config/m32r/m32r.c (init_reg_tables): Likewise.
2225 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
2226 enum to hold the modes.
2228 2014-04-29 Richard Biener <rguenther@suse.de>
2230 * dominance.c (free_dominance_info): Add overload with
2232 (dom_info_state): Likewise.
2233 (dom_info_available_p): Likewise.
2234 * basic-block.h (free_dominance_info, dom_info_state,
2235 dom_info_available_p): Declare overloads.
2236 * passes.c (execute_function_todo): Verify that verifiers
2237 don't change dominator info state. Drop dominator info
2238 for IPA pass invocations.
2239 * cgraph.c (release_function_body): Restore asserts that
2240 dominator information is released.
2242 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
2244 * doc/invoke.texi: Fix typo.
2245 * tree-vrp.c: Fix typos.
2246 * gimple.c (infer_nonnull_range): Reorder operands of an && condition.
2248 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2250 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
2252 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
2254 * config/aarch64/aarch64-builtins.c
2255 (aarch64_types_storestruct_lane_qualifiers): New.
2256 (TYPES_STORESTRUCT_LANE): Likewise.
2257 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
2258 (st3_lane): Likewise.
2259 (st4_lane): Likewise.
2260 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
2261 (vec_store_lanesci_lane<mode>): Likewise.
2262 (vec_store_lanesxi_lane<mode>): Likewise.
2263 (aarch64_st2_lane<VQ:mode>): Likewise.
2264 (aarch64_st3_lane<VQ:mode>): Likewise.
2265 (aarch64_st4_lane<VQ:mode>): Likewise.
2266 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
2267 * config/aarch64/arm_neon.h
2268 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
2269 use new macro arguments.
2270 (__ST3_LANE_FUNC): Likewise.
2271 (__ST4_LANE_FUNC): Likewise.
2272 * config/aarch64/iterators.md (V_TWO_ELEM): New.
2273 (V_THREE_ELEM): Likewise.
2274 (V_FOUR_ELEM): Likewise.
2276 2014-04-28 David Malcolm <dmalcolm@redhat.com>
2278 * doc/gimple.texi: Replace the description of the now-defunct
2279 union gimple_statement_d with a diagram showing the
2280 gimple_statement_base class hierarchy and its relationships to
2281 the GSS_ and GIMPLE_ enums.
2283 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
2285 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
2286 * config/aarch64/aarch64.c
2287 (aarch64_cannot_change_mode_class): Weaken conditions.
2288 (aarch64_modes_tieable_p): New.
2289 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
2291 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
2293 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2294 (loadsync_<mode>): Change mode.
2295 (load_quadpti, store_quadpti): New.
2296 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2297 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2299 2014-04-28 Martin Jambor <mjambor@suse.cz>
2301 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
2302 same alias type as the original statement.
2303 (subreplacement_assignment_data): New type.
2304 (handle_unscalarized_data_in_subtree): New type of parameter,
2305 generate new memory accesses with same alias type as the original
2307 (load_assign_lhs_subreplacements): Likewise.
2308 (sra_modify_constructor_assign): Generate new memory accesses with
2309 same alias type as the original statement.
2311 2014-04-28 Richard Biener <rguenther@suse.de>
2313 * tree-pass.h (TODO_verify_il): Define.
2314 (TODO_verify_all): Complete properly.
2315 * passes.c (execute_function_todo): Move existing loop-closed
2316 SSA verification under TODO_verify_il.
2317 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
2318 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
2319 Fix tree sharing issue.
2321 2014-04-28 Richard Biener <rguenther@suse.de>
2324 * builtins.def (DEF_C11_BUILTIN): Add.
2325 (BUILT_IN_ALIGNED_ALLOC): Likewise.
2326 * coretypes.h (enum function_class): Add function_c11_misc.
2327 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
2328 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
2329 (call_may_clobber_ref_p_1): Likewise.
2330 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
2331 (mark_all_reaching_defs_necessary_1): Likewise.
2332 (propagate_necessity): Likewise.
2333 (eliminate_unnecessary_stmts): Likewise.
2334 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
2336 2014-04-28 Richard Biener <rguenther@suse.de>
2338 * tree-vrp.c (vrp_var_may_overflow): Remove.
2339 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
2340 with overflow immediately bump to one before that value and
2341 let iteration figure out overflow status.
2343 2014-04-28 Richard Biener <rguenther@suse.de>
2345 * configure.ac: Do valgrind header checks unconditionally.
2346 Add --enable-valgrind-annotations.
2347 * system.h: Guard valgrind header inclusion with
2348 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
2349 * alloc-pool.c (pool_alloc, pool_free): Use
2350 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
2351 to guard possibly dead code.
2352 * config.in: Regenerated.
2353 * configure: Likewise.
2355 2014-04-28 Jeff Law <law@redhat.com>
2357 PR tree-optimization/60902
2358 * tree-ssa-threadedge.c
2359 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
2360 over real defs when invalidating outputs from statements that do not
2361 produce useful outputs for threading.
2363 2014-04-28 Richard Biener <rguenther@suse.de>
2365 PR tree-optimization/60979
2366 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
2367 SCOPs that end in a block with a successor with abnormal
2370 2014-04-28 Richard Biener <rguenther@suse.de>
2372 * tree-pass.h (execute_pass_list): Adjust prototype.
2373 * passes.c (pass_manager::execute_early_local_passes): Adjust.
2374 (do_per_function): Change callback signature, push all actual
2375 work to the callbals.
2376 (do_per_function_toporder): Likewise.
2377 (execute_function_dump): Adjust.
2378 (execute_function_todo): Likewise.
2379 (clear_last_verified): Likewise.
2380 (verify_curr_properties): Likewise.
2381 (update_properties_after_pass): Likewise.
2382 (execute_pass_list_1): Split out from ...
2383 (execute_pass_list): ... here. Adjust.
2384 (execute_ipa_pass_list): Likewise.
2385 * cgraphunit.c (cgraph_add_new_function): Adjust.
2386 (analyze_function): Likewise.
2387 (expand_function): Likewise.
2388 * cgraph.c (release_function_body): Free dominance info
2389 here instead of asserting it was magically freed elsewhere.
2391 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
2393 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
2394 * configure: Regenerate.
2395 * config/sparc/sparc.opt (muser-mode): New option.
2396 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
2398 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
2399 * doc/invoke.texi (SPARC options): Document -muser-mode.
2401 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
2403 * cselib.c (find_slot_memmode): Delete.
2404 (cselib_hasher): Change compare_type to a struct.
2405 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
2407 (preserve_constants_and_equivs): Adjust for new compare_type.
2408 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
2409 (wrap_constant): Delete.
2410 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
2412 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
2414 * doc/install.texi (Building with profile feedback): Remove
2417 2014-04-26 Tom de Vries <tom@codesourcery.com>
2419 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
2422 2014-04-25 Cary Coutant <ccoutant@google.com>
2425 * dwarf2out.c (should_move_die_to_comdat): A type definition
2426 can contain a subprogram definition, but don't move it to a
2428 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
2429 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
2431 (clone_tree_hash): Rename to...
2432 (clone_tree_partial): ...this; change callers. Copy
2433 DW_TAG_subprogram DIEs as declarations.
2434 (copy_decls_walk): Don't copy children of a declaration into a
2437 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
2440 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
2443 2014-04-25 Jiong Wang <jiong.wang@arm.com>
2445 * config/arm/predicates.md (call_insn_operand): Add long_call check.
2446 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
2448 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
2451 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2453 * config/arm/arm.c (arm_cortex_a8_tune): Initialise T16-related fields.
2455 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2457 PR tree-optimization/60930
2458 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
2459 creating a multiply candidate by folding two constant
2460 multiplicands when the result overflows.
2462 2014-04-25 Jakub Jelinek <jakub@redhat.com>
2464 PR tree-optimization/60960
2465 * tree-vect-generic.c (expand_vector_operation): Only call
2466 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
2468 2014-04-25 Tom de Vries <tom@codesourcery.com>
2470 * expr.c (clobber_reg_mode): New function.
2471 * expr.h (clobber_reg): New function.
2473 2014-04-25 Tom de Vries <tom@codesourcery.com>
2475 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
2478 2014-04-25 Radovan Obradovic <robradovic@mips.com>
2479 Tom de Vries <tom@codesourcery.com>
2481 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
2483 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
2484 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
2485 new argument to find_all_hard_reg_sets call.
2487 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2489 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
2490 Use HOST_WIDE_INT_C for mask literal.
2491 (aarch_rev16_shleft_mask_imm_p): Likewise.
2493 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
2496 * config/sparc/sparc.md (ashlsi3_extend): Delete.
2498 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
2500 PR preprocessor/56540
2501 * config/i386/i386-c.c (ix86_target_macros): Define
2502 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
2504 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2506 * configure.ac (tga_func): Remove.
2507 (LIB_TLS_SPEC): Remove.
2508 * configure: Regenerate.
2509 * config.in: Regenerate.
2510 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
2512 2014-04-25 Richard Biener <rguenther@suse.de>
2515 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
2516 call stmt use/clobber sets during stmt walk instead of
2517 walking the possibly incomplete set of caller edges.
2519 2014-04-25 Richard Biener <rguenther@suse.de>
2522 * passes.c (apply_ipa_transforms): Inline into only caller ...
2523 (execute_one_pass): ... here. Properly bring in function
2524 bodies for nodes we want to apply IPA transforms to.
2526 2014-04-24 Cong Hou <congh@google.com>
2528 PR tree-optimization/60896
2529 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
2530 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
2531 (vect_mark_pattern_stmts): Set the def type of all statements in
2532 PATTERN_DEF_SEQ as vect_internal_def.
2534 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
2536 * doc/extend.texi (PowerPC Built-in Functions): Document new
2537 powerpc extended divide, bcd, pack/unpack 128-bit, builtin functions.
2538 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
2540 * config/rs6000/predicates.md (const_0_to_3_operand): New
2541 predicate to match 0..3 integer constants.
2543 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
2544 to support adding miscellaneous builtin functions.
2545 (BU_DFP_MISC_2): Likewise.
2546 (BU_P7_MISC_1): Likewise.
2547 (BU_P7_MISC_2): Likewise.
2548 (BU_P8V_MISC_3): Likewise.
2549 (BU_MISC_1): Likewise.
2550 (BU_MISC_2): Likewise.
2551 (DIVWE): Add extended divide builtin functions.
2554 (DIVWEUO): Likewise.
2558 (DIVDEUO): Likewise.
2559 (DXEX): Add decimal floating-point builtin functions.
2562 (DDEDPDQ): Likewise.
2564 (DENBCDQ): Likewise.
2571 (CDTBCD): Add new BCD builtin functions.
2575 (BCDADD_LT): Likewise.
2576 (BCDADD_EQ): Likewise.
2577 (BCDADD_GT): Likewise.
2578 (BCDADD_OV): Likewise.
2580 (BCDSUB_LT): Likewise.
2581 (BCDSUB_EQ): Likewise.
2582 (BCDSUB_GT): Likewise.
2583 (BCDSUB_OV): Likewise.
2584 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
2585 (UNPACK_TD): Likewise.
2586 (PACK_TF): Likewise.
2587 (UNPACK_TF): Likewise.
2588 (UNPACK_TF_0): Likewise.
2589 (UNPACK_TF_1): Likewise.
2590 (PACK_V1TI): Likewise.
2591 (UNPACK_V1TI): Likewise.
2593 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
2594 support for decimal floating point builtin functions.
2595 (rs6000_expand_ternop_builtin): Add checks for the new builtin
2596 functions that take constant arguments.
2597 (rs6000_invalid_builtin): Add decimal floating point builtin support.
2598 (rs6000_init_builtins): Setup long double, _Decimal64, and
2599 _Decimal128 types for new builtin functions.
2600 (builtin_function_type): Set the unsigned flags appropriately for
2601 the new builtin functions.
2602 (rs6000_opt_masks): Add support for decimal floating point builtin
2605 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
2606 floating point builtin functions.
2607 (RS6000_BTM_COMMON): Likewise.
2608 (RS6000_BTI_long_double): Likewise.
2609 (RS6000_BTI_dfloat64): Likewise.
2610 (RS6000_BTI_dfloat128): Likewise.
2611 (long_double_type_internal_node): Likewise.
2612 (dfloat64_type_internal_node): Likewise.
2613 (dfloat128_type_internal_node): Likewise.
2615 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
2616 2.07 bcd arithmetic instructions.
2617 (UNSPEC_BCDSUB): Likewise.
2618 (UNSPEC_BCD_OVERFLOW): Likewise.
2619 (UNSPEC_BCD_ADD_SUB): Likewise.
2620 (bcd_add_sub): Likewise.
2621 (BCD_TEST): Likewise.
2622 (bcd<bcd_add_sub>): Likewise.
2623 (bcd<bcd_add_sub>_test): Likewise.
2624 (bcd<bcd_add_sub>_test2): Likewise.
2625 (bcd<bcd_add_sub>_<code>): Likewise.
2626 (peephole2 for combined bcd ops): Likewise.
2628 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
2629 decimal floating point builtin functions.
2630 (UNSPEC_DENBCD): Likewise.
2631 (UNSPEC_DXEX): Likewise.
2632 (UNSPEC_DIEX): Likewise.
2633 (UNSPEC_DSCLI): Likewise.
2634 (UNSPEC_DSCRI): Likewise.
2635 (D64_D128): Likewise.
2636 (dfp_suffix): Likewise.
2637 (dfp_ddedpd_<mode>): Likewise.
2638 (dfp_denbcd_<mode>): Likewise.
2639 (dfp_dxex_<mode>): Likewise.
2640 (dfp_diex_<mode>): Likewise.
2641 (dfp_dscli_<mode>): Likewise.
2642 (dfp_dscri_<mode>): Likewise.
2644 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
2646 (UNSPEC_CDTBCD): Likewise.
2647 (UNSPEC_CBCDTD): Likewise.
2648 (UNSPEC_DIVE): Add support for new extended divide builtin functions.
2649 (UNSPEC_DIVEO): Likewise.
2650 (UNSPEC_DIVEU): Likewise.
2651 (UNSPEC_DIVEUO): Likewise.
2652 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
2653 pack/unpack 128-bit types.
2654 (UNSPEC_PACK_128BIT): Likewise.
2655 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
2656 (udiv<mode>3): Use idiv_ldiv mode attribute.
2657 (div<mode>3): Likewise.
2658 (addg6s): Add new BCD builtin functions.
2661 (UNSPEC_DIV_EXTEND): Add support for new extended divide instructions.
2662 (div_extend): Likewise.
2663 (div<div_extend>_<mode>"): Likewise.
2664 (FP128_64): Add support for new builtin functions to pack/unpack
2666 (unpack<mode>): Likewise.
2667 (unpacktf_0): Likewise.
2668 (unpacktf_1): Likewise.
2669 (unpack<mode>_dm): Likewise.
2670 (unpack<mode>_nodm): Likewise.
2671 (pack<mode>): Likewise.
2672 (unpackv1ti): Likewise.
2673 (packv1ti): Likewise.
2675 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
2677 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
2680 2014-04-24 Jakub Jelinek <jakub@redhat.com>
2682 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
2683 * gimplify.c (omp_is_private): Change last argument's type to int.
2684 Only diagnose lastprivate if the simd argument is 1, only diagnose
2685 linear if the simd argument is 2.
2686 (gimplify_omp_for): Adjust omp_is_private callers. When adding
2687 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
2688 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
2689 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
2690 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
2691 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
2692 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
2693 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
2694 * tree-nested.c (convert_nonlocal_omp_clauses,
2695 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
2697 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
2700 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
2703 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
2705 * flag-types.h (enum ivar_visibility): Add.
2707 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
2709 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
2710 function * argument.
2712 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
2714 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
2716 2014-04-24 Radovan Obradovic <robradovic@mips.com>
2717 Tom de Vries <tom@codesourcery.com>
2719 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
2720 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
2722 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
2723 * emit-rtl.c (try_split): Same.
2725 2014-04-24 Radovan Obradovic <robradovic@mips.com>
2726 Tom de Vries <tom@codesourcery.com>
2728 * common.opt (fuse-caller-save): New option.
2730 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
2732 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of
2733 elements for big-endian.
2735 2014-04-24 Richard Biener <rguenther@suse.de>
2737 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
2738 during TER and instead use the sepops interface for expanding
2739 non-GIMPLE_SINGLE_RHS.
2741 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2743 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
2744 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
2746 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2748 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
2749 assembler 64-bit option.
2750 * configure: Regenerate.
2752 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2754 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
2755 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
2756 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
2757 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
2758 (TARGET_CRYPTO): Take TARGET_SIMD into account.
2760 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2762 * config/aarch64/aarch64-builtins.c
2763 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
2764 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
2765 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
2766 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
2768 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
2769 (Vrevsuff): New mode attribute.
2771 2014-04-24 Terry Guo <terry.guo@arm.com>
2773 * config/arm/arm.h (machine_function): Define variable
2774 after_arm_reorg here.
2775 * config/arm/arm.c (after_arm_reorg): Remove the definition.
2776 (arm_split_constant): Update the way to access variable
2779 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
2781 2014-04-23 Tom de Vries <tom@codesourcery.com>
2783 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
2785 2014-04-23 David Malcolm <dmalcolm@redhat.com>
2787 * is-a.h: Update comments to reflect the following changes to the
2788 "pointerness" of the API, making the template parameter match the
2789 return type, allowing use of is-a.h with typedefs of pointers.
2790 (is_a_helper::cast): Return a T rather then a pointer to a T, so
2791 that the return type matches the parameter to the is_a_helper.
2793 (dyn_cast): Likewise.
2795 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
2796 pointer from the is-a.h API.
2798 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
2799 (is_a_helper <cgraph_node *>::test): ...this, matching change to
2801 (is_a_helper <varpool_node>::test): Likewise, convert to...
2802 (is_a_helper <varpool_node *>::test): ...this.
2804 (varpool_first_variable): Update for removal of implicit pointer
2805 from the is-a.h API.
2806 (varpool_next_variable): Likewise.
2807 (varpool_first_static_initializer): Likewise.
2808 (varpool_next_static_initializer): Likewise.
2809 (varpool_first_defined_variable): Likewise.
2810 (varpool_next_defined_variable): Likewise.
2811 (cgraph_first_defined_function): Likewise.
2812 (cgraph_next_defined_function): Likewise.
2813 (cgraph_first_function): Likewise.
2814 (cgraph_next_function): Likewise.
2815 (cgraph_first_function_with_gimple_body): Likewise.
2816 (cgraph_next_function_with_gimple_body): Likewise.
2817 (cgraph_alias_target): Likewise.
2818 (varpool_alias_target): Likewise.
2819 (cgraph_function_or_thunk_node): Likewise.
2820 (varpool_variable_node): Likewise.
2821 (symtab_real_symbol_p): Likewise.
2822 * cgraphunit.c (referred_to_p): Likewise.
2823 (analyze_functions): Likewise.
2824 (handle_alias_pairs): Likewise.
2825 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
2826 * gimple-ssa.h (gimple_vuse_op): Likewise.
2827 (gimple_vdef_op): Likewise.
2828 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
2829 * gimple.c (gimple_build_asm_1): Likewise.
2830 (gimple_build_try): Likewise.
2831 (gimple_build_resx): Likewise.
2832 (gimple_build_eh_dispatch): Likewise.
2833 (gimple_build_omp_for): Likewise.
2834 (gimple_omp_for_set_clauses): Likewise.
2836 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
2837 (is_a_helper <gimple_statement_asm *>::test): ...this.
2838 (is_a_helper <gimple_statement_bind>::test): Convert to...
2839 (is_a_helper <gimple_statement_bind *>::test): ...this.
2840 (is_a_helper <gimple_statement_call>::test): Convert to...
2841 (is_a_helper <gimple_statement_call *>::test): ...this.
2842 (is_a_helper <gimple_statement_catch>::test): Convert to...
2843 (is_a_helper <gimple_statement_catch *>::test): ...this.
2844 (is_a_helper <gimple_statement_resx>::test): Convert to...
2845 (is_a_helper <gimple_statement_resx *>::test): ...this.
2846 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
2847 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
2848 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
2849 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
2850 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
2851 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
2852 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
2853 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
2854 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
2855 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
2856 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
2857 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
2858 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
2859 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
2860 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
2861 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
2862 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
2863 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
2864 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
2865 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
2866 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
2867 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
2868 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
2869 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
2870 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
2871 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
2872 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
2873 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
2874 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
2875 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
2876 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
2877 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
2878 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
2879 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
2880 (is_a_helper <gimple_statement_phi>::test): Convert to...
2881 (is_a_helper <gimple_statement_phi *>::test): ...this.
2882 (is_a_helper <gimple_statement_transaction>::test): Convert to...
2883 (is_a_helper <gimple_statement_transaction *>::test): ...this.
2884 (is_a_helper <gimple_statement_try>::test): Convert to...
2885 (is_a_helper <gimple_statement_try *>::test): ...this.
2886 (is_a_helper <gimple_statement_wce>::test): Convert to...
2887 (is_a_helper <gimple_statement_wce *>::test): ...this.
2888 (is_a_helper <const gimple_statement_asm>::test): Convert to...
2889 (is_a_helper <const gimple_statement_asm *>::test): ...this.
2890 (is_a_helper <const gimple_statement_bind>::test): Convert to...
2891 (is_a_helper <const gimple_statement_bind *>::test): ...this.
2892 (is_a_helper <const gimple_statement_call>::test): Convert to...
2893 (is_a_helper <const gimple_statement_call *>::test): ...this.
2894 (is_a_helper <const gimple_statement_catch>::test): Convert to...
2895 (is_a_helper <const gimple_statement_catch *>::test): ...this.
2896 (is_a_helper <const gimple_statement_resx>::test): Convert to...
2897 (is_a_helper <const gimple_statement_resx *>::test): ...this.
2898 (is_a_helper <const gimple_statement_eh_dispatch>::test): Convert to...
2899 (is_a_helper <const gimple_statement_eh_dispatch *>::test): ...this.
2900 (is_a_helper <const gimple_statement_eh_filter>::test): Convert to...
2901 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
2902 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
2904 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
2906 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
2908 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
2910 (is_a_helper <const gimple_statement_omp_return>::test): Convert to...
2911 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
2912 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
2914 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
2915 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
2917 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
2918 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
2919 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
2920 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert to...
2921 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
2922 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
2924 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
2925 (is_a_helper <const gimple_statement_omp_target>::test): Convert to...
2926 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
2927 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
2929 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
2930 (is_a_helper <const gimple_statement_omp_single>::test): Convert to...
2931 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
2932 (is_a_helper <const gimple_statement_omp_teams>::test): Convert to...
2933 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
2934 (is_a_helper <const gimple_statement_omp_task>::test): Convert to...
2935 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
2936 (is_a_helper <const gimple_statement_phi>::test): Convert to...
2937 (is_a_helper <const gimple_statement_phi *>::test): ...this.
2938 (is_a_helper <const gimple_statement_transaction>::test): Convert to...
2939 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
2940 (is_a_helper <const gimple_statement_with_ops>::test): Convert to...
2941 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
2942 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
2943 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
2944 (is_a_helper <const gimple_statement_with_memory_ops>::test): Convert
2946 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
2948 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert to...
2949 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
2951 (gimple_use_ops): Update for removal of implicit pointer from the
2953 (gimple_set_use_ops): Likewise.
2954 (gimple_vuse): Likewise.
2955 (gimple_vdef): Likewise.
2956 (gimple_vuse_ptr): Likewise.
2957 (gimple_vdef_ptr): Likewise.
2958 (gimple_set_vuse): Likewise.
2959 (gimple_set_vdef): Likewise.
2960 (gimple_omp_return_set_lhs): Likewise.
2961 (gimple_omp_return_lhs): Likewise.
2962 (gimple_omp_return_lhs_ptr): Likewise.
2963 (gimple_call_fntype): Likewise.
2964 (gimple_call_set_fntype): Likewise.
2965 (gimple_call_set_internal_fn): Likewise.
2966 (gimple_call_use_set): Likewise.
2967 (gimple_call_clobber_set): Likewise.
2968 (gimple_bind_vars): Likewise.
2969 (gimple_bind_set_vars): Likewise.
2970 (gimple_bind_body_ptr): Likewise.
2971 (gimple_bind_set_body): Likewise.
2972 (gimple_bind_add_stmt): Likewise.
2973 (gimple_bind_block): Likewise.
2974 (gimple_bind_set_block): Likewise.
2975 (gimple_asm_ninputs): Likewise.
2976 (gimple_asm_noutputs): Likewise.
2977 (gimple_asm_nclobbers): Likewise.
2978 (gimple_asm_nlabels): Likewise.
2979 (gimple_asm_input_op): Likewise.
2980 (gimple_asm_input_op_ptr): Likewise.
2981 (gimple_asm_output_op): Likewise.
2982 (gimple_asm_output_op_ptr): Likewise.
2983 (gimple_asm_set_output_op): Likewise.
2984 (gimple_asm_clobber_op): Likewise.
2985 (gimple_asm_set_clobber_op): Likewise.
2986 (gimple_asm_label_op): Likewise.
2987 (gimple_asm_set_label_op): Likewise.
2988 (gimple_asm_string): Likewise.
2989 (gimple_catch_types): Likewise.
2990 (gimple_catch_types_ptr): Likewise.
2991 (gimple_catch_handler_ptr): Likewise.
2992 (gimple_catch_set_types): Likewise.
2993 (gimple_catch_set_handler): Likewise.
2994 (gimple_eh_filter_types): Likewise.
2995 (gimple_eh_filter_types_ptr): Likewise.
2996 (gimple_eh_filter_failure_ptr): Likewise.
2997 (gimple_eh_filter_set_types): Likewise.
2998 (gimple_eh_filter_set_failure): Likewise.
2999 (gimple_eh_must_not_throw_fndecl): Likewise.
3000 (gimple_eh_must_not_throw_set_fndecl): Likewise.
3001 (gimple_eh_else_n_body_ptr): Likewise.
3002 (gimple_eh_else_e_body_ptr): Likewise.
3003 (gimple_eh_else_set_n_body): Likewise.
3004 (gimple_eh_else_set_e_body): Likewise.
3005 (gimple_try_eval_ptr): Likewise.
3006 (gimple_try_cleanup_ptr): Likewise.
3007 (gimple_try_set_eval): Likewise.
3008 (gimple_try_set_cleanup): Likewise.
3009 (gimple_wce_cleanup_ptr): Likewise.
3010 (gimple_wce_set_cleanup): Likewise.
3011 (gimple_phi_capacity): Likewise.
3012 (gimple_phi_num_args): Likewise.
3013 (gimple_phi_result): Likewise.
3014 (gimple_phi_result_ptr): Likewise.
3015 (gimple_phi_set_result): Likewise.
3016 (gimple_phi_arg): Likewise.
3017 (gimple_phi_set_arg): Likewise.
3018 (gimple_resx_region): Likewise.
3019 (gimple_resx_set_region): Likewise.
3020 (gimple_eh_dispatch_region): Likewise.
3021 (gimple_eh_dispatch_set_region): Likewise.
3022 (gimple_omp_critical_name): Likewise.
3023 (gimple_omp_critical_name_ptr): Likewise.
3024 (gimple_omp_critical_set_name): Likewise.
3025 (gimple_omp_for_clauses): Likewise.
3026 (gimple_omp_for_clauses_ptr): Likewise.
3027 (gimple_omp_for_set_clauses): Likewise.
3028 (gimple_omp_for_collapse): Likewise.
3029 (gimple_omp_for_index): Likewise.
3030 (gimple_omp_for_index_ptr): Likewise.
3031 (gimple_omp_for_set_index): Likewise.
3032 (gimple_omp_for_initial): Likewise.
3033 (gimple_omp_for_initial_ptr): Likewise.
3034 (gimple_omp_for_set_initial): Likewise.
3035 (gimple_omp_for_final): Likewise.
3036 (gimple_omp_for_final_ptr): Likewise.
3037 (gimple_omp_for_set_final): Likewise.
3038 (gimple_omp_for_incr): Likewise.
3039 (gimple_omp_for_incr_ptr): Likewise.
3040 (gimple_omp_for_set_incr): Likewise.
3041 (gimple_omp_for_pre_body_ptr): Likewise.
3042 (gimple_omp_for_set_pre_body): Likewise.
3043 (gimple_omp_parallel_clauses): Likewise.
3044 (gimple_omp_parallel_clauses_ptr): Likewise.
3045 (gimple_omp_parallel_set_clauses): Likewise.
3046 (gimple_omp_parallel_child_fn): Likewise.
3047 (gimple_omp_parallel_child_fn_ptr): Likewise.
3048 (gimple_omp_parallel_set_child_fn): Likewise.
3049 (gimple_omp_parallel_data_arg): Likewise.
3050 (gimple_omp_parallel_data_arg_ptr): Likewise.
3051 (gimple_omp_parallel_set_data_arg): Likewise.
3052 (gimple_omp_task_clauses): Likewise.
3053 (gimple_omp_task_clauses_ptr): Likewise.
3054 (gimple_omp_task_set_clauses): Likewise.
3055 (gimple_omp_task_child_fn): Likewise.
3056 (gimple_omp_task_child_fn_ptr): Likewise.
3057 (gimple_omp_task_set_child_fn): Likewise.
3058 (gimple_omp_task_data_arg): Likewise.
3059 (gimple_omp_task_data_arg_ptr): Likewise.
3060 (gimple_omp_task_set_data_arg): Likewise.
3061 (gimple_omp_taskreg_clauses): Likewise.
3062 (gimple_omp_taskreg_clauses_ptr): Likewise.
3063 (gimple_omp_taskreg_set_clauses): Likewise.
3064 (gimple_omp_taskreg_child_fn): Likewise.
3065 (gimple_omp_taskreg_child_fn_ptr): Likewise.
3066 (gimple_omp_taskreg_set_child_fn): Likewise.
3067 (gimple_omp_taskreg_data_arg): Likewise.
3068 (gimple_omp_taskreg_data_arg_ptr): Likewise.
3069 (gimple_omp_taskreg_set_data_arg): Likewise.
3070 (gimple_omp_task_copy_fn): Likewise.
3071 (gimple_omp_task_copy_fn_ptr): Likewise.
3072 (gimple_omp_task_set_copy_fn): Likewise.
3073 (gimple_omp_task_arg_size): Likewise.
3074 (gimple_omp_task_arg_size_ptr): Likewise.
3075 (gimple_omp_task_set_arg_size): Likewise.
3076 (gimple_omp_task_arg_align): Likewise.
3077 (gimple_omp_task_arg_align_ptr): Likewise.
3078 (gimple_omp_task_set_arg_align): Likewise.
3079 (gimple_omp_single_clauses): Likewise.
3080 (gimple_omp_single_clauses_ptr): Likewise.
3081 (gimple_omp_single_set_clauses): Likewise.
3082 (gimple_omp_target_clauses): Likewise.
3083 (gimple_omp_target_clauses_ptr): Likewise.
3084 (gimple_omp_target_set_clauses): Likewise.
3085 (gimple_omp_target_child_fn): Likewise.
3086 (gimple_omp_target_child_fn_ptr): Likewise.
3087 (gimple_omp_target_set_child_fn): Likewise.
3088 (gimple_omp_target_data_arg): Likewise.
3089 (gimple_omp_target_data_arg_ptr): Likewise.
3090 (gimple_omp_target_set_data_arg): Likewise.
3091 (gimple_omp_teams_clauses): Likewise.
3092 (gimple_omp_teams_clauses_ptr): Likewise.
3093 (gimple_omp_teams_set_clauses): Likewise.
3094 (gimple_omp_sections_clauses): Likewise.
3095 (gimple_omp_sections_clauses_ptr): Likewise.
3096 (gimple_omp_sections_set_clauses): Likewise.
3097 (gimple_omp_sections_control): Likewise.
3098 (gimple_omp_sections_control_ptr): Likewise.
3099 (gimple_omp_sections_set_control): Likewise.
3100 (gimple_omp_for_set_cond): Likewise.
3101 (gimple_omp_for_cond): Likewise.
3102 (gimple_omp_atomic_store_set_val): Likewise.
3103 (gimple_omp_atomic_store_val): Likewise.
3104 (gimple_omp_atomic_store_val_ptr): Likewise.
3105 (gimple_omp_atomic_load_set_lhs): Likewise.
3106 (gimple_omp_atomic_load_lhs): Likewise.
3107 (gimple_omp_atomic_load_lhs_ptr): Likewise.
3108 (gimple_omp_atomic_load_set_rhs): Likewise.
3109 (gimple_omp_atomic_load_rhs): Likewise.
3110 (gimple_omp_atomic_load_rhs_ptr): Likewise.
3111 (gimple_omp_continue_control_def): Likewise.
3112 (gimple_omp_continue_control_def_ptr): Likewise.
3113 (gimple_omp_continue_set_control_def): Likewise.
3114 (gimple_omp_continue_control_use): Likewise.
3115 (gimple_omp_continue_control_use_ptr): Likewise.
3116 (gimple_omp_continue_set_control_use): Likewise.
3117 (gimple_transaction_body_ptr): Likewise.
3118 (gimple_transaction_label): Likewise.
3119 (gimple_transaction_label_ptr): Likewise.
3120 (gimple_transaction_set_body): Likewise.
3121 (gimple_transaction_set_label): Likewise.
3123 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
3124 * ipa-inline-analysis.c (inline_write_summary): Likewise.
3125 * ipa-ref.c (ipa_record_reference): Likewise.
3126 * ipa-reference.c (analyze_function): Likewise.
3127 (ipa_reference_write_optimization_summary): Likewise.
3128 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
3129 (address_taken_from_non_vtable_p): Likewise.
3130 (comdat_can_be_unshared_p_1): Likewise.
3131 * lto-cgraph.c (lto_output_ref): Likewise.
3132 (add_references): Likewise.
3133 (compute_ltrans_boundary): Likewise.
3134 (output_symtab): Likewise.
3135 (input_ref): Likewise.
3136 (input_cgraph_1): Likewise.
3137 (output_cgraph_opt_summary): Likewise.
3138 * lto-streamer-out.c (lto_output): Likewise.
3139 (output_symbol_p): Likewise.
3140 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
3141 (lsei_start_function_in_partition): Likewise.
3142 (lsei_next_variable_in_partition): Likewise.
3143 (lsei_start_variable_in_partition): Likewise.
3144 * symtab.c (insert_to_assembler_name_hash): Likewise.
3145 (unlink_from_assembler_name_hash): Likewise.
3146 (symtab_unregister_node): Likewise.
3147 (symtab_remove_node): Likewise.
3148 (dump_symtab_node): Likewise.
3149 (verify_symtab_base): Likewise.
3150 (verify_symtab_node): Likewise.
3151 (symtab_make_decl_local): Likewise.
3152 (symtab_alias_ultimate_target): Likewise.
3153 (symtab_resolve_alias): Likewise.
3154 (symtab_get_symbol_partitioning_class): Likewise.
3155 * tree-phinodes.c (allocate_phi_node): Likewise.
3156 (reserve_phi_args_for_new_edge): Likewise.
3157 (remove_phi_args): Likewise.
3158 * varpool.c (varpool_node_for_asm): Likewise.
3159 (varpool_remove_unreferenced_decls): Likewise.
3161 2014-04-23 Jeff Law <law@redhat.com>
3163 PR tree-optimization/60902
3164 * tree-ssa-threadedge.c
3165 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
3166 invalidate outputs from statements that do not produce useful
3167 outputs for threading.
3169 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
3171 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
3172 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
3173 machine descriptions for Stack Smashing Protector.
3175 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
3177 * aarch64.md (<optab>_rol<mode>3): New pattern.
3178 (<optab>_rolsi3_uxtw): Likewise.
3179 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
3181 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
3183 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
3184 (arm_cortex_a12_tune): Likewise.
3186 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3188 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
3190 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3192 * config/arm/arm.md (arm_rev16si2): New pattern.
3193 (arm_rev16si2_alt): Likewise.
3194 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
3196 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3198 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
3199 (rev16<mode>2_alt): Likewise.
3200 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
3201 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
3202 (aarch_rev16_shleft_mask_imm_p): Likewise.
3203 (aarch_rev16_p_1): Likewise.
3204 (aarch_rev16_p): Likewise.
3205 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
3206 (aarch_rev16_shright_mask_imm_p): Likewise.
3207 (aarch_rev16_shleft_mask_imm_p): Likewise.
3209 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3211 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
3212 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
3214 (cortex_a53_extra_costs): Likewise.
3215 (cortex_a57_extra_costs): Likewise.
3216 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
3217 (cortexa7_extra_costs): Likewise.
3218 (cortexa8_extra_costs): Likewise.
3219 (cortexa12_extra_costs): Likewise.
3220 (cortexa15_extra_costs): Likewise.
3221 (v7m_extra_costs): Likewise.
3222 (arm_new_rtx_costs): Handle BSWAP.
3224 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3226 * config/arm/arm.c (cortexa8_extra_costs): New table.
3227 (arm_cortex_a8_tune): New tuning struct.
3228 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
3230 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3232 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
3234 2014-04-23 Richard Biener <rguenther@suse.de>
3236 * Makefile.in (OBJS): Remove loop-unswitch.o.
3237 * tree-pass.h (make_pass_rtl_unswitch): Remove.
3238 * passes.def (pass_rtl_unswitch): Likewise.
3239 * loop-init.c (gate_rtl_unswitch): Likewise.
3240 (rtl_unswitch): Likewise.
3241 (pass_data_rtl_unswitch): Likewise.
3242 (pass_rtl_unswitch): Likewise.
3243 (make_pass_rtl_unswitch): Likewise.
3244 * rtl.h (reversed_condition): Likewise.
3245 (compare_and_jump_seq): Likewise.
3246 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
3248 * loop-unroll.c (compare_and_jump_seq): Likewise.
3250 2014-04-23 Richard Biener <rguenther@suse.de>
3252 PR tree-optimization/60903
3253 * tree-ssa-loop-im.c (analyze_memory_references): Remove
3254 commented code block.
3255 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
3256 loop flags to newly created BBs and edges.
3258 2014-04-23 Nick Clifton <nickc@redhat.com>
3260 * config/msp430/msp430.c (msp430_handle_option): Move function
3262 (msp430_option_override): Simplify mcu and mcpu option handling.
3263 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
3264 support for -mhwmult command line option.
3265 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
3266 -mhwmult command line option.
3267 (msp430_hwmult_enabled): Delete.
3268 (msp43o_output_labelref): Add support for -mhwmult command line option.
3269 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
3270 (umulsidi3): Likewise.
3271 * config/msp430/msp430.opt (mmcu): Add Report attribute.
3272 (mcpu, mlarge, msmall): Likewise.
3273 (mhwmult): New option.
3274 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
3276 (msp430_is_f5_mcu): Remove prototype.
3277 (msp430_use_f5_series_hwmult): Add prototype.
3278 * config/msp430/msp430-opts.h: New file.
3279 * common/config/msp430: New directory.
3280 * common/config/msp430/msp430-common.c: New file.
3281 * config.gcc (msp430): Remove target_has_targetm_common.
3282 * doc/invoke.texi: Document -mhwmult command line option.
3284 2014-04-23 Nick Clifton <nickc@redhat.com>
3286 * config/i386/cygwin.h (ENDFILE_SPEC): Include
3287 default-manifest.o if it can be found in the search path.
3288 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
3290 2014-04-23 Terry Guo <terry.guo@arm.com>
3292 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
3294 2014-04-23 Richard Biener <rguenther@suse.de>
3297 * tree-inline.c (declare_return_variable): Use mark_addressable.
3299 2014-04-23 Richard Biener <rguenther@suse.de>
3302 * loop-init.c (loop_optimizer_init): Make sure to apply
3303 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
3305 2014-04-22 Jakub Jelinek <jakub@redhat.com>
3308 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
3310 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
3311 if flag_sanitize_undefined_trap_on_error.
3312 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
3313 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
3314 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
3315 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
3316 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
3317 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
3318 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
3319 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
3320 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
3321 * ubsan.c (ubsan_instrument_unreachable): Return
3322 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
3323 (ubsan_expand_null_ifn): Emit __builtin_trap ()
3324 if flag_sanitize_undefined_trap_on_error and
3325 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
3326 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
3327 instrument_bool_enum_load): Emit __builtin_trap () if
3328 flag_sanitize_undefined_trap_on_error and
3329 __builtin_handle_*_abort () if !flag_sanitize_recover.
3330 * doc/invoke.texi (-fsanitize-recover,
3331 -fsanitize-undefined-trap-on-error): Document.
3333 2014-04-22 Christian Bruel <christian.bruel@st.com>
3335 * config/sh/sh.md (mov<mode>): Replace movQIHI.
3336 Force immediates to SImode.
3338 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
3340 * config/nios2/nios2.md (UNSPEC_ROUND): New.
3342 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
3343 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
3344 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
3345 (nios2_fpu_insn): Add entry for round.
3346 (N2FPU_NO_ERRNO_P): Define.
3347 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
3349 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
3351 2014-04-22 Richard Henderson <rth@redhat.com>
3353 * config/aarch64/aarch64 (addti3, subti3): New expanders.
3354 (add<GPI>3_compare0): Remove leading * from name.
3355 (add<GPI>3_carryin): Likewise.
3356 (sub<GPI>3_compare0): Likewise.
3357 (sub<GPI>3_carryin): Likewise.
3358 (<su_optab>mulditi3): New expander.
3359 (multi3): New expander.
3360 (madd<GPI>): Remove leading * from name.
3362 2014-04-22 Martin Jambor <mjambor@suse.cz>
3364 * cgraphclones.c (cgraph_function_versioning): Copy
3365 ipa_transforms_to_apply instead of asserting it is empty.
3367 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
3370 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
3371 on count_exp to get mode.
3373 2014-04-22 Andrew Pinski <apinski@cavium.com>
3375 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
3376 Handle TLS for ILP32.
3377 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
3378 (tlsie_small_<mode>): this and handle PTR.
3379 (tlsie_small_sidi): New pattern.
3380 (tlsle_small): Change to an expand to handle ILP32.
3381 (tlsle_small_<mode>): New pattern.
3382 (tlsdesc_small): Rename to ...
3383 (tlsdesc_small_<mode>): this and handle PTR.
3385 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3387 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
3389 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
3391 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
3392 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
3393 (aarch64_types_signed_poly_qualifiers): Likewise.
3394 (aarch64_types_unsigned_signed_qualifiers): Likewise.
3395 (aarch64_types_poly_signed_qualifiers): Likewise.
3396 (TYPES_REINTERP_SS): Type macro added.
3397 (TYPES_REINTERP_SU): Likewise.
3398 (TYPES_REINTERP_SP): Likewise.
3399 (TYPES_REINTERP_US): Likewise.
3400 (TYPES_REINTERP_PS): Likewise.
3401 (aarch64_fold_builtin): New expression folding added.
3402 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
3403 Declarations removed.
3404 (REINTERP_SS): Declarations added.
3405 (REINTERP_US): Likewise.
3406 (REINTERP_PS): Likewise.
3407 (REINTERP_SU): Likewise.
3408 (REINTERP_SP): Likewise.
3409 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
3410 (vreinterpretq_p8_f64): Likewise.
3411 (vreinterpret_p16_f64): Likewise.
3412 (vreinterpretq_p16_f64): Likewise.
3413 (vreinterpret_f32_f64): Likewise.
3414 (vreinterpretq_f32_f64): Likewise.
3415 (vreinterpret_f64_f32): Likewise.
3416 (vreinterpret_f64_p8): Likewise.
3417 (vreinterpret_f64_p16): Likewise.
3418 (vreinterpret_f64_s8): Likewise.
3419 (vreinterpret_f64_s16): Likewise.
3420 (vreinterpret_f64_s32): Likewise.
3421 (vreinterpret_f64_s64): Likewise.
3422 (vreinterpret_f64_u8): Likewise.
3423 (vreinterpret_f64_u16): Likewise.
3424 (vreinterpret_f64_u32): Likewise.
3425 (vreinterpret_f64_u64): Likewise.
3426 (vreinterpretq_f64_f32): Likewise.
3427 (vreinterpretq_f64_p8): Likewise.
3428 (vreinterpretq_f64_p16): Likewise.
3429 (vreinterpretq_f64_s8): Likewise.
3430 (vreinterpretq_f64_s16): Likewise.
3431 (vreinterpretq_f64_s32): Likewise.
3432 (vreinterpretq_f64_s64): Likewise.
3433 (vreinterpretq_f64_u8): Likewise.
3434 (vreinterpretq_f64_u16): Likewise.
3435 (vreinterpretq_f64_u32): Likewise.
3436 (vreinterpretq_f64_u64): Likewise.
3437 (vreinterpret_s64_f64): Likewise.
3438 (vreinterpretq_s64_f64): Likewise.
3439 (vreinterpret_u64_f64): Likewise.
3440 (vreinterpretq_u64_f64): Likewise.
3441 (vreinterpret_s8_f64): Likewise.
3442 (vreinterpretq_s8_f64): Likewise.
3443 (vreinterpret_s16_f64): Likewise.
3444 (vreinterpretq_s16_f64): Likewise.
3445 (vreinterpret_s32_f64): Likewise.
3446 (vreinterpretq_s32_f64): Likewise.
3447 (vreinterpret_u8_f64): Likewise.
3448 (vreinterpretq_u8_f64): Likewise.
3449 (vreinterpret_u16_f64): Likewise.
3450 (vreinterpretq_u16_f64): Likewise.
3451 (vreinterpret_u32_f64): Likewise.
3452 (vreinterpretq_u32_f64): Likewise.
3454 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
3456 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
3457 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
3458 (vreinterpret_p8_s8): Likewise.
3459 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
3460 (vreinterpret_p8_s16): Likewise.
3461 (vreinterpret_p8_s32): Likewise.
3462 (vreinterpret_p8_s64): Likewise.
3463 (vreinterpret_p8_f32): Likewise.
3464 (vreinterpret_p8_u8): Likewise.
3465 (vreinterpret_p8_u16): Likewise.
3466 (vreinterpret_p8_u32): Likewise.
3467 (vreinterpret_p8_u64): Likewise.
3468 (vreinterpret_p8_p16): Likewise.
3469 (vreinterpretq_p8_s8): Likewise.
3470 (vreinterpretq_p8_s16): Likewise.
3471 (vreinterpretq_p8_s32): Likewise.
3472 (vreinterpretq_p8_s64): Likewise.
3473 (vreinterpretq_p8_f32): Likewise.
3474 (vreinterpretq_p8_u8): Likewise.
3475 (vreinterpretq_p8_u16): Likewise.
3476 (vreinterpretq_p8_u32): Likewise.
3477 (vreinterpretq_p8_u64): Likewise.
3478 (vreinterpretq_p8_p16): Likewise.
3479 (vreinterpret_p16_s8): Likewise.
3480 (vreinterpret_p16_s16): Likewise.
3481 (vreinterpret_p16_s32): Likewise.
3482 (vreinterpret_p16_s64): Likewise.
3483 (vreinterpret_p16_f32): Likewise.
3484 (vreinterpret_p16_u8): Likewise.
3485 (vreinterpret_p16_u16): Likewise.
3486 (vreinterpret_p16_u32): Likewise.
3487 (vreinterpret_p16_u64): Likewise.
3488 (vreinterpret_p16_p8): Likewise.
3489 (vreinterpretq_p16_s8): Likewise.
3490 (vreinterpretq_p16_s16): Likewise.
3491 (vreinterpretq_p16_s32): Likewise.
3492 (vreinterpretq_p16_s64): Likewise.
3493 (vreinterpretq_p16_f32): Likewise.
3494 (vreinterpretq_p16_u8): Likewise.
3495 (vreinterpretq_p16_u16): Likewise.
3496 (vreinterpretq_p16_u32): Likewise.
3497 (vreinterpretq_p16_u64): Likewise.
3498 (vreinterpretq_p16_p8): Likewise.
3499 (vreinterpret_f32_s8): Likewise.
3500 (vreinterpret_f32_s16): Likewise.
3501 (vreinterpret_f32_s32): Likewise.
3502 (vreinterpret_f32_s64): Likewise.
3503 (vreinterpret_f32_u8): Likewise.
3504 (vreinterpret_f32_u16): Likewise.
3505 (vreinterpret_f32_u32): Likewise.
3506 (vreinterpret_f32_u64): Likewise.
3507 (vreinterpret_f32_p8): Likewise.
3508 (vreinterpret_f32_p16): Likewise.
3509 (vreinterpretq_f32_s8): Likewise.
3510 (vreinterpretq_f32_s16): Likewise.
3511 (vreinterpretq_f32_s32): Likewise.
3512 (vreinterpretq_f32_s64): Likewise.
3513 (vreinterpretq_f32_u8): Likewise.
3514 (vreinterpretq_f32_u16): Likewise.
3515 (vreinterpretq_f32_u32): Likewise.
3516 (vreinterpretq_f32_u64): Likewise.
3517 (vreinterpretq_f32_p8): Likewise.
3518 (vreinterpretq_f32_p16): Likewise.
3519 (vreinterpret_s64_s8): Likewise.
3520 (vreinterpret_s64_s16): Likewise.
3521 (vreinterpret_s64_s32): Likewise.
3522 (vreinterpret_s64_f32): Likewise.
3523 (vreinterpret_s64_u8): Likewise.
3524 (vreinterpret_s64_u16): Likewise.
3525 (vreinterpret_s64_u32): Likewise.
3526 (vreinterpret_s64_u64): Likewise.
3527 (vreinterpret_s64_p8): Likewise.
3528 (vreinterpret_s64_p16): Likewise.
3529 (vreinterpretq_s64_s8): Likewise.
3530 (vreinterpretq_s64_s16): Likewise.
3531 (vreinterpretq_s64_s32): Likewise.
3532 (vreinterpretq_s64_f32): Likewise.
3533 (vreinterpretq_s64_u8): Likewise.
3534 (vreinterpretq_s64_u16): Likewise.
3535 (vreinterpretq_s64_u32): Likewise.
3536 (vreinterpretq_s64_u64): Likewise.
3537 (vreinterpretq_s64_p8): Likewise.
3538 (vreinterpretq_s64_p16): Likewise.
3539 (vreinterpret_u64_s8): Likewise.
3540 (vreinterpret_u64_s16): Likewise.
3541 (vreinterpret_u64_s32): Likewise.
3542 (vreinterpret_u64_s64): Likewise.
3543 (vreinterpret_u64_f32): Likewise.
3544 (vreinterpret_u64_u8): Likewise.
3545 (vreinterpret_u64_u16): Likewise.
3546 (vreinterpret_u64_u32): Likewise.
3547 (vreinterpret_u64_p8): Likewise.
3548 (vreinterpret_u64_p16): Likewise.
3549 (vreinterpretq_u64_s8): Likewise.
3550 (vreinterpretq_u64_s16): Likewise.
3551 (vreinterpretq_u64_s32): Likewise.
3552 (vreinterpretq_u64_s64): Likewise.
3553 (vreinterpretq_u64_f32): Likewise.
3554 (vreinterpretq_u64_u8): Likewise.
3555 (vreinterpretq_u64_u16): Likewise.
3556 (vreinterpretq_u64_u32): Likewise.
3557 (vreinterpretq_u64_p8): Likewise.
3558 (vreinterpretq_u64_p16): Likewise.
3559 (vreinterpret_s8_s16): Likewise.
3560 (vreinterpret_s8_s32): Likewise.
3561 (vreinterpret_s8_s64): Likewise.
3562 (vreinterpret_s8_f32): Likewise.
3563 (vreinterpret_s8_u8): Likewise.
3564 (vreinterpret_s8_u16): Likewise.
3565 (vreinterpret_s8_u32): Likewise.
3566 (vreinterpret_s8_u64): Likewise.
3567 (vreinterpret_s8_p8): Likewise.
3568 (vreinterpret_s8_p16): Likewise.
3569 (vreinterpretq_s8_s16): Likewise.
3570 (vreinterpretq_s8_s32): Likewise.
3571 (vreinterpretq_s8_s64): Likewise.
3572 (vreinterpretq_s8_f32): Likewise.
3573 (vreinterpretq_s8_u8): Likewise.
3574 (vreinterpretq_s8_u16): Likewise.
3575 (vreinterpretq_s8_u32): Likewise.
3576 (vreinterpretq_s8_u64): Likewise.
3577 (vreinterpretq_s8_p8): Likewise.
3578 (vreinterpretq_s8_p16): Likewise.
3579 (vreinterpret_s16_s8): Likewise.
3580 (vreinterpret_s16_s32): Likewise.
3581 (vreinterpret_s16_s64): Likewise.
3582 (vreinterpret_s16_f32): Likewise.
3583 (vreinterpret_s16_u8): Likewise.
3584 (vreinterpret_s16_u16): Likewise.
3585 (vreinterpret_s16_u32): Likewise.
3586 (vreinterpret_s16_u64): Likewise.
3587 (vreinterpret_s16_p8): Likewise.
3588 (vreinterpret_s16_p16): Likewise.
3589 (vreinterpretq_s16_s8): Likewise.
3590 (vreinterpretq_s16_s32): Likewise.
3591 (vreinterpretq_s16_s64): Likewise.
3592 (vreinterpretq_s16_f32): Likewise.
3593 (vreinterpretq_s16_u8): Likewise.
3594 (vreinterpretq_s16_u16): Likewise.
3595 (vreinterpretq_s16_u32): Likewise.
3596 (vreinterpretq_s16_u64): Likewise.
3597 (vreinterpretq_s16_p8): Likewise.
3598 (vreinterpretq_s16_p16): Likewise.
3599 (vreinterpret_s32_s8): Likewise.
3600 (vreinterpret_s32_s16): Likewise.
3601 (vreinterpret_s32_s64): Likewise.
3602 (vreinterpret_s32_f32): Likewise.
3603 (vreinterpret_s32_u8): Likewise.
3604 (vreinterpret_s32_u16): Likewise.
3605 (vreinterpret_s32_u32): Likewise.
3606 (vreinterpret_s32_u64): Likewise.
3607 (vreinterpret_s32_p8): Likewise.
3608 (vreinterpret_s32_p16): Likewise.
3609 (vreinterpretq_s32_s8): Likewise.
3610 (vreinterpretq_s32_s16): Likewise.
3611 (vreinterpretq_s32_s64): Likewise.
3612 (vreinterpretq_s32_f32): Likewise.
3613 (vreinterpretq_s32_u8): Likewise.
3614 (vreinterpretq_s32_u16): Likewise.
3615 (vreinterpretq_s32_u32): Likewise.
3616 (vreinterpretq_s32_u64): Likewise.
3617 (vreinterpretq_s32_p8): Likewise.
3618 (vreinterpretq_s32_p16): Likewise.
3619 (vreinterpret_u8_s8): Likewise.
3620 (vreinterpret_u8_s16): Likewise.
3621 (vreinterpret_u8_s32): Likewise.
3622 (vreinterpret_u8_s64): Likewise.
3623 (vreinterpret_u8_f32): Likewise.
3624 (vreinterpret_u8_u16): Likewise.
3625 (vreinterpret_u8_u32): Likewise.
3626 (vreinterpret_u8_u64): Likewise.
3627 (vreinterpret_u8_p8): Likewise.
3628 (vreinterpret_u8_p16): Likewise.
3629 (vreinterpretq_u8_s8): Likewise.
3630 (vreinterpretq_u8_s16): Likewise.
3631 (vreinterpretq_u8_s32): Likewise.
3632 (vreinterpretq_u8_s64): Likewise.
3633 (vreinterpretq_u8_f32): Likewise.
3634 (vreinterpretq_u8_u16): Likewise.
3635 (vreinterpretq_u8_u32): Likewise.
3636 (vreinterpretq_u8_u64): Likewise.
3637 (vreinterpretq_u8_p8): Likewise.
3638 (vreinterpretq_u8_p16): Likewise.
3639 (vreinterpret_u16_s8): Likewise.
3640 (vreinterpret_u16_s16): Likewise.
3641 (vreinterpret_u16_s32): Likewise.
3642 (vreinterpret_u16_s64): Likewise.
3643 (vreinterpret_u16_f32): Likewise.
3644 (vreinterpret_u16_u8): Likewise.
3645 (vreinterpret_u16_u32): Likewise.
3646 (vreinterpret_u16_u64): Likewise.
3647 (vreinterpret_u16_p8): Likewise.
3648 (vreinterpret_u16_p16): Likewise.
3649 (vreinterpretq_u16_s8): Likewise.
3650 (vreinterpretq_u16_s16): Likewise.
3651 (vreinterpretq_u16_s32): Likewise.
3652 (vreinterpretq_u16_s64): Likewise.
3653 (vreinterpretq_u16_f32): Likewise.
3654 (vreinterpretq_u16_u8): Likewise.
3655 (vreinterpretq_u16_u32): Likewise.
3656 (vreinterpretq_u16_u64): Likewise.
3657 (vreinterpretq_u16_p8): Likewise.
3658 (vreinterpretq_u16_p16): Likewise.
3659 (vreinterpret_u32_s8): Likewise.
3660 (vreinterpret_u32_s16): Likewise.
3661 (vreinterpret_u32_s32): Likewise.
3662 (vreinterpret_u32_s64): Likewise.
3663 (vreinterpret_u32_f32): Likewise.
3664 (vreinterpret_u32_u8): Likewise.
3665 (vreinterpret_u32_u16): Likewise.
3666 (vreinterpret_u32_u64): Likewise.
3667 (vreinterpret_u32_p8): Likewise.
3668 (vreinterpret_u32_p16): Likewise.
3669 (vreinterpretq_u32_s8): Likewise.
3670 (vreinterpretq_u32_s16): Likewise.
3671 (vreinterpretq_u32_s32): Likewise.
3672 (vreinterpretq_u32_s64): Likewise.
3673 (vreinterpretq_u32_f32): Likewise.
3674 (vreinterpretq_u32_u8): Likewise.
3675 (vreinterpretq_u32_u16): Likewise.
3676 (vreinterpretq_u32_u64): Likewise.
3677 (vreinterpretq_u32_p8): Likewise.
3678 (vreinterpretq_u32_p16): Likewise.
3680 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
3682 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
3684 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator extended.
3686 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
3687 (vqnegd_s64): Likewise.
3688 (vqabs_s64): Likewise.
3689 (vqabsd_s64): Likewise.
3691 2014-04-22 Richard Henderson <rth@redhat.com>
3693 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
3694 computation to the top of the loop.
3696 2014-04-22 Renlin <renlin.li@arm.com>
3697 Jiong Wang <jiong.wang@arm.com>
3699 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
3700 * config/aarch64/aarch64.c (aarch64_layout_frame)
3701 (aarch64_initial_elimination_offset): Likewise.
3703 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
3705 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
3708 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
3710 * machmode.h (bitwise_mode_for_mode): Declare.
3711 * stor-layout.h (bitwise_type_for_mode): Likewise.
3712 * stor-layout.c (bitwise_mode_for_mode): New function.
3713 (bitwise_type_for_mode): Likewise.
3714 * builtins.c (fold_builtin_memory_op): Use it instead of
3715 int_mode_for_mode and build_nonstandard_integer_type.
3717 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3719 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
3720 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
3721 (*-*-solaris2*): Simplify.
3722 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
3723 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
3724 *-*-solaris2.9* handling.
3726 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
3728 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
3729 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
3731 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
3732 * configure: Regenerate.
3734 * config/i386/sol2-9.h: Remove.
3736 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
3737 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
3738 Remove Solaris 9 references.
3740 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
3742 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
3743 (floatuns<GPI:mode><GPF:mode>2): Remove.
3744 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
3745 and floatuns conversions.
3746 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
3747 and floatuns conversions.
3748 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
3749 (w1,w2): New mode attributes for inequal width conversions.
3751 2014-04-22 Renlin Li <Renlin.Li@arm.com>
3753 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
3754 the output asm format.
3756 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
3758 * config/aarch64/aarch64-simd.md
3759 (aarch64_cm<optab>di): Always split.
3760 (*aarch64_cm<optab>di): New.
3761 (aarch64_cmtstdi): Always split.
3762 (*aarch64_cmtstdi): New.
3764 2014-04-22 Jakub Jelinek <jakub@redhat.com>
3766 PR tree-optimization/60823
3767 * omp-low.c (ipa_simd_modify_function_body): Go through
3768 all SSA_NAMEs and for those refering to vector arguments
3769 which are going to be replaced adjust SSA_NAME_VAR and,
3770 if it is a default definition, change it into a non-default
3771 definition assigned at the beginning of function from new_decl.
3772 (ipa_simd_modify_stmt_ops): Rewritten.
3773 * tree-dfa.c (set_ssa_default_def): When removing default def,
3774 check for NULL loc instead of NULL *loc.
3776 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3778 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
3779 restrictions on core registers for DImode values in Thumb2.
3781 2014-04-22 Ian Bolton <ian.bolton@arm.com>
3783 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
3784 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
3786 2014-04-22 Ian Bolton <ian.bolton@arm.com>
3788 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
3789 (*iordi_notzesidi_di): Likewise.
3790 (*iordi_notsesidi_di): Likewise.
3792 2014-04-22 Ian Bolton <ian.bolton@arm.com>
3794 * config/arm/arm-protos.h (tune_params): New struct members.
3795 * config/arm/arm.c: Initialise tune_params per processor.
3796 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
3797 for speed, based on new tune_params.
3799 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
3801 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro added.
3802 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added macro.
3803 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment corrected.
3804 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
3805 * config/aarch64/arm_neon.h (vrnd_f64): Added.
3806 (vrnda_f64): Likewise.
3807 (vrndi_f64): Likewise.
3808 (vrndm_f64): Likewise.
3809 (vrndn_f64): Likewise.
3810 (vrndp_f64): Likewise.
3811 (vrndx_f64): Likewise.
3813 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3815 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
3816 GET_MODE_SIZE argument is enum machine_mode.
3818 2014-04-22 Jakub Jelinek <jakub@redhat.com>
3821 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
3822 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
3824 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
3827 * asan.c (asan_emit_stack_protection): Force the base to align to
3828 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
3829 appropriate bits if STRICT_ALIGNMENT.
3830 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
3832 (expand_used_vars): Leave a space in the stack frame for alignment
3833 if STRICT_ALIGNMENT.
3835 2014-04-21 David Malcolm <dmalcolm@redhat.com>
3837 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
3839 (gimple_store_p): Likewise.
3840 (gimple_assign_load_p): Likewise.
3841 (gimple_assign_cast_p): Likewise.
3842 (gimple_clobber_p): Likewise.
3844 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
3845 rather than a gimple.
3846 (gimple_assign_cast_p): Likewise.
3848 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
3851 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
3852 If mode is DDmode and TARGET_E500_DOUBLE allow move.
3854 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
3855 more debug information for E500 if -mdebug=reg.
3857 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
3860 * config/i386/i386.c (ix86_expand_builtin)
3861 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
3862 register for target RTX.
3863 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
3865 2014-04-18 Cong Hou <congh@google.com>
3867 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
3868 the widen-mult pattern by handling two operands with different sizes,
3869 and operands whose size is smaller than half of the result type.
3871 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
3873 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
3874 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
3875 (do_estimate_edge_time): Compute it.
3876 * ipa-inline.c (want_inline_small_function_p): Bypass
3877 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
3879 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
3881 * ipa-inline.c (spec_rem): New static variable.
3882 (dump_overall_stats): New function.
3883 (dump_inline_stats): New function.
3885 2014-04-18 Richard Henderson <rth@redhat.com>
3887 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
3888 to GET_MODE_SIZE, not a reg_class_t.
3890 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3892 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
3893 (vsx_xxmrglw_<mode>): Likewise.
3895 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
3898 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
3899 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
3900 (rs6000_init_hard_regno_mode_ok): Likewise.
3902 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
3904 * ipa-inline.c (inline_small_functions): Account only non-cold
3906 * doc/invoke.texi (inline-unit-growth): Update documentation.
3908 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
3910 * config/rs6000/rs6000.md (addti3, subti3): New.
3912 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
3915 * config/i386/i386.c (ix86_expand_clear): Remove outdated
3916 comment. Check optimize_insn_for_size_p instead of
3917 optimize_insn_for_speed_p.
3919 2014-04-17 Martin Jambor <mjambor@suse.cz>
3921 * gimple-iterator.c (gsi_start_edge): New function.
3922 * gimple-iterator.h (gsi_start_edge): Declare.
3923 * tree-sra.c (single_non_eh_succ): New function.
3924 (disqualify_ops_if_throwing_stmt): Renamed to
3925 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
3926 having one non-EH successor BB.
3927 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
3928 generate loads into replacements.
3929 (sra_modify_assign): Likewise and and also use the simple path for
3931 (sra_modify_function_body): Commit statements on edges.
3933 2014-04-17 Richard Biener <rguenther@suse.de>
3936 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
3937 comparison results and add clarifying comment.
3939 2014-04-17 Jakub Jelinek <jakub@redhat.com>
3941 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
3942 (blank_mode): Initialize it.
3943 (emit_mode_size_inline, emit_mode_nunits_inline,
3944 emit_mode_inner_inline): New functions.
3945 (emit_insn_modes_h): Call them and surround their output with
3946 #if GCC_VERSION >= 4001 ... #endif.
3947 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
3948 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
3949 mode_* arrays if the argument is __builtin_constant_p.
3950 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
3951 is enum machine_mode.
3953 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3955 * passes.c (opt_pass::execute): Adjust.
3956 (pass_manager::execute_pass_mode_switching): Likewise.
3957 (early_local_passes::execute): Likewise.
3958 (execute_one_pass): Pass cfun to the pass's execute method.
3959 * tree-pass.h (opt_pass::execute): Add function * argument.
3960 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
3961 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
3962 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
3963 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
3964 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
3965 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
3966 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
3967 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
3968 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
3969 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
3970 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
3971 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
3972 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
3973 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
3974 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
3975 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
3976 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
3977 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
3978 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
3979 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3980 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
3981 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
3982 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
3983 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3984 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3985 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
3986 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
3987 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
3988 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
3991 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
3993 * passes.c (opt_pass::gate): Take function * argument.
3994 (gate_all_early_local_passes): Merge into
3995 (early_local_passes::gate): this.
3996 (gate_all_early_optimizations): Merge into
3997 (all_early_optimizations::gate): this.
3998 (gate_all_optimizations): Mege into
3999 (all_optimizations::gate): this.
4000 (gate_all_optimizations_g): Merge into
4001 (all_optimizations_g::gate): this.
4002 (gate_rest_of_compilation): Mege into
4003 (rest_of_compilation::gate): this.
4004 (gate_postreload): Merge into
4005 (postreload::gate): this.
4006 (dump_one_pass): Pass cfun to the pass's gate method.
4007 (execute_ipa_summary_passes): Likewise.
4008 (execute_one_pass): Likewise.
4009 (ipa_write_summaries_2): Likewise.
4010 (ipa_write_optimization_summaries_1): Likewise.
4011 (ipa_read_summaries_1): Likewise.
4012 (ipa_read_optimization_summaries_1): Likewise.
4013 (execute_ipa_stmt_fixups): Likewise.
4014 * tree-pass.h (opt_pass::gate): Add function * argument.
4015 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
4016 combine-stack-adj.c, combine.c, compare-elim.c,
4017 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
4018 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
4019 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
4020 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
4021 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
4022 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
4023 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
4024 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
4025 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
4026 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
4027 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
4028 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
4029 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
4030 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
4031 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
4032 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
4033 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
4034 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
4035 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
4036 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
4037 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
4038 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
4039 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
4040 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
4041 var-tracking.c, vtable-verify.c, web.c: Adjust.
4043 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
4045 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
4046 * configure: Regenerate.
4048 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
4050 * passes.c (dump_one_pass): don't check pass->has_gate.
4051 (execute_ipa_summary_passes): Likewise.
4052 (execute_one_pass): Likewise.
4053 (ipa_write_summaries_2): Likewise.
4054 (ipa_write_optimization_summaries_1): Likewise.
4055 (ipa_read_optimization_summaries_1): Likewise.
4056 (execute_ipa_stmt_fixups): Likewise.
4057 * tree-pass.h (pass_data::has_gate): Remove.
4058 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
4059 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
4060 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
4061 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
4062 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
4063 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
4064 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
4065 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
4066 gimple-low.c, gimple-ssa-isolate-paths.c,
4067 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
4068 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
4069 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
4070 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
4071 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
4072 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
4073 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
4074 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
4075 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
4076 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
4077 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
4078 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
4079 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
4080 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
4081 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
4082 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
4083 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
4084 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
4085 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
4086 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
4087 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
4088 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
4091 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
4093 * pass_manager.h (pass_manager::register_dump_files_1): Remove
4095 * passes.c (pass_manager::register_dump_files_1): Merge into
4096 (pass_manager::register_dump_files): this, and remove its handling of
4097 properties since the pass always has the properties anyway.
4098 (pass_manager::pass_manager): Adjust.
4100 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
4102 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
4103 * passes.c (pass_manager::register_dump_files_1): Remove dead code
4104 dealing with properties.
4105 (pass_manager::register_dump_files): Adjust.
4107 2014-03-20 Mark Wielaard <mjw@redhat.com>
4109 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
4110 then represent the bound as normal constant value.
4112 2014-04-17 Jakub Jelinek <jakub@redhat.com>
4115 Forward port from 4.8 branch
4116 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
4118 * config/i386/bmiintrin.h (_blsi_u32): New.
4122 (_blsmsk_u32): Ditto.
4123 (_blsmsk_u64): Ditto.
4124 (_tzcnt_u32): Ditto.
4125 (_tzcnt_u64): Ditto.
4127 2014-04-17 Kito Cheng <kito@0xlab.org>
4129 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
4131 2014-04-17 Richard Biener <rguenther@suse.de>
4134 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
4135 boolean results for comparisons.
4137 2014-04-17 Richard Biener <rguenther@suse.de>
4139 PR tree-optimization/60836
4140 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
4141 initial PHI args to be gimple values.
4143 2014-04-17 Richard Biener <rguenther@suse.de>
4145 PR tree-optimization/60841
4146 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
4147 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
4148 of stmts to SLP build.
4149 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
4150 (vect_analyze_slp): Likewise.
4151 (vect_analyze_slp_instance): Likewise.
4152 (vect_build_slp_tree): Limit overall SLP tree growth.
4153 * tree-vectorizer.h (vect_analyze_data_refs,
4154 vect_analyze_slp): Adjust prototypes.
4156 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
4158 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
4161 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
4163 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
4164 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
4165 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
4166 for TARGET_SLOW_PSHUFB
4168 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
4170 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
4171 * config/i386/i386.c (intel_cost): Ditto.
4173 2014-04-17 Joey Ye <joey.ye@arm.com>
4175 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
4177 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
4179 * opts.c (common_handle_option): Disable -fipa-reference coorectly
4182 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
4184 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
4185 (type_all_derivations_known_p): New predicate.
4186 (type_all_ctors_visible_p): New predicate.
4187 (type_possibly_instantiated_p): New predicate.
4188 (get_odr_type): Compute all_derivations_known.
4189 (dump_odr_type): Dump the flag.
4190 (maybe_record_type): Cleanup.
4191 (record_target_from_binfo): Add bases_to_consider array;
4192 record bases for types w/o instances and skip CXX destructor.
4193 (possible_polymorphic_call_targets_1): Add bases_to_consider
4194 and consider_construction parameters; check if type may have instance.
4195 (get_polymorphic_call_info): Set maybe_in_construction to true
4196 when we know nothing.
4197 (record_targets_from_bases): Skip CXX destructors; they are
4198 never called for types in construction.
4199 (possible_polymorphic_call_targets): Do not record target when
4200 type may not have instance.
4202 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
4205 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
4206 external aliases alive, too.
4208 2014-04-16 Andrew Pinski <apinski@cavium.com>
4210 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
4213 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
4215 * final.c (compute_alignments): Do not apply loop alignment to a block
4216 falling through to the exit.
4218 2014-04-16 Catherine Moore <clm@codesourcery.com>
4220 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
4221 Adjust constraints for microMIPS store patterns.
4223 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
4225 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
4227 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
4229 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
4230 (append_use): Run at -O0.
4231 (append_vdef): Likewise.
4232 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
4233 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
4235 2014-04-16 Jakub Jelinek <jakub@redhat.com>
4237 PR tree-optimization/60844
4238 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
4239 (propagate_op_to_single_use, remove_visited_stmt_chain,
4240 linearize_expr, repropagate_negates, reassociate_bb): Use it
4241 instead of gsi_remove.
4243 2014-04-16 Martin Jambor <mjambor@suse.cz>
4245 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
4246 ipa_transforms_to_apply.
4247 (cgraph_function_versioning): Assert that old_node has empty
4248 ipa_transforms_to_apply.
4249 * trans-mem.c (ipa_tm_create_version): Likewise.
4250 * tree-inline.c (tree_function_versioning): Do not duplicate
4251 ipa_transforms_to_apply.
4253 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4256 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
4258 Pass necessary as flags on 64-bit Solaris/x86.
4259 Use lowercase relocs for x86_64-*-*.
4260 * configure: Regenerate.
4262 2014-04-15 Jan Hubicka <jh@suse.cz>
4264 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
4265 (maybe_record_node, likely_target_p): Use it.
4267 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4270 Revert following patch
4272 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
4275 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
4276 software floating point or no floating point registers, do not
4277 allow any type in the FPRs. Eliminate a test for SPE SIMD types
4278 in GPRs that occurs after we tested for GPRs that would never be
4281 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
4282 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
4283 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
4284 specifically allow DDmode, since that does not use the SPE SIMD
4287 2014-03-21 Mark Wielaard <mjw@redhat.com>
4289 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
4290 as unsigned or int depending on type and value used.
4292 2014-04-15 Richard Biener <rguenther@suse.de>
4294 PR rtl-optimization/56965
4295 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
4296 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
4298 * alias.c (true_dependence_1): Do not call
4299 nonoverlapping_component_refs_p.
4300 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
4301 nonoverlapping_component_refs_p.
4302 (indirect_refs_may_alias_p): Likewise.
4304 2014-04-15 Teresa Johnson <tejohnson@google.com>
4306 * cfg.c (dump_bb_info): Fix flags check.
4307 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
4309 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4311 PR rtl-optimization/60663
4312 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
4315 2014-04-15 Richard Biener <rguenther@suse.de>
4317 * lto-streamer.h (LTO_major_version): Bump to 4.
4319 2014-04-15 Richard Biener <rguenther@suse.de>
4321 * common.opt (lto_partition_model): New enum.
4322 (flto-partition=): Merge separate options with a single with argument,
4323 add -flto-partition=one support.
4324 * flag-types.h (enum lto_partition_model): Declare.
4325 * opts.c (finish_options): Remove duplicate -flto-partition=
4327 * lto-wrapper.c (run_gcc): Adjust.
4329 2014-04-15 Richard Biener <rguenther@suse.de>
4331 * alias.c (ncr_compar): New function.
4332 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
4334 2014-04-15 Richard Biener <rguenther@suse.de>
4336 * alias.c (record_component_aliases): Do not walk BINFOs.
4338 2014-04-15 Richard Biener <rguenther@suse.de>
4340 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
4341 Add struct function argument and adjust.
4342 (find_func_aliases_for_call): Likewise.
4343 (find_func_aliases): Likewise.
4344 (find_func_clobbers): Likewise.
4345 (intra_create_variable_infos): Likewise.
4346 (compute_points_to_sets): Likewise.
4347 (ipa_pta_execute): Adjust. Do not push/pop cfun.
4349 2014-04-15 Richard Biener <rguenther@suse.de>
4351 * tree.c (iterative_hash_expr): Use enum tree_code_class
4352 to store TREE_CODE_CLASS.
4353 (tree_block): Likewise.
4354 (tree_set_block): Likewise.
4355 * tree.h (fold_build_pointer_plus_loc): Use
4356 convert_to_ptrofftype_loc.
4358 2014-04-15 Jakub Jelinek <jakub@redhat.com>
4361 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
4364 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
4366 * cfgloop.h (struct loop): Move force_vectorize down.
4367 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
4368 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
4369 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
4370 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
4371 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
4372 * tree-core.h (enum annot_expr_kind): Add new kind values.
4373 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
4374 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
4376 * tree.def (ANNOTATE_EXPR): Tweak comment.
4378 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
4380 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
4383 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
4385 * tree.h (TYPE_IDENTIFIER): Declare.
4386 * tree.c (subrange_type_for_debug_p): Use it.
4387 * godump.c (go_format_type): Likewise.
4388 * dwarf2out.c (is_cxx_auto, modified_type_die,
4389 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
4390 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
4392 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
4395 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
4397 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
4399 * config/i386/i386.c (examine_argument): Return bool. Return true if
4400 parameter should be passed in memory.
4401 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
4402 (construct_container): Update calls to examine_argument.
4403 (function_arg_advance_64): Ditto.
4404 (return_in_memory_32): Merge with ix86_return_in_memory.
4405 (return_in_memory_64): Ditto.
4406 (return_in_memory_ms_64): Ditto.
4408 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
4410 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
4411 * coverage.c (coverage_compute_profile_id): Handle externally visible
4414 2014-04-14 Martin Jambor <mjambor@suse.cz>
4416 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
4417 DECL_DISREGARD_INLINE_LIMITS functions.
4419 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
4422 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
4424 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
4427 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
4428 optimize_insn_for_speed_p instead of
4429 optimize_function_for_speed_p.
4431 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
4433 * doc/invoke.texi (free): Document AArch64.
4435 2014-04-14 Richard Biener <rguenther@suse.de>
4437 PR tree-optimization/60042
4438 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
4439 (insert_into_preds_of_block): Do not prevent PHI insertion
4440 for REFERENCE exprs here ...
4441 (eliminate_dom_walker::before_dom_children): ... but prevent
4442 their use here under similar conditions when applied to the
4443 IL after PRE optimizations.
4445 2014-04-14 Richard Biener <rguenther@suse.de>
4447 * passes.def: Move early points-to after early SRA.
4449 2014-04-14 Richard Biener <rguenther@suse.de>
4451 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
4452 check for which sign-changes we allow when forwarding
4453 a converted value into a switch.
4455 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
4457 * stor-layout.c (place_field): Finalize non-constant offset for the
4460 2014-04-14 Richard Biener <rguenther@suse.de>
4462 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
4464 (expand_switch_using_bit_tests_p): Likewise.
4465 (process_switch): Compute and pass on speed_p based on the
4467 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
4468 optimize_bb_for_speed_p.
4470 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
4472 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
4473 * function.h (struct function): Rename has_force_vect_loops into
4474 has_force_vectorize_loops.
4475 * lto-streamer-in.c (input_cfg): Adjust for renaming.
4476 (input_struct_function_base): Likewise.
4477 * lto-streamer-out.c (output_cfg): Likewise.
4478 (output_struct_function_base): Likewise.
4479 * omp-low.c (expand_omp_simd): Likewise.
4480 * tree-cfg.c (move_sese_region_to_fn): Likewise.
4481 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
4482 (version_loop_for_if_conversion): Likewise.
4483 (tree_if_conversion): Likewise.
4484 (main_tree_if_conversion): Likewise.
4485 (gate_tree_if_conversion): Likewise.
4486 * tree-inline.c (copy_loops): Likewise.
4487 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
4488 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
4489 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
4490 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
4491 * tree-vectorizer.c (vectorize_loops): Likewise.
4492 * tree-vectorizer.h (unlimited_cost_model): Likewise.
4494 2014-04-14 Richard Biener <rguenther@suse.de>
4497 * lto-streamer-out.c (wrap_refs): New function.
4498 (lto_output): Wrap symbol references in global initializes in
4499 type-preserving MEM_REFs.
4501 2014-04-14 Christian Bruel <christian.bruel@st.com>
4503 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
4505 2014-04-14 Christian Bruel <christian.bruel@st.com>
4507 * config/sh/sh.md (setmemqi): New expand pattern.
4508 * config/sh/sh.h (CLEAR_RATIO): Define.
4509 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
4510 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
4512 2014-04-14 Richard Biener <rguenther@suse.de>
4515 * fold-const.c (negate_expr_p): Don't negate directional rounding
4517 (fold_negate_expr): Likewise.
4519 2014-04-14 Richard Biener <rguenther@suse.de>
4521 PR tree-optimization/59817
4522 PR tree-optimization/60453
4523 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
4524 recursion to catch all CHRECs in the scalar evolution and restrict
4525 the predicate for the remains appropriately.
4527 2014-04-12 Catherine Moore <clm@codesourcery.com>
4529 * config/mips/constraints.md: Add new register constraint "kb".
4530 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
4531 (*movhi_internal): Likewise.
4532 (*movqi_internal): Likewise.
4533 * config/mips/mips.h (M16_STORE_REGS): New register class.
4534 (REG_CLASS_NAMES): Add M16_STORE_REGS.
4535 (REG_CLASS_CONTENTS): Likewise.
4536 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
4538 2014-04-11 Tobias Burnus <burnus@net-b.de>
4541 * doc/invoke.texi (-Wformat-signedness): Document it.
4542 (Wformat=2): Mention that this enables -Wformat-signedness.
4544 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
4546 * common/config/epiphany/epiphany-common.c
4547 (epiphany_option_optimization_table): Enable section anchors by
4548 default at -O1 or higher.
4549 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
4550 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
4551 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
4552 carries no extra cost.
4553 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
4554 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
4555 * config/epiphany/predicates.md (memclob_operand): New predicate.
4556 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
4557 Use memclob_operand predicate and X constraint for operand 3.
4559 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
4561 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
4562 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
4565 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
4567 PR rtl-optimization/60651
4568 * mode-switching.c (optimize_mode_switching): Make sure to emit
4569 sets of a lower numbered entity before sets of a higher numbered
4570 entity to a mode of the same or lower priority.
4571 When creating a seginfo for a basic block that starts with a code
4572 label, move the insertion point past the code label.
4573 (new_seginfo): Document and enforce requirement that
4574 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
4575 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
4576 * doc/tm.texi: Regenerate.
4578 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
4581 * config/arc/arc.c (arc_save_restore): Fix assert typo.
4583 2013-04-11 Jakub Jelinek <jakub@redhat.com>
4585 * BASE-VER: Set to 4.10.0.
4587 2014-04-11 Tobias Burnus <burnus@net-b.de>
4590 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
4591 * doc/gcc.texi (Service): Update description in the @menu
4592 * doc/invoke.texi (Option Summary): Remove misplaced and
4595 2014-04-11 Steve Ellcey <sellcey@mips.com>
4596 Jakub Jelinek <jakub@redhat.com>
4599 * expr.c (convert_move): Use emit_store_flag_force instead of
4600 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
4603 2014-04-11 Richard Biener <rguenther@suse.de>
4606 * varasm.c (assemble_alias): Avoid endless error reporting
4607 recursion by setting TREE_ASM_WRITTEN.
4609 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4611 * config/s390/s390.md: Add a splitter for NOT rtx.
4613 2014-04-11 Jakub Jelinek <jakub@redhat.com>
4615 PR rtl-optimization/60663
4616 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
4618 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
4619 Jakub Jelinek <jakub@redhat.com>
4622 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
4623 flag from decl_node to node.
4625 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4628 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
4629 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
4630 ameliorating the cases where it can be.
4632 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
4635 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
4637 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
4638 (loadsync_<mode>): Change mode.
4639 (load_quadpti, store_quadpti): New.
4640 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
4641 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
4642 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
4644 2014-04-09 Cong Hou <congh@google.com>
4647 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
4650 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4652 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
4653 instead of vnor to exploit possible fusion opportunity in the
4655 (altivec_expand_vec_perm_const_le): Likewise.
4657 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
4659 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
4660 (loadsync_<mode>): Change mode.
4661 (load_quadpti, store_quadpti): New.
4662 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
4663 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
4665 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
4668 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
4669 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
4670 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
4672 2014-04-08 Richard Biener <rguenther@suse.de>
4675 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
4676 a 64bit widest int print double-int similar to on HWI64 hosts.
4678 2014-04-08 Richard Biener <rguenther@suse.de>
4680 PR tree-optimization/60785
4681 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
4682 default defs properly.
4684 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
4686 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
4687 (Weffc++): Likewise.
4689 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
4691 * ipa-devirt.c (maybe_record_node): When node is not recorded,
4692 set completep to false rather than true.
4694 2014-04-07 Douglas B Rupp <rupp@adacore.com>
4697 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
4698 ARM_TARGET2_DWARF_FORMAT.
4700 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
4703 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
4704 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
4707 2014-04-07 Richard Biener <rguenther@suse.de>
4709 PR tree-optimization/60766
4710 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
4711 (may_eliminate_iv): Convert cand_value_at result to desired type.
4713 2014-04-07 Jason Merrill <jason@redhat.com>
4716 * common.opt (-fno-gnu-unique): Add.
4717 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
4719 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4721 * haifa-sched.c: Fix outdated function reference and minor
4722 grammar errors in introductory comment.
4724 2014-04-07 Richard Biener <rguenther@suse.de>
4727 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
4729 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
4731 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
4734 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
4735 size accounting for thunks.
4736 (pa_asm_output_mi_thunk): Use final_start_function() and
4737 final_end_function() to output function start and end directives.
4739 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
4741 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
4742 device specific ISA/ feature information. Remove short_sp and
4743 errata_skip ds. Add avr_device_specific_features enum to have device
4745 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
4746 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
4747 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
4748 updated device specific info.
4749 * config/avr/avr-mcus.def: Merge device specific details to
4750 dev_attribute field.
4751 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
4753 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
4754 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
4755 assembler if RMW isa supported by current device.
4756 * config/avr/genmultilib.awk: Update as device info structure changed.
4757 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
4759 2014-04-04 Cong Hou <congh@google.com>
4761 PR tree-optimization/60656
4762 * tree-vect-stmts.c (supportable_widening_operation):
4763 Fix a bug that elements in a vector with vect_used_by_reduction
4764 property are incorrectly reordered when the operation on it is not
4765 consistant with the one in reduction operation.
4767 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
4769 PR rtl-optimization/60155
4770 * gcse.c (record_set_data): New function.
4771 (single_set_gcse): New function.
4772 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
4773 (hoist_code): Likewise.
4774 (get_pressure_class_and_nregs): Likewise.
4776 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
4778 * explow.c (probe_stack_range): Emit a final optimization blockage.
4780 2014-04-04 Anthony Green <green@moxielogic.com>
4782 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
4785 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
4788 * lto-cgraph.c (input_overwrite_node): Check that partitioning
4789 flags are set only during streaming.
4790 * ipa.c (process_references, walk_polymorphic_call_targets,
4791 symtab_remove_unreachable_nodes): Drop bodies of always inline
4792 after early inlining.
4793 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
4795 2014-04-04 Jakub Jelinek <jakub@redhat.com>
4796 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4799 * dwarf2out.c (const_ok_for_output_1): Reject expressions
4802 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4805 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
4807 (cortex_a53_fdivd): Likewise.
4809 2014-04-04 Martin Jambor <mjambor@suse.cz>
4812 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
4814 * cgraph.c (clone_of_p): Also return true if thunks match.
4815 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
4816 cgraph_function_or_thunk_node and an obsolete comment.
4817 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
4819 (build_function_decl_skip_args): Likewise.
4820 (set_new_clone_decl_and_node_flags): New function.
4821 (duplicate_thunk_for_node): Likewise.
4822 (redirect_edge_duplicating_thunks): Likewise.
4823 (cgraph_clone_node): New parameter args_to_skip, pass it to
4824 redirect_edge_duplicating_thunks which is called instead of
4825 cgraph_redirect_edge_callee.
4826 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
4827 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
4829 2014-04-04 Jeff Law <law@redhat.com>
4832 * config/arm/predicates.md (const_int_I_operand): New predicate.
4833 (const_int_M_operand): Similarly.
4834 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
4836 (insv_t2, extv_reg, extzv_t2): Likewise.
4837 (load_multiple_with_writeback): Similarly for const_int_I_operand.
4838 (pop_multiple_with_writeback_and_return): Likewise.
4839 (vfp_pop_multiple_with_writeback): Likewise
4841 2014-04-04 Richard Biener <rguenther@suse.de>
4844 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
4845 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
4847 * gimplify.h (gimple_add_tmp_var_fn): Declare.
4848 * gimplify.c (gimple_add_tmp_var_fn): New function.
4849 * gimple-expr.h (create_tmp_reg_fn): Declare.
4850 * gimple-expr.c (create_tmp_reg_fn): New function.
4851 * gimple-low.c (record_vars_into): Don't change cfun.
4852 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
4853 code generation without cfun.
4855 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
4858 * Makefile.in (install-driver): Fix shell scripting.
4860 2014-04-03 Cong Hou <congh@google.com>
4862 PR tree-optimization/60505
4863 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
4864 threshold of number of iterations below which no vectorization
4866 * tree-vect-loop.c (new_loop_vec_info):
4867 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
4868 * tree-vect-loop.c (vect_analyze_loop_operations):
4869 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
4870 * tree-vect-loop.c (vect_transform_loop):
4871 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
4872 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
4873 of iterations of the loop and see if we should build the epilogue.
4875 2014-04-03 Richard Biener <rguenther@suse.de>
4877 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
4878 (streamer_tree_cache_create): Adjust.
4879 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
4880 to allow optional nodes array.
4881 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
4882 (streamer_tree_cache_append): Likewise.
4883 (streamer_tree_cache_create): Create nodes array optionally
4884 as specified by parameter.
4885 * lto-streamer-out.c (create_output_block): Avoid maintaining
4886 the node array in the writer cache.
4887 (DFS_write_tree): Remove assertion.
4888 (produce_asm_for_decls): Free the out decl state hash table early.
4889 * lto-streamer-in.c (lto_data_in_create): Adjust for
4890 streamer_tree_cache_create prototype change.
4892 2014-04-03 Richard Biener <rguenther@suse.de>
4894 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
4895 set TREE_CHAIN to NULL_TREE.
4897 2014-04-03 Richard Biener <rguenther@suse.de>
4899 PR tree-optimization/60740
4900 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
4901 over all GIMPLE_COND operands.
4903 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
4905 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
4906 (Weffc++): Remove Scott's numbering, merge lists and reference
4909 2014-04-03 Nick Clifton <nickc@redhat.com>
4911 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
4914 2014-04-03 Martin Jambor <mjambor@suse.cz>
4916 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
4917 mention gcc_unreachable before failing.
4918 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
4921 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
4924 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
4925 inconsistent code and instead mark the context inconsistent.
4926 (possible_polymorphic_call_targets): For inconsistent contexts
4927 return empty complete list.
4929 2014-04-02 Anthony Green <green@moxielogic.com>
4931 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
4932 (extendqisi2, extendhisi2): Define.
4933 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
4934 (WCHAR_TYPE): Change to unsigned int.
4936 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4938 PR tree-optimization/60733
4939 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
4940 insertion point for PHI candidates to be the end of the feeding
4941 block for the PHI argument.
4943 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
4945 PR rtl-optimization/60650
4946 * lra-constraints.c (process_alt_operands): Decrease reject for
4947 earlyclobber matching.
4949 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4951 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
4953 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4955 * config/spu/spu.c (pad_bb): Do not crash when the last
4956 insn is CODE_FOR_blockage.
4958 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4960 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
4961 lies outside the target mode.
4963 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
4966 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
4967 software floating point or no floating point registers, do not
4968 allow any type in the FPRs. Eliminate a test for SPE SIMD types
4969 in GPRs that occurs after we tested for GPRs that would never be
4972 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
4973 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
4974 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
4975 specifically allow DDmode, since that does not use the SPE SIMD
4978 2014-04-02 Richard Biener <rguenther@suse.de>
4981 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
4982 MODE_INTs. Properly use negv_optab.
4983 (expand_abs): Likewise.
4985 2014-04-02 Richard Biener <rguenther@suse.de>
4988 * Makefile.in (install-driver): Guard extra installs with special
4991 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
4993 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4996 2014-04-01 Richard Henderson <rth@redhat.com>
4999 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
5000 alternative enabled before register allocation.
5002 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
5004 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
5005 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
5007 (nios2_large_got_address): Remove unneeded 'sym' parameter.
5008 (nios2_got_address): Update nios2_large_got_address call site.
5009 (nios2_delegitimize_address): New function.
5010 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
5011 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
5012 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
5014 2014-04-01 Martin Husemann <martin@duskware.de>
5016 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
5019 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
5021 PR rtl-optimization/60604
5022 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
5023 check from register_operand.
5024 (register_operand): Redefine in terms of general_operand.
5025 (nonmemory_operand): Use register_operand for the non-constant cases.
5027 2014-04-01 Richard Biener <rguenther@suse.de>
5029 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
5031 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
5033 * doc/invoke.texi (mapp-regs): Clarify.
5035 2014-03-31 Ulrich Drepper <drepper@gmail.com>
5037 * config/i386/avx512fintrin.h (__v32hi): Define type.
5038 (__v64qi): Likewise.
5039 (_mm512_set1_epi8): Define.
5040 (_mm512_set1_epi16): Define.
5041 (_mm512_set4_epi32): Define.
5042 (_mm512_set4_epi64): Define.
5043 (_mm512_set4_pd): Define.
5044 (_mm512_set4_ps): Define.
5045 (_mm512_setr4_epi64): Define.
5046 (_mm512_setr4_epi32): Define.
5047 (_mm512_setr4_pd): Define.
5048 (_mm512_setr4_ps): Define.
5049 (_mm512_setzero_epi32): Define.
5051 2014-03-31 Martin Jambor <mjambor@suse.cz>
5054 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
5055 callsite_arguments_match_p. Updated all callers. Also check types of
5056 corresponding formal parameters and actual arguments.
5057 (not_all_callers_have_enough_arguments_p) Renamed to
5058 some_callers_have_mismatched_arguments_p.
5060 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
5062 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
5064 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
5067 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
5070 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
5072 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
5073 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
5075 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
5076 Use FMAMODE_NOVF512 mode iterator.
5077 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
5078 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
5079 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
5080 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
5082 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
5083 Use VF_128_256 mode iterator.
5084 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
5087 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
5089 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
5090 static chain if needed.
5092 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
5095 * lra-constraints.c (index_part_to_reg): New.
5096 (process_address): Use it.
5098 2014-03-27 Jeff Law <law@redhat.com>
5099 Jakub Jelinek <jakub@redhat.com>
5102 * expr.c (do_tablejump): Use simplify_gen_binary rather than
5103 gen_rtx_{PLUS,MULT} to build up the address expression.
5105 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
5106 creating non-canonical RTL.
5108 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
5111 * ipa-inline.c (want_inline_small_function_p): Short circuit large
5112 functions; reorganize to make cheap checks first.
5113 (inline_small_functions): Do not estimate growth when dumping;
5115 * ipa-inline.h (inline_summary): Add min_size.
5116 (growth_likely_positive): New function.
5117 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
5118 (set_cond_stmt_execution_predicate): Cleanup.
5119 (estimate_edge_size_and_time): Compute min_size.
5120 (estimate_calls_size_and_time): Likewise.
5121 (estimate_node_size_and_time): Likewise.
5122 (inline_update_overall_summary): Update min_size.
5123 (do_estimate_edge_time): Likewise.
5124 (do_estimate_edge_size): Update.
5125 (do_estimate_edge_hints): Update.
5126 (growth_likely_positive): New function.
5128 2014-03-28 Jakub Jelinek <jakub@redhat.com>
5131 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
5132 also if addr has VOIDmode.
5134 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5136 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
5137 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
5139 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
5140 instructions as well as AdvancedSIMD loads.
5142 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5144 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
5145 Use crypto_aese type.
5146 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
5147 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
5148 crypto_aese, crypto_aesmc. Move to types.md.
5149 * config/arm/types.md (crypto_aes): Split into crypto_aese,
5151 * config/arm/iterators.md (crypto_type): Likewise.
5153 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
5155 * cgraph.c: Include expr.h and tree-dfa.h.
5156 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
5159 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
5162 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
5163 regs from checking multi-reg pseudos.
5165 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5167 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
5169 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
5171 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
5172 if it would clobber the stack pointer, even temporarily.
5174 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
5176 * mode-switching.c: Make small adjustments to the top comment.
5178 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
5180 * config/rs6000/constraints.md (wD constraint): New constraint to
5181 match the constant integer to get the top DImode/DFmode out of a
5182 vector in a VSX register.
5184 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
5185 match the constant integer to get the top DImode/DFmode out of a
5186 vector in a VSX register.
5188 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
5191 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
5194 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
5195 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
5197 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
5198 Optimize vec_extract of 64-bit values, where the value being
5199 extracted is in the top word, where we can use scalar
5200 instructions. Add direct move and store support. Combine the big
5201 endian/little endian vector select load support into a single insn.
5202 (vsx_extract_<mode>_internal1): Likewise.
5203 (vsx_extract_<mode>_internal2): Likewise.
5204 (vsx_extract_<mode>_load): Likewise.
5205 (vsx_extract_<mode>_store): Likewise.
5206 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
5207 combined into vsx_extract_<mode>_load.
5208 (vsx_extract_<mode>_one_le): Likewise.
5210 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
5211 define the top 64-bit vector element.
5213 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
5216 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
5217 Document vec_vbpermq builtin.
5220 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
5221 enable use of xxsldwi and xxpermdi builtin functions.
5222 (vec_xxpermdi): Likewise.
5224 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
5225 Document use of vec_xxsldwi and vec_xxpermdi builtins.
5227 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
5229 PR rtl-optimization/60650
5230 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
5232 (find_spills_for): New.
5233 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
5234 Spill all pseudos on the second iteration.
5236 2014-03-27 Marek Polacek <polacek@redhat.com>
5239 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
5242 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5244 * config/s390/s390.c (s390_can_use_return_insn): Check for
5245 call-saved FPRs on 31 bit.
5247 2014-03-27 Jakub Jelinek <jakub@redhat.com>
5250 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
5251 if they need regimplification, just drop them instead of
5252 calling gimple_regimplify_operands on them.
5254 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
5257 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
5258 (aarch64_frame_pointer_required): Adjust logic.
5259 (aarch64_can_eliminate): Adjust logic.
5260 (aarch64_override_options_after_change): Adjust logic.
5262 2014-03-27 Dehao Chen <dehao@google.com>
5264 * ipa-inline.c (early_inliner): Update node's inline info.
5266 2014-03-26 Dehao Chen <dehao@google.com>
5268 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
5269 compiler inserted conditional jumps for NAN float check.
5271 2014-03-26 Jakub Jelinek <jakub@redhat.com>
5273 * ubsan.h (ubsan_create_data): Change second argument's type
5274 to const location_t *.
5275 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
5277 (ubsan_create_data): Change second argument to const location_t *PLOC.
5278 Create Loc field whenever PLOC is non-NULL.
5279 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
5280 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
5284 * real.c (real_to_integer2): Change type of low to UHWI.
5286 2014-03-26 Tobias Burnus <burnus@net-b.de>
5288 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
5289 (CILK_SELF_SPECS): New define.
5290 (driver_self_specs): Use it.
5292 2014-03-26 Richard Biener <rguenther@suse.de>
5294 * tree-pretty-print.c (percent_K_format): Implement special
5295 case for LTO and its stripped down BLOCK tree.
5297 2014-03-26 Jakub Jelinek <jakub@redhat.com>
5300 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
5302 * tree-vrp.c (simplify_internal_call_using_ranges): If only
5303 one range is range_int_cst_p, but not both, at least optimize
5304 addition/subtraction of 0 and multiplication by 0 or 1.
5305 * gimple-fold.c (gimple_fold_call): Fold
5306 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
5307 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
5308 INTEGER_CSTs, try to fold at least x * 0 and y - y.
5310 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
5312 PR rtl-optimization/60452
5313 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
5314 <case REG>: Return 1 for invalid offsets from the frame pointer.
5316 2014-03-26 Marek Polacek <polacek@redhat.com>
5319 * doc/extend.texi (C Extensions): Mention variable-length arrays in
5322 2014-03-26 Marek Polacek <polacek@redhat.com>
5325 * doc/extend.texi (Designated Inits): Describe what happens to omitted
5328 2014-03-26 Marek Polacek <polacek@redhat.com>
5331 * ira-color.c (update_conflict_hard_regno_costs): Perform the
5332 multiplication in unsigned type.
5334 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
5336 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
5338 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
5340 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
5342 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
5345 * cif-code.def (UNREACHABLE) New code.
5346 * ipa-inline.c (inline_small_functions): Skip edges to
5347 __builtlin_unreachable.
5348 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
5349 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
5350 predicate to __bulitin_unreachable.
5351 (set_cond_stmt_execution_predicate): Fix issue when
5352 invert_tree_comparison returns ERROR_MARK.
5353 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
5354 propagate to inline clones.
5355 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
5357 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
5358 * cgraphclones.c (cgraph_clone_node): If call destination is already
5359 ureachable, do not redirect it back.
5360 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
5363 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
5365 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
5366 Do not modify inline clones.
5368 2014-03-25 Jakub Jelinek <jakub@redhat.com>
5370 * config/i386/i386.md (general_sext_operand): New mode attr.
5371 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
5372 don't generate (sign_extend (const_int)).
5373 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
5374 operands[2]. Use We constraint instead of <i> and
5375 <general_sext_operand> predicate instead of <general_operand>.
5376 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
5377 * config/i386/constraints.md (We): New constraint.
5378 * config/i386/predicates.md (x86_64_sext_operand,
5379 sext_operand): New predicates.
5381 2014-03-25 Martin Jambor <mjambor@suse.cz>
5384 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
5385 inconsistent devirtualizations to __builtin_unreachable.
5387 2014-03-25 Marek Polacek <polacek@redhat.com>
5390 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
5392 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
5394 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
5395 order of elements for big-endian.
5397 2014-03-25 Richard Biener <rguenther@suse.de>
5400 * gimplify-me.c (gimple_regimplify_operands): Update the
5403 2014-03-25 Martin Jambor <mjambor@suse.cz>
5406 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
5407 (lto_output_varpool_node): Likewise.
5408 (input_overwrite_node): Likewise.
5409 (input_varpool_node): Likewise.
5411 2014-03-25 Richard Biener <rguenther@suse.de>
5413 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
5414 (run_gcc): Likewise.
5416 2014-03-25 Jakub Jelinek <jakub@redhat.com>
5418 * combine.c (simplify_compare_const): Add MODE argument.
5419 Handle mode_width 0 as very large mode_width.
5420 (try_combine, simplify_comparison): Adjust callers.
5422 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
5423 type to avoid signed integer overflow.
5424 * explow.c (plus_constant): Likewise.
5426 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
5428 * doc/generic.texi: Correct typos.
5430 2014-03-24 Tobias Burnus <burnus@net-b.de>
5432 * doc/invoke.texi (-flto): Expand section about
5433 using static libraries with LTO.
5435 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5437 PR rtl-optimization/60501
5438 * optabs.def (addptr3_optab): New optab.
5439 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
5440 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
5441 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
5443 * lra.c (emit_add3_insn): Use the addptr pattern if available.
5445 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
5447 2014-03-24 Ulrich Drepper <drepper@gmail.com>
5449 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
5452 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
5453 (_mm256_undefined_ps): Define.
5454 (_mm256_undefined_pd): Define.
5455 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
5456 (_mm_undefined_pd): Define.
5457 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
5458 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
5459 (_mm512_undefined_ps): Define.
5460 (_mm512_undefined_pd): Define.
5461 Use _mm*_undefined_*.
5462 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
5464 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
5466 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
5467 (lshr_simd): DI mode added.
5468 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
5469 (aarch64_ushr_simddi): Likewise.
5470 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
5471 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
5472 (vshrd_n_u64): Likewise.
5474 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5476 * Makefile.in (s-macro_list): Depend on cc1.
5478 2014-03-23 Teresa Johnson <tejohnson@google.com>
5480 * ipa-utils.c (ipa_print_order): Use specified dump file.
5482 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
5484 PR rtl-optimization/60601
5485 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
5487 * gcc.c (eval_spec_function): Initialize save_growing_value.
5489 2014-03-22 Jakub Jelinek <jakub@redhat.com>
5492 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
5493 code == MINUS_EXPR, never swap op0 with op1.
5495 * toplev.c (init_local_tick): Avoid signed integer multiplication
5497 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
5498 shift by first operand's bitsize.
5500 2014-03-21 Jakub Jelinek <jakub@redhat.com>
5503 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
5505 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
5506 TARGET_ISA_64BIT_P(x).
5508 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5510 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
5511 pattern for vector nor instead of subtract from splat(-1).
5512 (altivec_expand_vec_perm_const_le): Likewise.
5514 2014-03-21 Richard Henderson <rth@twiddle.net>
5517 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
5518 related insns after epilogue_completed.
5520 2014-03-21 Martin Jambor <mjambor@suse.cz>
5523 * cgraph.h (symtab_node): New flag body_removed.
5524 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
5525 when removing bodies.
5526 * symtab.c (dump_symtab_base): Dump body_removed flag.
5527 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
5528 had their bodies removed.
5530 2014-03-21 Martin Jambor <mjambor@suse.cz>
5533 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
5536 2014-03-21 Richard Biener <rguenther@suse.de>
5538 PR tree-optimization/60577
5539 * tree-core.h (struct tree_base): Document nothrow_flag use
5540 in VAR_DECL_NONALIASED.
5541 * tree.h (VAR_DECL_NONALIASED): New.
5542 (may_be_aliased): Adjust.
5543 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
5545 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
5547 * expr.c (expand_expr_real_1): Remove outdated comment.
5549 2014-03-20 Jakub Jelinek <jakub@redhat.com>
5552 * ira.c (adjust_cleared_regs): Call copy_rtx on
5553 *reg_equiv[REGNO (loc)].src_p before passing it to
5554 simplify_replace_fn_rtx.
5557 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
5558 into CONST, put pic register as first operand of PLUS. Use
5559 gen_const_mem for both 32-bit and 64-bit PIC got loads.
5561 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5563 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
5565 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
5567 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
5568 around for store forwarding issue in the FPU on the UT699.
5569 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
5570 loads and operations if -mfix-ut699 is specified.
5571 (divtf3_hq): Tweak attribute.
5572 (sqrttf2_hq): Likewise.
5574 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
5576 * calls.c (store_one_arg): Remove incorrect const qualification on the
5577 type of the temporary.
5578 * cfgexpand.c (expand_return): Likewise.
5579 * expr.c (expand_constructor): Likewise.
5580 (expand_expr_real_1): Likewise.
5582 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
5584 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
5587 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
5590 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
5592 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
5594 * config/arm/aarch-common-protos.h
5595 (alu_cost_table): Fix spelling of "extend".
5596 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
5598 2014-03-19 Richard Biener <rguenther@suse.de>
5601 * tree-core.h (tree_type_common): Re-order pointer members
5602 to reduce recursion depth during GC walks.
5604 2014-03-19 Marek Polacek <polacek@redhat.com>
5607 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
5608 before accessing it.
5610 2014-03-19 Richard Biener <rguenther@suse.de>
5613 * lto-streamer-in.c (input_function): In WPA stage do not drop
5616 2014-03-19 Jakub Jelinek <jakub@redhat.com>
5618 PR tree-optimization/60559
5619 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
5620 with build_zero_cst assignment.
5622 2014-03-18 Kai Tietz <ktietz@redhat.com>
5624 PR rtl-optimization/56356
5625 * sdbout.c (sdbout_parms): Verify that parms'
5626 incoming argument is valid.
5627 (sdbout_reg_parms): Likewise.
5629 2014-03-18 Richard Henderson <rth@redhat.com>
5632 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
5633 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
5634 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
5636 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
5638 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
5639 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
5640 Italicize plugin event names in description. Explain that
5641 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
5642 Remind that no GCC functions should be called after PLUGIN_FINISH.
5643 Explain what pragmas with expansion are.
5645 2014-03-18 Martin Liska <mliska@suse.cz>
5647 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
5648 gimple call statement is update.
5649 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
5650 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
5652 2014-03-18 Jakub Jelinek <jakub@redhat.com>
5655 * ubsan.c (ubsan_instrument_unreachable): Call
5656 initialize_sanitizer_builtins.
5657 (ubsan_pass): Likewise.
5660 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
5661 varpool_finalize_decl instead of rest_of_decl_compilation.
5663 2014-03-18 Richard Biener <rguenther@suse.de>
5665 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
5666 by using bitmap_and_compl instead of bitmap_and_compl_into.
5667 (df_rd_transfer_function): Likewise.
5669 2014-03-18 Richard Biener <rguenther@suse.de>
5671 * doc/lto.texi (fresolution): Fix typo.
5673 2014-03-18 Richard Biener <rguenther@suse.de>
5675 * doc/invoke.texi (flto): Update for changes in 4.9.
5677 2014-03-18 Richard Biener <rguenther@suse.de>
5679 * doc/loop.texi: Remove section on the removed lambda framework.
5680 Update loop docs with recent changes in preserving loop structure.
5682 2014-03-18 Richard Biener <rguenther@suse.de>
5684 * doc/lto.texi (-fresolution): Document.
5686 2014-03-18 Richard Biener <rguenther@suse.de>
5688 * doc/contrib.texi: Adjust my name.
5690 2014-03-18 Jakub Jelinek <jakub@redhat.com>
5693 * internal-fn.c: Include diagnostic-core.h.
5694 (expand_BUILTIN_EXPECT): New function.
5695 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
5696 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
5697 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
5698 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
5700 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
5701 Revert 3 argument __builtin_expect code.
5702 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
5703 * gimple-fold.c (gimple_fold_call): Likewise.
5704 * tree.h (fold_builtin_expect): New prototype.
5705 * builtins.c (build_builtin_expect_predicate): Add predictor
5706 argument, if non-NULL, create 3 argument __builtin_expect.
5707 (fold_builtin_expect): No longer static. Add ARG2 argument,
5708 pass it through to build_builtin_expect_predicate.
5709 (fold_builtin_2): Adjust caller.
5710 (fold_builtin_3): Handle BUILT_IN_EXPECT.
5711 * internal-fn.def (BUILTIN_EXPECT): New.
5713 2014-03-18 Tobias Burnus <burnus@net-b.de>
5716 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
5717 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
5718 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
5720 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
5723 * predict.c (combine_predictions_for_bb): Fix up formatting.
5724 (expr_expected_value_1, expr_expected_value): Add predictor argument,
5725 fill what it points to if non-NULL.
5726 (tree_predict_by_opcode): Adjust caller, use the predictor.
5727 * predict.def (PRED_COMPARE_AND_SWAP): Add.
5729 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
5731 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
5732 proper constant for the store mode.
5734 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
5736 * symtab.c (change_decl_assembler_name): Fix transparent alias
5739 2014-03-16 Renlin Li <Renlin.Li@arm.com>
5741 * config/aarch64/aarch64.c: Correct the comments about the
5742 aarch64 stack layout.
5744 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
5746 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
5747 check for GF_OMP_FOR_KIND_FOR.
5749 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
5751 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
5752 ymm and zmm register names.
5754 2014-03-17 Jakub Jelinek <jakub@redhat.com>
5757 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
5758 note creation for the 2010-08-31 changes.
5760 2014-03-17 Marek Polacek <polacek@redhat.com>
5763 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
5764 as -fno-tree-loop-vectorize.
5765 (expand_omp_simd): Likewise.
5767 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
5769 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
5770 (eligible_for_call_delay): New prototype.
5771 * config/sparc/sparc.c (tls_call_delay): Rename into...
5772 (eligible_for_call_delay): ...this. Return false if the instruction
5773 cannot be put in the delay slot of a branch.
5774 (eligible_for_restore_insn): Simplify.
5775 (eligible_for_return_delay): Return false if the instruction cannot be
5776 put in the delay slot of a branch and simplify.
5777 (eligible_for_sibcall_delay): Return false if the instruction cannot be
5778 put in the delay slot of a branch.
5779 * config/sparc/sparc.md (fix_ut699): New attribute.
5780 (tls_call_delay): Delete.
5781 (in_call_delay): Reimplement.
5782 (eligible_for_sibcall_delay): Rename into...
5783 (in_sibcall_delay): ...this.
5784 (eligible_for_return_delay): Rename into...
5785 (in_return_delay): ...this.
5786 (in_branch_delay): Reimplement.
5787 (in_uncond_branch_delay): Delete.
5788 (in_annul_branch_delay): Delete.
5790 2014-03-14 Richard Henderson <rth@redhat.com>
5793 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
5794 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
5795 (*floathi<X87MODEF>2_i387_with_temp): Remove.
5796 (floathi splitters): Remove.
5797 (float<SWI48x>xf2): New pattern.
5798 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
5799 code that tried to handle DImode for 32-bit, but which was excluded
5800 by the pattern's condition. Drop allocation of stack temporary.
5801 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
5802 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
5803 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
5804 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
5805 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
5806 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
5807 (*float<SWI48><MODEF>2_sse_interunit): Remove.
5808 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
5809 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
5810 (*float<SWI48x><X87MODEF>2_i387): Remove.
5811 (all float _with_temp splitters): Remove.
5812 (*float<SWI48x><MODEF>2_i387): New pattern.
5813 (*float<SWI48><MODEF>2_sse): New pattern.
5814 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
5815 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
5817 2014-03-14 Jakub Jelinek <jakub@redhat.com>
5818 Marek Polacek <polacek@redhat.com>
5821 * common.opt (dump_base_name_prefixed): New Variable.
5822 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
5823 if x_dump_base_name_prefixed is already set, set it at the end.
5825 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
5827 PR rtl-optimization/60508
5828 * lra-constraints.c (get_reload_reg): Add new parameter
5830 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
5831 Pass the new parameter values.
5833 2014-03-14 Richard Biener <rguenther@suse.de>
5835 * common.opt: Revert unintented changes from r205065.
5838 2014-03-14 Richard Biener <rguenther@suse.de>
5841 * cfghooks.c (split_block): Properly adjust all loops the
5842 block was a latch of.
5844 2014-03-14 Martin Jambor <mjambor@suse.cz>
5847 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
5850 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
5853 * config/avr/avr.c (avr_set_current_function): Pass function name
5854 through default_strip_name_encoding before sanity checking instead
5855 of skipping the first char of the assembler name.
5857 2014-03-13 Richard Henderson <rth@redhat.com>
5860 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
5861 (ix86_force_to_memory, ix86_free_from_memory): Remove.
5862 * config/i386/i386-protos.h: Likewise.
5863 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
5864 in the expander instead of a splitter.
5865 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
5866 any possibility of requiring a memory.
5867 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
5868 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
5869 (fp branch splitters): Update for ix86_split_fp_branch.
5870 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
5871 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
5872 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
5873 (*fop_<MODEF>_2_i387): Remove f/r alternative.
5874 (*fop_<MODEF>_3_i387): Likewise.
5875 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
5876 (splitters for the fop_* register patterns): Remove.
5877 (fscalexf4_i387): Rename from *fscalexf4_i387.
5878 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
5880 2014-03-13 Jakub Jelinek <jakub@redhat.com>
5882 PR tree-optimization/59779
5883 * tree-dfa.c (get_ref_base_and_extent): Use double_int
5884 type for bitsize and maxsize instead of HOST_WIDE_INT.
5886 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
5888 PR rtl-optimization/57320
5889 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
5890 the CFG after thread_prologue_and_epilogue_insns.
5892 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
5894 PR rtl-optimization/57189
5895 * lra-constraints.c (process_alt_operands): Disfavor spilling
5898 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
5900 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
5902 2014-03-13 Jakub Jelinek <jakub@redhat.com>
5904 PR tree-optimization/59025
5906 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
5907 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
5909 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
5912 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
5913 calls of avr_out_plus_1.
5915 2014-03-13 Bin Cheng <bin.cheng@arm.com>
5917 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
5918 BB's single pred and update the father loop's latch info later.
5920 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
5922 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
5926 (VEC_base): Likewise.
5927 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
5928 registers, we need to swap double words in little endian mode.
5930 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
5931 to be a container mode for 128-bit integer operations added in ISA
5932 2.07. Unlike TImode and PTImode, the preferred register set is
5933 the Altivec/VMX registers for the 128-bit operations.
5935 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
5937 (rs6000_split_128bit_ok_p): Likewise.
5939 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
5940 macros for creating ISA 2.07 normal and overloaded builtin
5941 functions with 3 arguments.
5942 (BU_P8V_OVERLOAD_3): Likewise.
5943 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
5944 for use as overloaded functions.
5945 (VPERM_1TI_UNS): Likewise.
5946 (VSEL_1TI): Likewise.
5947 (VSEL_1TI_UNS): Likewise.
5948 (ST_INTERNAL_1ti): Likewise.
5949 (LD_INTERNAL_1ti): Likewise.
5950 (XXSEL_1TI): Likewise.
5951 (XXSEL_1TI_UNS): Likewise.
5952 (VPERM_1TI): Likewise.
5953 (VPERM_1TI_UNS): Likewise.
5954 (XXPERMDI_1TI): Likewise.
5955 (SET_1TI): Likewise.
5956 (LXVD2X_V1TI): Likewise.
5957 (STXVD2X_V1TI): Likewise.
5958 (VEC_INIT_V1TI): Likewise.
5959 (VEC_SET_V1TI): Likewise.
5960 (VEC_EXT_V1TI): Likewise.
5961 (EQV_V1TI): Likewise.
5962 (NAND_V1TI): Likewise.
5963 (ORC_V1TI): Likewise.
5964 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
5965 added in ISA 2.07. Add both normal 'altivec' builtins, and the
5967 (VADDUQM): Likewise.
5968 (VSUBCUQ): Likewise.
5969 (VADDEUQM): Likewise.
5970 (VADDECUQ): Likewise.
5971 (VSUBEUQM): Likewise.
5972 (VSUBECUQ): Likewise.
5974 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
5975 __int128_t and __uint128_t types.
5976 (__uint128_type): Likewise.
5977 (altivec_categorize_keyword): Add support for vector __int128_t,
5978 vector __uint128_t, vector __int128, and vector unsigned __int128
5979 as a container type for TImode operations that need to be done in
5980 VSX/Altivec registers.
5981 (rs6000_macro_to_expand): Likewise.
5982 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
5983 to support 128-bit integer instructions vaddcuq, vadduqm,
5984 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
5985 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
5987 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
5988 for V1TImode, and set up preferences to use VSX/Altivec registers.
5989 Setup VSX reload handlers.
5990 (rs6000_debug_reg_global): Likewise.
5991 (rs6000_init_hard_regno_mode_ok): Likewise.
5992 (rs6000_preferred_simd_mode): Likewise.
5993 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
5994 (easy_altivec_constant): Likewise.
5995 (output_vec_const_move): Likewise.
5996 (rs6000_expand_vector_set): Convert V1TImode set and extract to
5998 (rs6000_expand_vector_extract): Likewise.
5999 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
6001 (rs6000_const_vec): Add support for V1TImode.
6002 (rs6000_emit_le_vsx_load): Swap double words when loading or
6003 storing TImode/V1TImode.
6004 (rs6000_emit_le_vsx_store): Likewise.
6005 (rs6000_emit_le_vsx_move): Likewise.
6006 (rs6000_emit_move): Add support for V1TImode.
6007 (altivec_expand_ld_builtin): Likewise.
6008 (altivec_expand_st_builtin): Likewise.
6009 (altivec_expand_vec_init_builtin): Likewise.
6010 (altivec_expand_builtin): Likewise.
6011 (rs6000_init_builtins): Add support for V1TImode type. Add
6012 support for ISA 2.07 128-bit integer builtins. Define type names
6013 for the VSX/Altivec vector types.
6014 (altivec_init_builtins): Add support for overloaded vector
6015 functions with V1TImode type.
6016 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
6017 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
6019 (rs6000_split_128bit_ok_p): Likewise.
6020 (rs6000_handle_altivec_attribute): Create V1TImode from vector
6021 __int128_t and vector __uint128_t.
6023 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
6024 and mode attributes.
6031 (VS_scalar): Likewise.
6032 (VS_double): Likewise.
6033 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
6035 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
6036 we support the ISA 2.07 128-bit integer arithmetic instructions.
6037 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
6038 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
6039 and TImode types for use with the builtin functions.
6040 (V1TI_type_node): Likewise.
6041 (unsigned_V1TI_type_node): Likewise.
6042 (intTI_type_internal_node): Likewise.
6043 (uintTI_type_internal_node): Likewise.
6045 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
6046 128-bit builtin functions.
6047 (UNSPEC_VADDEUQM): Likewise.
6048 (UNSPEC_VADDECUQ): Likewise.
6049 (UNSPEC_VSUBCUQ): Likewise.
6050 (UNSPEC_VSUBEUQM): Likewise.
6051 (UNSPEC_VSUBECUQ): Likewise.
6052 (VM): Add V1TImode to vector mode iterators.
6054 (VI_unit): Likewise.
6055 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
6056 (altivec_vaddcuq): Likewise.
6057 (altivec_vsubuqm): Likewise.
6058 (altivec_vsubcuq): Likewise.
6059 (altivec_vaddeuqm): Likewise.
6060 (altivec_vaddecuq): Likewise.
6061 (altivec_vsubeuqm): Likewise.
6062 (altivec_vsubecuq): Likewise.
6064 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
6066 (BOOL_128): Likewise.
6067 (BOOL_REGS_OUTPUT): Likewise.
6068 (BOOL_REGS_OP1): Likewise.
6069 (BOOL_REGS_OP2): Likewise.
6070 (BOOL_REGS_UNARY): Likewise.
6071 (BOOL_REGS_AND_CR0): Likewise.
6073 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
6074 128-bit integer builtin support.
6075 (vec_vadduqm): Likewise.
6076 (vec_vaddecuq): Likewise.
6077 (vec_vaddeuqm): Likewise.
6078 (vec_vsubecuq): Likewise.
6079 (vec_vsubeuqm): Likewise.
6080 (vec_vsubcuq): Likewise.
6081 (vec_vsubuqm): Likewise.
6083 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
6084 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
6085 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
6086 128-bit integer add/subtract to ISA 2.07.
6088 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
6090 * config/arc/arc.c (arc_predicate_delay_insns):
6091 Fix third argument passed to conditionalize_nonjump.
6093 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
6095 * config/aarch64/aarch64-builtins.c
6096 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
6097 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
6098 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
6099 instead of __builtin_lfloor.
6100 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
6102 2014-03-12 Jakub Jelinek <jakub@redhat.com>
6104 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
6105 (tree_ssa_ifcombine_bb_1): New function.
6106 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
6107 is an empty forwarder block to then_bb or vice versa and then_bb
6108 and else_bb are effectively swapped.
6110 2014-03-12 Christian Bruel <christian.bruel@st.com>
6113 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
6114 REG_CFA_DEF_CFA note.
6115 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
6116 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
6118 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
6120 PR tree-optimization/60454
6121 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
6123 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6125 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
6126 Do not define target_cpu_default2 to generic.
6127 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
6128 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
6129 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
6131 2014-03-12 Jakub Jelinek <jakub@redhat.com>
6132 Marc Glisse <marc.glisse@inria.fr>
6134 PR tree-optimization/60502
6135 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
6136 instead of build_low_bits_mask.
6138 2014-03-12 Jakub Jelinek <jakub@redhat.com>
6141 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
6142 if there are multiple uses, but op doesn't live on E edge.
6143 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
6144 clobber stmts before __builtin_unreachable.
6146 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
6148 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
6149 hard_frame_pointer_rtx.
6150 * cse.c (cse_insn): Remove volatile check.
6151 * cselib.c (cselib_process_insn): Likewise.
6152 * dse.c (scan_insn): Likewise.
6154 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
6156 * config/arc/arc.c (conditionalize_nonjump): New function,
6158 (arc_ifcvt): ... this.
6159 (arc_predicate_delay_insns): Use it.
6161 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
6163 * config/arc/predicates.md (extend_operand): During/after reload,
6164 allow const_int_operand.
6165 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
6166 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
6167 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
6169 (umulsi3_highpart_i): Likewise.
6171 2014-03-11 Richard Biener <rguenther@suse.de>
6173 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
6174 Add asserts to guard possible wrong-code bugs.
6176 2014-03-11 Richard Biener <rguenther@suse.de>
6178 PR tree-optimization/60429
6179 PR tree-optimization/60485
6180 * tree-ssa-structalias.c (set_union_with_increment): Properly
6181 take into account all fields that overlap the shifted vars.
6182 (do_sd_constraint): Likewise.
6183 (do_ds_constraint): Likewise.
6184 (get_constraint_for_ptr_offset): Likewise.
6186 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
6188 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
6189 (nios2_compute_frame_layout):
6190 Add calculation of cfun->machine->fp_save_offset.
6191 (nios2_expand_prologue): Correct setting of frame pointer register
6193 (nios2_expand_epilogue): Update recovery of stack pointer from
6194 frame pointer accordingly.
6195 (nios2_initial_elimination_offset): Update calculation of offset
6196 for eliminating to HARD_FRAME_POINTER_REGNUM.
6198 2014-03-10 Jakub Jelinek <jakub@redhat.com>
6201 * ipa.c (symtab_remove_unreachable_nodes): Don't call
6202 cgraph_get_create_node on VAR_DECLs.
6204 2014-03-10 Richard Biener <rguenther@suse.de>
6207 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
6209 2014-03-08 Douglas B Rupp <rupp@gnat.com>
6211 * config/vms/vms.opt (vms_float_format): New variable.
6213 2014-03-08 Tobias Burnus <burnus@net-b.de>
6215 * doc/invoke.texi (-fcilkplus): Update implementation status.
6217 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
6218 Richard Biener <rguenther@suse.de>
6220 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
6221 consistently accross all TUs.
6222 (run_gcc): Enable -fshort-double automatically at link at link-time
6223 and disallow override.
6225 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
6228 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
6229 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
6230 if they can't be used.
6232 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6234 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
6235 for Solaris 11/x86 ld.
6236 * configure: Regenerate.
6238 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6240 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
6241 (LIB_TLS_SPEC): Save as ld_tls_libs.
6242 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
6243 (HAVE_AS_IX86_TLSLDM): New test.
6244 * configure, config.in: Regenerate.
6245 * config/i386/i386.c (legitimize_tls_address): Fall back to
6246 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
6247 cannot support TLS_MODEL_LOCAL_DYNAMIC.
6248 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
6249 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
6251 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
6253 * common.opt (fira-loop-pressure): Mark as optimization.
6255 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
6257 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
6258 an OpenMP mappable type.
6260 2014-03-06 Matthias Klose <doko@ubuntu.com>
6262 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
6263 MULTILIB_OSDIRNAMES is not defined.
6265 2014-03-06 Jakub Jelinek <jakub@redhat.com>
6266 Meador Inge <meadori@codesourcery.com>
6269 * config/arm/arm.c (arm_tls_symbol_p): Remove.
6270 (arm_legitimize_address): Call legitimize_tls_address for any
6271 arm_tls_referenced_p expression, handle constant addend. Call it
6272 before testing for !TARGET_ARM.
6273 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
6275 2014-03-06 Richard Biener <rguenther@suse.de>
6281 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
6283 * tree-streamer.c (record_common_node): Assert we don't record
6284 nodes with type double.
6285 (preload_common_node): Skip type double, complex double and double
6286 pointer since it is now frontend dependent due to fshort-double option.
6288 2014-03-06 Richard Biener <rguenther@suse.de>
6290 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
6291 or -fno-lto is specified and the linker has full plugin support.
6292 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
6293 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
6294 * lto-wrapper.c (merge_and_complain): Merge compile-time
6295 optimization levels.
6296 (run_gcc): And pass it through to the link options.
6298 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
6302 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
6304 * cselib.c (remove_useless_values): Skip to avoid quadratic
6305 behavior if the condition moved from...
6306 (cselib_process_insn): ... here holds.
6308 2014-03-05 Jakub Jelinek <jakub@redhat.com>
6311 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
6312 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
6315 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
6316 (TM_H): Add x86-tune.def.
6318 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6320 * config/aarch64/aarch64.c (generic_tunings):
6321 Use cortexa57_extra_costs.
6323 2014-03-05 Jakub Jelinek <jakub@redhat.com>
6326 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
6327 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
6328 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
6331 2014-03-04 Heiher <r@hev.cc>
6333 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
6334 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
6336 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
6338 * config/i386/predicates.md (const2356_operand): Change to ...
6339 (const2367_operand): ... this.
6340 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
6342 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6343 (*avx512pf_scatterpf<mode>sf): Ditto.
6344 (avx512pf_scatterpf<mode>df): Ditto.
6345 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6346 (*avx512pf_scatterpf<mode>df): Ditto.
6347 * config/i386/i386.c (ix86_expand_builtin): Update
6348 incorrect hint operand error message.
6350 2014-03-04 Richard Biener <rguenther@suse.de>
6352 * lto-section-in.c (lto_get_section_data): Fix const cast.
6354 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
6356 * tree-streamer.c (record_common_node): Assert we don't record
6357 nodes with type double.
6358 (preload_common_node): Skip type double, complex double and double
6359 pointer since it is now frontend dependent due to fshort-double option.
6361 2014-03-04 Richard Biener <rguenther@suse.de>
6364 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
6365 (lto_input_toplevel_asms): Likewise.
6366 * lto-section-in.c (lto_get_section_data): Instead do it here
6369 2014-03-04 Richard Biener <rguenther@suse.de>
6371 PR tree-optimization/60382
6372 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
6373 dead PHIs a reduction.
6375 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
6377 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
6379 (_mm_prefetch): Move out of GCC target("sse") pragma.
6380 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
6381 GCC target("prfchw") pragma.
6382 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
6384 * config/i386/i386.c (ix86_option_override_internal): Enable
6385 -mprfchw with -mprefetchwt1.
6387 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
6389 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
6392 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
6394 * opts.h (CL_PCH_IGNORE): Define.
6395 * targhooks.c (option_affects_pch_p):
6396 Return false for options that have CL_PCH_IGNORE set.
6397 * opt-functions.awk: Process PchIgnore.
6398 * doc/options.texi: Document PchIgnore.
6400 * config/arc/arc.opt (misize): Add PchIgnore property.
6402 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6404 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
6405 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
6406 constraint on constants to permit them being loaded into
6407 GENERAL_REGS or BASE_REGS.
6409 2014-03-03 Nick Clifton <nickc@redhat.com>
6411 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
6412 anti-cacnonical alternatives.
6413 (negandhi3_real): New pattern.
6414 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
6416 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
6418 * config/avr/avr-mcus.def: Remove atxmega16x1.
6419 * config/avr/avr-tables.opt: Regenerate.
6420 * config/avr/t-multilib: Regenerate.
6421 * doc/avr-mmcu.texi: Regenerate.
6423 2014-03-03 Tobias Grosser <tobias@grosser.es>
6424 Mircea Namolaru <mircea.namolaru@inria.fr>
6426 PR tree-optimization/58028
6427 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
6430 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6432 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
6433 not handled by recognizers.
6435 2014-03-03 Jakub Jelinek <jakub@redhat.com>
6438 * function.c (expand_function_end): Don't emit
6439 clobber_return_register sequence if clobber_after is a BARRIER.
6440 * cfgexpand.c (construct_exit_block): Append instructions before
6441 return_label to prev_bb.
6443 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6445 * config/rs6000/constraints.md: Document reserved use of "wc".
6447 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
6450 * ipa.c (function_and_variable_visibility): When dissolving comdat
6451 group, also set all symbols to local.
6453 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
6458 2013-12-14 Jan Hubicka <jh@suse.cz>
6460 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
6462 2014-03-02 Jon Beniston <jon@beniston.com>
6468 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
6469 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
6470 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
6471 (simple_return, *simple_return): New patterns
6472 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
6473 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
6475 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
6477 * dwarf2out.c (gen_subprogram_die): Tidy.
6479 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
6482 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
6483 (*mov_t_msb_neg_negc): ... this new insn.
6485 2014-02-28 Jason Merrill <jason@redhat.com>
6488 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
6491 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
6494 * dwarf2out.c (decltype_auto_die): New static.
6495 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
6496 (gen_type_die_with_usage): Handle 'decltype(auto)'.
6497 (is_cxx_auto): Likewise.
6499 2014-02-28 Ian Bolton <ian.bolton@arm.com>
6501 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
6502 we are not using general regs only.
6504 2014-02-28 Richard Biener <rguenther@suse.de>
6507 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
6508 previous fix and only allow to remove trivial pre-headers
6509 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
6510 (remove_forwarder_block): Properly update the latch of a loop.
6512 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
6515 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
6516 (cselib_preserved_hash_table): New.
6517 (preserve_constants_and_equivs): Move preserved vals to it.
6518 (cselib_find_slot): Look it up first.
6519 (cselib_init): Initialize it.
6520 (cselib_finish): Release it.
6521 (dump_cselib_table): Dump it.
6523 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
6526 * cselib.c (remove_useless_values): Skip to avoid quadratic
6527 behavior if the condition moved from...
6528 (cselib_process_insn): ... here holds.
6530 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
6533 * var-tracking.c (vt_initialize): Apply the same condition to
6534 preserve the CFA base value.
6536 2014-02-28 Joey Ye <joey.ye@arm.com>
6539 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
6540 if reload in progress or completed.
6542 2014-02-28 Tobias Burnus <burnus@net-b.de>
6545 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
6548 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
6550 * doc/tm.texi.in (Condition Code Status): Update documention for
6551 relative locations of cc0-setter and cc0-user.
6553 2014-02-27 Jeff Law <law@redhat.com>
6555 PR rtl-optimization/52714
6556 * combine.c (try_combine): When splitting an unrecognized PARALLEL
6557 into two independent simple sets, if I3 is a jump, ensure the
6558 pattern we place into I3 is a (set (pc) ...).
6560 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
6561 Jeff Law <law@redhat.com>
6563 PR rtl-optimization/49847
6564 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
6565 are in different blocks.
6566 * doc/tm.texi (Condition Code Status): Update documention for
6567 relative locations of cc0-setter and cc0-user.
6569 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
6572 * lra.c (lra_emit_add): Check SUBREG too.
6574 2014-02-27 Andreas Schwab <schwab@suse.de>
6576 * config/m68k/m68k.c (m68k_option_override): Disable
6577 -flive-range-shrinkage for classic m68k.
6578 (m68k_override_options_after_change): Likewise.
6580 2014-02-27 Marek Polacek <polacek@redhat.com>
6583 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
6584 -Wmaybe-uninitialized.
6586 2014-02-27 Alan Modra <amodra@gmail.com>
6589 * reload1.c (emit_input_reload_insns): When reload_override_in,
6590 set old to rl->in_reg when rl->in_reg is a subreg.
6592 2014-02-26 Richard Biener <rguenther@suse.de>
6595 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
6597 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
6599 * common/config/i386/predicates.md (const1256_operand): Remove.
6600 (const2356_operand): New.
6601 (const_1_to_2_operand): Remove.
6602 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
6603 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
6604 (*avx512pf_gatherpf<mode>sf): Ditto.
6605 (avx512pf_gatherpf<mode>df): Ditto.
6606 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6607 (*avx512pf_gatherpf<mode>df): Ditto.
6608 (avx512pf_scatterpf<mode>sf): Ditto.
6609 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6610 (*avx512pf_scatterpf<mode>sf): Ditto.
6611 (avx512pf_scatterpf<mode>df): Ditto.
6612 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6613 (*avx512pf_scatterpf<mode>df): Ditto.
6614 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
6616 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
6618 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
6619 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
6620 (_mm512_mask_testn_epi64_mask): Move to ...
6621 * config/i386/avx512cdintrin.h: Here.
6622 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
6623 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
6624 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
6625 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
6626 TARGET_AVX512F from TARGET_AVX512CD.
6628 2014-02-26 Richard Biener <rguenther@suse.de>
6631 * ipa.c (walk_polymorphic_call_targets): Properly guard
6632 call to inline_update_overall_summary.
6634 2014-02-26 Bin Cheng <bin.cheng@arm.com>
6637 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
6638 and latches only if requested. Fix latch if it is removed.
6639 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
6640 LOOPS_HAVE_PREHEADERS.
6642 2014-02-25 Andrew Pinski <apinski@cavium.com>
6644 * builtins.c (expand_builtin_thread_pointer): Create a new target
6645 when the target is NULL.
6647 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
6649 PR rtl-optimization/60317
6650 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
6651 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
6652 * lra-assigns.c: Include params.h.
6653 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
6654 other reload pseudos considerations.
6656 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6658 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
6659 to use canonical form for nor<mode>3.
6661 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6664 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
6667 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
6669 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
6670 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
6671 (ix86_handle_option): Handle OPT_mprefetchwt1.
6672 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
6673 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
6675 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
6676 OPTION_MASK_ISA_PREFETCHWT1.
6677 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
6678 (PTA_PREFETCHWT1): New.
6679 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
6680 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
6681 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
6682 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
6683 (*prefetch_avx512pf_<mode>_: Change into ...
6684 (*prefetch_prefetchwt1_<mode>: This.
6685 * config/i386/i386.opt (mprefetchwt1): New.
6686 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
6687 (_mm_prefetch): Handle intent to write.
6688 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
6690 2014-02-25 Richard Biener <rguenther@suse.de>
6693 * emit-rtl.c (mem_attrs_htab): Remove.
6694 (mem_attrs_htab_hash): Likewise.
6695 (mem_attrs_htab_eq): Likewise.
6696 (set_mem_attrs): Always allocate new mem-attrs when something changed.
6697 (init_emit_once): Do not allocate mem_attrs_htab.
6699 2014-02-25 Richard Biener <rguenther@suse.de>
6702 * lto-opts.c (lto_write_options): Output non-explicit conservative
6703 -fwrapv, -fno-trapv and -fno-strict-overflow.
6704 * lto-wrapper.c (merge_and_complain): Handle merging those options.
6705 (run_gcc): And pass them through.
6707 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
6709 * sel-sched.c (calculate_new_fences): New parameter ptime.
6710 Calculate it as a maximum over all fence cycles.
6711 (sel_sched_region_2): Adjust the call to calculate_new_fences.
6712 Print the final schedule timing when sched_verbose.
6714 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
6716 PR rtl-optimization/60292
6717 * sel-sched.c (fill_vec_av_set): Do not reset target availability
6718 bit fot the fence instruction.
6720 2014-02-24 Alangi Derick <alangiderick@gmail.com>
6722 * calls.h: Fix typo in comment.
6724 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
6726 * config/pa/pa.c (pa_output_move_double): Don't valididate when
6727 adjusting offsetable addresses.
6729 2014-02-24 Guozhi Wei <carrot@google.com>
6731 * sparseset.h (sparseset_pop): Fix the wrong index.
6733 2014-02-24 Walter Lee <walt@tilera.com>
6735 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
6736 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
6738 * common/config/tilegx/tilegx-common.c
6739 (TARGET_DEFAULT_TARGET_FLAGS): Define.
6740 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
6742 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
6743 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
6744 (tilegx_gimplify_va_arg_expr): Handle big endian.
6745 (tilegx_expand_unaligned_load): Ditto.
6746 (tilegx_expand_unaligned_store): Ditto.
6747 (TARGET_RETURN_IN_MSB): New.
6748 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
6749 (TARGET_ENDIAN_DEFAULT): New.
6750 (TARGET_BIG_ENDIAN): Handle big endian.
6751 (BYTES_BIG_ENDIAN): Ditto.
6752 (WORDS_BIG_ENDIAN): Ditto.
6753 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
6756 * config/tilegx/tilegx.md (extv): Handle big endian.
6758 (insn_st<n>): Ditto.
6759 (insn_st<n>_add<bitsuffix>): Ditto.
6760 (insn_stnt<n>): Ditto.
6761 (insn_stnt<n>_add<bitsuffix>):Ditto.
6762 (vec_interleave_highv8qi): Handle big endian.
6763 (vec_interleave_highv8qi_be): New.
6764 (vec_interleave_highv8qi_le): New.
6765 (insn_v1int_h): Handle big endian.
6766 (vec_interleave_lowv8qi): Handle big endian.
6767 (vec_interleave_lowv8qi_be): New.
6768 (vec_interleave_lowv8qi_le): New.
6769 (insn_v1int_l): Handle big endian.
6770 (vec_interleave_highv4hi): Handle big endian.
6771 (vec_interleave_highv4hi_be): New.
6772 (vec_interleave_highv4hi_le): New.
6773 (insn_v2int_h): Handle big endian.
6774 (vec_interleave_lowv4hi): Handle big endian.
6775 (vec_interleave_lowv4hi_be): New.
6776 (vec_interleave_lowv4hi_le): New.
6777 (insn_v2int_l): Handle big endian.
6778 (vec_interleave_highv2si): Handle big endian.
6779 (vec_interleave_highv2si_be): New.
6780 (vec_interleave_highv2si_le): New.
6781 (insn_v4int_h): Handle big endian.
6782 (vec_interleave_lowv2si): Handle big endian.
6783 (vec_interleave_lowv2si_be): New.
6784 (vec_interleave_lowv2si_le): New.
6785 (insn_v4int_l): Handle big endian.
6786 * config/tilegx/tilegx.opt (mbig-endian): New option.
6787 (mlittle-endian): New option.
6788 * doc/install.texi: Document tilegxbe-linux.
6789 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
6791 2014-02-24 Martin Jambor <mjambor@suse.cz>
6794 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
6795 there are no parameter descriptors.
6797 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
6799 PR rtl-optimization/60268
6800 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
6801 initialization to ...
6802 (sched_rgn_init): ... here.
6803 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
6805 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
6807 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
6810 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
6812 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
6815 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
6817 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
6818 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
6820 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
6822 * config/microblaze/predicates.md: Add cmp_op predicate.
6823 * config/microblaze/microblaze.md: Add branch_compare instruction
6824 which uses cmp_op predicate and emits cmp insn before branch.
6825 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
6826 to microblaze_expand_conditional_branch and consolidate logic.
6827 (microblaze_expand_conditional_branch): emit branch_compare
6828 insn instead of handling cmp op separate from branch insn.
6830 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6832 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
6835 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6837 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
6838 define_insn with define_expand and new define_insn
6839 *altivec_lve<VI_char>x_internal.
6840 (altivec_stve<VI_char>x): Replace define_insn with define_expand
6841 and new define_insn *altivec_stve<VI_char>x_internal.
6842 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
6844 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
6846 (altivec_expand_stvex_be): New function.
6848 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
6850 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
6851 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
6852 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
6853 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
6855 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
6858 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
6859 instead of emit_move_insn.
6861 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6863 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
6865 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
6866 gen_altivec_vsumsws.
6868 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6870 * config/rs6000/altivec.md (altivec_lvxl): Rename as
6871 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
6872 (altivec_lvxl_<mode>): New define_expand incorporating
6873 -maltivec=be semantics where needed.
6874 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
6875 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
6876 semantics where needed.
6877 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
6878 (altivec_stvx_<mode>): New define_expand incorporating
6879 -maltivec=be semantics where needed.
6880 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
6881 VM2 iterator instead of V4SI.
6882 (altivec_stvxl_<mode>): New define_expand incorporating
6883 -maltivec=be semantics where needed.
6884 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
6885 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
6886 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
6887 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
6888 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
6889 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
6890 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
6891 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
6892 ALTIVEC_BUILTIN_STVXL.
6893 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
6894 (altivec_expand_stvx_be): Likewise.
6895 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
6896 (altivec_expand_lvx_be): Likewise.
6897 (altivec_expand_stvx_be): Likewise.
6898 (altivec_expand_builtin): Add cases for
6899 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
6900 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
6901 (altivec_init_builtins): Add definitions for
6902 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
6903 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
6905 2014-02-21 Catherine Moore <clm@codesourcery.com>
6907 * doc/invoke.texi (mvirt, mno-virt): Document.
6908 * config/mips/mips.opt (mvirt): New option.
6909 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
6911 2014-02-21 Richard Biener <rguenther@suse.de>
6913 PR tree-optimization/60276
6914 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
6915 (STMT_VINFO_MIN_NEG_DIST): New macro.
6916 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
6917 STMT_VINFO_MIN_NEG_DIST.
6918 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
6919 made for negative dependence distances still hold.
6921 2014-02-21 Richard Biener <rguenther@suse.de>
6924 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
6925 DECL_INITIAL for globals not in the current function context.
6927 2014-02-21 Jakub Jelinek <jakub@redhat.com>
6929 PR tree-optimization/56490
6930 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
6931 * tree-ssa-uninit.c: Include params.h.
6932 (compute_control_dep_chain): Add num_calls argument, return false
6933 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
6934 num_calls to recursive call.
6935 (find_predicates): Change dep_chain into normal array,
6936 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
6937 variable and adjust compute_control_dep_chain caller.
6938 (find_def_preds): Likewise.
6940 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
6942 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
6943 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
6945 2014-02-21 Nick Clifton <nickc@redhat.com>
6947 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
6948 (pushhi1): Likewise.
6949 (popqi1): Add mode to pre_dec.
6952 2014-02-21 Jakub Jelinek <jakub@redhat.com>
6954 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
6955 mode for mask of V8SFmode permutation.
6957 2014-02-20 Richard Henderson <rth@redhat.com>
6960 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
6961 a new pseudo for OLDVAL.
6963 2014-02-20 Jakub Jelinek <jakub@redhat.com>
6966 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
6967 gen_reg_rtx if d->testing_p.
6968 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
6969 if d->testing_p and we will certainly return true.
6970 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
6973 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
6975 * emit-rtl.c (gen_reg_rtx): Assert that
6976 crtl->emit.regno_pointer_align_length is non-zero.
6978 2014-02-20 Richard Henderson <rth@redhat.com>
6981 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
6982 on failure the store back into EXPECT.
6984 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
6985 Sandra Loosemore <sandra@codesourcery.com>
6987 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
6988 * config/nios2/nios2.c (nios2_function_profiler): Add
6989 -fPIC (flag_pic == 2) support.
6990 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
6991 (nios2_large_offset_p): New function.
6992 (nios2_unspec_reloc_p): Move up position, update to use
6993 nios2_large_offset_p.
6994 (nios2_unspec_address): Remove function.
6995 (nios2_unspec_offset): New function.
6996 (nios2_large_got_address): New function.
6997 (nios2_got_address): Add large offset support.
6998 (nios2_legitimize_tls_address): Update usage of removed and new
7000 (nios2_symbol_binds_local_p): New function.
7001 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
7002 (nios2_legitimize_address): Update to use nios2_large_offset_p.
7003 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
7004 (nios2_print_operand): Merge H/L processing, add hiadj/lo
7005 processing for (const (unspec ...)).
7006 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
7008 2014-02-20 Richard Biener <rguenther@suse.de>
7010 * tree-cfg.c (replace_uses_by): Mark altered BBs before
7011 doing the substitution.
7012 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
7014 2014-02-20 Martin Jambor <mjambor@suse.cz>
7017 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
7018 info when checking whether lattices are bottom.
7020 2014-02-20 Richard Biener <rguenther@suse.de>
7023 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
7026 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
7029 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
7030 parameter specifying the scaling.
7031 (inline_call): Update.
7032 (want_inline_recursively): Guard division by zero.
7033 (recursive_inlining): Update.
7034 * ipa-inline.h (clone_inlined_nodes): Update.
7036 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
7039 * config/i386/i386.c (classify_argument): Pass structures of size
7040 64 bytes or less in register.
7042 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
7043 Kirill Yukhin <kirill.yukhin@intel.com>
7045 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
7046 (_mm_rcp28_round_ss): Ditto.
7047 (_mm_rsqrt28_round_sd): Ditto.
7048 (_mm_rsqrt28_round_ss): Ditto.
7049 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
7050 (_mm_rcp14_round_ss): Ditto.
7051 (_mm_rsqrt14_round_sd): Ditto.
7052 (_mm_rsqrt14_round_ss): Ditto.
7053 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
7054 the first input operand, get rid of match_dup.
7055 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
7057 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
7059 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
7060 operand as the first input operand, set type attribute.
7061 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
7063 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
7064 operand as the first input operand, set type attribute.
7066 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7068 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
7071 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
7074 * config/i386/i386.c (construct_container): Remove TFmode check
7075 for X86_64_INTEGER_CLASS.
7077 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
7080 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
7081 only when -Wpsabi is enabled.
7083 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
7086 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
7087 passing arrays in registers are the same as for structs, so remove the
7088 special case for them.
7090 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
7092 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
7093 destination type, extract only the valid bits if the source type is not
7094 integral and has a different mode.
7096 2014-02-19 Richard Biener <rguenther@suse.de>
7099 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
7102 2014-02-19 Richard Biener <rguenther@suse.de>
7105 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
7106 (ipa_modify_call_arguments): Emit an argument load explicitely and
7107 preserve virtual SSA form there and for the replacement call.
7108 Do not update SSA form nor free dominance info.
7110 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
7112 * ipa.c (function_and_variable_visibility): Also clear WEAK
7113 flag when disolving COMDAT_GROUP.
7115 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
7117 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
7118 * ipa-prop.c (ipa_set_jf_known_type): Return early when
7120 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
7121 do more sanity checks.
7122 (detect_type_change): Return true when giving up early.
7123 (compute_complex_assign_jump_func): Fix type parameter of
7124 ipa_set_ancestor_jf.
7125 (compute_complex_ancestor_jump_func): Likewise.
7126 (update_jump_functions_after_inlining): Fix updating of
7128 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
7130 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
7132 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
7133 inline clones when edge disappears.
7135 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
7138 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
7139 Split 64-bit moves into 2 patterns. Do not allow the use of
7140 direct move for TDmode in little endian, since the decimal value
7141 has little endian bytes within a word, but the 64-bit pieces are
7142 ordered in a big endian fashion, and normal subreg's of TDmode are
7144 (mov<mode>_64bit_dm): Likewise.
7145 (movtd_64bit_nodm): Likewise.
7147 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
7149 PR tree-optimization/60174
7150 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
7151 statement of an SSA_NAME that occurs in an abnormal PHI node.
7153 2014-02-18 Jakub Jelinek <jakub@redhat.com>
7156 * final.c (SEEN_BB): Remove.
7157 (SEEN_NOTE, SEEN_EMITTED): Renumber.
7158 (final_scan_insn): Don't force_source_line on second
7159 NOTE_INSN_BASIC_BLOCK.
7161 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
7164 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
7165 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
7166 (type_natural_mode): Warn ABI change when %zmm register is not
7167 available for AVX512F vector value passing.
7169 2014-02-18 Kai Tietz <ktietz@redhat.com>
7172 * config/i386/i386.c (ix86_expand_prologue): Use value in
7173 rax register as displacement when restoring %r10 or %rax.
7174 Fix wrong offset when restoring both registers.
7176 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
7178 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
7179 assertion with conditional return.
7181 2014-02-18 Jakub Jelinek <jakub@redhat.com>
7182 Uros Bizjak <ubizjak@gmail.com>
7185 * config/i386/driver-i386.c (host_detect_local_cpu): If
7186 YMM state is not saved by the OS, also clear has_f16c. Move
7187 CPUID 0x80000001 handling before YMM state saving checking.
7189 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
7191 PR rtl-optimization/58960
7192 * haifa-sched.c (alloc_global_sched_pressure_data): New,
7193 factored out from ...
7194 (sched_init): ... here.
7195 (free_global_sched_pressure_data): New, factored out from ...
7196 (sched_finish): ... here.
7197 * sched-int.h (free_global_sched_pressure_data): Declare.
7198 * sched-rgn.c (nr_regions_initial): New static global.
7199 (haifa_find_rgns): Initialize it.
7200 (schedule_region): Disable sched-pressure for the newly
7203 2014-02-17 Richard Biener <rguenther@suse.de>
7205 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
7206 release SSA defs of pattern stmts.
7208 2014-02-17 Richard Biener <rguenther@suse.de>
7210 * tree-inline.c (expand_call_inline): Release the virtual
7211 operand defined by the call we are about to inline.
7213 2014-02-17 Richard Biener <rguenther@suse.de>
7215 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
7217 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
7218 Ilya Tocar <ilya.tocar@intel.com>
7220 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
7221 arguments order in builtin.
7222 (_mm512_permutexvar_epi64): Ditto.
7223 (_mm512_mask_permutexvar_epi64): Ditto
7224 (_mm512_maskz_permutexvar_epi32): Ditto
7225 (_mm512_permutexvar_epi32): Ditto
7226 (_mm512_mask_permutexvar_epi32): Ditto
7228 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7230 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
7231 (p8_vmrgow): Likewise.
7233 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7235 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
7238 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
7241 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
7242 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
7243 into 64-bit and 32-bit moves. On 64-bit moves, add support for
7244 using direct move instructions on ISA 2.07. Also adjust
7245 instruction length for 64-bit.
7246 (mov<mode>_64bit, TFmode/TDmode): Likewise.
7247 (mov<mode>_32bit, TFmode/TDmode): Likewise.
7249 2014-02-15 Alan Modra <amodra@gmail.com>
7253 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
7254 find_replacement on parts of insn rtl that might be reloaded.
7256 2014-02-15 Richard Biener <rguenther@suse.de>
7258 PR tree-optimization/60183
7259 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
7260 (tree_ssa_phiprop): Calculate and free post-dominators.
7262 2014-02-14 Jeff Law <law@redhat.com>
7264 PR rtl-optimization/60131
7265 * ree.c (get_extended_src_reg): New function.
7266 (combine_reaching_defs): Use it rather than assuming location of REG.
7267 (find_and_remove_re): Verify first operand of extension is
7268 a REG before adding the insns to the copy list.
7270 2014-02-14 Roland McGrath <mcgrathr@google.com>
7272 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
7273 * configure: Regenerated.
7274 * config.in: Regenerated.
7275 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
7276 instead of ASM_SHORT.
7278 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
7279 Richard Earnshaw <rearnsha@arm.com>
7281 PR rtl-optimization/59535
7282 * lra-constraints.c (process_alt_operands): Encourage alternative
7283 when unassigned pseudo class is superset of the alternative class.
7284 (inherit_reload_reg): Don't inherit when optimizing for code size.
7285 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
7286 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
7287 modes not less than 4 for Thumb1.
7289 2014-02-14 Kyle McMartin <kyle@redhat.com>
7292 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
7294 2014-02-14 Richard Biener <rguenther@suse.de>
7296 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
7297 (get_frame_arg): Drop the assert with langhook types_compatible_p.
7298 Do not strip INDIRECT_REFs.
7300 2014-02-14 Richard Biener <rguenther@suse.de>
7303 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
7304 DECL_FUNCTION_SPECIFIC_TARGET.
7305 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
7306 * tree-streamer-out.c (pack_ts_target_option): Remove.
7307 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
7308 (write_ts_function_decl_tree_pointers): Do not stream
7309 DECL_FUNCTION_SPECIFIC_TARGET.
7310 * tree-streamer-in.c (unpack_ts_target_option): Remove.
7311 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
7312 (lto_input_ts_function_decl_tree_pointers): Do not stream
7313 DECL_FUNCTION_SPECIFIC_TARGET.
7315 2014-02-14 Jakub Jelinek <jakub@redhat.com>
7317 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
7318 (get_initial_def_for_induction, vectorizable_induction): Ignore
7319 debug stmts when looking for exit_phi.
7320 (vectorizable_live_operation): Fix up condition.
7322 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
7324 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
7325 nreverse() because it changes the content of original tree list.
7327 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
7329 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
7330 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
7332 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
7334 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
7335 GNU coding standards.
7337 2014-02-13 Jakub Jelinek <jakub@redhat.com>
7340 * dwarf2out.c (gen_subprogram_die): Don't call
7341 add_calling_convention_attribute if subr_die is old_die.
7343 2014-02-13 Sharad Singhai <singhai@google.com>
7345 * doc/optinfo.texi: Fix order of nodes.
7347 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
7349 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
7350 operands[2], not operands[3].
7352 2014-02-13 Richard Biener <rguenther@suse.de>
7355 * doc/install.texi (ISL): Update recommended version to 0.12.2,
7356 mention the possibility of an in-tree build.
7357 (CLooG): Update recommended version to 0.18.1, mention the
7358 possibility of an in-tree build and clarify that the ISL
7359 bundled with CLooG does not work.
7361 2014-02-13 Jakub Jelinek <jakub@redhat.com>
7364 * expr.c (compress_float_constant): If x is a hard register,
7365 extend into a pseudo and then move to x.
7367 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
7369 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
7370 caused by bad second argument to warning_at() with -mhotpatch and
7371 nested functions (e.g. with gfortran).
7373 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
7375 * opts.c (option_name): Remove "enabled by default" rider.
7377 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
7379 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
7381 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
7382 Uros Bizjak <ubizjak@gmail.com>
7385 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
7386 * configure: Regenerated.
7388 2014-02-12 Richard Biener <rguenther@suse.de>
7390 * vec.c (vec_prefix::calculate_allocation): Move as
7391 inline variant to vec.h.
7392 (vec_prefix::calculate_allocation_1): New out-of-line version.
7393 * vec.h (vec_prefix::calculate_allocation_1): Declare.
7394 (vec_prefix::m_has_auto_buf): Rename to ...
7395 (vec_prefix::m_using_auto_storage): ... this.
7396 (vec_prefix::calculate_allocation): Inline the easy cases
7397 and dispatch to calculate_allocation_1 which doesn't need the
7399 (va_heap::reserve): Use gcc_checking_assert.
7400 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
7401 m_using_auto_storage.
7402 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
7404 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
7405 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
7406 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
7408 2014-02-12 Richard Biener <rguenther@suse.de>
7410 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
7411 when we found a dependence.
7413 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
7415 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
7417 (maybe_fold_stmt): ... into this new function.
7418 * omp-low.c (lower_omp): Update comment.
7420 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
7423 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
7426 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
7428 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
7429 identifiers in comments.
7430 (cortexa53_extra_costs): Likewise.
7431 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
7432 (cortexa7_extra_costs): Likewise.
7433 (cortexa12_extra_costs): Likewise.
7434 (cortexa15_extra_costs): Likewise.
7435 (v7m_extra_costs): Likewise.
7437 2014-02-12 Richard Biener <rguenther@suse.de>
7440 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
7441 of posix_memalign being successful.
7442 (lower_stmt): Restrict lowering of posix_memalign to when
7443 -ftree-bit-ccp is enabled.
7445 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
7447 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
7449 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
7451 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
7453 PR rtl-optimization/60116
7454 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
7455 other_insn once the combination has been validated.
7457 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
7460 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
7462 * ipa-devirt.c: Include demangle.h
7463 (odr_violation_reported): New static variable.
7464 (add_type_duplicate): Update odr_violations.
7465 (maybe_record_node): Add completep parameter; update it.
7466 (record_target_from_binfo): Add COMPLETEP parameter;
7467 update it as needed.
7468 (possible_polymorphic_call_targets_1): Likewise.
7469 (struct polymorphic_call_target_d): Add nonconstruction_targets;
7470 rename FINAL to COMPLETE.
7471 (record_targets_from_bases): Sanity check we found the binfo;
7472 fix COMPLETEP updating.
7473 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
7474 parameter, fix computing of COMPLETEP.
7475 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
7476 at LTO time do demangling.
7477 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
7478 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
7480 (gimple_get_virt_method_for_binfo): Likewise.
7481 * gimple-fold.h (gimple_get_virt_method_for_binfo,
7482 gimple_get_virt_method_for_vtable): Update prototypes.
7484 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
7487 * genautomata.c (add_presence_absence): Fix typo with
7488 {final_}presence_list.
7490 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
7493 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
7494 for VSX/Altivec vectors that land in GPR registers.
7496 2014-02-11 Richard Henderson <rth@redhat.com>
7497 Jakub Jelinek <jakub@redhat.com>
7500 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
7501 around drhs if type conversion to lacc->type is not useless.
7503 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7505 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
7507 (cortex-a57.cortex-a53): Likewise.
7508 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
7510 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7512 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
7515 2014-02-11 Renlin Li <Renlin.Li@arm.com>
7517 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
7518 add_options_for_arm_vfp3.
7520 2014-02-11 Jeff Law <law@redhat.com>
7523 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
7524 object with an undesirable mode.
7526 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7529 * config/i386/sol2-9.h: New file.
7530 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
7531 *-*-solaris2.9*): Use it.
7533 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
7535 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
7536 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
7538 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
7540 * config/microblaze/microblaze.c: Extend mcpu version format
7542 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
7544 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
7546 2014-02-10 Richard Henderson <rth@redhat.com>
7549 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
7550 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
7551 ms-abi vs -mno-accumulate-outgoing-args.
7552 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
7553 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
7556 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
7559 * cfgexpand.c (expand_asm_operands): Attach source location to
7560 ASM_INPUT rtx objects.
7561 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
7563 2014-02-10 Nick Clifton <nickc@redhat.com>
7565 * config/mn10300/mn10300.c (popcount): New function.
7566 (mn10300_expand_prologue): Include saved registers in stack usage
7569 2014-02-10 Jeff Law <law@redhat.com>
7572 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
7573 when changing the SET_DEST of a prior insn to avoid an input reload.
7575 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
7577 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
7578 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
7579 -mcall-openbsd, or -mcall-linux.
7580 (CC1_ENDIAN_BIG_SPEC): Remove.
7581 (CC1_ENDIAN_LITTLE_SPEC): Remove.
7582 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
7583 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
7584 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
7585 and %cc1_endian_default.
7586 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
7588 2014-02-10 Richard Biener <rguenther@suse.de>
7590 PR tree-optimization/60115
7591 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
7592 MEM_REF handling. Properly verify that the accesses are not
7593 out of the objects bound.
7595 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7597 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
7600 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
7602 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
7603 proper constants and fix formatting.
7604 (possible_polymorphic_call_targets): Fix formatting.
7606 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
7607 Ilya Tocar <ilya.tocar@intel.com>
7609 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
7610 (_mm512_loadu_epi32): Renamed into...
7611 (_mm512_loadu_si512): This.
7612 (_mm512_storeu_epi32): Renamed into...
7613 (_mm512_storeu_si512): This.
7614 (_mm512_maskz_ceil_ps): Removed.
7615 (_mm512_maskz_ceil_pd): Ditto.
7616 (_mm512_maskz_floor_ps): Ditto.
7617 (_mm512_maskz_floor_pd): Ditto.
7618 (_mm512_floor_round_ps): Ditto.
7619 (_mm512_floor_round_pd): Ditto.
7620 (_mm512_ceil_round_ps): Ditto.
7621 (_mm512_ceil_round_pd): Ditto.
7622 (_mm512_mask_floor_round_ps): Ditto.
7623 (_mm512_mask_floor_round_pd): Ditto.
7624 (_mm512_mask_ceil_round_ps): Ditto.
7625 (_mm512_mask_ceil_round_pd): Ditto.
7626 (_mm512_maskz_floor_round_ps): Ditto.
7627 (_mm512_maskz_floor_round_pd): Ditto.
7628 (_mm512_maskz_ceil_round_ps): Ditto.
7629 (_mm512_maskz_ceil_round_pd): Ditto.
7630 (_mm512_expand_pd): Ditto.
7631 (_mm512_expand_ps): Ditto.
7632 * config/i386/i386.c (ix86_builtins): Remove
7633 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
7634 (bdesc_args): Ditto.
7635 * config/i386/predicates.md (const1256_operand): New.
7636 (const_1_to_2_operand): Ditto.
7637 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
7638 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
7639 (*avx512pf_gatherpf<mode>sf): Ditto.
7640 (avx512pf_gatherpf<mode>df): Ditto.
7641 (*avx512pf_gatherpf<mode>df_mask): Ditto.
7642 (*avx512pf_gatherpf<mode>df): Ditto.
7643 (avx512pf_scatterpf<mode>sf): Ditto.
7644 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
7645 (*avx512pf_scatterpf<mode>sf): Ditto.
7646 (avx512pf_scatterpf<mode>df): Ditto.
7647 (*avx512pf_scatterpf<mode>df_mask): Ditto.
7648 (*avx512pf_scatterpf<mode>df): Ditto.
7649 (avx512f_expand<mode>): Removed.
7650 (<shift_insn><mode>3<mask_name>): Change predicate type.
7652 2014-02-08 Jakub Jelinek <jakub@redhat.com>
7654 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
7655 not at the end of datarefs vector use ordered_remove to avoid
7656 reordering datarefs vector.
7659 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
7660 mark local addressable non-static vars as GOVD_PRIVATE
7661 instead of GOVD_LOCAL.
7662 * omp-low.c (lower_omp_for): Move gimple_bind_vars
7663 and BLOCK_VARS of gimple_bind_block to new_stmt rather
7667 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
7668 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
7669 assume_aligned or alloc_align attributes.
7670 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
7671 arguments. Handle also assume_aligned and alloc_align attributes.
7672 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
7673 calls to functions with assume_aligned or alloc_align attributes.
7674 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
7676 2014-02-08 Terry Guo <terry.guo@arm.com>
7678 * doc/invoke.texi: Document ARM -march=armv7e-m.
7680 2014-02-08 Jakub Jelinek <jakub@redhat.com>
7682 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
7683 flag on __cilkrts_rethrow builtin.
7686 * ipa-cp.c (determine_versionability): Fail at -O0
7687 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
7688 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
7691 2014-02-04 Jakub Jelinek <jakub@redhat.com>
7694 * tree-inline.c (copy_forbidden): Fail for
7695 __attribute__((optimize (0))) functions.
7697 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
7699 * varpool.c: Include pointer-set.h.
7700 (varpool_remove_unreferenced_decls): Variables in other partitions
7701 will not be output; be however careful to not lose information
7704 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
7706 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
7707 lookup in the vtable constructor.
7709 2014-02-07 Jeff Law <law@redhat.com>
7712 * config/m68k/m68k.md (ashldi_extsi): Turn into a
7713 define_insn_and_split.
7715 * ipa-inline.c (inline_small_functions): Fix typos.
7717 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7719 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
7720 (s390_can_use_return_insn): Declare.
7721 * config/s390/s390.h (EPILOGUE_USES): Define.
7722 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
7724 (s390_chunkify_start): Handle return JUMP_LABELs.
7725 (s390_early_mach): Emit a main_pool instruction on the entry edge.
7726 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
7727 (s390_can_use_return_insn): New functions.
7728 (s390_fix_long_loop_prediction): Handle conditional returns.
7729 (TARGET_SET_UP_BY_PROLOGUE): Define.
7730 * config/s390/s390.md (ANY_RETURN): New code iterator.
7731 (*creturn, *csimple_return, return, simple_return): New patterns.
7733 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7735 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
7736 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
7737 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
7738 REG_CFA_RESTORE list when deciding not to restore a register.
7740 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7742 * config/s390/s390.c: Include tree-pass.h and context.h.
7743 (s390_early_mach): New function, split out from...
7744 (s390_emit_prologue): ...here.
7745 (pass_data_s390_early_mach): New pass structure.
7746 (pass_s390_early_mach): New class.
7747 (s390_option_override): Create and register early_mach pass.
7748 Move to end of file.
7750 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7752 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
7753 to match for the exit block.
7755 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7757 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
7758 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
7759 Reject misaligned operands.
7761 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7763 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
7765 2014-02-07 Richard Biener <rguenther@suse.de>
7768 * gimple-low.c (lower_builtin_posix_memalign): New function.
7769 (lower_stmt): Call it to lower posix_memalign in a way
7770 to make alignment info accessible.
7772 2014-02-07 Jakub Jelinek <jakub@redhat.com>
7775 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
7776 __builtin_setjmp_receiver.
7778 2014-02-07 Richard Biener <rguenther@suse.de>
7781 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
7782 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
7783 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
7784 Handle BUILT_IN_POSIX_MEMALIGN.
7785 (find_func_clobbers): Likewise.
7786 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
7787 (call_may_clobber_ref_p_1): Likewise.
7789 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7792 * ipa-devirt.c (record_target_from_binfo): Remove overactive
7795 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7798 * lto-cgraph.c (lto_output_node): Use
7799 symtab_get_symbol_partitioning_class.
7800 (lto_output_varpool_node): likewise.
7801 (symtab_get_symbol_partitioning_class): Move here from
7803 * cgraph.h (symbol_partitioning_class): Likewise.
7804 (symtab_get_symbol_partitioning_class): Declare.
7806 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7808 * ggc.h (ggc_internal_cleared_alloc): New macro.
7809 * vec.h (vec_safe_copy): Handle memory stats.
7810 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
7811 * target-globals.c (save_target_globals): Likewise.
7813 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
7816 * expr.c (emit_move_resolve_push): Export; be bit more selective
7817 on when to clear alias set.
7818 * expr.h (emit_move_resolve_push): Declare.
7819 * function.h (struct function): Add tail_call_marked.
7820 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
7821 * config/i386/i386-protos.h (ix86_expand_push): Remove.
7822 * config/i386/i386.md (TImode move expander): De not call
7824 (FP push expanders): Preserve memory attributes.
7825 * config/i386/sse.md (push<mode>1): Remove.
7826 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
7827 (ix86_expand_push): Remove.
7828 * config/i386/mmx.md (push<mode>1): Remove.
7830 2014-02-06 Jakub Jelinek <jakub@redhat.com>
7832 PR rtl-optimization/60030
7833 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
7834 lopart with paradoxical subreg before shifting it up by hprec.
7836 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7838 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
7839 Remove extra newline at end of file.
7840 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
7841 (arm_issue_rate): Handle cortexa57.
7842 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
7843 (cortex-a57.cortex-a53): Likewise.
7845 2014-02-06 Jakub Jelinek <jakub@redhat.com>
7848 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
7849 don't record in REG_FRAME_RELATED_EXPR registers not set in that
7851 (arm_expand_prologue): Adjust all callers.
7852 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
7853 info, registers also at the lowest numbered registers side. Use
7854 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
7858 * var-tracking.c (adjust_mems): Before adding a SET to
7859 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
7861 2014-02-06 Alan Modra <amodra@gmail.com>
7864 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
7865 change SDmode to DDmode when lra_in_progress.
7867 2014-02-06 Jakub Jelinek <jakub@redhat.com>
7870 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
7871 free_data_ref on the dr first, and before goto again also set dr
7872 to the next dr. For simd_lane_access, free old datarefs[i] before
7873 overwriting it. For get_vectype_for_scalar_type failure, don't
7874 free_data_ref if simd_lane_access.
7876 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
7879 * tree.h (opts_for_fn): New inline function.
7880 (opt_for_fn): Define.
7881 * config/i386/i386.c (ix86_function_regparm): Use
7882 opt_for_fn (decl, optimize) instead of optimize.
7884 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
7886 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
7887 for SYMBOL_REF in large memory model.
7889 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7891 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
7893 (cortex-a57): Likewise.
7894 (cortex-a57.cortex-a53): Likewise.
7896 2014-02-06 Yury Gribov <y.gribov@samsung.com>
7897 Kugan Vivekanandarajah <kuganv@linaro.org>
7899 * config/arm/arm.c (arm_vector_alignment_reachable): Check
7901 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
7903 2014-02-06 Richard Biener <rguenther@suse.de>
7905 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
7906 set_loop_copy and initialize_original_copy_tables.
7908 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
7910 * config/aarch64/aarch64-simd.md
7911 (aarch64_ashr_simddi): Change QI to SI.
7913 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
7914 Jakub Jelinek <jakub@redhat.com>
7917 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
7920 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7922 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
7923 CODE_FOR_altivec_vpku[hw]um to
7924 CODE_FOR_altivec_vpku[hw]um_direct.
7925 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
7926 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
7927 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
7928 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
7930 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7932 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
7933 generation for -maltivec=be.
7934 (altivec_vsumsws): Simplify redundant test.
7936 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7938 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
7939 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
7940 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
7941 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
7942 gen_altivec_vpkuwum.
7943 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
7945 (altivec_vpks<VI_char>ss): Likewise.
7946 (altivec_vpks<VI_char>us): Likewise.
7947 (altivec_vpku<VI_char>us): Likewise.
7948 (altivec_vpku<VI_char>um): Likewise.
7949 (altivec_vpku<VI_char>um_direct): New (copy of
7950 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
7952 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
7953 target is little endian and -maltivec=be is not specified.
7954 (*altivec_vupkhs<VU_char>_direct): New (copy of
7955 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
7956 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
7957 target is little endian and -maltivec=be is not specified.
7958 (*altivec_vupkls<VU_char>_direct): New (copy of
7959 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
7960 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
7961 little endian and -maltivec=be is not specified.
7962 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
7963 little endian and -maltivec=be is not specified.
7965 2014-02-05 Richard Henderson <rth@redhat.com>
7968 * combine-stack-adj.c: Revert r206943.
7969 * sched-int.h (struct deps_desc): Add last_args_size.
7970 * sched-deps.c (init_deps): Initialize it.
7971 (sched_analyze_insn): Add OUTPUT dependencies between insns that
7972 contain REG_ARGS_SIZE notes.
7974 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
7976 * lto-cgraph.c (asm_nodes_output): Make global.
7977 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
7978 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
7979 (driver_handle_option): Handle OPT_fwpa.
7981 2014-02-05 Jakub Jelinek <jakub@redhat.com>
7984 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
7985 a comment typo and formatting issue. If odr_hash hasn't been
7986 created, return vNULL and set *completep to false.
7989 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
7990 bb with no successors.
7992 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
7995 * doc/invoke.texi (-march): Clarify documentation for ARM.
7999 2014-02-05 Richard Biener <rguenther@suse.de>
8001 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
8002 when not vectorizing because of too many alias checks.
8003 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
8004 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
8006 2014-02-05 Nick Clifton <nickc@redhat.com>
8008 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
8009 accept extended registers in any mode when compiling for the MN10300.
8011 2014-02-05 Yury Gribov <y.gribov@samsung.com>
8013 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
8014 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
8015 sanitization attributes.
8016 (can_inline_edge_p): Likewise.
8017 (sanitize_attrs_match_for_inline_p): New function.
8019 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
8021 * ipa-prop.c (detect_type_change): Shor circuit testing of
8022 type changes on THIS pointer.
8024 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
8027 * config/pa/pa.c (legitimize_tls_address): Return original address
8028 if not passed a SYMBOL_REF rtx.
8029 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
8031 (pa_emit_move_sequence): Simplify TLS source operands.
8032 (pa_legitimate_constant_p): Reject all TLS constants.
8033 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
8034 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
8036 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
8038 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
8039 groups when we know they are controlled by LTO.
8040 * varasm.c (default_binds_local_p_1): If object is in other partition,
8041 it will be resolved locally.
8043 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
8045 * config/host-linux.c (linux_gt_pch_use_address): Don't
8046 use SSIZE_MAX because it is not always defined.
8048 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
8051 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
8052 threshold for pseudo splitting.
8053 (update_ebb_live_info): Process call argument hard registers and
8054 hard registers from insn definition too.
8055 (max_small_class_regs_num): New constant.
8056 (inherit_in_ebb): Update live hard regs through EBBs. Update
8057 reloads_num only for small register classes. Don't split for
8060 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
8063 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
8066 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
8068 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
8071 2014-02-04 Marek Polacek <polacek@redhat.com>
8073 * gdbinit.in (pel): Define.
8075 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
8077 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
8080 2014-02-04 Richard Biener <rguenther@suse.de>
8083 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
8084 in function context local.
8085 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
8086 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
8087 similar to LTO_imported_decl_ref.
8089 2014-02-04 Jakub Jelinek <jakub@redhat.com>
8091 PR tree-optimization/60002
8092 * cgraphclones.c (build_function_decl_skip_args): Clear
8095 PR tree-optimization/60023
8096 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
8097 false to gsi_replace.
8098 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
8099 has been in some EH region and vec_stmt could throw, add
8100 vec_stmt into the same EH region.
8101 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
8102 has no lhs, ignore it.
8103 * internal-fn.c (expand_MASK_LOAD): Likewise.
8106 * tree-inline.c (copy_forbidden): Fail for
8107 __attribute__((optimize (0))) functions.
8110 * omp-low.c (simd_clone_struct_copy): If from->inbranch
8111 is set, copy one less argument.
8112 (expand_simd_clones): Don't subtract clone_info->inbranch
8113 from simd_clone_struct_alloc argument.
8115 PR rtl-optimization/57915
8116 * recog.c (simplify_while_replacing): If all unary/binary/relational
8117 operation arguments are constant, attempt to simplify those.
8120 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
8121 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
8123 2014-02-04 Richard Biener <rguenther@suse.de>
8125 PR tree-optimization/60012
8126 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
8127 TBAA disambiguation to all DDRs.
8129 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8132 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
8133 (LINK_SPEC): Use it for -shared, -shared-libgcc.
8135 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
8138 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
8140 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
8142 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
8143 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
8145 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
8148 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
8149 to figure out targets of polymorphic calls with known decl.
8150 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
8151 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
8152 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
8153 (get_polymorphic_call_info): ... here.
8154 (get_polymorphic_call_info_from_invariant): New function.
8156 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
8158 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
8159 lookup via vtable pointer; check for type consistency
8160 and turn inconsitent facts into UNREACHABLE.
8161 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
8162 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
8163 type inconsistent querries; return UNREACHABLE instead.
8165 2014-02-03 Richard Henderson <rth@twiddle.net>
8168 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
8169 already processed this node.
8170 (normalize_one_pred_1): Pass along mark_set.
8171 (normalize_one_pred): Create and destroy a pointer_set_t.
8172 (normalize_one_pred_chain): Likewise.
8174 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
8176 PR gcov-profile/58602
8177 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
8179 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
8182 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
8183 -fno-devirtualize; try to devirtualize by the knowledge of
8184 virtual table pointer given by aggregate propagation.
8185 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
8186 (ipa_print_node_jump_functions): Dump also offset that
8187 is relevant for polymorphic calls.
8188 (determine_known_aggregate_parts): Add arg_type parameter; use it
8189 instead of determining the type from pointer type.
8190 (ipa_compute_jump_functions_for_edge): Update call of
8191 determine_known_aggregate_parts.
8192 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
8193 (gimple_get_virt_method_for_binfo): ... here; simplify using
8194 vtable_pointer_value_to_vtable.
8195 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
8196 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
8197 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
8198 (vtable_pointer_value_to_vtable): Break out from ...; handle also
8200 (vtable_pointer_value_to_binfo): ... here.
8201 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
8203 2014-02-03 Teresa Johnson <tejohnson@google.com>
8205 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
8206 redef of outer loop index variable.
8208 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
8212 * doc/extend.texi (Function Attributes): Typo.
8214 2014-02-03 Cong Hou <congh@google.com>
8216 PR tree-optimization/60000
8217 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
8218 if the vectorized statement is a store. A store statement can only
8219 appear at the end of pattern statements.
8221 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
8223 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
8224 (ix86_option_override_internal): Default long double to 64-bit for
8225 32-bit Bionic and to 128-bit for 64-bit Bionic.
8227 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
8228 TARGET_LONG_DOUBLE_128 is true.
8229 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
8231 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
8232 (mlong-double-64): Negate -mlong-double-128.
8233 (mlong-double-128): New option.
8235 * config/i386/i386-c.c (ix86_target_macros): Define
8236 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
8238 * doc/invoke.texi: Document -mlong-double-128.
8240 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
8242 PR rtl-optimization/60024
8243 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
8245 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
8247 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
8249 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
8251 PR rtl-optimization/57662
8252 * sel-sched.c (code_motion_path_driver): Do not mark already not
8253 existing blocks in the visiting bitmap.
8255 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
8257 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
8258 on the insn being emitted.
8260 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
8261 Will Deacon <will.deacon@arm.com>
8263 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
8265 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8267 * config/arm/arm-tables.opt: Regenerate.
8269 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8271 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
8272 for vector types other than V16QImode.
8273 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
8274 define_expand, and call altivec_expand_vec_perm_le when producing
8275 code with little endian element order.
8276 (*altivec_vperm_<mode>_internal): New insn having previous
8277 behavior of altivec_vperm_<mode>.
8278 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
8279 altivec_expand_vec_perm_le when producing code with little endian
8281 (*altivec_vperm_<mode>_uns_internal): New insn having previous
8282 behavior of altivec_vperm_<mode>_uns.
8284 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8286 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
8287 (altivec_vsumsws): Add handling for -maltivec=be with a little
8289 (altivec_vsumsws_direct): New.
8290 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
8291 gen_altivec_vsumsws.
8293 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
8295 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
8296 vtable_pointer_value_to_binfo): New functions.
8297 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
8298 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
8300 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
8302 * config/nios2/nios2.md (load_got_register): Initialize GOT
8303 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
8304 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
8306 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
8308 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
8309 preserverd by passthrough, do not propagate the type.
8311 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
8313 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
8314 (mips_atomic_assign_expand_fenv): New function.
8315 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
8317 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
8319 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
8320 (__builtin_mips_set_fcsr): Likewise.
8321 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
8322 MIPS_USI_FTYPE_VOID.
8323 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
8324 (mips16_expand_set_fcsr): Likewise.
8325 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
8326 (mips16_set_fcsr_stub): Likewise.
8327 (mips16_get_fcsr_one_only_stub): New class.
8328 (mips16_set_fcsr_one_only_stub): Likewise.
8329 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
8330 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
8331 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
8332 (hard_float): New availability predicate.
8333 (mips_builtins): Add get_fcsr and set_fcsr.
8334 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
8335 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
8336 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
8337 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
8338 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
8341 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
8343 * config/mips/mips.c (mips_one_only_stub): New class.
8344 (mips_need_mips16_rdhwr_p): Replace with...
8345 (mips16_rdhwr_stub): ...this new variable.
8346 (mips16_stub_call_address): New function.
8347 (mips16_rdhwr_one_only_stub): New class.
8348 (mips_expand_thread_pointer): Use mips16_stub_call_address.
8349 (mips_output_mips16_rdhwr): Delete.
8350 (mips_finish_stub): New function.
8351 (mips_code_end): Use it to handle rdhwr stubs.
8353 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
8356 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
8357 when calculating size of integer atomic types.
8359 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
8361 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
8363 2014-02-01 Jakub Jelinek <jakub@redhat.com>
8365 PR tree-optimization/60003
8366 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
8367 * profile.c (branch_prob): Use gimple_call_builtin_p
8368 to check for BUILT_IN_SETJMP_RECEIVER.
8369 * tree-inline.c (copy_bb): Call notice_special_calls.
8371 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
8374 * lra-constraints.c (process_alt_operands): Update reload_sum only
8377 2014-01-31 Richard Henderson <rth@redhat.com>
8380 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
8381 until after else_eh is processed.
8383 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
8385 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
8386 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
8387 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
8388 in smmintrin.h, remove them.
8389 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
8390 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
8391 * config/i386/i386.md (ROUND_SAE): Fix value.
8392 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
8393 (const48_operand): New.
8394 * config/i386/subst.md (round), (round_expand): Use
8395 const_4_or_8_to_11_operand.
8396 (round_saeonly), (round_saeonly_expand): Use const48_operand.
8398 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
8400 * config/i386/constraints.md (Yk): Swap meaning with k.
8401 * config/i386/i386.md (movhi_internal): Change Yk to k.
8402 (movqi_internal): Ditto.
8403 (*k<logic><mode>): Ditto.
8406 (kandn<mode>): Ditto.
8407 (*<code>hi_1): Ditto.
8408 (*<code>qi_1): Ditto.
8409 (kxnor<mode>): Ditto.
8410 (kortestzhi): Ditto.
8411 (kortestchi): Ditto.
8413 (*one_cmplhi2_1): Ditto.
8414 (*one_cmplqi2_1): Ditto.
8415 * config/i386/sse.md (): Change k to Yk.
8416 (avx512f_load<mode>_mask): Ditto.
8417 (avx512f_blendm<mode>): Ditto.
8418 (avx512f_store<mode>_mask): Ditto.
8419 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
8420 (avx512f_storedqu<mode>_mask): Ditto.
8421 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
8423 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
8424 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
8425 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
8426 (avx512f_maskcmp<mode>3): Ditto.
8427 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
8428 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
8429 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
8430 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
8431 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
8432 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
8433 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
8434 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
8435 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
8436 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
8437 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
8438 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
8439 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
8440 (vec_extract_lo_<mode>_maskm): Ditto.
8441 (vec_extract_hi_<mode>_maskm): Ditto.
8442 (avx512f_vternlog<mode>_mask): Ditto.
8443 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
8444 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
8445 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
8446 (avx512f_<code>v8div16qi2_mask): Ditto.
8447 (avx512f_<code>v8div16qi2_mask_store): Ditto.
8448 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
8449 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
8450 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
8451 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
8452 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
8453 (*avx512pf_gatherpf<mode>df_mask): Ditto.
8454 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
8455 (*avx512pf_scatterpf<mode>df_mask): Ditto.
8456 (avx512cd_maskb_vec_dupv8di): Ditto.
8457 (avx512cd_maskw_vec_dupv16si): Ditto.
8458 (avx512f_vpermi2var<mode>3_maskz): Ditto.
8459 (avx512f_vpermi2var<mode>3_mask): Ditto.
8460 (avx512f_vpermi2var<mode>3_mask): Ditto.
8461 (avx512f_vpermt2var<mode>3_maskz): Ditto.
8462 (*avx512f_gathersi<mode>): Ditto.
8463 (*avx512f_gathersi<mode>_2): Ditto.
8464 (*avx512f_gatherdi<mode>): Ditto.
8465 (*avx512f_gatherdi<mode>_2): Ditto.
8466 (*avx512f_scattersi<mode>): Ditto.
8467 (*avx512f_scatterdi<mode>): Ditto.
8468 (avx512f_compress<mode>_mask): Ditto.
8469 (avx512f_compressstore<mode>_mask): Ditto.
8470 (avx512f_expand<mode>_mask): Ditto.
8471 * config/i386/subst.md (mask): Change k to Yk.
8472 (mask_scalar_merge): Ditto.
8475 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
8477 * doc/extend.texi (Vector Extensions): Document ?: in C++.
8479 2014-01-31 Richard Biener <rguenther@suse.de>
8482 * builtins.c (fold_builtin_memory_op): Make sure to not
8483 use a floating-point mode or a boolean or enumeral type for
8486 2014-01-30 DJ Delorie <dj@redhat.com>
8488 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
8489 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
8490 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
8491 whenever main() has an epilogue.
8493 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8495 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
8496 unused variable "field".
8497 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
8498 (vsx_mergeh_<mode>): Likewise.
8499 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
8500 (altivec_vmrghh): Likewise.
8501 (altivec_vmrghw): Likewise.
8502 (altivec_vmrglb): Likewise.
8503 (altivec_vmrglh): Likewise.
8504 (altivec_vmrglw): Likewise.
8505 (altivec_vspltb): Add missing uses.
8506 (altivec_vsplth): Likewise.
8507 (altivec_vspltw): Likewise.
8508 (altivec_vspltsf): Likewise.
8510 2014-01-30 Jakub Jelinek <jakub@redhat.com>
8513 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
8514 frame related instructions.
8516 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
8518 PR rtl-optimization/59959
8519 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
8520 any reload of register whose subreg is invalid.
8522 2014-01-30 Jakub Jelinek <jakub@redhat.com>
8524 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
8525 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
8526 Add missing return type - void.
8528 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8530 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
8531 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
8532 remove element index adjustment for endian (now handled in vsx.md
8534 (altivec_expand_vec_perm_const): Use
8535 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
8536 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
8537 (vsx_xxspltw_<mode>): Adjust element index for little endian.
8538 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
8539 define_expand and a new define_insn *altivec_vspltb_internal;
8540 adjust for -maltivec=be on a little endian target.
8541 (altivec_vspltb_direct): New.
8542 (altivec_vsplth): Divide into a define_expand and a new
8543 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
8544 little endian target.
8545 (altivec_vsplth_direct): New.
8546 (altivec_vspltw): Divide into a define_expand and a new
8547 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
8548 little endian target.
8549 (altivec_vspltw_direct): New.
8550 (altivec_vspltsf): Divide into a define_expand and a new
8551 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
8552 a little endian target.
8554 2014-01-30 Richard Biener <rguenther@suse.de>
8556 PR tree-optimization/59993
8557 * tree-ssa-forwprop.c (associate_pointerplus): Check we
8558 can propagate form the earlier stmt and avoid the transform
8559 when the intermediate result is needed.
8561 2014-01-30 Alangi Derick <alangiderick@gmail.com>
8563 * README.Portability: Fix typo.
8565 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
8567 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
8568 comparison_operator with ordered_comparison_operator.
8570 2014-01-30 Nick Clifton <nickc@redhat.com>
8572 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
8573 Rename to mn10300_store_multiple_regs.
8574 * config/mn10300/mn10300.c: Likewise.
8575 * config/mn10300/mn10300.md (store_movm): Fix typo: call
8576 store_multiple_regs.
8577 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
8578 Call mn10300_store_multiple_regs.
8580 2014-01-30 Nick Clifton <nickc@redhat.com>
8581 DJ Delorie <dj@redhat.com>
8583 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
8584 %fp 2 to keep registers after it properly word-aligned.
8585 (rl78_alloc_physical_registers_umul): Handle the case where both
8586 input operands are the same.
8588 2014-01-30 Richard Biener <rguenther@suse.de>
8590 PR tree-optimization/59903
8591 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
8594 2014-01-30 Jason Merrill <jason@redhat.com>
8597 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
8600 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
8602 2014-01-30 Richard Biener <rguenther@suse.de>
8604 PR tree-optimization/59951
8605 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
8607 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
8610 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
8611 SFmode to DFmode case.
8613 2014-01-29 DJ Delorie <dj@redhat.com>
8615 * config/msp430/msp430.opt (-minrt): New.
8616 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
8618 (ENDFILE_SPEC): Likewise.
8620 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
8622 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
8623 (estimate_function_body_sizes): Use it.
8625 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
8628 * dwarf2out.c (is_cxx_auto): New.
8629 (is_base_type): Use it.
8630 (gen_type_die_with_usage): Likewise.
8632 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8634 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
8635 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
8636 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
8637 -maltivec=be with LE targets.
8638 (vsx_mergeh_<mode>): Likewise.
8639 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
8640 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
8641 (altivec_vmrghb): Replace with define_expand and new
8642 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
8643 (altivec_vmrghb_direct): New define_insn.
8644 (altivec_vmrghh): Replace with define_expand and new
8645 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
8646 (altivec_vmrghh_direct): New define_insn.
8647 (altivec_vmrghw): Replace with define_expand and new
8648 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
8649 (altivec_vmrghw_direct): New define_insn.
8650 (*altivec_vmrghsf): Adjust for endianness.
8651 (altivec_vmrglb): Replace with define_expand and new
8652 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
8653 (altivec_vmrglb_direct): New define_insn.
8654 (altivec_vmrglh): Replace with define_expand and new
8655 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
8656 (altivec_vmrglh_direct): New define_insn.
8657 (altivec_vmrglw): Replace with define_expand and new
8658 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
8659 (altivec_vmrglw_direct): New define_insn.
8660 (*altivec_vmrglsf): Adjust for endianness.
8661 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
8662 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
8663 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
8664 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
8665 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
8666 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
8667 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
8668 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
8670 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
8672 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
8673 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
8676 2014-01-29 Richard Biener <rguenther@suse.de>
8678 PR tree-optimization/58742
8679 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
8680 associate_pointerplus_align.
8681 (associate_pointerplus_diff): New function.
8682 (associate_pointerplus): Likewise. Call associate_pointerplus_align
8683 and associate_pointerplus_diff.
8685 2014-01-29 Richard Biener <rguenther@suse.de>
8687 * lto-streamer.h (LTO_major_version): Bump to 3.
8688 (LTO_minor_version): Reset to 0.
8690 2014-01-29 Renlin Li <Renlin.Li@arm.com>
8692 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
8693 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
8694 (arm_file_start): Generate correct asm header for armv7ve.
8695 * config/arm/bpabi.h: Add multilib support for armv7ve.
8696 * config/arm/driver-arm.c: Change the architectures of cortex-a7
8697 and cortex-a15 to armv7ve.
8698 * config/arm/t-aprofile: Add multilib support for armv7ve.
8699 * doc/invoke.texi: Document -march=armv7ve.
8701 2014-01-29 Richard Biener <rguenther@suse.de>
8703 PR tree-optimization/58742
8704 * tree-ssa-forwprop.c (associate_plusminus): Return true
8705 if we changed sth, defer EH cleanup to ...
8706 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
8707 (simplify_mult): New function.
8709 2014-01-29 Jakub Jelinek <jakub@redhat.com>
8712 PR tree-optimization/59920
8713 * tree.c (build_common_builtin_nodes): Remove
8714 __builtin_setjmp_dispatcher initialization.
8715 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
8716 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
8717 instead of gsi_after_labels + manually skipping debug stmts.
8718 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
8719 ignore bbs with IFN_ABNORMAL_DISPATCHER.
8720 * tree-inline.c (copy_edges_for_bb): Remove
8721 can_make_abnormal_goto argument, instead add abnormal_goto_dest
8722 argument. Ignore computed_goto_p stmts. Don't call
8723 make_abnormal_goto_edges. If a call might need abnormal edges
8724 for non-local gotos, see if it already has an edge to
8725 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
8726 with true argument, don't do anything then, otherwise add
8727 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
8728 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
8730 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
8731 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
8732 (lower_stmt): Don't set data->calls_builtin_setjmp.
8733 (lower_builtin_setjmp): Adjust comment.
8734 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
8735 * tree-cfg.c (found_computed_goto): Remove.
8736 (factor_computed_gotos): Remove.
8737 (make_goto_expr_edges): Return bool, true for computed gotos.
8738 Don't call make_abnormal_goto_edges.
8739 (build_gimple_cfg): Don't set found_computed_goto, don't call
8740 factor_computed_gotos.
8741 (computed_goto_p): No longer static.
8742 (make_blocks): Don't set found_computed_goto.
8743 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
8744 (make_edges): If make_goto_expr_edges returns true, push bb
8745 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
8746 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
8747 vector. Record mapping between bbs and OpenMP regions if there
8748 are any, adjust make_gimple_omp_edges caller. Call
8749 handle_abnormal_edges.
8750 (make_abnormal_goto_edges): Remove.
8751 * tree-cfg.h (make_abnormal_goto_edges): Remove.
8752 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
8753 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
8754 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
8755 * internal-fn.def (ABNORMAL_DISPATCHER): New.
8756 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
8757 filling *region also set *region_idx to (*region)->entry->index.
8760 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
8761 For REGs set ORIGINAL_REGNO.
8763 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
8765 * doc/md.texi: Mention that a target shouldn't implement
8766 vec_widen_(s|u)mul_even/odd pair if it is less efficient
8769 2014-01-29 Jakub Jelinek <jakub@redhat.com>
8771 PR tree-optimization/59594
8772 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
8773 a copy of the datarefs vector rather than the vector itself.
8775 2014-01-28 Jason Merrill <jason@redhat.com>
8778 * dwarf2out.c (auto_die): New static.
8779 (gen_type_die_with_usage): Handle C++1y 'auto'.
8780 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
8783 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
8786 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
8787 (SPEC_X32): Likewise.
8788 (SPEC_64): Likewise.
8789 * config/i386/i386.c (ix86_option_override_internal): Turn off
8790 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
8792 (x86_file_start): Output .code16gcc for TARGET_16BIT.
8793 * config/i386/i386.h (TARGET_16BIT): New macro.
8794 (TARGET_16BIT_P): Likewise.
8795 * config/i386/i386.opt: Add m16.
8796 * doc/invoke.texi: Document -m16.
8798 2014-01-28 Jakub Jelinek <jakub@redhat.com>
8800 PR preprocessor/59935
8801 * input.c (location_get_source_line): Bail out on when line number
8802 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
8804 2014-01-28 Richard Biener <rguenther@suse.de>
8806 PR tree-optimization/58742
8807 * tree-ssa-forwprop.c (associate_plusminus): Handle
8808 pointer subtraction of the form (T)(P + A) - (T)P.
8810 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8812 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
8815 2014-01-28 Richard Biener <rguenther@suse.de>
8818 2014-01-28 Richard Biener <rguenther@suse.de>
8820 PR rtl-optimization/45364
8821 PR rtl-optimization/59890
8822 * var-tracking.c (local_get_addr_clear_given_value): Handle
8823 already cleared slot.
8824 (val_reset): Handle not allocated local_get_addr_cache.
8825 (vt_find_locations): Use post-order on the inverted CFG.
8827 2014-01-28 Richard Biener <rguenther@suse.de>
8829 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
8831 2014-01-28 Richard Biener <rguenther@suse.de>
8833 PR rtl-optimization/45364
8834 PR rtl-optimization/59890
8835 * var-tracking.c (local_get_addr_clear_given_value): Handle
8836 already cleared slot.
8837 (val_reset): Handle not allocated local_get_addr_cache.
8838 (vt_find_locations): Use post-order on the inverted CFG.
8840 2014-01-28 Alan Modra <amodra@gmail.com>
8842 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
8843 * configure.ac <recursive call for build != host>: Define
8844 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
8845 and LD_FOR_BUILD too.
8846 * configure: Regenerate.
8848 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
8850 * config/i386/i386.c (get_builtin_code_for_version): Separate
8851 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
8852 Broadwell from Haswell.
8854 2014-01-27 Steve Ellcey <sellcey@mips.com>
8856 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
8857 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
8858 * config/mips/mips.c (mips_option_override): Change setting
8860 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
8861 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
8862 Change from Mask to Var.
8864 2014-01-27 Jeff Law <law@redhat.com>
8866 * ipa-inline.c (inline_small_functions): Fix typo.
8868 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
8870 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
8871 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
8872 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
8873 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
8874 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
8875 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
8876 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
8877 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
8878 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
8879 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
8880 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
8881 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
8882 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
8883 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
8884 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
8885 (_mm512_storeu_epi64): Ditto.
8886 (_mm512_cmpge_epi32_mask): Ditto.
8887 (_mm512_cmpge_epu32_mask): Ditto.
8888 (_mm512_cmpge_epi64_mask): Ditto.
8889 (_mm512_cmpge_epu64_mask): Ditto.
8890 (_mm512_cmple_epi32_mask): Ditto.
8891 (_mm512_cmple_epu32_mask): Ditto.
8892 (_mm512_cmple_epi64_mask): Ditto.
8893 (_mm512_cmple_epu64_mask): Ditto.
8894 (_mm512_cmplt_epi32_mask): Ditto.
8895 (_mm512_cmplt_epu32_mask): Ditto.
8896 (_mm512_cmplt_epi64_mask): Ditto.
8897 (_mm512_cmplt_epu64_mask): Ditto.
8898 (_mm512_cmpneq_epi32_mask): Ditto.
8899 (_mm512_cmpneq_epu32_mask): Ditto.
8900 (_mm512_cmpneq_epi64_mask): Ditto.
8901 (_mm512_cmpneq_epu64_mask): Ditto.
8902 (_mm512_expand_pd): Ditto.
8903 (_mm512_expand_ps): Ditto.
8904 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
8905 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
8906 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
8907 * config/i386/i386.c (ix86_builtins): Add
8908 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
8909 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
8910 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
8911 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
8912 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
8913 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
8914 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
8915 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
8916 IX86_BUILTIN_PMOVUSQW512_MEM.
8917 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
8918 __builtin_ia32_pmovsqd512mem_mask,
8919 __builtin_ia32_pmovqd512mem_mask,
8920 __builtin_ia32_pmovusqw512mem_mask,
8921 __builtin_ia32_pmovsqw512mem_mask,
8922 __builtin_ia32_pmovqw512mem_mask,
8923 __builtin_ia32_pmovusdw512mem_mask,
8924 __builtin_ia32_pmovsdw512mem_mask,
8925 __builtin_ia32_pmovdw512mem_mask,
8926 __builtin_ia32_pmovqb512mem_mask,
8927 __builtin_ia32_pmovusqb512mem_mask,
8928 __builtin_ia32_pmovsqb512mem_mask,
8929 __builtin_ia32_pmovusdb512mem_mask,
8930 __builtin_ia32_pmovsdb512mem_mask,
8931 __builtin_ia32_pmovdb512mem_mask.
8932 (bdesc_args): Add __builtin_ia32_expanddf512,
8933 __builtin_ia32_expandsf512.
8934 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
8935 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
8936 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
8937 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
8938 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
8939 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
8940 (avx512f_<code>v8div16qi2_mask_store): This.
8941 (avx512f_expand<mode>): New.
8943 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
8945 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
8947 (_mm512_mask_prefetch_i64gather_pd): Ditto.
8948 (_mm512_prefetch_i32scatter_pd): Ditto.
8949 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
8950 (_mm512_prefetch_i64scatter_pd): Ditto.
8951 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
8952 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
8953 (_mm512_mask_prefetch_i64gather_ps): Ditto.
8954 (_mm512_prefetch_i32scatter_ps): Ditto.
8955 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
8956 (_mm512_prefetch_i64scatter_ps): Ditto.
8957 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
8958 * config/i386/i386-builtin-types.def: Define
8959 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
8960 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
8961 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
8962 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
8963 IX86_BUILTIN_SCATTERPFQPD.
8964 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
8965 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
8966 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
8967 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
8968 __builtin_ia32_scatterpfqps.
8969 (ix86_expand_builtin): Expand new built-ins.
8970 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
8971 fix memory access data type.
8972 (*avx512pf_gatherpf<mode>_mask): Ditto.
8973 (*avx512pf_gatherpf<mode>): Ditto.
8974 (avx512pf_scatterpf<mode>): Ditto.
8975 (*avx512pf_scatterpf<mode>_mask): Ditto.
8976 (*avx512pf_scatterpf<mode>): Ditto.
8977 (GATHER_SCATTER_SF_MEM_MODE): New.
8978 (avx512pf_gatherpf<mode>df): Ditto.
8979 (*avx512pf_gatherpf<mode>df_mask): Ditto.
8980 (*avx512pf_scatterpf<mode>df): Ditto.
8982 2014-01-27 Jakub Jelinek <jakub@redhat.com>
8985 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
8986 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
8989 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
8991 * common/config/arm/arm-common.c
8992 (arm_rewrite_mcpu): Handle multiple names.
8994 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
8996 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
8998 * gimple-builder.h (create_gimple_tmp): Delete.
9000 2014-01-27 Christian Bruel <christian.bruel@st.com>
9002 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
9005 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
9007 * config/pa/pa.md (call): Generate indirect long calls to non-local
9008 functions when outputing 32-bit code.
9009 (call_value): Likewise except for special call to buggy powf function.
9011 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
9012 portable runtime and PIC indirect calls.
9013 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
9014 and PIC call sequences. Use ldo instead of blr to set return register
9015 in PIC call sequence.
9017 2014-01-25 Walter Lee <walt@tilera.com>
9019 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
9020 avoid clobbering a live register.
9022 2014-01-25 Walter Lee <walt@tilera.com>
9024 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
9025 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
9026 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
9027 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
9029 2014-01-25 Walter Lee <walt@tilera.com>
9031 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
9032 arguments on even registers.
9033 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
9035 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
9036 (BIGGEST_ALIGNMENT): Ditto.
9037 (BIGGEST_FIELD_ALIGNMENT): Ditto.
9039 2014-01-25 Walter Lee <walt@tilera.com>
9041 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
9042 insns before bundling.
9043 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
9045 2014-01-25 Walter Lee <walt@tilera.com>
9047 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
9048 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
9049 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
9051 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
9053 * config/mips/constraints.md (kl): Delete.
9054 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
9055 define expands, using...
9056 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
9057 instructions for MIPS16.
9058 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
9059 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
9061 2014-01-25 Walter Lee <walt@tilera.com>
9063 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
9067 2014-01-25 Walter Lee <walt@tilera.com>
9069 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
9070 (TARGET_EXPAND_TO_RTL_HOOK): Define.
9072 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
9074 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
9077 2014-01-25 Jakub Jelinek <jakub@redhat.com>
9079 * print-rtl.c (in_call_function_usage): New var.
9080 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
9081 EXPR_LIST mode as mode and not as reg note name.
9084 * cfgloopmanip.c (copy_loop_info): If
9085 loop->warned_aggressive_loop_optimizations, make sure
9086 the flag is set in target loop too.
9088 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
9090 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
9092 * builtins.def: Likewise.
9093 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
9094 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
9095 * ira.c (ira_setup_eliminable_regset): Likewise.
9096 * omp-low.c (gate_expand_omp): Likewise.
9097 (execute_lower_omp): Likewise.
9098 (diagnose_sb_0): Likewise.
9099 (gate_diagnose_omp_blocks): Likewise.
9100 (simd_clone_clauses_extract): Likewise.
9103 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9105 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
9106 correction for little endian...
9107 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
9110 2014-01-24 Jeff Law <law@redhat.com>
9112 PR tree-optimization/59919
9113 * tree-vrp.c (find_assert_locations_1): Do not register asserts
9114 for non-returning calls.
9116 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
9118 * common/config/aarch64/aarch64-common.c
9119 (aarch64_rewrite_mcpu): Handle multiple names.
9120 * config/aarch64/aarch64.h
9121 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
9123 2014-01-24 Dodji Seketeli <dodji@redhat.com>
9125 * input.c (add_file_to_cache_tab): Handle the case where fopen
9128 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
9131 * config/i386/i386.md (pushsf splitter): Get stack adjustment
9132 from push operand if code of push isn't PRE_DEC.
9134 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
9137 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
9138 -mquad-memory-atomic. Update -mquad-memory documentation to say
9139 it is only used for non-atomic loads/stores.
9141 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
9142 -mquad-memory or -mquad-memory-atomic switches.
9144 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
9145 -mquad-memory-atomic to ISA 2.07 support.
9147 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
9148 to separate support of normal quad word memory operations (ldq, stq)
9149 from the atomic quad word memory operations.
9151 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
9152 support to separate non-atomic quad word operations from atomic
9153 quad word operations. Disable non-atomic quad word operations in
9154 little endian mode so that we don't have to swap words after the
9155 load and before the store.
9156 (quad_load_store_p): Add comment about atomic quad word support.
9157 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
9158 options printed with -mdebug=reg.
9160 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
9161 -mquad-memory-atomic as the test for whether we have quad word
9162 atomic instructions.
9163 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
9164 or -mp8-vector are used, allow byte/half-word atomic operations.
9166 * config/rs6000/sync.md (load_lockedti): Insure that the address
9167 is a proper indexed or indirect address for the lqarx instruction.
9168 On little endian systems, swap the hi/lo registers after the lqarx
9170 (load_lockedpti): Use indexed_or_indirect_operand predicate to
9171 insure the address is valid for the lqarx instruction.
9172 (store_conditionalti): Insure that the address is a proper indexed
9173 or indirect address for the stqcrx. instruction. On little endian
9174 systems, swap the hi/lo registers before doing the stqcrx.
9176 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
9177 insure the address is valid for the stqcrx. instruction.
9179 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
9180 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
9181 type of quad memory support is available.
9183 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
9186 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
9187 there is a danger of looping.
9189 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
9191 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
9192 force flag_ira_loop_pressure if set via command line.
9194 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
9196 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
9197 (ashr_simd): New builtin handling DI mode.
9198 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
9199 (aarch64_sshr_simddi): New match pattern.
9200 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
9201 (vshrd_n_s64): Likewise.
9202 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
9204 2014-01-23 Nick Clifton <nickc@redhat.com>
9206 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
9207 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
9208 favour of mcu specific scripts.
9209 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
9212 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
9213 Alex Velenko <Alex.Velenko@arm.com>
9215 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
9216 (vaddv_s16): Likewise.
9217 (vaddv_s32): Likewise.
9218 (vaddv_u8): Likewise.
9219 (vaddv_u16): Likewise.
9220 (vaddv_u32): Likewise.
9221 (vaddvq_s8): Likewise.
9222 (vaddvq_s16): Likewise.
9223 (vaddvq_s32): Likewise.
9224 (vaddvq_s64): Likewise.
9225 (vaddvq_u8): Likewise.
9226 (vaddvq_u16): Likewise.
9227 (vaddvq_u32): Likewise.
9228 (vaddvq_u64): Likewise.
9229 (vaddv_f32): Likewise.
9230 (vaddvq_f32): Likewise.
9231 (vaddvq_f64): Likewise.
9232 (vmaxv_f32): Likewise.
9233 (vmaxv_s8): Likewise.
9234 (vmaxv_s16): Likewise.
9235 (vmaxv_s32): Likewise.
9236 (vmaxv_u8): Likewise.
9237 (vmaxv_u16): Likewise.
9238 (vmaxv_u32): Likewise.
9239 (vmaxvq_f32): Likewise.
9240 (vmaxvq_f64): Likewise.
9241 (vmaxvq_s8): Likewise.
9242 (vmaxvq_s16): Likewise.
9243 (vmaxvq_s32): Likewise.
9244 (vmaxvq_u8): Likewise.
9245 (vmaxvq_u16): Likewise.
9246 (vmaxvq_u32): Likewise.
9247 (vmaxnmv_f32): Likewise.
9248 (vmaxnmvq_f32): Likewise.
9249 (vmaxnmvq_f64): Likewise.
9250 (vminv_f32): Likewise.
9251 (vminv_s8): Likewise.
9252 (vminv_s16): Likewise.
9253 (vminv_s32): Likewise.
9254 (vminv_u8): Likewise.
9255 (vminv_u16): Likewise.
9256 (vminv_u32): Likewise.
9257 (vminvq_f32): Likewise.
9258 (vminvq_f64): Likewise.
9259 (vminvq_s8): Likewise.
9260 (vminvq_s16): Likewise.
9261 (vminvq_s32): Likewise.
9262 (vminvq_u8): Likewise.
9263 (vminvq_u16): Likewise.
9264 (vminvq_u32): Likewise.
9265 (vminnmv_f32): Likewise.
9266 (vminnmvq_f32): Likewise.
9267 (vminnmvq_f64): Likewise.
9269 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
9271 * config/aarch64/aarch64-simd.md
9272 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
9273 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
9274 (*aarch64_mul3_elt<mode>): Likewise.
9275 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
9276 (*aarch64_mul3_elt_to_64v2df): Likewise.
9277 (*aarch64_mla_elt<mode>): Likewise.
9278 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
9279 (*aarch64_mls_elt<mode>): Likewise.
9280 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
9281 (*aarch64_fma4_elt<mode>): Likewise.
9282 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
9283 (*aarch64_fma4_elt_to_64v2df): Likewise.
9284 (*aarch64_fnma4_elt<mode>): Likewise.
9285 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
9286 (*aarch64_fnma4_elt_to_64v2df): Likewise.
9287 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
9288 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
9289 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
9290 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
9291 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
9292 (aarch64_sqdmull_lane<mode>_internal): Likewise.
9293 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
9295 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
9297 * config/aarch64/aarch64-simd.md
9298 (aarch64_be_checked_get_lane<mode>): New define_expand.
9299 * config/aarch64/aarch64-simd-builtins.def
9300 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
9301 New builtin definition.
9302 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
9303 Use new safe be builtin.
9305 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
9307 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
9309 (aarch64_be_st1<mode>): Likewise.
9310 (aarch_ld1<VALL:mode>): Define_expand modified.
9311 (aarch_st1<VALL:mode>): Likewise.
9312 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
9313 (UNSPEC_ST1): Likewise.
9315 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
9317 * config/microblaze/microblaze.md: Add trap insn and attribute
9319 2014-01-23 Dodji Seketeli <dodji@redhat.com>
9321 PR preprocessor/58580
9322 * input.h (location_get_source_line): Take an additional line_size
9324 (void diagnostics_file_cache_fini): Declare new function.
9325 * input.c (struct fcache): New type.
9326 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
9327 New static constants.
9328 (diagnostic_file_cache_init, total_lines_num)
9329 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
9330 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
9331 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
9332 (get_next_line, read_next_line, goto_next_line, read_line_num):
9333 New static function definitions.
9334 (diagnostic_file_cache_fini): New function.
9335 (location_get_source_line): Take an additional output line_len
9336 parameter. Re-write using lookup_or_add_file_to_cache_tab and
9338 * diagnostic.c (diagnostic_finish): Call
9339 diagnostic_file_cache_fini.
9340 (adjust_line): Take an additional input parameter for the length
9341 of the line, rather than calculating it with strlen.
9342 (diagnostic_show_locus): Adjust the use of
9343 location_get_source_line and adjust_line with respect to their new
9344 signature. While displaying a line now, do not stop at the first
9345 null byte. Rather, display the zero byte as a space and keep
9346 going until we reach the size of the line.
9347 * Makefile.in: Add vec.o to OBJS-libcommon
9349 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
9350 Ilya Tocar <ilya.tocar@intel.com>
9352 * config/i386/avx512fintrin.h (_mm512_kmov): New.
9353 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
9354 (__builtin_ia32_kmov16): Ditto.
9355 * config/i386/i386.md (UNSPEC_KMOV): New.
9358 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
9360 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
9361 (_mm512_storeu_si512): Ditto.
9363 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
9366 * rtl.h (get_referenced_operands): Declare.
9367 * recog.c (get_referenced_operands): New function.
9368 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
9369 operands have been referenced when recording LO_SUM references.
9371 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
9373 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
9375 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
9377 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
9378 Enable for generic and recent AMD targets.
9380 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
9382 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
9383 ARG_SIZE note when adjustment was eliminated.
9385 2014-01-22 Jeff Law <law@redhat.com>
9387 PR tree-optimization/59597
9388 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
9389 in file. Accept new argument REGISTERING and use it to modify
9390 dump output appropriately.
9391 (register_jump_thread): Corresponding changes.
9392 (mark_threaded_blocks): Reinstate code to cancel unprofitable
9393 thread paths involving joiner blocks. Add code to dump cancelled
9394 jump threading paths.
9396 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
9398 PR rtl-optimization/59477
9399 * lra-constraints.c (inherit_in_ebb): Process call for living hard
9400 regs. Update reloads_num and potential_reload_hard_regs for all insns.
9402 2014-01-22 Tom Tromey <tromey@redhat.com>
9404 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
9406 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
9408 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
9410 PR rtl-optimization/59896
9411 * lra-constraints.c (process_alt_operands): Check unused note for
9412 matched operands of insn with no output reloads.
9414 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
9416 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
9417 (mips_move_from_gpr_cost): Likewise.
9419 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
9421 PR rtl-optimization/59858
9422 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
9423 ira_class_hard_regs_num.
9424 (process_alt_operands): Increase reject for dying matched operand.
9426 2014-01-21 Jakub Jelinek <jakub@redhat.com>
9429 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
9430 smaller than size, perform several stores or loads and stores
9431 at dst + count - size to store or copy all of size bytes, rather
9432 than just last modesize bytes.
9434 2014-01-20 DJ Delorie <dj@redhat.com>
9436 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
9437 that CLOBBERs are REGs before propogating their values.
9439 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
9442 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
9443 (cgraph_inline_failed_type): New function.
9444 * cgraph.h (DEFCIFCODE): Add type.
9445 (cgraph_inline_failed_type_t): New enum.
9446 (cgraph_inline_failed_type): New prototype.
9447 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
9448 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
9449 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
9450 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
9451 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
9452 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
9453 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
9454 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
9455 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
9456 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
9457 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
9458 OPTIMIZATION_MISMATCH.
9459 * tree-inline.c (expand_call_inline): Emit errors during
9460 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
9462 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
9465 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
9466 mode attribute in insn output.
9468 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
9470 * output.h (output_constant): Delete.
9471 * varasm.c (output_constant): Make private.
9473 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
9475 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
9477 2014-01-20 Jakub Jelinek <jakub@redhat.com>
9480 * tree.h (fold_builtin_strcat): New prototype.
9481 * builtins.c (fold_builtin_strcat): No longer static. Add len
9482 argument, if non-NULL, don't call c_strlen. Optimize
9483 directly into __builtin_memcpy instead of __builtin_strcpy.
9484 (fold_builtin_2): Adjust fold_builtin_strcat caller.
9485 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
9487 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
9489 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
9490 for SImode_address_operand operands, having only a REG argument.
9492 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
9494 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
9495 loader name using mbig-endian.
9496 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
9498 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
9500 * doc/invoke.texi (-march): Clarify documentation for AArch64.
9504 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
9506 * config/aarch64/aarch64-protos.h
9507 (aarch64_cannot_change_mode_class_ptr): Declare.
9508 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
9509 aarch64_cannot_change_mode_class_ptr): New.
9510 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
9511 backend hook aarch64_cannot_change_mode_class.
9513 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
9515 * common/config/aarch64/aarch64-common.c
9516 (aarch64_handle_option): Don't handle any option order logic here.
9517 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
9518 selected_cpu, warn on architecture version mismatch.
9519 (aarch64_override_options): Fix parsing order for option strings.
9521 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
9522 Iain Sandoe <iain@codesourcery.com>
9525 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
9526 warning. Amend comment to reflect current functionality.
9528 2014-01-20 Richard Biener <rguenther@suse.de>
9531 * builtins.c (fold_builtin_strcat): Remove case better handled
9532 by tree-ssa-strlen.c.
9534 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
9536 * config/aarch64/aarch64.opt
9537 (mcpu, march, mtune): Make case-insensitive.
9539 2014-01-20 Jakub Jelinek <jakub@redhat.com>
9542 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
9543 if operands[1] is a REG or ZERO_EXTEND of a REG.
9545 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
9547 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
9549 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
9551 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
9552 long non-pic millicode calls.
9554 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
9556 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
9558 2014-01-19 Kito Cheng <kito@0xlab.org>
9560 * builtins.c (expand_movstr): Check movstr expand done or fail.
9562 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
9563 H.J. Lu <hongjiu.lu@intel.com>
9566 * config/i386/i386.md (*lea<mode>): Zero-extend return register
9567 to DImode for zero-extended addresses.
9569 2014-01-19 Jakub Jelinek <jakub@redhat.com>
9571 PR rtl-optimization/57763
9572 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
9573 on the new indirect jump_insn and increment LABEL_NUSES (label).
9575 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
9579 * config.gcc (x86_archs): New variable.
9580 (x86_64_archs): Likewise.
9581 (x86_cpus): Likewise.
9582 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
9583 --with-arch/--with-cpu= options.
9584 Support --with-arch=/--with-cpu={nehalem,westmere,
9585 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
9587 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
9589 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
9590 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
9592 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
9594 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
9596 2014-01-18 Jakub Jelinek <jakub@redhat.com>
9599 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
9600 clear cpp_get_options (parse_in)->warn_unused_macros for
9601 ix86_target_macros_internal with cpp_define.
9603 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
9605 * jump.c (delete_related_insns): Keep (use (insn))s.
9606 * reorg.c (redundant_insn): Check for barriers too.
9608 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
9610 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
9612 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
9614 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
9615 call to $$dyncall when TARGET_LONG_CALLS is true.
9617 2014-01-17 Jeff Law <law@redhat.com>
9619 * ree.c (combine_set_extension): Temporarily disable test for
9620 changing number of hard registers.
9622 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
9625 * ipa-inline-analysis.c (inline_free_summary):
9626 Do not free summary of aliases.
9628 2014-01-17 Jakub Jelinek <jakub@redhat.com>
9631 * gimplify.c (gimplify_expr): Use create_tmp_var
9632 instead of create_tmp_var_raw. If cond doesn't have
9633 integral type, don't add the IFN_ANNOTATE builtin at all.
9635 2014-01-17 Martin Jambor <mjambor@suse.cz>
9638 * ipa-cp.c (prev_edge_clone): New variable.
9639 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
9640 Also resize prev_edge_clone vector.
9641 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
9642 (ipcp_edge_removal_hook): New function.
9643 (ipcp_driver): Register ipcp_edge_removal_hook.
9645 2014-01-17 Andrew Pinski <apinski@cavium.com>
9646 Steve Ellcey <sellcey@mips.com>
9649 * config/mips/mips.c (mips_print_operand): Check operand mode instead
9652 2014-01-17 Jeff Law <law@redhat.com>
9655 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
9656 so that pass_ccp runs first.
9658 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
9660 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
9661 (ix86_adjust_cost): Use !TARGET_XXX.
9662 (do_reorder_for_imul): Likewise.
9663 (swap_top_of_ready_list): Likewise.
9664 (ix86_sched_reorder): Likewise.
9666 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
9668 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
9669 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
9670 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
9671 (intel_memset): New. Duplicate slm_memset.
9672 (intel_cost): New. Duplicate slm_cost.
9673 (m_INTEL): New macro.
9674 (processor_target_table): Add "intel".
9675 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
9676 with PROCESSOR_INTEL for "intel".
9677 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
9678 PROCESSOR_SILVERMONT.
9679 (ix86_issue_rate): Likewise.
9680 (ix86_adjust_cost): Likewise.
9681 (ia32_multipass_dfa_lookahead): Likewise.
9682 (swap_top_of_ready_list): Likewise.
9683 (ix86_sched_reorder): Likewise.
9684 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
9685 instead of TARGET_OPT_AGU.
9686 * config/i386/i386.h (TARGET_INTEL): New.
9687 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
9688 (processor_type): Add PROCESSOR_INTEL.
9689 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
9690 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
9692 2014-01-17 Marek Polacek <polacek@redhat.com>
9695 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
9698 2014-01-17 Richard Biener <rguenther@suse.de>
9700 PR tree-optimization/46590
9701 * opts.c (default_options_table): Add entries for
9702 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
9703 all enabled at -O1 but not for -Og.
9704 * common.opt (fbranch-count-reg): Remove Init(1).
9705 (fmove-loop-invariants): Likewise.
9706 (ftree-pta): Likewise.
9708 2014-01-17 Jakub Jelinek <jakub@redhat.com>
9710 * config/i386/i386.c (ix86_data_alignment): For compatibility with
9711 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
9712 decls to at least the GCC 4.8 used alignments.
9715 * tree-nested.c (convert_nonlocal_reference_stmt,
9716 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
9717 of GIMPLE_BIND stmts, adjust associated decls.
9719 2014-01-17 Richard Biener <rguenther@suse.de>
9721 PR tree-optimization/46590
9722 * vec.h (vec<>::bseach): New member function implementing
9723 binary search according to C89 bsearch.
9724 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
9725 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
9726 bitmap pointer again. Make accesses_in_loop a flat array.
9727 (mem_ref_obstack): New global.
9728 (outermost_indep_loop): Adjust for mem_ref->stored changes.
9729 (mark_ref_stored): Likewise.
9730 (ref_indep_loop_p_2): Likewise.
9731 (set_ref_stored_in_loop): New helper function.
9732 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
9733 (memref_free): Adjust.
9734 (record_mem_ref_loc): Simplify.
9735 (gather_mem_refs_stmt): Adjust.
9736 (sort_locs_in_loop_postorder_cmp): New function.
9737 (analyze_memory_references): Sort accesses_in_loop after
9738 loop postorder number.
9739 (find_ref_loc_in_loop_cmp): New function.
9740 (for_all_locs_in_loop): Find relevant cluster of locs in
9741 accesses_in_loop and iterate without recursion.
9742 (execute_sm): Avoid uninit warning.
9743 (struct ref_always_accessed): Simplify.
9744 (ref_always_accessed::operator ()): Likewise.
9745 (ref_always_accessed_p): Likewise.
9746 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
9747 loop postorder numbers here.
9748 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
9751 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
9754 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
9755 on decls for which assemble_alias has been called.
9757 2014-01-17 Nick Clifton <nickc@redhat.com>
9759 * config/msp430/msp430.opt: (mcpu): New option.
9760 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
9761 (msp430_option_override): Parse target_cpu. If the MCU name
9762 matches a generic string, clear target_mcu.
9763 (msp430_attr): Allow numeric interrupt values up to 63.
9764 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
9765 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
9767 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
9769 * config/msp430/msp430.md (popm): Use %J rather than %I.
9770 (addsi3): Use msp430_nonimmediate_operand for operand 2.
9771 (addhi_cy_i): Use immediate_operand for operand 2.
9772 * doc/invoke.texi: Document -mcpu option.
9774 2014-01-17 Richard Biener <rguenther@suse.de>
9776 PR rtl-optimization/38518
9777 * df.h (df_analyze_loop): Declare.
9778 * df-core.c: Include cfgloop.h.
9779 (df_analyze_1): Split out main part of df_analyze.
9780 (df_analyze): Adjust.
9781 (loop_inverted_post_order_compute): New function.
9782 (loop_post_order_compute): Likewise.
9783 (df_analyze_loop): New function avoiding whole-function
9785 * loop-invariant.c (find_defs): Use df_analyze_loop.
9786 (find_invariants): Adjust.
9787 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
9789 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
9791 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
9792 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
9794 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
9796 * ipa-ref.c (ipa_remove_stmt_references): Fix references
9797 traversal when removing references.
9799 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
9802 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
9804 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
9807 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
9808 pushing a reload for an autoinc when we had previously reloaded an
9809 inner part of the address.
9811 2014-01-16 Jakub Jelinek <jakub@redhat.com>
9813 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
9815 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
9816 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
9817 when not giving up or versioning for alias only because of
9819 (vect_analyze_data_ref_dependences): Set to true.
9820 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
9822 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
9823 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
9827 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
9830 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
9831 operand 0 predicate for gathers, use a new pseudo as subtarget.
9833 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
9836 * lra-constraints.c (process_alt_operands): Add printing debug info.
9837 Check absence of input/output reloads for matched operands too.
9839 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
9841 PR rtl-optimization/59835
9842 * ira.c (ira_init_register_move_cost): Increase cost for
9845 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
9847 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
9849 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
9852 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
9853 non-register objects. Use gen_(high/low)part more consistently.
9856 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
9859 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
9860 endian support, remove tests for WORDS_BIG_ENDIAN.
9861 (p8_mfvsrd_3_<mode>): Likewise.
9862 (reload_gpr_from_vsx<mode>): Likewise.
9863 (reload_gpr_from_vsxsf): Likewise.
9864 (p8_mfvsrd_4_disf): Likewise.
9866 2014-01-16 Richard Biener <rguenther@suse.de>
9868 PR rtl-optimization/46590
9869 * lcm.c (compute_antinout_edge): Use postorder iteration.
9870 (compute_laterin): Use inverted postorder iteration.
9872 2014-01-16 Nick Clifton <nickc@redhat.com>
9875 * varasm.c (output_constant): Return the number of bytes actually
9877 (output_constructor_array_range): Update the field size with the
9878 number of bytes emitted by output_constant.
9879 (output_constructor_regular_field): Likewise. Also do not
9880 complain if the total number of bytes emitted is now greater
9881 than the expected fieldpos.
9882 * output.h (output_constant): Update prototype and descriptive comment.
9884 2014-01-16 Marek Polacek <polacek@redhat.com>
9887 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
9888 it is error_mark_node.
9890 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
9892 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
9893 VALID_AVX256_REG_OR_OI_MODE.
9895 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
9897 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
9898 current procedure should be profiled.
9900 2014-01-15 Andrew Pinski <apinski@cavium.com>
9902 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
9903 of moving from/to the STACK_REG register class.
9905 2014-01-15 Richard Henderson <rth@redhat.com>
9908 * reginfo.c (global_regs_decl): Globalize.
9909 * rtl.h (global_regs_decl): Declare.
9910 * ira.c (do_reload): Diagnose frame_pointer_needed and it
9911 reserved via global_regs.
9913 2014-01-15 Teresa Johnson <tejohnson@google.com>
9915 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
9917 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
9919 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
9920 and vmulosh rather than call gen_vec_widen_smult_*.
9921 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
9922 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
9923 (vec_widen_smult_even_v16qi): Likewise.
9924 (vec_widen_umult_even_v8hi): Likewise.
9925 (vec_widen_smult_even_v8hi): Likewise.
9926 (vec_widen_umult_odd_v16qi): Likewise.
9927 (vec_widen_smult_odd_v16qi): Likewise.
9928 (vec_widen_umult_odd_v8hi): Likewise.
9929 (vec_widen_smult_odd_v8hi): Likewise.
9930 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
9931 vmuloub rather than call gen_vec_widen_umult_*.
9932 (vec_widen_umult_lo_v16qi): Likewise.
9933 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
9934 vmulosb rather than call gen_vec_widen_smult_*.
9935 (vec_widen_smult_lo_v16qi): Likewise.
9936 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
9937 rather than call gen_vec_widen_umult_*.
9938 (vec_widen_umult_lo_v8hi): Likewise.
9939 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
9940 rather than call gen_vec_widen_smult_*.
9941 (vec_widen_smult_lo_v8hi): Likewise.
9943 2014-01-15 Jeff Law <law@redhat.com>
9945 PR tree-optimization/59747
9946 * ree.c (find_and_remove_re): Properly handle case where a second
9947 eliminated extension requires widening a copy created for elimination
9948 of a prior extension.
9949 (combine_set_extension): Ensure that the number of hard regs needed
9950 for a destination register does not change when we widen it.
9952 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
9954 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
9955 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
9956 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
9957 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
9958 (avr-*-rtems*): Likewise.
9959 (bfin*-rtems*): Likewise.
9960 (moxie-*-rtems*): Likewise.
9961 (h8300-*-rtems*): Likewise.
9962 (i[34567]86-*-rtems*): Likewise.
9963 (lm32-*-rtems*): Likewise.
9964 (m32r-*-rtems*): Likewise.
9965 (m68k-*-rtems*): Likewise.
9966 (microblaze*-*-rtems*): Likewise.
9967 (mips*-*-rtems*): Likewise.
9968 (powerpc-*-rtems*): Likewise.
9969 (sh-*-rtems*): Likewise.
9970 (sparc-*-rtems*): Likewise.
9971 (sparc64-*-rtems*): Likewise.
9972 (v850-*-rtems*): Likewise.
9973 (m32c-*-rtems*): Likewise.
9975 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
9977 PR rtl-optimization/59511
9978 * ira.c (ira_init_register_move_cost): Use memory costs for some
9979 cases of register move cost calculations.
9980 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
9981 instead of BB frequency.
9982 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
9983 * lra-assigns.c (find_hard_regno_for): Ditto.
9985 2014-01-15 Richard Biener <rguenther@suse.de>
9987 PR tree-optimization/59822
9988 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
9989 (vectorizable_load): Use it to hoist defs of uses of invariant
9990 loads out of the loop.
9992 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
9993 Kugan Vivekanandarajah <kuganv@linaro.org>
9996 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
9999 2014-01-15 Richard Biener <rguenther@suse.de>
10001 PR rtl-optimization/59802
10002 * lcm.c (compute_available): Use inverted postorder to seed
10003 the initial worklist.
10005 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10008 * config/s390/s390.c (s390_preferred_reload_class): Don't return
10009 ADDR_REGS for invalid symrefs in non-PIC code.
10011 2014-01-15 Jakub Jelinek <jakub@redhat.com>
10014 * builtins.c (determine_block_size): Initialize *probable_max_size
10015 even if len_rtx is CONST_INT.
10017 2014-01-14 Andrew Pinski <apinski@cavium.com>
10019 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
10020 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
10021 (cortexa53_tunings): Likewise.
10022 (aarch64_sched_issue_rate): New function.
10023 (TARGET_SCHED_ISSUE_RATE): Define.
10025 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
10027 * ira-costs.c (find_costs_and_classes): Add missed
10028 ira_init_register_move_cost_if_necessary.
10030 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
10033 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
10035 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
10038 * config/i386/i386.c (type_natural_mode): Add a bool parameter
10039 to indicate if type is used for function return value. Warn ABI
10040 change if the vector mode isn't available for function return value.
10041 (ix86_function_arg_advance): Pass false to type_natural_mode.
10042 (ix86_function_arg): Likewise.
10043 (ix86_gimplify_va_arg): Likewise.
10044 (function_arg_32): Don't warn ABI change.
10045 (ix86_function_value): Pass true to type_natural_mode.
10046 (ix86_return_in_memory): Likewise.
10047 (ix86_struct_value_rtx): Removed.
10048 (TARGET_STRUCT_VALUE_RTX): Likewise.
10050 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
10052 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
10053 converting a conditional jump into a conditional return.
10055 2014-01-14 Richard Biener <rguenther@suse.de>
10057 PR tree-optimization/58921
10058 PR tree-optimization/59006
10059 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
10060 hoisting invariant stmts.
10061 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
10062 invariant loads on the preheader edge if possible.
10064 2014-01-14 Joey Ye <joey.ye@arm.com>
10066 * doc/plugin.texi (Building GCC plugins): Update to C++.
10068 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
10070 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
10071 (_mm_rcp28_round_ss): Ditto.
10072 (_mm_rsqrt28_round_sd): Ditto.
10073 (_mm_rsqrt28_round_ss): Ditto.
10074 (_mm_rcp28_sd): Ditto.
10075 (_mm_rcp28_ss): Ditto.
10076 (_mm_rsqrt28_sd): Ditto.
10077 (_mm_rsqrt28_ss): Ditto.
10078 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
10079 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
10080 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
10081 (IX86_BUILTIN_RCP28SD): Ditto.
10082 (IX86_BUILTIN_RCP28SS): Ditto.
10083 (IX86_BUILTIN_RSQRT28SD): Ditto.
10084 (IX86_BUILTIN_RSQRT28SS): Ditto.
10085 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
10086 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
10087 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
10088 (ix86_expand_special_args_builtin): Expand new FTYPE.
10089 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
10090 (srcp14<mode>): Make insn unary.
10091 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
10092 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
10093 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
10094 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
10095 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
10096 Fix rounding: make it SAE only.
10097 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
10099 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
10101 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
10102 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
10103 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
10104 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
10105 (round_saeonly_mask_scalar_operand4): Ditto.
10106 (round_saeonly_mask_scalar_op3): Ditto.
10107 (round_saeonly_mask_scalar_op4): Ditto.
10109 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10111 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
10112 Implement -maltivec=be for vec_insert and vec_extract.
10114 2014-01-10 DJ Delorie <dj@redhat.com>
10116 * config/msp430/msp430.md (call_internal): Don't allow memory
10117 references with SP as the base register.
10118 (call_value_internal): Likewise.
10119 * config/msp430/constraints.md (Yc): New. For memory references
10120 that don't use SP as a base register.
10122 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
10123 "an integer without a # prefix"
10124 * config/msp430/msp430.md (epilogue_helper): Use it.
10126 2014-01-13 Jakub Jelinek <jakub@redhat.com>
10129 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
10130 AVX512F gather builtins.
10131 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
10132 on gather decls with INTEGER_TYPE masktype.
10133 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
10134 directly into the builtin rather than hoisting it before loop.
10136 PR tree-optimization/59387
10137 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
10138 (scev_const_prop): If folded_casts and type has undefined overflow,
10139 use force_gimple_operand instead of force_gimple_operand_gsi and
10140 for each added stmt if it is assign with
10141 arith_code_with_undefined_signed_overflow, call
10142 rewrite_to_defined_overflow.
10143 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
10144 gimple-fold.h instead.
10145 (arith_code_with_undefined_signed_overflow,
10146 rewrite_to_defined_overflow): Moved to ...
10147 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
10148 rewrite_to_defined_overflow): ... here. No longer static.
10149 Include gimplify-me.h.
10150 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
10151 rewrite_to_defined_overflow): New prototypes.
10153 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10155 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
10157 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
10159 * builtins.c (get_object_alignment_2): Minor tweak.
10160 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
10162 2014-01-13 Christian Bruel <christian.bruel@st.com>
10164 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
10165 optimized non constant lengths.
10167 2014-01-13 Jakub Jelinek <jakub@redhat.com>
10170 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
10171 load as __atomic_load_N if possible.
10173 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
10175 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
10177 (rs6000_expand_builtin): Adjust call.
10179 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
10182 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
10183 * config/rs6000/rs6000.c: Include target-globals.h.
10184 (rs6000_set_current_function): Instead of doing target_reinit
10185 unconditionally, use save_target_globals_default_opts and
10186 restore_target_globals.
10188 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
10190 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
10191 (rs6000_expand_builtin): Handle mffs and mtfsf.
10192 (rs6000_init_builtins): Define mffs and mtfsf.
10193 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
10194 (rs6000_mffs): New pattern.
10195 (rs6000_mtfsf): New pattern.
10197 2014-01-11 Bin Cheng <bin.cheng@arm.com>
10199 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
10200 Start narrowing with START. Apply candidate-use pair
10201 and check overall cost in narrowing.
10202 (iv_ca_prune): Pass new argument.
10204 2014-01-10 Jeff Law <law@redhat.com>
10206 PR middle-end/59743
10207 * ree.c (combine_reaching_defs): Ensure the defining statement
10208 occurs before the extension when optimizing extensions with
10209 different source and destination hard registers.
10211 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
10214 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
10215 vtables into the type inheritance graph.
10217 2014-01-10 Jakub Jelinek <jakub@redhat.com>
10219 PR rtl-optimization/59754
10220 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
10221 modes in the REGNO != REGNO case.
10223 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10225 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
10227 2014-01-10 Jakub Jelinek <jakub@redhat.com>
10229 PR tree-optimization/59745
10230 * tree-predcom.c (tree_predictive_commoning_loop): Call
10231 free_affine_expand_cache if giving up because components is NULL.
10233 * target-globals.c (save_target_globals): Allocate < 4KB structs using
10234 GC in payload of target_globals struct instead of allocating them on
10235 the heap and the larger structs separately using GC.
10236 * target-globals.h (struct target_globals): Make regs, hard_regs,
10237 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
10238 of GTY((skip)) and change type to void *.
10239 (reset_target_globals): Cast loads from those fields to corresponding
10242 2014-01-10 Steve Ellcey <sellcey@mips.com>
10245 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
10246 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
10247 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
10249 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
10252 * aarch64-modes.def (CC_Zmode): New flags mode.
10253 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
10254 represents an equality.
10255 (aarch64_get_condition_code): Handle CC_Zmode.
10256 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
10258 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10260 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
10261 extraction in good case.
10263 2014-01-10 Richard Biener <rguenther@suse.de>
10265 PR tree-optimization/59374
10266 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
10267 checking after SLP discovery. Mark stmts not participating
10268 in any SLP instance properly.
10270 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10272 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
10273 when handling a SET rtx.
10275 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10277 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
10278 (cortex-a57): Likewise.
10279 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
10281 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10283 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
10284 non-iwmmxt builtins.
10286 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
10290 * ipa-devirt.c record_target_from_binfo): Take as argument
10291 stack of binfos and lookup matching one for virtual inheritance.
10292 (possible_polymorphic_call_targets_1): Update.
10294 2014-01-10 Huacai Chen <chenhc@lemote.com>
10296 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
10297 kernel strings for Loongson-2E/2F/3A.
10299 2014-01-10 Jakub Jelinek <jakub@redhat.com>
10301 PR middle-end/59670
10302 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
10303 is_gimple_call before calling gimple_call_internal_p.
10305 2014-01-09 Steve Ellcey <sellcey@mips.com>
10307 * Makefile.in (TREE_FLOW_H): Remove.
10308 (TREE_SSA_H): Add file names from tree-flow.h.
10309 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
10310 * tree.h: Remove tree-flow.h reference.
10311 * hash-table.h: Remove tree-flow.h reference.
10312 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
10313 reference with tree-ssa-loop.h.
10315 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10317 * doc/invoke.texi: Add -maltivec={be,le} options, and document
10318 default element-order behavior for -maltivec.
10319 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
10320 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
10321 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
10322 when targeting big endian, at least for now.
10323 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
10325 2014-01-09 Jakub Jelinek <jakub@redhat.com>
10327 PR middle-end/47735
10328 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
10329 var satisfies use_register_for_decl, just take into account type
10330 alignment, rather than decl alignment.
10332 PR tree-optimization/59622
10333 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
10334 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
10335 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
10336 Don't devirtualize for inplace at all. For targets.length () == 1,
10337 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
10339 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
10341 * config/i386/i386.md (cpu): Remove the unused btver1.
10343 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
10345 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
10347 2014-01-09 Jakub Jelinek <jakub@redhat.com>
10350 * tree-core.h (struct target_globals): New forward declaration.
10351 (struct tree_target_option): Add globals field.
10352 * tree.h (TREE_TARGET_GLOBALS): Define.
10353 (prepare_target_option_nodes_for_pch): New prototype.
10354 * target-globals.h (struct target_globals): Define even if
10355 !SWITCHABLE_TARGET.
10356 * tree.c (prepare_target_option_node_for_pch,
10357 prepare_target_option_nodes_for_pch): New functions.
10358 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
10359 * config/i386/i386.c: Include target-globals.h.
10360 (ix86_set_current_function): Instead of doing target_reinit
10361 unconditionally, use save_target_globals_default_opts and
10362 restore_target_globals.
10364 2014-01-09 Richard Biener <rguenther@suse.de>
10366 PR tree-optimization/59715
10367 * tree-cfg.h (split_critical_edges): Declare.
10368 * tree-cfg.c (split_critical_edges): Export.
10369 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
10371 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
10373 * cfgexpand.c (expand_stack_vars): Optionally disable
10374 asan stack protection.
10375 (expand_used_vars): Likewise.
10376 (partition_stack_vars): Likewise.
10377 * asan.c (asan_emit_stack_protection): Optionally disable
10378 after return stack usage.
10379 (instrument_derefs): Optionally disable memory access instrumentation.
10380 (instrument_builtin_call): Likewise.
10381 (instrument_strlen_call): Likewise.
10382 (asan_protect_global): Optionally disable global variables protection.
10383 * doc/invoke.texi: Added doc for new options.
10384 * params.def: Added new options.
10385 * params.h: Likewise.
10387 2014-01-09 Jakub Jelinek <jakub@redhat.com>
10389 PR rtl-optimization/59724
10390 * ifcvt.c (cond_exec_process_if_block): Don't call
10391 flow_find_head_matching_sequence with 0 longest_match.
10392 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
10393 non-active insns if !stop_after.
10394 (try_head_merge_bb): Revert 2014-01-07 changes.
10396 2014-01-08 Jeff Law <law@redhat.com>
10398 * ree.c (get_sub_rtx): New function, extracted from...
10399 (merge_def_and_ext): Here.
10400 (combine_reaching_defs): Use get_sub_rtx.
10402 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
10404 * cgraph.h (varpool_variable_node): Do not choke on null node.
10406 2014-01-08 Catherine Moore <clm@codesourcery.com>
10408 * config/mips/mips.md (simple_return): Attempt to use JRC
10410 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
10412 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
10414 PR rtl-optimization/59137
10415 * reorg.c (steal_delay_list_from_target): Call update_block for
10417 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
10419 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10421 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
10422 two duplicate entries.
10424 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
10427 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
10429 * config/mips/mips.c (mips_truncated_op_cost): New function.
10430 (mips_rtx_costs): Adjust test for BADDU.
10431 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
10433 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
10435 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
10436 (*baddu_si): ...this new pattern.
10438 2014-01-08 Jakub Jelinek <jakub@redhat.com>
10441 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
10443 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
10445 PR middle-end/57748
10446 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
10448 (expand_expr, expand_normal): Adjust.
10449 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
10450 inner_reference_p. Use inner_reference_p to expand inner references.
10451 (store_expr): Adjust.
10452 * cfgexpand.c (expand_call_stmt): Adjust.
10454 2014-01-08 Rong Xu <xur@google.com>
10456 * gcov-io.c (gcov_var): Move from gcov-io.h.
10457 (gcov_position): Ditto.
10458 (gcov_is_error): Ditto.
10459 (gcov_rewrite): Ditto.
10460 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
10461 only part to libgcc/libgcov.h.
10463 2014-01-08 Marek Polacek <polacek@redhat.com>
10465 PR middle-end/59669
10466 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
10468 2014-01-08 Marek Polacek <polacek@redhat.com>
10471 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
10473 2014-01-08 Jakub Jelinek <jakub@redhat.com>
10475 PR rtl-optimization/59649
10476 * stor-layout.c (get_mode_bounds): For BImode return
10477 0 and STORE_FLAG_VALUE.
10479 2014-01-08 Richard Biener <rguenther@suse.de>
10481 PR middle-end/59630
10482 * gimple.h (is_gimple_builtin_call): Remove.
10483 (gimple_builtin_call_types_compatible_p): New.
10484 (gimple_call_builtin_p): New overload.
10485 * gimple.c (is_gimple_builtin_call): Remove.
10486 (validate_call): Rename to ...
10487 (gimple_builtin_call_types_compatible_p): ... this and export. Also
10488 check return types.
10489 (validate_type): New static function.
10490 (gimple_call_builtin_p): New overload and adjust.
10491 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
10492 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
10493 (gimple_fold_stmt_to_constant_1): Likewise.
10494 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
10496 2014-01-08 Richard Biener <rguenther@suse.de>
10498 PR middle-end/59471
10499 * gimplify.c (gimplify_expr): Gimplify register-register type
10500 VIEW_CONVERT_EXPRs to separate stmts.
10502 2014-01-07 Jeff Law <law@redhat.com>
10504 PR middle-end/53623
10505 * ree.c (combine_set_extension): Handle case where source
10506 and destination registers in an extension insn are different.
10507 (combine_reaching_defs): Allow source and destination registers
10508 in extension to be different under limited circumstances.
10509 (add_removable_extension): Remove restriction that the
10510 source and destination registers in the extension are the same.
10511 (find_and_remove_re): Emit a copy from the extension's
10512 destination to its source after the defining insn if
10513 the source and destination registers are different.
10515 PR middle-end/59285
10516 * ifcvt.c (merge_if_block): If we are merging a block with more than
10517 one successor with a block with no successors, remove any BARRIER
10518 after the second block.
10520 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
10522 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
10524 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
10527 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
10528 for 14-bit register offsets when INT14_OK_STRICT is false.
10530 2014-01-07 Roland Stigge <stigge@antcom.de>
10531 Michael Meissner <meissner@linux.vnet.ibm.com>
10534 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
10535 Only check TFmode for SPE constants. Don't check TImode or TDmode.
10537 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
10539 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
10542 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
10544 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
10545 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
10546 rtx is const0_rtx or not.
10548 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
10551 * target-globals.c (save_target_globals): Remove this_fn_optab
10553 * toplev.c: Include optabs.h.
10554 (target_reinit): Temporarily restore the global options if another
10555 set of options are in force.
10557 2014-01-07 Jakub Jelinek <jakub@redhat.com>
10559 PR rtl-optimization/58668
10560 * cfgcleanup.c (flow_find_cross_jump): Don't count
10561 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
10562 to determine what is counted.
10563 (flow_find_head_matching_sequence): Use active_insn_p to determine
10565 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
10567 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
10568 determine what is counted.
10570 PR tree-optimization/59643
10571 * tree-predcom.c (split_data_refs_to_components): If one dr is
10572 read and one write, determine_offset fails and the write isn't
10573 in the bad component, just put the read into the bad component.
10575 2014-01-07 Mike Stump <mikestump@comcast.net>
10576 Jakub Jelinek <jakub@redhat.com>
10579 * tree-core.h (struct tree_optimization_option): Change optabs
10580 type from unsigned char * to void *.
10581 * optabs.c (init_tree_optimization_optabs): Adjust
10582 TREE_OPTIMIZATION_OPTABS initialization.
10584 2014-01-06 Jakub Jelinek <jakub@redhat.com>
10587 * config/i386/i386.h (struct machine_function): Add
10588 no_drap_save_restore field.
10589 * config/i386/i386.c (ix86_save_reg): Use
10590 !cfun->machine->no_drap_save_restore instead of
10591 crtl->stack_realign_needed.
10592 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
10593 this function clears frame_pointer_needed. Set
10594 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
10595 and DRAP reg is needed.
10597 2014-01-06 Marek Polacek <polacek@redhat.com>
10600 * doc/implement-c.texi: Mention that other integer types are
10601 permitted as bit-field types in strictly conforming mode.
10603 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
10605 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
10606 is newly allocated.
10608 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
10610 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
10612 2014-01-06 Martin Jambor <mjambor@suse.cz>
10615 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
10617 * ipa-prop.c (ipa_print_node_params): Fix indentation.
10619 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
10623 * var-tracking.c (add_stores): Preserve the value of the source even if
10624 we don't record the store.
10626 2014-01-06 Terry Guo <terry.guo@arm.com>
10628 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
10630 2014-01-05 Iain Sandoe <iain@codesourcery.com>
10633 * config/darwin.c (darwin_function_section): Adjust return values to
10634 correspond to optimisation changes made in r206070.
10636 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
10638 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
10639 from prefetch_block tune setting.
10640 (nocona_cost): Correct size of prefetch block to 64.
10642 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
10644 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
10645 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
10646 used to save the static chain register in the computation of the offset
10647 from which the FP registers need to be restored.
10649 2014-01-04 Jakub Jelinek <jakub@redhat.com>
10651 PR tree-optimization/59519
10652 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
10653 ICE if get_current_def (current_new_name) is already non-NULL, as long
10654 as it is a phi result of some other phi in *new_exit_bb that has
10657 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
10658 or vmovdqu* for misaligned_operand.
10659 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
10660 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
10661 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
10662 aligned_mem for AVX512F masked aligned load and store builtins and for
10663 non-temporal moves.
10665 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
10667 PR tree-optimization/59651
10668 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
10669 Address range for negative step should be added by TYPE_SIZE_UNIT.
10671 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
10673 * config/m68k/m68k.c (handle_move_double): Handle pushes with
10674 overlapping registers also for registers other than the stack pointer.
10676 2014-01-03 Marek Polacek <polacek@redhat.com>
10679 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
10682 2014-01-03 Jakub Jelinek <jakub@redhat.com>
10685 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
10688 * config/i386/i386.md (MODE_SIZE): New mode attribute.
10689 (push splitter): Use <P:MODE_SIZE> instead of
10690 GET_MODE_SIZE (<P:MODE>mode).
10691 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
10692 (mov -1, reg peephole2): Likewise.
10693 * config/i386/sse.md (*mov<mode>_internal,
10694 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
10695 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
10696 *<code><mode>3, *andnot<mode>3<mask_name>,
10697 <mask_codefor><code><mode>3<mask_name>): Likewise.
10698 * config/i386/subst.md (mask_mode512bit_condition,
10699 sd_mask_mode512bit_condition): Likewise.
10701 2014-01-02 Xinliang David Li <davidxl@google.com>
10703 PR tree-optimization/59303
10704 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
10705 (dump_predicates): Better output format.
10706 (pred_equal_p): New function.
10707 (is_neq_relop_p): Ditto.
10708 (is_neq_zero_form_p): Ditto.
10709 (pred_expr_equal_p): Ditto.
10710 (pred_neg_p): Ditto.
10711 (simplify_pred): Ditto.
10712 (simplify_preds_2): Ditto.
10713 (simplify_preds_3): Ditto.
10714 (simplify_preds_4): Ditto.
10715 (simplify_preds): Ditto.
10716 (push_pred): Ditto.
10717 (push_to_worklist): Ditto.
10718 (get_pred_info_from_cmp): Ditto.
10719 (is_degenerated_phi): Ditto.
10720 (normalize_one_pred_1): Ditto.
10721 (normalize_one_pred): Ditto.
10722 (normalize_one_pred_chain): Ditto.
10723 (normalize_preds): Ditto.
10724 (normalize_cond_1): Remove function.
10725 (normalize_cond): Ditto.
10726 (is_gcond_subset_of): Ditto.
10727 (is_subset_of_any): Ditto.
10728 (is_or_set_subset_of): Ditto.
10729 (is_and_set_subset_of): Ditto.
10730 (is_norm_cond_subset_of): Ditto.
10731 (pred_chain_length_cmp): Ditto.
10732 (convert_control_dep_chain_into_preds): Type change.
10733 (find_predicates): Ditto.
10734 (find_def_preds): Ditto.
10735 (destroy_predicates_vecs): Ditto.
10736 (find_matching_predicates_in_rest_chains): Ditto.
10737 (use_pred_not_overlap_with_undef_path_pred): Ditto.
10738 (is_pred_expr_subset): Ditto.
10739 (is_pred_chain_subset_of): Ditto.
10740 (is_included_in): Ditto.
10741 (is_superset_of): Ditto.
10743 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
10745 Update copyright years.
10747 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
10749 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
10750 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
10751 config/arc/arc.md, config/arc/arc.opt,
10752 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
10753 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
10754 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
10755 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
10756 config/linux-protos.h, config/linux.c, config/winnt-c.c,
10757 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
10758 vtable-verify.c, vtable-verify.h: Use the standard form for the
10761 2014-01-02 Tobias Burnus <burnus@net-b.de>
10763 * gcc.c (process_command): Update copyright notice dates.
10764 * gcov-dump.c: Ditto.
10766 * doc/cpp.texi: Bump @copying's copyright year.
10767 * doc/cppinternals.texi: Ditto.
10768 * doc/gcc.texi: Ditto.
10769 * doc/gccint.texi: Ditto.
10770 * doc/gcov.texi: Ditto.
10771 * doc/install.texi: Ditto.
10772 * doc/invoke.texi: Ditto.
10774 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
10776 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
10778 2014-01-01 Jakub Jelinek <jakub@redhat.com>
10780 * config/i386/sse.md (*mov<mode>_internal): Guard
10781 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
10783 PR rtl-optimization/59647
10784 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
10785 new_rtx into UNSIGNED_FLOAT rtxes.
10787 Copyright (C) 2014 Free Software Foundation, Inc.
10789 Copying and distribution of this file, with or without modification,
10790 are permitted in any medium without royalty provided the copyright
10791 notice and this notice are preserved.
projects / gcc.git / blob