1 2013-08-23 Jan Hubicka <jh@suse.cz>
3 * ipa-utils.h (method_class_type): Declare.
4 * ipa-devirt.c (method_class_type): Export.
6 * cgraphunit.c (analyze_functions): Do basic devirtualization;
7 do not walk base classes of anonymous types.
9 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
11 PR rtl-optimization/58220
13 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
14 handle SEQUENCE insns properly.
16 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
18 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
20 (pp_newline_and_indent): Likewise.
21 (pp_separate_with): Likewise.
22 * pretty-print.c (pp_newline_and_flush): Define.
23 (pp_newline_and_indent): Likewise.
24 (pp_separate_with): Likewise.
26 2013-08-23 Jakub Jelinek <jakub@redhat.com>
29 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
30 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
32 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
34 * gcc/config/i386/predicates.md (ext_sse_reg_operand): New.
35 * gcc/config/i386/i386.md (*movti_internal): Use
36 predicate to determine if EVEX is needed.
37 (*movsi_internal): Ditto.
38 (*movdf_internal): Ditto.
39 (*movsf_internal): Ditto.
40 * gcc/config/i386/mmx.md (*mov<mode>_internal): Ditto.
42 2013-08-23 Jakub Jelinek <jakub@redhat.com>
44 PR tree-optimization/58209
45 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
46 (find_tail_calls): Give up for pointer result types if m is non-NULL.
47 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
48 emit POINTER_PLUS_EXPR.
49 (create_tailcall_accumulator): For pointer result type accumulate in
52 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
54 * configure.ac: Add backslashes missing from the last change.
55 * configure: Regenerate.
57 2013-08-22 Jan Hubicka <jh@susue.cz>
59 * ipa.c (function_and_variable_visibility): First remember function
60 was global and then make it local.
62 2013-08-22 Julian Brown <julian@codesourcery.com>
64 * configure.ac: Add aarch64 to list of arches which use "nop" in
66 * configure: Regenerate.
68 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
70 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
71 gnu_libc_has_function.
72 * config/s390/tpf.h: Likewise.
74 2013-08-22 Jan Hubicka <jh@susue.cz>
76 * timevar.c (validate_phases): Add cast.
78 2013-08-22 Jan Hubicka <jh@susue.cz>
80 * timevar.c (validate_phases): Use size_t for memory.
81 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
83 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
85 * pretty-print.h (output_buffer::output_buffer): Declare.
86 (pretty_printer::pretty_printer): Likewise.
87 (pp_construct): Remove.
88 * pretty-print.c (output_buffer::output_buffer): Define.
89 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
90 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
91 (print_gimple_expr): Likewise.
92 (print_gimple_seq): Likewise.
93 (gimple_dump_bb): Likewise.
94 * sched-vis.c (dump_value_slim): Likewise.
95 (dump_insn_slim): Likewise.
96 (dump_rtl_slim): Likewise.
97 (str_pattern_slim): Likewise.
98 * tree-mudflap.c (mf_varname_tree): Likewise.
99 * graph.c (print_graph_cfg): Likewise.
100 (start_graph_dump): Likewise.
101 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
103 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
104 pretty printer initialization.
105 * coretypes.h (diagnostic_context): Remove superflous type alias
107 (pretty_printer): Likewise. Declare directly as a class.
108 (pretty_print_info): Remove declaration as class.
109 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
110 and pp_clear_output_area.
111 (asan_add_global): Likewise.
113 2013-08-22 Jan Hubicka <jh@suse.cz>
115 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
116 * ipa-utils.h (update_type_inheritance_graph): Declare.
117 (possible_polymorphic_call_target_p): Declare.
118 (possible_polymorphic_call_target_p): New.
119 * ipa-devirt.c: Update toplevel comments.
120 (cached_polymorphic_call_targets): Move up.
121 (odr_type_d): Move ID down.
122 (polymorphic_type_binfo_p): Update comment.
123 (odr_hasher::remove): Likewise;
124 (get_odr_type): Set anonymous_namespace.
125 (dump_odr_type): Dump it.
126 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
127 (maybe_record_node): Record node in cached_polymorphic_call_targets.
128 (record_binfo): Add comment.
129 (free_polymorphic_call_targets_hash): Do not ICE when cache is not built.
130 (devirt_node_removal_hook): Do not iCE when cache is freed.
131 (possible_polymorphic_call_target_p): New predicate.
132 (update_type_inheritance_graph): New function.
134 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
135 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
136 Sergey Lega <sergey.s.lega@intel.com>
137 Anna Tikhonova <anna.tikhonova@intel.com>
138 Ilya Tocar <ilya.tocar@intel.com>
139 Andrey Turetskiy <andrey.turetskiy@intel.com>
140 Ilya Verbin <ilya.verbin@intel.com>
141 Kirill Yukhin <kirill.yukhin@intel.com>
142 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
144 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
145 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
146 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
147 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
148 (OPTION_MASK_ISA_AVX2_UNSET): Update.
149 (OPTION_MASK_ISA_AVX512F_UNSET): New.
150 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
151 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
152 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
153 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
154 OPT_mavx512pf, OPT_mavx512er cases.
155 * config/i386/constraints.md (v): New constraint.
156 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
157 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
159 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
160 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
161 * config/i386/i386-c.c (ix86_target_macros_internal):
162 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
164 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
165 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
166 * config/i386/i386.c (regclass_map, dbx_register_map)
167 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
168 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
169 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
171 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
172 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
173 -mavx512pf options. Fix formatting.
174 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
175 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
176 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
177 -mavx512cd, -mavx512pf options.
178 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
179 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
180 (ix86_preferred_output_reload_class): Replace SSE_REGS with
182 (ix86_hard_regno_mode_ok): Support 512-bit registers.
183 (ix86_set_reg_reg_cost): Ditto.
184 (x86_order_regs_for_local_alloc): Ditto.
185 (MAX_VECT_LEN): Extend to 64-byte.
186 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
187 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
188 (TARGET_AVX512ER, TARGET_AVX512CD): New.
189 (BIGGEST_ALIGNMENT): Extend to 512-bits.
190 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
191 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
192 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
193 (SSE_REG_MODE_P): Support new modes.
194 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
195 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
196 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
197 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
198 (REG_CLASS_CONTENTS): Add new registers.
199 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
200 (EXT_REX_SSE_REGNO_P): New.
201 (HI_REGISTER_NAMES): Add new registers.
202 * config/i386/i386.md: Define constants for new registers.
203 (mode): Add new 512-bit modes.
204 (prefix): Support evex prefix.
205 (isa): Support avx512f, noavx512f, fma_avx512f.
206 (ssemodesuffix): Add new 512-bit modes.
208 (*movxi_internal_avx512f): Ditto.
209 (*movdi_internal): Replace constraint "x" with the new constraint "v".
211 (*movsi_internal): Likewise.
212 (*movdf_internal): Likewise.
213 (*movsf_internal): Likewise.
214 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
215 (<code><mode>3): Likewise.
216 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd): New.
217 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
218 with the new constraint "v".
219 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
221 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
222 with the new constraint "v".
223 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
224 (<sse2>_storedqu<avxsizesuffix>): Likewise.
225 (*<plusminus_insn><mode>3): Likewise.
226 (<sse>_vm<plusminus_insn><mode>3): Likewise.
227 (*mul<mode>3): Likewise.
228 (<sse>_vmmul<mode>3): Likewise.
229 (<sse>_div<mode>3): Likewise.
230 (<sse>_vmdiv<mode>3): Likewise.
231 (<sse>_sqrt<mode>2): Likewise.
232 (<sse>_vmsqrt<mode>2): Likewise.
233 (*<code><mode>3_finite): Likewise.
234 (*<code><mode>3) <smaxmin>: Likewise.
235 (<sse>_vm<code><mode>3): Likewise.
236 (*<code><mode>3) <any_logic>: Likewise.
237 (*fma_fmadd_<mode>): Likewise.
238 (*fma_fmsub_<mode>): Likewise.
239 (*fma_fnmadd_<mode>): Likewise.
240 (*fma_fnmsub_<mode>): Likewise.
241 (*fma_fmaddsub_<mode>): Likewise.
242 (*fma_fmsubadd_<mode>): Likewise.
243 (*fmai_fmadd_<mode>): Likewise.
244 (*fmai_fmsub_<mode>): Likewise.
245 (*fmai_fnmadd_<mode>): Likewise.
246 (*fmai_fnmsub_<mode>): Likewise.
247 (sse_cvtsi2ss): Likewise.
248 (sse_cvtsi2ssq): Likewise.
249 (sse_cvtss2si): Likewise.
250 (sse_cvtss2si_2): Likewise.
251 (sse_cvtss2siq): Likewise.
252 (sse_cvtss2siq_2): Likewise.
253 (sse_cvttss2si): Likewise.
254 (sse_cvtss2siq_2): Likewise.
255 (float<sseintvecmodelower><mode>2): Likewise.
256 (sse2_cvtsd2si_2): Likewise.
257 (sse2_cvtsd2siq_2): Likewise.
258 (*<plusminus_insn><mode>3): Likewise.
259 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
260 (*<sse4_1_avx2>_mul<mode>3): Likewise.
261 (ashr<mode>3): Likewise.
262 (<shift_insn><mode>3): Likewise.
263 (avx2_<code><mode>3): Likewise.
264 (*avx2_<code><mode>3): Likewise.
265 (*andnot<mode>3): Likewise.
266 (*<code><mode>3) <any_logic>: Likewise.
267 (abs<mode>2): Likewise.
268 (avx2_permvar<mode>): Likewise.
269 (avx2_perm<mode>_1): Likewise.
270 (*avx_vpermilp<mode>): Likewise.
271 (avx_vpermilvar<mode>3): Likewise.
272 (avx2_ashrv<mode>): Likewise.
273 (avx2_<shift_insn>v<mode>): Likewise.
274 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
276 * doc/rtl.texi: Document XImode.
278 2013-08-21 Jeff Law <law@redhat.com>
280 * tree-flow.h (register_jump_thread): Pass vector of edges
281 instead of each important edge.
282 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
283 thread path into a vector and pass that to register_jump_thread.
284 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
285 passed in edge vector to the current 3-edge form.
288 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
290 * dce.c (fini_dce): Call df_analyze again just in case
291 delete_unmarked_insns removed anything.
293 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
295 * reload.h (struct reg_equivs): Rename to ..
296 (struct reg_equivs_s): .. this.
298 2013-08-20 Martin Liska <marxin.liska@gmail.com>
300 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
302 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
304 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest
307 2013-08-21 Jeff Law <law@redhat.com>
309 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
310 simplify assignments too. If the RHS collapses to a singleton
311 range, then return the value for the range.
313 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
315 * config/i386/sse.md (V16): Rename to...
317 (mov<mode>): Update iterator name.
318 (*mov<mode>_internal): Ditto.
319 (push<mode>1): Ditto.
320 (movmisalign<mode>): Ditto.
322 2013-08-20 Jan Hubicka <jh@suse.cz>
325 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
326 entry for direct edges.
327 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
329 2013-08-20 David Malcolm <dmalcolm@redhat.com>
331 Revert my last two changes, r201865 and r201864:
334 2013-08-20 David Malcolm <dmalcolm@redhat.com>
336 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
337 instances can own GC refs.
339 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
340 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
341 (gcc::context::gt_pch_nx): Likewise.
342 (gcc::context::gt_pch_nx): Likewise.
343 * ggc.h (gt_ggc_mx <T>): New.
344 (gt_pch_nx_with_op <T>): New.
345 (gt_pch_nx <T>): New.
346 * passes.c (opt_pass::gt_ggc_mx): New.
347 (opt_pass::gt_pch_nx): New.
348 (opt_pass::gt_pch_nx_with_op): New.
349 (pass_manager::gt_ggc_mx): New.
350 (pass_manager::gt_pch_nx): New.
351 (pass_manager::gt_pch_nx_with_op): New.
352 (pass_manager::operator new): Use
353 ggc_internal_cleared_alloc_stat rather than xcalloc.
354 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
355 (pass_manager::gt_ggc_mx): New.
356 (pass_manager::gt_pch_nx): New.
357 (pass_manager::gt_pch_nx_with_op): New.
358 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
359 (opt_pass::operator new): New.
360 (opt_pass::gt_ggc_mx): New.
361 (opt_pass::gt_pch_nx): New.
362 (opt_pass::gt_pch_nx_with_op): New.
365 2013-08-20 David Malcolm <dmalcolm@redhat.com>
367 * Makefile.in (GTFILES): Add context.h.
368 * context.c (gcc::context::operator new): New.
369 (gcc::context::gt_ggc_mx): New.
370 (gcc::context::gt_pch_nx): New.
371 (gcc::context::gt_pch_nx): New.
372 * context.h (gcc::context): Add GTY((user)) marking.
373 (gcc::context::operator new): New.
374 (gcc::context::gt_ggc_mx): New.
375 (gcc::context::gt_pch_nx): New.
376 (gcc::context::gt_pch_nx): New.
377 (g): Add GTY marking.
378 (gt_ggc_mx (gcc::context *)): New.
379 (gt_pch_nx (gcc::context *)): New.
380 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
382 * gengtype.c (open_base_files) <ifiles>: Add context.h.
384 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
386 * dce.c (fini_dce): Call df_analyze again just in case
387 delete_unmarked_insns removed anything.
389 2013-08-20 Teresa Johnson <tejohnson@google.com>
391 PR rtl-optimizations/57451
392 * final.c (reemit_insn_block_notes): Prevent lexical blocks
393 from crossing split section boundaries.
395 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
397 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
398 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
399 with MULTLIB_DEFAULTS.
401 2013-08-20 Nick Clifton <nickc@redhat.com>
403 * target.def (narrow_volatile_bitfield): Note that the default
404 value is false, not !TARGET_STRICT_ALIGN.
405 * doc/tm.texi: Regenerate.
407 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
409 Fix LIB_SPEC for systems without libpthread.
411 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
412 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
414 * config/i386/linux-common.h: Likewise.
415 * config/mips/linux-common.h: Likewise.
417 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
419 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
422 2013-08-20 David Malcolm <dmalcolm@redhat.com>
424 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
425 instances can own GC refs.
427 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
428 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
429 (gcc::context::gt_pch_nx): Likewise.
430 (gcc::context::gt_pch_nx): Likewise.
431 * ggc.h (gt_ggc_mx <T>): New.
432 (gt_pch_nx_with_op <T>): New.
433 (gt_pch_nx <T>): New.
434 * passes.c (opt_pass::gt_ggc_mx): New.
435 (opt_pass::gt_pch_nx): New.
436 (opt_pass::gt_pch_nx_with_op): New.
437 (pass_manager::gt_ggc_mx): New.
438 (pass_manager::gt_pch_nx): New.
439 (pass_manager::gt_pch_nx_with_op): New.
440 (pass_manager::operator new): Use
441 ggc_internal_cleared_alloc_stat rather than xcalloc.
442 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
443 (pass_manager::gt_ggc_mx): New.
444 (pass_manager::gt_pch_nx): New.
445 (pass_manager::gt_pch_nx_with_op): New.
446 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
447 (opt_pass::operator new): New.
448 (opt_pass::gt_ggc_mx): New.
449 (opt_pass::gt_pch_nx): New.
450 (opt_pass::gt_pch_nx_with_op): New.
452 2013-08-20 David Malcolm <dmalcolm@redhat.com>
454 * Makefile.in (GTFILES): Add context.h.
455 * context.c (gcc::context::operator new): New.
456 (gcc::context::gt_ggc_mx): New.
457 (gcc::context::gt_pch_nx): New.
458 (gcc::context::gt_pch_nx): New.
459 * context.h (gcc::context): Add GTY((user)) marking.
460 (gcc::context::operator new): New.
461 (gcc::context::gt_ggc_mx): New.
462 (gcc::context::gt_pch_nx): New.
463 (gcc::context::gt_pch_nx): New.
464 (g): Add GTY marking.
465 (gt_ggc_mx (gcc::context *)): New.
466 (gt_pch_nx (gcc::context *)): New.
467 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
469 * gengtype.c (open_base_files) <ifiles>: Add context.h.
471 2013-08-20 Alan Modra <amodra@gmail.com>
474 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
475 (rs6000_emit_epilogue): Likewise.
477 2013-08-19 Dehao Chen (dehao@google.com)
479 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
481 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
482 Jakub Jelinek <jakub@redhat.com>
484 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
485 (BUILT_IN_FABSD64): Likewise.
486 (BUILT_IN_FABSD128): Likewise.
487 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
488 (fold_builtin_1): Likewise.
489 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping destination
491 (*abstd2_fpr): Likewise.
492 (*nabstd2_fpr): Likewise.
494 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
496 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
499 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
501 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
504 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
506 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
507 * builtins.c (default_libc_has_function): New.
508 (gnu_libc_has_function): Ditto.
509 (no_c99_libc_has_function): Ditto.
510 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
511 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
512 (fold_builtin_sincos): Likewise.
513 (fold_builtin_cexp): Likewise.
514 * builtins.def (DEF_C94_BUILTIN): Likewise.
515 (DEF_C99_BUILTIN): Likewise.
516 (DEF_C99_C90RES_BUILTIN): Likewise.
517 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
518 definitions to using this define.
519 * config/darwin-protos.h (darwin_libc_has_function): New.
520 * config/darwin.c: (darwin_libc_has_function: Ditto.
521 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
522 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
523 * config/darwin.h: Ditto.
524 * config/elfos.h: Ditto.
525 * config/freebsd.h: Ditto.
526 * config/i386/cygming.h: Ditto.
527 * config/i386/djgpp.h: Ditto.
528 * config/i386/i386-interix.h: Ditto.
529 * config/microblaze/microblaze.h: Ditto.
530 * config/mmix/mmix.h: Ditto.
531 * config/gnu-user.h: Ditto.
532 * config/ia64/hpux.h: Ditto.
533 * config/pa/pa-hpux.h: Ditto.
534 * config/pdp11/pdp11.h: Ditto.
535 * config/picochip/picochip.h: Ditto.
536 * config/linux.h: Ditto.
537 * config/netbsd.h: Ditto.
538 * config/openbsd.h: Ditto.
539 * config/rs6000/aix43.h: Ditto.
540 * config/rs6000/aix51.h: Ditto.
541 * config/rs6000/aix52.h: Ditto.
542 * config/rs6000/aix53.h: Ditto.
543 * config/rs6000/aix61.h: Ditto.
544 * config/rs6000/darwin.h: Ditto.
545 * config/rs6000/linux.h: Ditto.
546 * config/rs6000/linux64.h: Ditto.
547 * config/s390/tpf.h: Ditto.
548 * config/sol2-10.h: Ditto.
549 * config/sol2.h: Ditto.
550 * config/vms/vms.h: Ditto.
551 * config/vxworks.h: Ditto.
552 * config/linux-android.c (linux_android_libc_has_function):
553 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
554 * config/linux-protos.h (linux_android_libc_has_function):
556 * config/i386/i386.c (ix86_libc_has_function): New.
557 * config/i386/i386-protos.h
558 (ix86_libc_has_function): New declaration.
559 * config/i386/i386.md
560 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
561 ("isinf<mode>2): Likewise.
562 * convert.c (convert_to_integer): Using new target hook
563 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
564 TARGET_C99_FUNCTIONS.
565 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
566 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
567 * coretypes.h (function_class): New enum for different
568 classes of functions.
569 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
570 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
571 (TARGET_HAS_SINCOS): Likewise.
572 (TARGET_LIBC_HAS_FUNCTION): New.
573 * doc/tm.texi: Regenerated.
574 * targhooks.h (default_libc_has_function): New declaration.
575 (no_c99_libc_has_function): Ditto.
576 (gnu_libc_has_function): Ditto.
577 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
578 and TARGET_HAS_SINCOS.
580 2013-08-18 Jan Hubicka <jh@suse.cz>
582 * Makeifle-in (ipa-devirt.o): New.
583 (GTFILES): Add ipa-utils.h and ipa-devirt.c
584 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
585 (analyze_functions): Look into possible targets of polymorphic call.
586 * dumpfile.c (dump_files): Add type-inheritance dump.
587 * dumpfile.h (TDI_inheritance): New.
588 * ipa-devirt.c: New file.
589 * ipa-utils.h (odr_type_d): Forward declare.
590 (odr_type): New type.
591 (build_type_inheritance_graph): Declare.
592 (possible_polymorphic_call_targets): Declare and introduce inline
593 variant when only edge is pased.
594 (dump_possible_polymorphic_call_targets): Likewise.
595 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
596 * tree.c (type_in_anonymous_namespace_p): Break out from ...
597 (types_same_for_odr): ... here.
598 * tree.h (type_in_anonymous_namespace_p): Declare.
600 2013-08-18 Jakub Jelinek <jakub@redhat.com>
602 PR tree-optimization/58006
603 * tree-parloops.c (take_address_of): Don't ICE if get_name
605 (eliminate_local_variables_stmt): Remove clobber stmts.
607 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
609 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
610 error message is issued for an alias to undefined symbol.
612 2013-08-18 Jan Hubicka <jh@suse.cz>
614 * cgraph.c (cgraph_create_indirect_edge): Discover
615 polymorphic calls and record basic info into indirect_info.
616 * gimple-fold.c (gimple_fold_call): When doing BINFO based
617 devirtualization, ignore objc function calls.
618 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
619 call with no parm index info.
620 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
621 * tree.c (virtual_method_call_p): New function.
622 * tree.h (virtual_method_call_p): Declare.
624 2013-08-16 Jan Hubicka <jh@suse.cz>
627 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
629 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
631 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
634 2013-08-16 David Malcolm <dmalcolm@redhat.com>
636 * gengtype.c (type_for_name): Add special-case support for
637 locating types within the "gcc::" namespace.
638 (open_base_files): Emit a "using namespace gcc" directive.
640 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
643 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
644 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
646 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
647 array instead of each individual operand as a separate argument.
648 (emit_fusion_gpr_load): Likewise.
649 (expand_fusion_gpr_load): Add new function declaration.
651 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
652 signature to have the operands passed as an array, instead of as
653 separate arguments. Allow ZERO_EXTEND to be in the memory
654 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
655 depend on the register live/dead flags when peepholes are run.
656 (expand_fusion_gpr_load): New function to be called from the
657 peephole2 pass, to change the register that addis sets to be the
659 (emit_fusion_gpr_load): Change the calling signature to have the
660 operands passed as an array, instead of as separate arguments.
661 Allow ZERO_EXTEND to be in the memory address, and also
662 SIGN_EXTEND if -mpower8-fusion-sign.
664 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
666 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
667 adjust the register addis loads up in the peephole2 pass. Do not
668 depend on the register live/dead state when the peephole pass is
671 2013-08-16 David Malcolm <dmalcolm@redhat.com>
673 * gengtype.c (create_user_defined_type): Ensure that the kind
674 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
675 declaration is seen before the GTY((user)) marking.
677 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
680 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
682 2013-08-16 Jan Hubicka <jh@suse.cz>
684 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
685 arugment expected_type.
686 (gimple_fold_call): Use it.
687 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
688 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
689 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
690 (try_make_edge_direct_virtual_call): Likewise.
691 * tree.c (obj_type_ref_class): New.
692 * tree.h (obj_type_ref_class): Use it.
694 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
696 * sched-vis.c (rtl_slim_pp_initialized): Remove.
697 (rtl_slim_pp): Likewise.
698 (init_rtl_slim_pretty_print): Likewise.
699 (dump_value_slim): Don't call it. Use local pretty printer.
700 (dump_insn_slim): Likewise.
701 (dump_rtl_slim): Likewise.
702 (str_pattern_slim): Likewise.
703 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
706 2013-08-16 Jakub Jelinek <jakub@redhat.com>
708 PR tree-optimization/58164
709 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
710 walk gimple_goto_dest of GIMPLE_GOTO.
712 PR tree-optimization/58165
713 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
714 bi_call must be the last stmt in a bb, don't split_block, instead
715 use fallthru edge from it and give up if there is none.
716 Release conds vector when returning early.
718 2013-08-14 Xinliang David Li <davidxl@google.com>
720 * config/i386/i386.c (ix86_option_override_internal):
721 Remove unused variable and field.
723 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
726 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
727 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
728 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
729 and Linux, correct BLKmode alignment when 128-bit alignment is
730 required and compatibility flag is not set.
731 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
732 for zero-size arguments when compatibility flag is not set.
734 2013-08-14 Jakub Jelinek <jakub@redhat.com>
736 PR tree-optimization/58145
737 * tree-sra.c (build_ref_for_offset): If prev_base has
738 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
740 2013-08-14 Xinliang David Li <davidxl@google.com>
742 * config/i386/i386.c (ix86_option_override_internal):
743 Fix uninitialized variable error.
745 2013-08-14 Xinliang David Li <davidxl@google.com>
747 * config/i386/i386.opt: Define two new options.
748 * config/i386/x86-tune.def: Add arch selector field in macros.
749 * config/i386/i386.h: Adjust macro definition.
750 * config/i386/i386.c (ix86_option_override_internal):
752 (parse_mtune_ctrl_str): New function.
753 (set_ix86_tune_features): New function.
754 (ix86_function_specific_restore): Call the new helper function.
756 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
758 PR rtl-optimization/57662
759 * sel-sched.c (code_motion_process_successors): When the current insn
760 is removed after the recursive traversal, break from the loop.
761 Add comments and debug printouts.
763 2013-08-14 Jakub Jelinek <jakub@redhat.com>
764 Alexandre Oliva <aoliva@redhat.com>
767 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
768 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
769 there also UNSPEC_PLTOFF.
771 2013-08-14 Marek Polacek <polacek@redhat.com>
773 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
774 NUM_CONDITIONS bit positions.
776 2013-08-13 Cary Coutant <ccoutant@google.com>
778 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
779 (attr_checksum): Hash vector contents instead of pointer.
780 (attr_checksum_ordered): Likewise.
782 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
784 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
785 when Pmode != word_mode. Add length_address attribute.
786 (sse3_monitor_<mode>): Merge from sse3_monitor and
787 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
788 Pmode != word_mode. Update insn length attribute.
789 * config/i386/i386.c (ix86_option_override_internal): Update
790 ix86_gen_monitor selection for merged sse3_monitor insn.
792 2013-08-13 Julian Brown <julian@codesourcery.com>
794 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
795 perform invalid legitimization on greater-than-word-size modes for
798 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
800 * ira.c (setup_class_translate_array): Use aclass instead of cl
801 for classes not fully covered by allocno classes.
803 2013-08-13 Jakub Jelinek <jakub@redhat.com>
805 PR tree-optimization/57661
806 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
807 * tree-inline.c (tree_function_versioning): Initialize it.
808 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
809 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
810 that is not being copied.
813 * asan.c (instrument_strlen_call): Fix typo in comment.
814 Use char * type even for the lhs of POINTER_PLUS_EXPR.
816 2013-08-13 Steve Ellcey <sellcey@mips.com>
818 * config/mips/mips.md (prefetch): Use lw instead of ld on
819 loongson in 32bit mode.
821 2013-08-13 Nick Clifton <nickc@redhat.com>
823 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
825 2013-08-13 Jan Hubicka <jh@suse.cz>
827 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
828 introduced edge; fix typo in sanity check.
829 (cgraph_resolve_speculation): Export; improve diagnostic.
830 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
831 speculation at type mismatch.
832 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
833 (cgraph_resolve_speculation): Declare.
834 (symtab_can_be_discarded): New function.
835 * value-prof.c (gimple_ic_transform): Remove actual transform code.
836 * ipa-inline-transform.c (speculation_removed): New global var.
837 (clone_inlined_nodes): See if speculation can be removed.
838 (inline_call): If speculations was removed, we growths may not match.
839 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
840 (speculation_useful_p): New function.
841 (resolve_noninline_speculation): New function.
842 (inline_small_functions): Resolve useless speculations.
843 * ipa-inline.h (speculation_useful_p): Declare
844 * ipa.c (can_replace_by_local_alias): Simplify.
845 (ipa_profile): Produce speculative calls in non-lto, too;
846 add simple cost model; produce local aliases.
848 2013-08-13 David Malcolm <dmalcolm@redhat.com>
850 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
853 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
855 * config/i386/i386.c (ix86_function_versions): Use error + inform.
857 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
859 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
860 iterator instead of X87MODEF.
862 2013-08-12 Perez Read <netfirewall@gmail.com>
865 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
866 operand 0 for intel asm alternative.
867 (*movabs<mode>_2): Ditto for operand 1.
869 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
871 * config/aarch64/arm_none.h
872 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
874 2013-08-12 Nick Clifton <nickc@redhat.com>
876 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
878 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
880 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
881 expand for QI/HImode operand to produce more effictive code for
882 unsigned char(short) --> float(double) conversion.
884 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
886 * doc/invoke.texi: Mention that -ftls-model does not force the final
889 2013-08-12 Marek Polacek <polacek@redhat.com>
890 Marc Glisse <marc.glisse@inria.fr>
892 PR tree-optimization/57980
893 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
894 when creating -1 constant.
896 2013-08-10 Jan Hubicka <jh@suse.cz>
898 Workaround binutils PR14342.
899 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
900 (gimple_init_edge_profiler): Likewise.
901 (gimple_gen_ic_func_profiler): Likewise.
903 2013-08-09 Jan Hubicka <jh@suse.cz>
905 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
907 2013-08-09 Xinliang David Li <davidxl@google.com>
909 * config/i386/stringop.def: New file.
910 * config/i386/stringop.opt: New file.
911 * config/i386/i386-opts.h: Include stringopt.def.
912 * config/i386/i386.opt: Include stringopt.opt.
913 * config/i386/i386.c (ix86_option_override_internal):
914 Override default size based stringop inline strategies with options.
915 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
918 2013-08-09 Jan Hubicka <jh@suse.cz>
920 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
922 2013-08-09 Jan Hubicka <jh@suse.cz>
924 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
926 (dump_cgraph_node): Dump profile-id.
927 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
928 and common_target_probability.
929 * lto-cgraph.c (lto_output_edge): Stream common targets.
930 (lto_output_node): Stream profile ids.
931 (input_node): Stream profile ids.
932 (input_edge): Stream common targets.
933 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
934 * ipa.c: Include value-prof.h
935 (ipa_profile_generate_summary): Turn indirect call statement histograms
937 (ipa_profile): Turn common targets into speculative edges.
939 2013-08-09 Jan Hubicka <jh@suse.cz>
941 * cgraph.h (cgraph_node): Add profile_id.
942 * value-prof.c (cgraph_node_map): Turn into pointer_map.
943 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
944 (del_node_map): Update.
945 (find_func_by_funcdef_no): Replace by ...
946 (find_func_by_profile_id): ... this one.
947 (gimple_ic_transform): Do not remove useful histograms when
948 speculation is not done; dump info when indirect call removal
950 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
951 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
952 (__gcov_indirect_call_profiler_v2): .. this one.
953 * profile.h (init_node_map): Update.
954 * coverage.c (coverage_compute_profile_id): New function.
955 * coverage.h (coverage_compute_profile_id): Declare.
956 * tree-profile.c (init_ic_make_global_vars): Make
957 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
958 (gimple_init_edge_profiler): Update prototype of
959 __gcov_indirect_call_profiler.
960 (gimple_gen_ic_func_profiler): Simplify.
961 (tree_profiling): Use init_node_map
963 2013-08-09 Jan Hubicka <jh@suse.cz>
965 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
966 non-speculative refs.
967 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
968 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
969 (cgraph_set_call_stmt): Likewise.
970 (cgraph_create_edge_1): Fix release checking compilatoin;
972 (cgraph_free_edge): Free indirect info.
973 (cgraph_turn_edge_to_speculative): New function.
974 (cgraph_speculative_call_info): New function.
975 (cgraph_make_edge_direct): Return direct edge; handle speculation.
976 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
977 (dump_cgraph_node): Dump speculation.
978 (verify_edge_count_and_frequency): Accept speculative edges.
979 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
980 (verify_cgraph_node): Handle speculation.
981 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
982 (cgraph_set_call_stmt): Update prototype.
983 (cgraph_make_edge_direct): Update prototype.
984 (cgraph_speculative_call_info): Declare.
985 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
986 to change; update call of ipa_find_references.
987 * ipa-ref.c (ipa_record_reference): Fix return value; clear
988 lto_stmt_uid and speculative flags.
989 (ipa_dump_references): Dump speculation.
990 (ipa_clone_references): Clone speculative flag.
991 (ipa_clone_referring): Likewise.
992 (ipa_clone_ref): New function.
993 (ipa_find_reference): Look into lto_stmt_uids
994 (ipa_clear_stmts_in_references): Do not clear speculative calls.
995 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
996 (ipa_find_reference): Update declaration.
997 (ipa_clone_ref): Declare.
998 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
999 stream speculative flag.
1000 (lto_output_ref): Stream statements uids and speculation.
1001 (input_ref): Likewise.
1002 (input_edge): Stream speuclation.
1003 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
1004 (cgraph_set_call_stmt_including_clones): Handle speculation.
1005 * ipa-inline.c (heap_edge_removal_hook): New function.
1006 (inline_small_functions): Register it.
1007 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
1008 also initialize refs.
1009 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
1011 (try_make_edge_direct_simple_call): Likewise.
1012 (try_make_edge_direct_simple_call): Likewise.
1013 (update_indirect_edges_after_inlining): Likewise.
1014 (remove_described_reference): Look proper lto_stmt_uid.
1015 (propagate_controlled_uses): Likewise.
1016 (propagate_controlled_uses): Liekwise.
1017 * tree-inline.c (copy_bb): Copy speculative edges.
1018 (redirect_all_calls): New function.
1019 (copy_cfg_body): Do redirection after loop info is updated.
1020 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
1022 2013-08-09 Jan Hubicka <jh@suse.cz>
1024 * lto-streamer-out.c (output_function): Renumber PHIs.
1025 * lto-streamer-in.c (input_function): Likewise.
1027 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
1029 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
1030 (get_lane_unsigned): Likewise.
1031 (dup_lane_scalar): Likewise.
1032 (get_lane): enable for VALL.
1033 * config/aarch64/aarch64-simd.md
1034 (aarch64_dup_lane_scalar<mode>): Remove.
1035 (aarch64_get_lane_signed<mode>): Likewise.
1036 (aarch64_get_lane_unsigned<mode>): Likewise.
1037 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
1038 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
1039 (aarch64_get_lane<mode>): Enable for all vector modes.
1040 (aarch64_get_lanedi): Remove misleading constraints.
1041 * config/aarch64/arm_neon.h
1042 (__aarch64_vget_lane_any): Define.
1043 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
1044 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
1045 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
1046 * config/aarch64/iterators.md (VDQQH): New.
1050 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
1052 * configure.ac: Add GAS check for LEON instructions on SPARC.
1053 * configure: Regenerate.
1054 * config.in: Likewise.
1055 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
1057 * config/sparc/sparc.opt (LEON, LEON3): New masks.
1058 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
1060 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
1061 (AS_LEON_FLAG): New macro.
1062 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
1063 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
1064 Deal with LEON and LEON3 for the memory model.
1065 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
1066 (atomic_compare_and_swap<mode>_1): Likewise.
1067 (*atomic_compare_and_swap<mode>_1): Likewise.
1069 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1071 * config/arm/neon.md (vcond): Fix floating-point vector
1072 comparisons against 0.
1074 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
1076 * lra-constraints.c (emit_spill_move): Remove assert.
1077 (process_alt_operands): Add more debugging
1078 output. Increase reject for spilling into memory. Decrease
1079 reject for reloading scratch.
1080 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
1082 2013-08-08 Steve Ellcey <sellcey@mips.com>
1084 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
1085 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
1086 micromips incompatible. Add nan2008.
1087 (MULTILIB_DIRNAMES): Add nan2008.
1088 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
1089 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
1090 and micromips incompatible. Add nan2008.
1091 (MULTILIB_DIRNAMES): Add nan2008.
1092 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
1094 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
1096 PR rtl-optimization/58079
1097 * combine.c (combine_simplify_rtx): Avoid using SUBST if
1098 simplify_comparison has widened a comparison with an integer.
1100 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1102 * config/arm/neon.md (movmisalign<mode>): Disable when we
1103 don't allow unaligned accesses.
1104 (*movmisalign<mode>_neon_store): Likewise.
1105 (*movmisalign<mode>_neon_load): Likewise.
1106 (*movmisalign<mode>_neon_store): Likewise.
1107 (*movmisalign<mode>_neon_load): Likewise.
1109 2013-08-08 Jan Hubicka <jh@suse.cz>
1111 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
1112 (make_pass_rebuild_cgraph_edges): Also clear references.
1113 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
1114 * ipa-inline-transform.c (inline_transform): Remove all references
1116 * cgraphunit.c (expand_function): Remove all references after
1118 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
1119 (ipa_find_reference): Rewrite to iterator.
1120 (remove_stmt_references): Likewise.
1121 (ipa_clear_stmts_in_references): New function.
1122 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
1123 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
1125 * ipa-split.c (split_function): Remove references in split function.
1127 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
1130 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
1131 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
1133 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
1136 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
1137 suggested mode for the assignment isn't compatible with the
1140 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
1143 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
1145 2013-08-07 Xinliang David Li <davidxl@google.com>
1147 * config/i386/i386.opt: New option -mtune-ctrl=.
1148 * config/i386/x86-tune.def: New file.
1149 * config/i386/i386.h: include x86-tune.def.
1150 * config/i386/i386.c (ix86_option_override_internal):
1151 Parsing -mtune-ctrl= option and set tune features.
1153 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
1156 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
1157 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
1160 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
1162 * rtl.h (update_alignments): Declare.
1163 * final.c (grow_label_align): New function extracted from...
1164 (shorten_branches): ...here. Call it.
1165 (update_alignments): New function.
1166 * reorg.c (sibling_labels): New variable.
1167 (get_label_before): Add SIBLING parameter. If it is non-zero, push
1168 the new label along with it onto the sibling_labels vector.
1169 (fill_simple_delay_slots): Adjust call to get_label_before.
1170 (fill_slots_from_thread): Likewise.
1171 (relax_delay_slots): Likewise.
1172 (make_return_insns): Likewise.
1173 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
1175 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
1177 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
1178 document its semantics.
1179 (diagnostic_report_diagnostic): Adjust accordingly.
1181 2013-08-07 David Malcolm <dmalcolm@redhat.com>
1183 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
1184 (sparc_option_override): ...and port to new C++ pass API.
1185 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
1187 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
1189 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
1191 2013-08-06 Caroline Tice <cmtice@google.com>
1193 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
1194 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
1195 * tree-pass.h: Add pass_vtable_verify.
1196 * varasm.c (assemble_variable): Add code to properly set the comdat
1197 section and name for the .vtable_map_vars section.
1198 (assemble_vtyv_preinit_initializer): New function.
1199 (default_sectin_type_flags): Make sure .vtable_map_vars section has
1201 * output.h: Add function decl for assemble_vtv_preinit_initializer.
1202 * vtable-verify.c: New file.
1203 * vtable-verify.h: New file.
1204 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
1205 initialiation levels.
1206 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
1207 * passes.def: Insert pass_vtable_verify.
1208 * aclocal.m4: Reorder includes.
1209 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
1210 -fvtv-counts options.
1211 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
1212 as appropriate, if -fvtable-verify=... is used.
1213 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
1214 -fvtable-verify=... is used.
1215 * Makefile.in (OBJS): Add vtable-verify.o to list.
1216 (vtable-verify.o): Add new build rule.
1217 (GTFILES): Add vtable-verify.c to list.
1218 * common.opt (fvtable-verify=): New flag.
1219 (vtv_priority): Values for fvtable-verify= flag.
1220 (fvtv-counts): New flag.
1221 (fvtv-debug): New flag.
1222 * tree.h (save_vtable_map_decl): New extern function decl.
1224 2013-08-07 David Malcolm <dmalcolm@redhat.com>
1226 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
1227 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
1228 (pass_data_rl78_devirt): ...new pass_data instance and...
1229 (make_pass_rl78_devirt): ...new function.
1230 (rl78_asm_file_start): Port pass registration to new C++ API.
1232 2013-08-07 David Malcolm <dmalcolm@redhat.com>
1234 * coretypes.h (rtl_opt_pass): Add.
1235 (gcc::context): Add.
1236 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
1237 (epiphany_init): Port to new C++ pass API.
1238 (epiphany_optimize_mode_switching): Likewise.
1239 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
1240 (pass_manager::get_pass_mode_switching): New.
1241 (pass_manager::get_pass_peephole2): New.
1242 * mode-switching.c (pass_mode_switching): Add clone method.
1243 * recog.c (pass_peephole2): Add clone method.
1244 (pass_split_all_insns): Add clone method.
1246 2013-08-06 David Malcolm <dmalcolm@redhat.com>
1248 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
1249 (mips_option_override): ...here, porting to new C++ API for passes.
1251 2013-08-06 Jan Hubicka <jh@suse.cz>
1253 * cgraph.c (cgraph_get_body): New function based on lto.c
1255 * cgraph.h (cgraph_get_body): Declare.
1256 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
1258 * cgraphunit.c (expand_function): Get body prior expanding.
1259 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
1260 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
1262 * passes.c (do_per_function_toporder): Get body.
1263 * tree-inline.c (expand_call_inline): Get body prior inlining it.
1264 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
1266 2013-08-06 Martin Jambor <mjambor@suse.cz>
1269 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
1270 is not re-finalized. Rename second parameter to no_collect.
1272 2013-08-06 Martin Jambor <mjambor@suse.cz>
1275 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
1276 MEM_REF has proper alignment information.
1278 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
1281 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
1283 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
1284 rtx (*) (rtx, ...) with insn_gen_fn.
1285 * genoutput.c (output_insn_data): Cast gen_? function pointers to
1286 insn_gen_fn::stored_funcptr. Add initializer braces.
1288 2013-08-05 David Malcolm <dmalcolm@redhat.com>
1290 Rewrite how instances of passes are cloned to remove assumptions
1291 about their sizes (thus allowing pass subclasses to have
1292 additional data fields, albeit non-GC-managed ones at this point).
1294 * passes.c (make_pass_instance): Now that passes have clone
1295 methods, rewrite this function to eliminate XNEW and memcpy
1296 calls that used hardcoded sizes. Since this function no longer
1297 creates pass instances, rename it to...
1298 (add_pass_instance): ...this. Document the old way that passes were
1299 numbered and flagged, and rework this function to continue using it.
1300 (next_pass_1): Add an initial_pass argument for use by
1302 (position_pass): When adding multiple instances of a pass, use
1303 the pass's clone method, rather than relying on the XNEW/memcpy
1304 within the former make_pass_instance (now add_pass_instance).
1305 (pass_manager::pass_manager): When invoking next_pass_1, also supply
1306 the initial instance of the current pass within the pass manager.
1308 2013-08-05 David Malcolm <dmalcolm@redhat.com>
1310 This is the automated part of the conversion of passes from C
1311 structs to C++ classes.
1313 Patch autogenerated by refactor_passes.py from
1314 https://github.com/davidmalcolm/gcc-refactoring-scripts
1315 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
1317 * asan.c (pass_asan): Convert from a global struct to a subclass of
1318 gimple_opt_pass along with...
1319 (pass_data_asan): ...new pass_data instance and...
1320 (make_pass_asan): ...new function.
1321 (pass_asan_O0): Convert from a global struct to a subclass of
1322 gimple_opt_pass along with...
1323 (pass_data_asan_O0): ...new pass_data instance and...
1324 (make_pass_asan_O0): ...new function.
1325 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
1326 subclass of rtl_opt_pass along with...
1327 (pass_data_inc_dec): ...new pass_data instance and...
1328 (make_pass_inc_dec): ...new function.
1329 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
1330 a subclass of rtl_opt_pass along with...
1331 (pass_data_reorder_blocks): ...new pass_data instance and...
1332 (make_pass_reorder_blocks): ...new function.
1333 (pass_duplicate_computed_gotos): Convert from a global struct to a
1334 subclass of rtl_opt_pass along with...
1335 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
1336 (make_pass_duplicate_computed_gotos): ...new function.
1337 (pass_partition_blocks): Convert from a global struct to a subclass of
1338 rtl_opt_pass along with...
1339 (pass_data_partition_blocks): ...new pass_data instance and...
1340 (make_pass_partition_blocks): ...new function.
1341 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
1342 struct to a subclass of rtl_opt_pass along with...
1343 (pass_data_branch_target_load_optimize1): ...new pass_data instance
1345 (make_pass_branch_target_load_optimize1): ...new function.
1346 (pass_branch_target_load_optimize2): Convert from a global struct to a
1347 subclass of rtl_opt_pass along with...
1348 (pass_data_branch_target_load_optimize2): ...new pass_data instance
1350 (make_pass_branch_target_load_optimize2): ...new function.
1351 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
1352 of rtl_opt_pass along with...
1353 (pass_data_jump): ...new pass_data instance and...
1354 (make_pass_jump): ...new function.
1355 (pass_jump2): Convert from a global struct to a subclass of
1356 rtl_opt_pass along with...
1357 (pass_data_jump2): ...new pass_data instance and...
1358 (make_pass_jump2): ...new function.
1359 * cfgexpand.c (pass_expand): Convert from a global struct to a
1360 subclass of rtl_opt_pass along with...
1361 (pass_data_expand): ...new pass_data instance and...
1362 (make_pass_expand): ...new function.
1363 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
1364 of rtl_opt_pass along with...
1365 (pass_data_free_cfg): ...new pass_data instance and...
1366 (make_pass_free_cfg): ...new function.
1367 (pass_into_cfg_layout_mode): Convert from a global struct to a
1368 subclass of rtl_opt_pass along with...
1369 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
1370 (make_pass_into_cfg_layout_mode): ...new function.
1371 (pass_outof_cfg_layout_mode): Convert from a global struct to a
1372 subclass of rtl_opt_pass along with...
1373 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
1374 (make_pass_outof_cfg_layout_mode): ...new function.
1375 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
1376 struct to a subclass of gimple_opt_pass along with...
1377 (pass_data_build_cgraph_edges): ...new pass_data instance and...
1378 (make_pass_build_cgraph_edges): ...new function.
1379 (pass_rebuild_cgraph_edges): Convert from a global struct to a
1380 subclass of gimple_opt_pass along with...
1381 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
1382 (make_pass_rebuild_cgraph_edges): ...new function.
1383 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
1384 subclass of gimple_opt_pass along with...
1385 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
1387 (make_pass_remove_cgraph_callee_edges): ...new function.
1388 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
1389 struct to a subclass of rtl_opt_pass along with...
1390 (pass_data_stack_adjustments): ...new pass_data instance and...
1391 (make_pass_stack_adjustments): ...new function.
1392 * combine.c (pass_combine): Convert from a global struct to a subclass
1393 of rtl_opt_pass along with...
1394 (pass_data_combine): ...new pass_data instance and...
1395 (make_pass_combine): ...new function.
1396 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
1397 global struct to a subclass of rtl_opt_pass along with...
1398 (pass_data_compare_elim_after_reload): ...new pass_data instance
1400 (make_pass_compare_elim_after_reload): ...new function.
1401 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
1402 of rtl_opt_pass along with...
1403 (pass_data_rtl_cprop): ...new pass_data instance and...
1404 (make_pass_rtl_cprop): ...new function.
1405 * cse.c (pass_cse): Convert from a global struct to a subclass of
1406 rtl_opt_pass along with...
1407 (pass_data_cse): ...new pass_data instance and...
1408 (make_pass_cse): ...new function.
1409 (pass_cse2): Convert from a global struct to a subclass of
1410 rtl_opt_pass along with...
1411 (pass_data_cse2): ...new pass_data instance and...
1412 (make_pass_cse2): ...new function.
1413 (pass_cse_after_global_opts): Convert from a global struct to a
1414 subclass of rtl_opt_pass along with...
1415 (pass_data_cse_after_global_opts): ...new pass_data instance and...
1416 (make_pass_cse_after_global_opts): ...new function.
1417 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
1418 of rtl_opt_pass along with...
1419 (pass_data_ud_rtl_dce): ...new pass_data instance and...
1420 (make_pass_ud_rtl_dce): ...new function.
1421 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
1422 rtl_opt_pass along with...
1423 (pass_data_fast_rtl_dce): ...new pass_data instance and...
1424 (make_pass_fast_rtl_dce): ...new function.
1425 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
1426 a subclass of rtl_opt_pass along with...
1427 (pass_data_df_initialize_opt): ...new pass_data instance and...
1428 (make_pass_df_initialize_opt): ...new function.
1429 (pass_df_initialize_no_opt): Convert from a global struct to a
1430 subclass of rtl_opt_pass along with...
1431 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
1432 (make_pass_df_initialize_no_opt): ...new function.
1433 (pass_df_finish): Convert from a global struct to a subclass of
1434 rtl_opt_pass along with...
1435 (pass_data_df_finish): ...new pass_data instance and...
1436 (make_pass_df_finish): ...new function.
1437 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
1438 rtl_opt_pass along with...
1439 (pass_data_rtl_dse1): ...new pass_data instance and...
1440 (make_pass_rtl_dse1): ...new function.
1441 (pass_rtl_dse2): Convert from a global struct to a subclass of
1442 rtl_opt_pass along with...
1443 (pass_data_rtl_dse2): ...new pass_data instance and...
1444 (make_pass_rtl_dse2): ...new function.
1445 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
1446 subclass of rtl_opt_pass along with...
1447 (pass_data_dwarf2_frame): ...new pass_data instance and...
1448 (make_pass_dwarf2_frame): ...new function.
1449 * except.c (pass_set_nothrow_function_flags): Convert from a global
1450 struct to a subclass of rtl_opt_pass along with...
1451 (pass_data_set_nothrow_function_flags): ...new pass_data instance
1453 (make_pass_set_nothrow_function_flags): ...new function.
1454 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
1455 subclass of rtl_opt_pass along with...
1456 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
1458 (make_pass_convert_to_eh_region_ranges): ...new function.
1459 * final.c (pass_compute_alignments): Convert from a global struct to a
1460 subclass of rtl_opt_pass along with...
1461 (pass_data_compute_alignments): ...new pass_data instance and...
1462 (make_pass_compute_alignments): ...new function.
1463 (pass_final): Convert from a global struct to a subclass of
1464 rtl_opt_pass along with...
1465 (pass_data_final): ...new pass_data instance and...
1466 (make_pass_final): ...new function.
1467 (pass_shorten_branches): Convert from a global struct to a subclass of
1468 rtl_opt_pass along with...
1469 (pass_data_shorten_branches): ...new pass_data instance and...
1470 (make_pass_shorten_branches): ...new function.
1471 (pass_clean_state): Convert from a global struct to a subclass of
1472 rtl_opt_pass along with...
1473 (pass_data_clean_state): ...new pass_data instance and...
1474 (make_pass_clean_state): ...new function.
1475 * function.c (pass_instantiate_virtual_regs): Convert from a global
1476 struct to a subclass of rtl_opt_pass along with...
1477 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
1478 (make_pass_instantiate_virtual_regs): ...new function.
1479 (pass_leaf_regs): Convert from a global struct to a subclass of
1480 rtl_opt_pass along with...
1481 (pass_data_leaf_regs): ...new pass_data instance and...
1482 (make_pass_leaf_regs): ...new function.
1483 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
1484 subclass of rtl_opt_pass along with...
1485 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
1487 (make_pass_thread_prologue_and_epilogue): ...new function.
1488 (pass_match_asm_constraints): Convert from a global struct to a
1489 subclass of rtl_opt_pass along with...
1490 (pass_data_match_asm_constraints): ...new pass_data instance and...
1491 (make_pass_match_asm_constraints): ...new function.
1492 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
1493 subclass of rtl_opt_pass along with...
1494 (pass_data_rtl_fwprop): ...new pass_data instance and...
1495 (make_pass_rtl_fwprop): ...new function.
1496 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
1497 rtl_opt_pass along with...
1498 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
1499 (make_pass_rtl_fwprop_addr): ...new function.
1500 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
1501 rtl_opt_pass along with...
1502 (pass_data_rtl_pre): ...new pass_data instance and...
1503 (make_pass_rtl_pre): ...new function.
1504 (pass_rtl_hoist): Convert from a global struct to a subclass of
1505 rtl_opt_pass along with...
1506 (pass_data_rtl_hoist): ...new pass_data instance and...
1507 (make_pass_rtl_hoist): ...new function.
1508 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
1509 subclass of gimple_opt_pass along with...
1510 (pass_data_lower_cf): ...new pass_data instance and...
1511 (make_pass_lower_cf): ...new function.
1512 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
1513 from a global struct to a subclass of gimple_opt_pass along with...
1514 (pass_data_strength_reduction): ...new pass_data instance and...
1515 (make_pass_strength_reduction): ...new function.
1516 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
1517 of rtl_opt_pass along with...
1518 (pass_data_rtl_ifcvt): ...new pass_data instance and...
1519 (make_pass_rtl_ifcvt): ...new function.
1520 (pass_if_after_combine): Convert from a global struct to a subclass of
1521 rtl_opt_pass along with...
1522 (pass_data_if_after_combine): ...new pass_data instance and...
1523 (make_pass_if_after_combine): ...new function.
1524 (pass_if_after_reload): Convert from a global struct to a subclass of
1525 rtl_opt_pass along with...
1526 (pass_data_if_after_reload): ...new pass_data instance and...
1527 (make_pass_if_after_reload): ...new function.
1528 * init-regs.c (pass_initialize_regs): Convert from a global struct to
1529 a subclass of rtl_opt_pass along with...
1530 (pass_data_initialize_regs): ...new pass_data instance and...
1531 (make_pass_initialize_regs): ...new function.
1532 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
1533 of ipa_opt_pass_d along with...
1534 (pass_data_ipa_cp): ...new pass_data instance and...
1535 (make_pass_ipa_cp): ...new function.
1536 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
1537 global struct to a subclass of gimple_opt_pass along with...
1538 (pass_data_inline_parameters): ...new pass_data instance and...
1539 (make_pass_inline_parameters): ...new function.
1540 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
1541 subclass of gimple_opt_pass along with...
1542 (pass_data_early_inline): ...new pass_data instance and...
1543 (make_pass_early_inline): ...new function.
1544 (pass_ipa_inline): Convert from a global struct to a subclass of
1545 ipa_opt_pass_d along with...
1546 (pass_data_ipa_inline): ...new pass_data instance and...
1547 (make_pass_ipa_inline): ...new function.
1548 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
1549 struct to a subclass of gimple_opt_pass along with...
1550 (pass_data_local_pure_const): ...new pass_data instance and...
1551 (make_pass_local_pure_const): ...new function.
1552 (pass_ipa_pure_const): Convert from a global struct to a subclass of
1553 ipa_opt_pass_d along with...
1554 (pass_data_ipa_pure_const): ...new pass_data instance and...
1555 (make_pass_ipa_pure_const): ...new function.
1556 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
1557 to a subclass of ipa_opt_pass_d along with...
1558 (pass_data_ipa_reference): ...new pass_data instance and...
1559 (make_pass_ipa_reference): ...new function.
1560 * ipa-split.c (pass_split_functions): Convert from a global struct to
1561 a subclass of gimple_opt_pass along with...
1562 (pass_data_split_functions): ...new pass_data instance and...
1563 (make_pass_split_functions): ...new function.
1564 (pass_feedback_split_functions): Convert from a global struct to a
1565 subclass of gimple_opt_pass along with...
1566 (pass_data_feedback_split_functions): ...new pass_data instance and...
1567 (make_pass_feedback_split_functions): ...new function.
1568 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
1569 global struct to a subclass of simple_ipa_opt_pass along with...
1570 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
1572 (make_pass_ipa_function_and_variable_visibility): ...new function.
1573 (pass_ipa_free_inline_summary): Convert from a global struct to a
1574 subclass of simple_ipa_opt_pass along with...
1575 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
1576 (make_pass_ipa_free_inline_summary): ...new function.
1577 (pass_ipa_whole_program_visibility): Convert from a global struct to a
1578 subclass of ipa_opt_pass_d along with...
1579 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
1581 (make_pass_ipa_whole_program_visibility): ...new function.
1582 (pass_ipa_profile): Convert from a global struct to a subclass of
1583 ipa_opt_pass_d along with...
1584 (pass_data_ipa_profile): ...new pass_data instance and...
1585 (make_pass_ipa_profile): ...new function.
1586 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
1587 ipa_opt_pass_d along with...
1588 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
1589 (make_pass_ipa_cdtor_merge): ...new function.
1590 * ira.c (pass_ira): Convert from a global struct to a subclass of
1591 rtl_opt_pass along with...
1592 (pass_data_ira): ...new pass_data instance and...
1593 (make_pass_ira): ...new function.
1594 (pass_reload): Convert from a global struct to a subclass of
1595 rtl_opt_pass along with...
1596 (pass_data_reload): ...new pass_data instance and...
1597 (make_pass_reload): ...new function.
1598 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
1599 subclass of rtl_opt_pass along with...
1600 (pass_data_cleanup_barriers): ...new pass_data instance and...
1601 (make_pass_cleanup_barriers): ...new function.
1602 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
1603 of rtl_opt_pass along with...
1604 (pass_data_loop2): ...new pass_data instance and...
1605 (make_pass_loop2): ...new function.
1606 (pass_rtl_loop_init): Convert from a global struct to a subclass of
1607 rtl_opt_pass along with...
1608 (pass_data_rtl_loop_init): ...new pass_data instance and...
1609 (make_pass_rtl_loop_init): ...new function.
1610 (pass_rtl_loop_done): Convert from a global struct to a subclass of
1611 rtl_opt_pass along with...
1612 (pass_data_rtl_loop_done): ...new pass_data instance and...
1613 (make_pass_rtl_loop_done): ...new function.
1614 (pass_rtl_move_loop_invariants): Convert from a global struct to a
1615 subclass of rtl_opt_pass along with...
1616 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
1617 (make_pass_rtl_move_loop_invariants): ...new function.
1618 (pass_rtl_unswitch): Convert from a global struct to a subclass of
1619 rtl_opt_pass along with...
1620 (pass_data_rtl_unswitch): ...new pass_data instance and...
1621 (make_pass_rtl_unswitch): ...new function.
1622 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
1623 subclass of rtl_opt_pass along with...
1624 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
1626 (make_pass_rtl_unroll_and_peel_loops): ...new function.
1627 (pass_rtl_doloop): Convert from a global struct to a subclass of
1628 rtl_opt_pass along with...
1629 (pass_data_rtl_doloop): ...new pass_data instance and...
1630 (make_pass_rtl_doloop): ...new function.
1631 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
1632 a subclass of rtl_opt_pass along with...
1633 (pass_data_lower_subreg): ...new pass_data instance and...
1634 (make_pass_lower_subreg): ...new function.
1635 (pass_lower_subreg2): Convert from a global struct to a subclass of
1636 rtl_opt_pass along with...
1637 (pass_data_lower_subreg2): ...new pass_data instance and...
1638 (make_pass_lower_subreg2): ...new function.
1639 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
1640 struct to a subclass of ipa_opt_pass_d along with...
1641 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
1642 (make_pass_ipa_lto_gimple_out): ...new function.
1643 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
1644 of ipa_opt_pass_d along with...
1645 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
1646 (make_pass_ipa_lto_finish_out): ...new function.
1647 * mode-switching.c (pass_mode_switching): Convert from a global struct
1648 to a subclass of rtl_opt_pass along with...
1649 (pass_data_mode_switching): ...new pass_data instance and...
1650 (make_pass_mode_switching): ...new function.
1651 * modulo-sched.c (pass_sms): Convert from a global struct to a
1652 subclass of rtl_opt_pass along with...
1653 (pass_data_sms): ...new pass_data instance and...
1654 (make_pass_sms): ...new function.
1655 * omp-low.c (pass_expand_omp): Convert from a global struct to a
1656 subclass of gimple_opt_pass along with...
1657 (pass_data_expand_omp): ...new pass_data instance and...
1658 (make_pass_expand_omp): ...new function.
1659 (pass_lower_omp): Convert from a global struct to a subclass of
1660 gimple_opt_pass along with...
1661 (pass_data_lower_omp): ...new pass_data instance and...
1662 (make_pass_lower_omp): ...new function.
1663 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
1664 of gimple_opt_pass along with...
1665 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
1666 (make_pass_diagnose_omp_blocks): ...new function.
1667 * passes.c (pass_early_local_passes): Convert from a global struct to
1668 a subclass of simple_ipa_opt_pass along with...
1669 (pass_data_early_local_passes): ...new pass_data instance and...
1670 (make_pass_early_local_passes): ...new function.
1671 (pass_all_early_optimizations): Convert from a global struct to a
1672 subclass of gimple_opt_pass along with...
1673 (pass_data_all_early_optimizations): ...new pass_data instance and...
1674 (make_pass_all_early_optimizations): ...new function.
1675 (pass_all_optimizations): Convert from a global struct to a subclass
1676 of gimple_opt_pass along with...
1677 (pass_data_all_optimizations): ...new pass_data instance and...
1678 (make_pass_all_optimizations): ...new function.
1679 (pass_all_optimizations_g): Convert from a global struct to a subclass
1680 of gimple_opt_pass along with...
1681 (pass_data_all_optimizations_g): ...new pass_data instance and...
1682 (make_pass_all_optimizations_g): ...new function.
1683 (pass_rest_of_compilation): Convert from a global struct to a subclass
1684 of rtl_opt_pass along with...
1685 (pass_data_rest_of_compilation): ...new pass_data instance and...
1686 (make_pass_rest_of_compilation): ...new function.
1687 (pass_postreload): Convert from a global struct to a subclass of
1688 rtl_opt_pass along with...
1689 (pass_data_postreload): ...new pass_data instance and...
1690 (make_pass_postreload): ...new function.
1691 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
1692 subclass of rtl_opt_pass along with...
1693 (pass_data_gcse2): ...new pass_data instance and...
1694 (make_pass_gcse2): ...new function.
1695 * postreload.c (pass_postreload_cse): Convert from a global struct to
1696 a subclass of rtl_opt_pass along with...
1697 (pass_data_postreload_cse): ...new pass_data instance and...
1698 (make_pass_postreload_cse): ...new function.
1699 * predict.c (pass_profile): Convert from a global struct to a subclass
1700 of gimple_opt_pass along with...
1701 (pass_data_profile): ...new pass_data instance and...
1702 (make_pass_profile): ...new function.
1703 (pass_strip_predict_hints): Convert from a global struct to a subclass
1704 of gimple_opt_pass along with...
1705 (pass_data_strip_predict_hints): ...new pass_data instance and...
1706 (make_pass_strip_predict_hints): ...new function.
1707 * recog.c (pass_peephole2): Convert from a global struct to a subclass
1708 of rtl_opt_pass along with...
1709 (pass_data_peephole2): ...new pass_data instance and...
1710 (make_pass_peephole2): ...new function.
1711 (pass_split_all_insns): Convert from a global struct to a subclass of
1712 rtl_opt_pass along with...
1713 (pass_data_split_all_insns): ...new pass_data instance and...
1714 (make_pass_split_all_insns): ...new function.
1715 (pass_split_after_reload): Convert from a global struct to a subclass
1716 of rtl_opt_pass along with...
1717 (pass_data_split_after_reload): ...new pass_data instance and...
1718 (make_pass_split_after_reload): ...new function.
1719 (pass_split_before_regstack): Convert from a global struct to a
1720 subclass of rtl_opt_pass along with...
1721 (pass_data_split_before_regstack): ...new pass_data instance and...
1722 (make_pass_split_before_regstack): ...new function.
1723 (pass_split_before_sched2): Convert from a global struct to a subclass
1724 of rtl_opt_pass along with...
1725 (pass_data_split_before_sched2): ...new pass_data instance and...
1726 (make_pass_split_before_sched2): ...new function.
1727 (pass_split_for_shorten_branches): Convert from a global struct to a
1728 subclass of rtl_opt_pass along with...
1729 (pass_data_split_for_shorten_branches): ...new pass_data instance
1731 (make_pass_split_for_shorten_branches): ...new function.
1732 * ree.c (pass_ree): Convert from a global struct to a subclass of
1733 rtl_opt_pass along with...
1734 (pass_data_ree): ...new pass_data instance and...
1735 (make_pass_ree): ...new function.
1736 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
1737 subclass of rtl_opt_pass along with...
1738 (pass_data_stack_regs): ...new pass_data instance and...
1739 (make_pass_stack_regs): ...new function.
1740 (pass_stack_regs_run): Convert from a global struct to a subclass of
1741 rtl_opt_pass along with...
1742 (pass_data_stack_regs_run): ...new pass_data instance and...
1743 (make_pass_stack_regs_run): ...new function.
1744 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
1745 subclass of rtl_opt_pass along with...
1746 (pass_data_cprop_hardreg): ...new pass_data instance and...
1747 (make_pass_cprop_hardreg): ...new function.
1748 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
1749 subclass of rtl_opt_pass along with...
1750 (pass_data_reginfo_init): ...new pass_data instance and...
1751 (make_pass_reginfo_init): ...new function.
1752 * regmove.c (pass_regmove): Convert from a global struct to a subclass
1753 of rtl_opt_pass along with...
1754 (pass_data_regmove): ...new pass_data instance and...
1755 (make_pass_regmove): ...new function.
1756 * regrename.c (pass_regrename): Convert from a global struct to a
1757 subclass of rtl_opt_pass along with...
1758 (pass_data_regrename): ...new pass_data instance and...
1759 (make_pass_regrename): ...new function.
1760 * reorg.c (pass_delay_slots): Convert from a global struct to a
1761 subclass of rtl_opt_pass along with...
1762 (pass_data_delay_slots): ...new pass_data instance and...
1763 (make_pass_delay_slots): ...new function.
1764 (pass_machine_reorg): Convert from a global struct to a subclass of
1765 rtl_opt_pass along with...
1766 (pass_data_machine_reorg): ...new pass_data instance and...
1767 (make_pass_machine_reorg): ...new function.
1768 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
1769 of rtl_opt_pass along with...
1770 (pass_data_sched): ...new pass_data instance and...
1771 (make_pass_sched): ...new function.
1772 (pass_sched2): Convert from a global struct to a subclass of
1773 rtl_opt_pass along with...
1774 (pass_data_sched2): ...new pass_data instance and...
1775 (make_pass_sched2): ...new function.
1776 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
1777 to a subclass of rtl_opt_pass along with...
1778 (pass_data_stack_ptr_mod): ...new pass_data instance and...
1779 (make_pass_stack_ptr_mod): ...new function.
1780 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
1781 to a subclass of rtl_opt_pass along with...
1782 (pass_data_rtl_store_motion): ...new pass_data instance and...
1783 (make_pass_rtl_store_motion): ...new function.
1784 * tracer.c (pass_tracer): Convert from a global struct to a subclass
1785 of gimple_opt_pass along with...
1786 (pass_data_tracer): ...new pass_data instance and...
1787 (make_pass_tracer): ...new function.
1788 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
1789 to a subclass of gimple_opt_pass along with...
1790 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
1791 (make_pass_diagnose_tm_blocks): ...new function.
1792 (pass_lower_tm): Convert from a global struct to a subclass of
1793 gimple_opt_pass along with...
1794 (pass_data_lower_tm): ...new pass_data instance and...
1795 (make_pass_lower_tm): ...new function.
1796 (pass_tm_init): Convert from a global struct to a subclass of
1797 gimple_opt_pass along with...
1798 (pass_data_tm_init): ...new pass_data instance and...
1799 (make_pass_tm_init): ...new function.
1800 (pass_tm_mark): Convert from a global struct to a subclass of
1801 gimple_opt_pass along with...
1802 (pass_data_tm_mark): ...new pass_data instance and...
1803 (make_pass_tm_mark): ...new function.
1804 (pass_tm_edges): Convert from a global struct to a subclass of
1805 gimple_opt_pass along with...
1806 (pass_data_tm_edges): ...new pass_data instance and...
1807 (make_pass_tm_edges): ...new function.
1808 (pass_tm_memopt): Convert from a global struct to a subclass of
1809 gimple_opt_pass along with...
1810 (pass_data_tm_memopt): ...new pass_data instance and...
1811 (make_pass_tm_memopt): ...new function.
1812 (pass_ipa_tm): Convert from a global struct to a subclass of
1813 simple_ipa_opt_pass along with...
1814 (pass_data_ipa_tm): ...new pass_data instance and...
1815 (make_pass_ipa_tm): ...new function.
1816 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
1817 subclass of gimple_opt_pass along with...
1818 (pass_data_call_cdce): ...new pass_data instance and...
1819 (make_pass_call_cdce): ...new function.
1820 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
1821 subclass of gimple_opt_pass along with...
1822 (pass_data_build_cfg): ...new pass_data instance and...
1823 (make_pass_build_cfg): ...new function.
1824 (pass_split_crit_edges): Convert from a global struct to a subclass of
1825 gimple_opt_pass along with...
1826 (pass_data_split_crit_edges): ...new pass_data instance and...
1827 (make_pass_split_crit_edges): ...new function.
1828 (pass_warn_function_return): Convert from a global struct to a
1829 subclass of gimple_opt_pass along with...
1830 (pass_data_warn_function_return): ...new pass_data instance and...
1831 (make_pass_warn_function_return): ...new function.
1832 (pass_warn_function_noreturn): Convert from a global struct to a
1833 subclass of gimple_opt_pass along with...
1834 (pass_data_warn_function_noreturn): ...new pass_data instance and...
1835 (make_pass_warn_function_noreturn): ...new function.
1836 (pass_warn_unused_result): Convert from a global struct to a subclass
1837 of gimple_opt_pass along with...
1838 (pass_data_warn_unused_result): ...new pass_data instance and...
1839 (make_pass_warn_unused_result): ...new function.
1840 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
1841 a subclass of gimple_opt_pass along with...
1842 (pass_data_merge_phi): ...new pass_data instance and...
1843 (make_pass_merge_phi): ...new function.
1844 * tree-complex.c (pass_lower_complex): Convert from a global struct to
1845 a subclass of gimple_opt_pass along with...
1846 (pass_data_lower_complex): ...new pass_data instance and...
1847 (make_pass_lower_complex): ...new function.
1848 (pass_lower_complex_O0): Convert from a global struct to a subclass of
1849 gimple_opt_pass along with...
1850 (pass_data_lower_complex_O0): ...new pass_data instance and...
1851 (make_pass_lower_complex_O0): ...new function.
1852 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
1853 subclass of gimple_opt_pass along with...
1854 (pass_data_lower_eh): ...new pass_data instance and...
1855 (make_pass_lower_eh): ...new function.
1856 (pass_refactor_eh): Convert from a global struct to a subclass of
1857 gimple_opt_pass along with...
1858 (pass_data_refactor_eh): ...new pass_data instance and...
1859 (make_pass_refactor_eh): ...new function.
1860 (pass_lower_resx): Convert from a global struct to a subclass of
1861 gimple_opt_pass along with...
1862 (pass_data_lower_resx): ...new pass_data instance and...
1863 (make_pass_lower_resx): ...new function.
1864 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
1865 of gimple_opt_pass along with...
1866 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
1867 (make_pass_lower_eh_dispatch): ...new function.
1868 (pass_cleanup_eh): Convert from a global struct to a subclass of
1869 gimple_opt_pass along with...
1870 (pass_data_cleanup_eh): ...new pass_data instance and...
1871 (make_pass_cleanup_eh): ...new function.
1872 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
1873 to a subclass of simple_ipa_opt_pass along with...
1874 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
1875 (make_pass_ipa_lower_emutls): ...new function.
1876 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
1877 a subclass of gimple_opt_pass along with...
1878 (pass_data_if_conversion): ...new pass_data instance and...
1879 (make_pass_if_conversion): ...new function.
1880 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
1881 subclass of gimple_opt_pass along with...
1882 (pass_data_build_ssa): ...new pass_data instance and...
1883 (make_pass_build_ssa): ...new function.
1884 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
1885 global struct to a subclass of gimple_opt_pass along with...
1886 (pass_data_loop_distribution): ...new pass_data instance and...
1887 (make_pass_loop_distribution): ...new function.
1888 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
1889 subclass of gimple_opt_pass along with...
1890 (pass_data_mudflap_1): ...new pass_data instance and...
1891 (make_pass_mudflap_1): ...new function.
1892 (pass_mudflap_2): Convert from a global struct to a subclass of
1893 gimple_opt_pass along with...
1894 (pass_data_mudflap_2): ...new pass_data instance and...
1895 (make_pass_mudflap_2): ...new function.
1896 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
1897 subclass of gimple_opt_pass along with...
1898 (pass_data_mudflap_1): ...new pass_data instance and...
1899 (make_pass_mudflap_1): ...new function.
1900 (pass_mudflap_2): Convert from a global struct to a subclass of
1901 gimple_opt_pass along with...
1902 (pass_data_mudflap_2): ...new pass_data instance and...
1903 (make_pass_mudflap_2): ...new function.
1904 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
1905 gimple_opt_pass along with...
1906 (pass_data_nrv): ...new pass_data instance and...
1907 (make_pass_nrv): ...new function.
1908 (pass_return_slot): Convert from a global struct to a subclass of
1909 gimple_opt_pass along with...
1910 (pass_data_return_slot): ...new pass_data instance and...
1911 (make_pass_return_slot): ...new function.
1912 * tree-object-size.c (pass_object_sizes): Convert from a global struct
1913 to a subclass of gimple_opt_pass along with...
1914 (pass_data_object_sizes): ...new pass_data instance and...
1915 (make_pass_object_sizes): ...new function.
1916 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
1917 global struct to a subclass of gimple_opt_pass along with...
1918 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
1920 (make_pass_cleanup_cfg_post_optimizing): ...new function.
1921 (pass_fixup_cfg): Convert from a global struct to a subclass of
1922 gimple_opt_pass along with...
1923 (pass_data_fixup_cfg): ...new pass_data instance and...
1924 (make_pass_fixup_cfg): ...new function.
1925 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
1926 (make_pass_mudflap_1): ...new function.
1927 (pass_mudflap_2): Replace declaration with that of...
1928 (make_pass_mudflap_2): ...new function.
1929 (pass_asan): Replace declaration with that of...
1930 (make_pass_asan): ...new function.
1931 (pass_asan_O0): Replace declaration with that of...
1932 (make_pass_asan_O0): ...new function.
1933 (pass_tsan): Replace declaration with that of...
1934 (make_pass_tsan): ...new function.
1935 (pass_tsan_O0): Replace declaration with that of...
1936 (make_pass_tsan_O0): ...new function.
1937 (pass_lower_cf): Replace declaration with that of...
1938 (make_pass_lower_cf): ...new function.
1939 (pass_refactor_eh): Replace declaration with that of...
1940 (make_pass_refactor_eh): ...new function.
1941 (pass_lower_eh): Replace declaration with that of...
1942 (make_pass_lower_eh): ...new function.
1943 (pass_lower_eh_dispatch): Replace declaration with that of...
1944 (make_pass_lower_eh_dispatch): ...new function.
1945 (pass_lower_resx): Replace declaration with that of...
1946 (make_pass_lower_resx): ...new function.
1947 (pass_build_cfg): Replace declaration with that of...
1948 (make_pass_build_cfg): ...new function.
1949 (pass_early_tree_profile): Replace declaration with that of...
1950 (make_pass_early_tree_profile): ...new function.
1951 (pass_cleanup_eh): Replace declaration with that of...
1952 (make_pass_cleanup_eh): ...new function.
1953 (pass_sra): Replace declaration with that of...
1954 (make_pass_sra): ...new function.
1955 (pass_sra_early): Replace declaration with that of...
1956 (make_pass_sra_early): ...new function.
1957 (pass_early_ipa_sra): Replace declaration with that of...
1958 (make_pass_early_ipa_sra): ...new function.
1959 (pass_tail_recursion): Replace declaration with that of...
1960 (make_pass_tail_recursion): ...new function.
1961 (pass_tail_calls): Replace declaration with that of...
1962 (make_pass_tail_calls): ...new function.
1963 (pass_tree_loop): Replace declaration with that of...
1964 (make_pass_tree_loop): ...new function.
1965 (pass_tree_loop_init): Replace declaration with that of...
1966 (make_pass_tree_loop_init): ...new function.
1967 (pass_lim): Replace declaration with that of...
1968 (make_pass_lim): ...new function.
1969 (pass_tree_unswitch): Replace declaration with that of...
1970 (make_pass_tree_unswitch): ...new function.
1971 (pass_predcom): Replace declaration with that of...
1972 (make_pass_predcom): ...new function.
1973 (pass_iv_canon): Replace declaration with that of...
1974 (make_pass_iv_canon): ...new function.
1975 (pass_scev_cprop): Replace declaration with that of...
1976 (make_pass_scev_cprop): ...new function.
1977 (pass_empty_loop): Replace declaration with that of...
1978 (make_pass_empty_loop): ...new function.
1979 (pass_record_bounds): Replace declaration with that of...
1980 (make_pass_record_bounds): ...new function.
1981 (pass_graphite): Replace declaration with that of...
1982 (make_pass_graphite): ...new function.
1983 (pass_graphite_transforms): Replace declaration with that of...
1984 (make_pass_graphite_transforms): ...new function.
1985 (pass_if_conversion): Replace declaration with that of...
1986 (make_pass_if_conversion): ...new function.
1987 (pass_loop_distribution): Replace declaration with that of...
1988 (make_pass_loop_distribution): ...new function.
1989 (pass_vectorize): Replace declaration with that of...
1990 (make_pass_vectorize): ...new function.
1991 (pass_slp_vectorize): Replace declaration with that of...
1992 (make_pass_slp_vectorize): ...new function.
1993 (pass_complete_unroll): Replace declaration with that of...
1994 (make_pass_complete_unroll): ...new function.
1995 (pass_complete_unrolli): Replace declaration with that of...
1996 (make_pass_complete_unrolli): ...new function.
1997 (pass_parallelize_loops): Replace declaration with that of...
1998 (make_pass_parallelize_loops): ...new function.
1999 (pass_loop_prefetch): Replace declaration with that of...
2000 (make_pass_loop_prefetch): ...new function.
2001 (pass_iv_optimize): Replace declaration with that of...
2002 (make_pass_iv_optimize): ...new function.
2003 (pass_tree_loop_done): Replace declaration with that of...
2004 (make_pass_tree_loop_done): ...new function.
2005 (pass_ch): Replace declaration with that of...
2006 (make_pass_ch): ...new function.
2007 (pass_ccp): Replace declaration with that of...
2008 (make_pass_ccp): ...new function.
2009 (pass_phi_only_cprop): Replace declaration with that of...
2010 (make_pass_phi_only_cprop): ...new function.
2011 (pass_build_ssa): Replace declaration with that of...
2012 (make_pass_build_ssa): ...new function.
2013 (pass_build_alias): Replace declaration with that of...
2014 (make_pass_build_alias): ...new function.
2015 (pass_build_ealias): Replace declaration with that of...
2016 (make_pass_build_ealias): ...new function.
2017 (pass_dominator): Replace declaration with that of...
2018 (make_pass_dominator): ...new function.
2019 (pass_dce): Replace declaration with that of...
2020 (make_pass_dce): ...new function.
2021 (pass_dce_loop): Replace declaration with that of...
2022 (make_pass_dce_loop): ...new function.
2023 (pass_cd_dce): Replace declaration with that of...
2024 (make_pass_cd_dce): ...new function.
2025 (pass_call_cdce): Replace declaration with that of...
2026 (make_pass_call_cdce): ...new function.
2027 (pass_merge_phi): Replace declaration with that of...
2028 (make_pass_merge_phi): ...new function.
2029 (pass_split_crit_edges): Replace declaration with that of...
2030 (make_pass_split_crit_edges): ...new function.
2031 (pass_pre): Replace declaration with that of...
2032 (make_pass_pre): ...new function.
2033 (pass_profile): Replace declaration with that of...
2034 (make_pass_profile): ...new function.
2035 (pass_strip_predict_hints): Replace declaration with that of...
2036 (make_pass_strip_predict_hints): ...new function.
2037 (pass_lower_complex_O0): Replace declaration with that of...
2038 (make_pass_lower_complex_O0): ...new function.
2039 (pass_lower_complex): Replace declaration with that of...
2040 (make_pass_lower_complex): ...new function.
2041 (pass_lower_vector): Replace declaration with that of...
2042 (make_pass_lower_vector): ...new function.
2043 (pass_lower_vector_ssa): Replace declaration with that of...
2044 (make_pass_lower_vector_ssa): ...new function.
2045 (pass_lower_omp): Replace declaration with that of...
2046 (make_pass_lower_omp): ...new function.
2047 (pass_diagnose_omp_blocks): Replace declaration with that of...
2048 (make_pass_diagnose_omp_blocks): ...new function.
2049 (pass_expand_omp): Replace declaration with that of...
2050 (make_pass_expand_omp): ...new function.
2051 (pass_expand_omp_ssa): Replace declaration with that of...
2052 (make_pass_expand_omp_ssa): ...new function.
2053 (pass_object_sizes): Replace declaration with that of...
2054 (make_pass_object_sizes): ...new function.
2055 (pass_strlen): Replace declaration with that of...
2056 (make_pass_strlen): ...new function.
2057 (pass_fold_builtins): Replace declaration with that of...
2058 (make_pass_fold_builtins): ...new function.
2059 (pass_stdarg): Replace declaration with that of...
2060 (make_pass_stdarg): ...new function.
2061 (pass_early_warn_uninitialized): Replace declaration with that of...
2062 (make_pass_early_warn_uninitialized): ...new function.
2063 (pass_late_warn_uninitialized): Replace declaration with that of...
2064 (make_pass_late_warn_uninitialized): ...new function.
2065 (pass_cse_reciprocals): Replace declaration with that of...
2066 (make_pass_cse_reciprocals): ...new function.
2067 (pass_cse_sincos): Replace declaration with that of...
2068 (make_pass_cse_sincos): ...new function.
2069 (pass_optimize_bswap): Replace declaration with that of...
2070 (make_pass_optimize_bswap): ...new function.
2071 (pass_optimize_widening_mul): Replace declaration with that of...
2072 (make_pass_optimize_widening_mul): ...new function.
2073 (pass_warn_function_return): Replace declaration with that of...
2074 (make_pass_warn_function_return): ...new function.
2075 (pass_warn_function_noreturn): Replace declaration with that of...
2076 (make_pass_warn_function_noreturn): ...new function.
2077 (pass_cselim): Replace declaration with that of...
2078 (make_pass_cselim): ...new function.
2079 (pass_phiopt): Replace declaration with that of...
2080 (make_pass_phiopt): ...new function.
2081 (pass_forwprop): Replace declaration with that of...
2082 (make_pass_forwprop): ...new function.
2083 (pass_phiprop): Replace declaration with that of...
2084 (make_pass_phiprop): ...new function.
2085 (pass_tree_ifcombine): Replace declaration with that of...
2086 (make_pass_tree_ifcombine): ...new function.
2087 (pass_dse): Replace declaration with that of...
2088 (make_pass_dse): ...new function.
2089 (pass_nrv): Replace declaration with that of...
2090 (make_pass_nrv): ...new function.
2091 (pass_rename_ssa_copies): Replace declaration with that of...
2092 (make_pass_rename_ssa_copies): ...new function.
2093 (pass_sink_code): Replace declaration with that of...
2094 (make_pass_sink_code): ...new function.
2095 (pass_fre): Replace declaration with that of...
2096 (make_pass_fre): ...new function.
2097 (pass_check_data_deps): Replace declaration with that of...
2098 (make_pass_check_data_deps): ...new function.
2099 (pass_copy_prop): Replace declaration with that of...
2100 (make_pass_copy_prop): ...new function.
2101 (pass_vrp): Replace declaration with that of...
2102 (make_pass_vrp): ...new function.
2103 (pass_uncprop): Replace declaration with that of...
2104 (make_pass_uncprop): ...new function.
2105 (pass_return_slot): Replace declaration with that of...
2106 (make_pass_return_slot): ...new function.
2107 (pass_reassoc): Replace declaration with that of...
2108 (make_pass_reassoc): ...new function.
2109 (pass_rebuild_cgraph_edges): Replace declaration with that of...
2110 (make_pass_rebuild_cgraph_edges): ...new function.
2111 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
2112 (make_pass_remove_cgraph_callee_edges): ...new function.
2113 (pass_build_cgraph_edges): Replace declaration with that of...
2114 (make_pass_build_cgraph_edges): ...new function.
2115 (pass_local_pure_const): Replace declaration with that of...
2116 (make_pass_local_pure_const): ...new function.
2117 (pass_tracer): Replace declaration with that of...
2118 (make_pass_tracer): ...new function.
2119 (pass_warn_unused_result): Replace declaration with that of...
2120 (make_pass_warn_unused_result): ...new function.
2121 (pass_diagnose_tm_blocks): Replace declaration with that of...
2122 (make_pass_diagnose_tm_blocks): ...new function.
2123 (pass_lower_tm): Replace declaration with that of...
2124 (make_pass_lower_tm): ...new function.
2125 (pass_tm_init): Replace declaration with that of...
2126 (make_pass_tm_init): ...new function.
2127 (pass_tm_mark): Replace declaration with that of...
2128 (make_pass_tm_mark): ...new function.
2129 (pass_tm_memopt): Replace declaration with that of...
2130 (make_pass_tm_memopt): ...new function.
2131 (pass_tm_edges): Replace declaration with that of...
2132 (make_pass_tm_edges): ...new function.
2133 (pass_split_functions): Replace declaration with that of...
2134 (make_pass_split_functions): ...new function.
2135 (pass_feedback_split_functions): Replace declaration with that of...
2136 (make_pass_feedback_split_functions): ...new function.
2137 (pass_strength_reduction): Replace declaration with that of...
2138 (make_pass_strength_reduction): ...new function.
2139 (pass_ipa_lower_emutls): Replace declaration with that of...
2140 (make_pass_ipa_lower_emutls): ...new function.
2141 (pass_ipa_function_and_variable_visibility): Replace declaration with
2143 (make_pass_ipa_function_and_variable_visibility): ...new function.
2144 (pass_ipa_tree_profile): Replace declaration with that of...
2145 (make_pass_ipa_tree_profile): ...new function.
2146 (pass_early_local_passes): Replace declaration with that of...
2147 (make_pass_early_local_passes): ...new function.
2148 (pass_ipa_whole_program_visibility): Replace declaration with that
2150 (make_pass_ipa_whole_program_visibility): ...new function.
2151 (pass_ipa_lto_gimple_out): Replace declaration with that of...
2152 (make_pass_ipa_lto_gimple_out): ...new function.
2153 (pass_ipa_increase_alignment): Replace declaration with that of...
2154 (make_pass_ipa_increase_alignment): ...new function.
2155 (pass_ipa_inline): Replace declaration with that of...
2156 (make_pass_ipa_inline): ...new function.
2157 (pass_ipa_free_lang_data): Replace declaration with that of...
2158 (make_pass_ipa_free_lang_data): ...new function.
2159 (pass_ipa_free_inline_summary): Replace declaration with that of...
2160 (make_pass_ipa_free_inline_summary): ...new function.
2161 (pass_ipa_cp): Replace declaration with that of...
2162 (make_pass_ipa_cp): ...new function.
2163 (pass_ipa_reference): Replace declaration with that of...
2164 (make_pass_ipa_reference): ...new function.
2165 (pass_ipa_pure_const): Replace declaration with that of...
2166 (make_pass_ipa_pure_const): ...new function.
2167 (pass_ipa_pta): Replace declaration with that of...
2168 (make_pass_ipa_pta): ...new function.
2169 (pass_ipa_lto_finish_out): Replace declaration with that of...
2170 (make_pass_ipa_lto_finish_out): ...new function.
2171 (pass_ipa_tm): Replace declaration with that of...
2172 (make_pass_ipa_tm): ...new function.
2173 (pass_ipa_profile): Replace declaration with that of...
2174 (make_pass_ipa_profile): ...new function.
2175 (pass_ipa_cdtor_merge): Replace declaration with that of...
2176 (make_pass_ipa_cdtor_merge): ...new function.
2177 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
2179 (make_pass_cleanup_cfg_post_optimizing): ...new function.
2180 (pass_init_datastructures): Replace declaration with that of...
2181 (make_pass_init_datastructures): ...new function.
2182 (pass_fixup_cfg): Replace declaration with that of...
2183 (make_pass_fixup_cfg): ...new function.
2184 (pass_expand): Replace declaration with that of...
2185 (make_pass_expand): ...new function.
2186 (pass_instantiate_virtual_regs): Replace declaration with that of...
2187 (make_pass_instantiate_virtual_regs): ...new function.
2188 (pass_rtl_fwprop): Replace declaration with that of...
2189 (make_pass_rtl_fwprop): ...new function.
2190 (pass_rtl_fwprop_addr): Replace declaration with that of...
2191 (make_pass_rtl_fwprop_addr): ...new function.
2192 (pass_jump): Replace declaration with that of...
2193 (make_pass_jump): ...new function.
2194 (pass_jump2): Replace declaration with that of...
2195 (make_pass_jump2): ...new function.
2196 (pass_lower_subreg): Replace declaration with that of...
2197 (make_pass_lower_subreg): ...new function.
2198 (pass_cse): Replace declaration with that of...
2199 (make_pass_cse): ...new function.
2200 (pass_fast_rtl_dce): Replace declaration with that of...
2201 (make_pass_fast_rtl_dce): ...new function.
2202 (pass_ud_rtl_dce): Replace declaration with that of...
2203 (make_pass_ud_rtl_dce): ...new function.
2204 (pass_rtl_dce): Replace declaration with that of...
2205 (make_pass_rtl_dce): ...new function.
2206 (pass_rtl_dse1): Replace declaration with that of...
2207 (make_pass_rtl_dse1): ...new function.
2208 (pass_rtl_dse2): Replace declaration with that of...
2209 (make_pass_rtl_dse2): ...new function.
2210 (pass_rtl_dse3): Replace declaration with that of...
2211 (make_pass_rtl_dse3): ...new function.
2212 (pass_rtl_cprop): Replace declaration with that of...
2213 (make_pass_rtl_cprop): ...new function.
2214 (pass_rtl_pre): Replace declaration with that of...
2215 (make_pass_rtl_pre): ...new function.
2216 (pass_rtl_hoist): Replace declaration with that of...
2217 (make_pass_rtl_hoist): ...new function.
2218 (pass_rtl_store_motion): Replace declaration with that of...
2219 (make_pass_rtl_store_motion): ...new function.
2220 (pass_cse_after_global_opts): Replace declaration with that of...
2221 (make_pass_cse_after_global_opts): ...new function.
2222 (pass_rtl_ifcvt): Replace declaration with that of...
2223 (make_pass_rtl_ifcvt): ...new function.
2224 (pass_into_cfg_layout_mode): Replace declaration with that of...
2225 (make_pass_into_cfg_layout_mode): ...new function.
2226 (pass_outof_cfg_layout_mode): Replace declaration with that of...
2227 (make_pass_outof_cfg_layout_mode): ...new function.
2228 (pass_loop2): Replace declaration with that of...
2229 (make_pass_loop2): ...new function.
2230 (pass_rtl_loop_init): Replace declaration with that of...
2231 (make_pass_rtl_loop_init): ...new function.
2232 (pass_rtl_move_loop_invariants): Replace declaration with that of...
2233 (make_pass_rtl_move_loop_invariants): ...new function.
2234 (pass_rtl_unswitch): Replace declaration with that of...
2235 (make_pass_rtl_unswitch): ...new function.
2236 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
2237 (make_pass_rtl_unroll_and_peel_loops): ...new function.
2238 (pass_rtl_doloop): Replace declaration with that of...
2239 (make_pass_rtl_doloop): ...new function.
2240 (pass_rtl_loop_done): Replace declaration with that of...
2241 (make_pass_rtl_loop_done): ...new function.
2242 (pass_web): Replace declaration with that of...
2243 (make_pass_web): ...new function.
2244 (pass_cse2): Replace declaration with that of...
2245 (make_pass_cse2): ...new function.
2246 (pass_df_initialize_opt): Replace declaration with that of...
2247 (make_pass_df_initialize_opt): ...new function.
2248 (pass_df_initialize_no_opt): Replace declaration with that of...
2249 (make_pass_df_initialize_no_opt): ...new function.
2250 (pass_reginfo_init): Replace declaration with that of...
2251 (make_pass_reginfo_init): ...new function.
2252 (pass_inc_dec): Replace declaration with that of...
2253 (make_pass_inc_dec): ...new function.
2254 (pass_stack_ptr_mod): Replace declaration with that of...
2255 (make_pass_stack_ptr_mod): ...new function.
2256 (pass_initialize_regs): Replace declaration with that of...
2257 (make_pass_initialize_regs): ...new function.
2258 (pass_combine): Replace declaration with that of...
2259 (make_pass_combine): ...new function.
2260 (pass_if_after_combine): Replace declaration with that of...
2261 (make_pass_if_after_combine): ...new function.
2262 (pass_ree): Replace declaration with that of...
2263 (make_pass_ree): ...new function.
2264 (pass_partition_blocks): Replace declaration with that of...
2265 (make_pass_partition_blocks): ...new function.
2266 (pass_match_asm_constraints): Replace declaration with that of...
2267 (make_pass_match_asm_constraints): ...new function.
2268 (pass_regmove): Replace declaration with that of...
2269 (make_pass_regmove): ...new function.
2270 (pass_split_all_insns): Replace declaration with that of...
2271 (make_pass_split_all_insns): ...new function.
2272 (pass_fast_rtl_byte_dce): Replace declaration with that of...
2273 (make_pass_fast_rtl_byte_dce): ...new function.
2274 (pass_lower_subreg2): Replace declaration with that of...
2275 (make_pass_lower_subreg2): ...new function.
2276 (pass_mode_switching): Replace declaration with that of...
2277 (make_pass_mode_switching): ...new function.
2278 (pass_sms): Replace declaration with that of...
2279 (make_pass_sms): ...new function.
2280 (pass_sched): Replace declaration with that of...
2281 (make_pass_sched): ...new function.
2282 (pass_ira): Replace declaration with that of...
2283 (make_pass_ira): ...new function.
2284 (pass_reload): Replace declaration with that of...
2285 (make_pass_reload): ...new function.
2286 (pass_clean_state): Replace declaration with that of...
2287 (make_pass_clean_state): ...new function.
2288 (pass_branch_prob): Replace declaration with that of...
2289 (make_pass_branch_prob): ...new function.
2290 (pass_value_profile_transformations): Replace declaration with that
2292 (make_pass_value_profile_transformations): ...new function.
2293 (pass_postreload_cse): Replace declaration with that of...
2294 (make_pass_postreload_cse): ...new function.
2295 (pass_gcse2): Replace declaration with that of...
2296 (make_pass_gcse2): ...new function.
2297 (pass_split_after_reload): Replace declaration with that of...
2298 (make_pass_split_after_reload): ...new function.
2299 (pass_branch_target_load_optimize1): Replace declaration with that
2301 (make_pass_branch_target_load_optimize1): ...new function.
2302 (pass_thread_prologue_and_epilogue): Replace declaration with that
2304 (make_pass_thread_prologue_and_epilogue): ...new function.
2305 (pass_stack_adjustments): Replace declaration with that of...
2306 (make_pass_stack_adjustments): ...new function.
2307 (pass_peephole2): Replace declaration with that of...
2308 (make_pass_peephole2): ...new function.
2309 (pass_if_after_reload): Replace declaration with that of...
2310 (make_pass_if_after_reload): ...new function.
2311 (pass_regrename): Replace declaration with that of...
2312 (make_pass_regrename): ...new function.
2313 (pass_cprop_hardreg): Replace declaration with that of...
2314 (make_pass_cprop_hardreg): ...new function.
2315 (pass_reorder_blocks): Replace declaration with that of...
2316 (make_pass_reorder_blocks): ...new function.
2317 (pass_branch_target_load_optimize2): Replace declaration with that
2319 (make_pass_branch_target_load_optimize2): ...new function.
2320 (pass_leaf_regs): Replace declaration with that of...
2321 (make_pass_leaf_regs): ...new function.
2322 (pass_split_before_sched2): Replace declaration with that of...
2323 (make_pass_split_before_sched2): ...new function.
2324 (pass_compare_elim_after_reload): Replace declaration with that of...
2325 (make_pass_compare_elim_after_reload): ...new function.
2326 (pass_sched2): Replace declaration with that of...
2327 (make_pass_sched2): ...new function.
2328 (pass_stack_regs): Replace declaration with that of...
2329 (make_pass_stack_regs): ...new function.
2330 (pass_stack_regs_run): Replace declaration with that of...
2331 (make_pass_stack_regs_run): ...new function.
2332 (pass_df_finish): Replace declaration with that of...
2333 (make_pass_df_finish): ...new function.
2334 (pass_compute_alignments): Replace declaration with that of...
2335 (make_pass_compute_alignments): ...new function.
2336 (pass_duplicate_computed_gotos): Replace declaration with that of...
2337 (make_pass_duplicate_computed_gotos): ...new function.
2338 (pass_variable_tracking): Replace declaration with that of...
2339 (make_pass_variable_tracking): ...new function.
2340 (pass_free_cfg): Replace declaration with that of...
2341 (make_pass_free_cfg): ...new function.
2342 (pass_machine_reorg): Replace declaration with that of...
2343 (make_pass_machine_reorg): ...new function.
2344 (pass_cleanup_barriers): Replace declaration with that of...
2345 (make_pass_cleanup_barriers): ...new function.
2346 (pass_delay_slots): Replace declaration with that of...
2347 (make_pass_delay_slots): ...new function.
2348 (pass_split_for_shorten_branches): Replace declaration with that of...
2349 (make_pass_split_for_shorten_branches): ...new function.
2350 (pass_split_before_regstack): Replace declaration with that of...
2351 (make_pass_split_before_regstack): ...new function.
2352 (pass_convert_to_eh_region_ranges): Replace declaration with that
2354 (make_pass_convert_to_eh_region_ranges): ...new function.
2355 (pass_shorten_branches): Replace declaration with that of...
2356 (make_pass_shorten_branches): ...new function.
2357 (pass_set_nothrow_function_flags): Replace declaration with that of...
2358 (make_pass_set_nothrow_function_flags): ...new function.
2359 (pass_dwarf2_frame): Replace declaration with that of...
2360 (make_pass_dwarf2_frame): ...new function.
2361 (pass_final): Replace declaration with that of...
2362 (make_pass_final): ...new function.
2363 (pass_rtl_seqabstr): Replace declaration with that of...
2364 (make_pass_rtl_seqabstr): ...new function.
2365 (pass_release_ssa_names): Replace declaration with that of...
2366 (make_pass_release_ssa_names): ...new function.
2367 (pass_early_inline): Replace declaration with that of...
2368 (make_pass_early_inline): ...new function.
2369 (pass_inline_parameters): Replace declaration with that of...
2370 (make_pass_inline_parameters): ...new function.
2371 (pass_update_address_taken): Replace declaration with that of...
2372 (make_pass_update_address_taken): ...new function.
2373 (pass_convert_switch): Replace declaration with that of...
2374 (make_pass_convert_switch): ...new function.
2375 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
2376 to a subclass of simple_ipa_opt_pass along with...
2377 (pass_data_ipa_tree_profile): ...new pass_data instance and...
2378 (make_pass_ipa_tree_profile): ...new function.
2379 * tree-sra.c (pass_sra_early): Convert from a global struct to a
2380 subclass of gimple_opt_pass along with...
2381 (pass_data_sra_early): ...new pass_data instance and...
2382 (make_pass_sra_early): ...new function.
2383 (pass_sra): Convert from a global struct to a subclass of
2384 gimple_opt_pass along with...
2385 (pass_data_sra): ...new pass_data instance and...
2386 (make_pass_sra): ...new function.
2387 (pass_early_ipa_sra): Convert from a global struct to a subclass of
2388 gimple_opt_pass along with...
2389 (pass_data_early_ipa_sra): ...new pass_data instance and...
2390 (make_pass_early_ipa_sra): ...new function.
2391 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
2392 subclass of gimple_opt_pass along with...
2393 (pass_data_ccp): ...new pass_data instance and...
2394 (make_pass_ccp): ...new function.
2395 (pass_fold_builtins): Convert from a global struct to a subclass of
2396 gimple_opt_pass along with...
2397 (pass_data_fold_builtins): ...new pass_data instance and...
2398 (make_pass_fold_builtins): ...new function.
2399 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
2400 subclass of gimple_opt_pass along with...
2401 (pass_data_copy_prop): ...new pass_data instance and...
2402 (make_pass_copy_prop): ...new function.
2403 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
2404 global struct to a subclass of gimple_opt_pass along with...
2405 (pass_data_rename_ssa_copies): ...new pass_data instance and...
2406 (make_pass_rename_ssa_copies): ...new function.
2407 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
2408 subclass of gimple_opt_pass along with...
2409 (pass_data_dce): ...new pass_data instance and...
2410 (make_pass_dce): ...new function.
2411 (pass_dce_loop): Convert from a global struct to a subclass of
2412 gimple_opt_pass along with...
2413 (pass_data_dce_loop): ...new pass_data instance and...
2414 (make_pass_dce_loop): ...new function.
2415 (pass_cd_dce): Convert from a global struct to a subclass of
2416 gimple_opt_pass along with...
2417 (pass_data_cd_dce): ...new pass_data instance and...
2418 (make_pass_cd_dce): ...new function.
2419 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
2420 subclass of gimple_opt_pass along with...
2421 (pass_data_dominator): ...new pass_data instance and...
2422 (make_pass_dominator): ...new function.
2423 (pass_phi_only_cprop): Convert from a global struct to a subclass of
2424 gimple_opt_pass along with...
2425 (pass_data_phi_only_cprop): ...new pass_data instance and...
2426 (make_pass_phi_only_cprop): ...new function.
2427 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
2428 subclass of gimple_opt_pass along with...
2429 (pass_data_dse): ...new pass_data instance and...
2430 (make_pass_dse): ...new function.
2431 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
2432 a subclass of gimple_opt_pass along with...
2433 (pass_data_forwprop): ...new pass_data instance and...
2434 (make_pass_forwprop): ...new function.
2435 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
2436 struct to a subclass of gimple_opt_pass along with...
2437 (pass_data_tree_ifcombine): ...new pass_data instance and...
2438 (make_pass_tree_ifcombine): ...new function.
2439 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
2440 subclass of gimple_opt_pass along with...
2441 (pass_data_ch): ...new pass_data instance and...
2442 (make_pass_ch): ...new function.
2443 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
2444 subclass of gimple_opt_pass along with...
2445 (pass_data_tree_loop): ...new pass_data instance and...
2446 (make_pass_tree_loop): ...new function.
2447 (pass_tree_loop_init): Convert from a global struct to a subclass of
2448 gimple_opt_pass along with...
2449 (pass_data_tree_loop_init): ...new pass_data instance and...
2450 (make_pass_tree_loop_init): ...new function.
2451 (pass_lim): Convert from a global struct to a subclass of
2452 gimple_opt_pass along with...
2453 (pass_data_lim): ...new pass_data instance and...
2454 (make_pass_lim): ...new function.
2455 (pass_tree_unswitch): Convert from a global struct to a subclass of
2456 gimple_opt_pass along with...
2457 (pass_data_tree_unswitch): ...new pass_data instance and...
2458 (make_pass_tree_unswitch): ...new function.
2459 (pass_predcom): Convert from a global struct to a subclass of
2460 gimple_opt_pass along with...
2461 (pass_data_predcom): ...new pass_data instance and...
2462 (make_pass_predcom): ...new function.
2463 (pass_vectorize): Convert from a global struct to a subclass of
2464 gimple_opt_pass along with...
2465 (pass_data_vectorize): ...new pass_data instance and...
2466 (make_pass_vectorize): ...new function.
2467 (pass_graphite): Convert from a global struct to a subclass of
2468 gimple_opt_pass along with...
2469 (pass_data_graphite): ...new pass_data instance and...
2470 (make_pass_graphite): ...new function.
2471 (pass_graphite_transforms): Convert from a global struct to a subclass
2472 of gimple_opt_pass along with...
2473 (pass_data_graphite_transforms): ...new pass_data instance and...
2474 (make_pass_graphite_transforms): ...new function.
2475 (pass_check_data_deps): Convert from a global struct to a subclass of
2476 gimple_opt_pass along with...
2477 (pass_data_check_data_deps): ...new pass_data instance and...
2478 (make_pass_check_data_deps): ...new function.
2479 (pass_iv_canon): Convert from a global struct to a subclass of
2480 gimple_opt_pass along with...
2481 (pass_data_iv_canon): ...new pass_data instance and...
2482 (make_pass_iv_canon): ...new function.
2483 (pass_scev_cprop): Convert from a global struct to a subclass of
2484 gimple_opt_pass along with...
2485 (pass_data_scev_cprop): ...new pass_data instance and...
2486 (make_pass_scev_cprop): ...new function.
2487 (pass_record_bounds): Convert from a global struct to a subclass of
2488 gimple_opt_pass along with...
2489 (pass_data_record_bounds): ...new pass_data instance and...
2490 (make_pass_record_bounds): ...new function.
2491 (pass_complete_unroll): Convert from a global struct to a subclass of
2492 gimple_opt_pass along with...
2493 (pass_data_complete_unroll): ...new pass_data instance and...
2494 (make_pass_complete_unroll): ...new function.
2495 (pass_complete_unrolli): Convert from a global struct to a subclass of
2496 gimple_opt_pass along with...
2497 (pass_data_complete_unrolli): ...new pass_data instance and...
2498 (make_pass_complete_unrolli): ...new function.
2499 (pass_parallelize_loops): Convert from a global struct to a subclass
2500 of gimple_opt_pass along with...
2501 (pass_data_parallelize_loops): ...new pass_data instance and...
2502 (make_pass_parallelize_loops): ...new function.
2503 (pass_loop_prefetch): Convert from a global struct to a subclass of
2504 gimple_opt_pass along with...
2505 (pass_data_loop_prefetch): ...new pass_data instance and...
2506 (make_pass_loop_prefetch): ...new function.
2507 (pass_iv_optimize): Convert from a global struct to a subclass of
2508 gimple_opt_pass along with...
2509 (pass_data_iv_optimize): ...new pass_data instance and...
2510 (make_pass_iv_optimize): ...new function.
2511 (pass_tree_loop_done): Convert from a global struct to a subclass of
2512 gimple_opt_pass along with...
2513 (pass_data_tree_loop_done): ...new pass_data instance and...
2514 (make_pass_tree_loop_done): ...new function.
2515 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
2516 struct to a subclass of gimple_opt_pass along with...
2517 (pass_data_cse_reciprocals): ...new pass_data instance and...
2518 (make_pass_cse_reciprocals): ...new function.
2519 (pass_cse_sincos): Convert from a global struct to a subclass of
2520 gimple_opt_pass along with...
2521 (pass_data_cse_sincos): ...new pass_data instance and...
2522 (make_pass_cse_sincos): ...new function.
2523 (pass_optimize_bswap): Convert from a global struct to a subclass of
2524 gimple_opt_pass along with...
2525 (pass_data_optimize_bswap): ...new pass_data instance and...
2526 (make_pass_optimize_bswap): ...new function.
2527 (pass_optimize_widening_mul): Convert from a global struct to a
2528 subclass of gimple_opt_pass along with...
2529 (pass_data_optimize_widening_mul): ...new pass_data instance and...
2530 (make_pass_optimize_widening_mul): ...new function.
2531 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
2532 subclass of gimple_opt_pass along with...
2533 (pass_data_phiopt): ...new pass_data instance and...
2534 (make_pass_phiopt): ...new function.
2535 (pass_cselim): Convert from a global struct to a subclass of
2536 gimple_opt_pass along with...
2537 (pass_data_cselim): ...new pass_data instance and...
2538 (make_pass_cselim): ...new function.
2539 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
2540 subclass of gimple_opt_pass along with...
2541 (pass_data_phiprop): ...new pass_data instance and...
2542 (make_pass_phiprop): ...new function.
2543 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
2544 subclass of gimple_opt_pass along with...
2545 (pass_data_pre): ...new pass_data instance and...
2546 (make_pass_pre): ...new function.
2547 (pass_fre): Convert from a global struct to a subclass of
2548 gimple_opt_pass along with...
2549 (pass_data_fre): ...new pass_data instance and...
2550 (make_pass_fre): ...new function.
2551 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
2552 subclass of gimple_opt_pass along with...
2553 (pass_data_reassoc): ...new pass_data instance and...
2554 (make_pass_reassoc): ...new function.
2555 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
2556 subclass of gimple_opt_pass along with...
2557 (pass_data_sink_code): ...new pass_data instance and...
2558 (make_pass_sink_code): ...new function.
2559 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
2560 subclass of gimple_opt_pass along with...
2561 (pass_data_strlen): ...new pass_data instance and...
2562 (make_pass_strlen): ...new function.
2563 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
2564 struct to a subclass of gimple_opt_pass along with...
2565 (pass_data_build_alias): ...new pass_data instance and...
2566 (make_pass_build_alias): ...new function.
2567 (pass_build_ealias): Convert from a global struct to a subclass of
2568 gimple_opt_pass along with...
2569 (pass_data_build_ealias): ...new pass_data instance and...
2570 (make_pass_build_ealias): ...new function.
2571 (pass_ipa_pta): Convert from a global struct to a subclass of
2572 simple_ipa_opt_pass along with...
2573 (pass_data_ipa_pta): ...new pass_data instance and...
2574 (make_pass_ipa_pta): ...new function.
2575 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
2576 subclass of gimple_opt_pass along with...
2577 (pass_data_uncprop): ...new pass_data instance and...
2578 (make_pass_uncprop): ...new function.
2579 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
2580 global struct to a subclass of gimple_opt_pass along with...
2581 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
2582 (make_pass_late_warn_uninitialized): ...new function.
2583 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
2584 to a subclass of gimple_opt_pass along with...
2585 (pass_data_init_datastructures): ...new pass_data instance and...
2586 (make_pass_init_datastructures): ...new function.
2587 (pass_early_warn_uninitialized): Convert from a global struct to a
2588 subclass of gimple_opt_pass along with...
2589 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
2590 (make_pass_early_warn_uninitialized): ...new function.
2591 (pass_update_address_taken): Convert from a global struct to a
2592 subclass of gimple_opt_pass along with...
2593 (pass_data_update_address_taken): ...new pass_data instance and...
2594 (make_pass_update_address_taken): ...new function.
2595 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
2596 struct to a subclass of gimple_opt_pass along with...
2597 (pass_data_release_ssa_names): ...new pass_data instance and...
2598 (make_pass_release_ssa_names): ...new function.
2599 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
2600 subclass of gimple_opt_pass along with...
2601 (pass_data_stdarg): ...new pass_data instance and...
2602 (make_pass_stdarg): ...new function.
2603 * tree-switch-conversion.c (pass_convert_switch): Convert from a
2604 global struct to a subclass of gimple_opt_pass along with...
2605 (pass_data_convert_switch): ...new pass_data instance and...
2606 (make_pass_convert_switch): ...new function.
2607 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
2608 to a subclass of gimple_opt_pass along with...
2609 (pass_data_tail_recursion): ...new pass_data instance and...
2610 (make_pass_tail_recursion): ...new function.
2611 (pass_tail_calls): Convert from a global struct to a subclass of
2612 gimple_opt_pass along with...
2613 (pass_data_tail_calls): ...new pass_data instance and...
2614 (make_pass_tail_calls): ...new function.
2615 * tree-vect-generic.c (pass_lower_vector): Convert from a global
2616 struct to a subclass of gimple_opt_pass along with...
2617 (pass_data_lower_vector): ...new pass_data instance and...
2618 (make_pass_lower_vector): ...new function.
2619 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
2620 gimple_opt_pass along with...
2621 (pass_data_lower_vector_ssa): ...new pass_data instance and...
2622 (make_pass_lower_vector_ssa): ...new function.
2623 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
2624 to a subclass of gimple_opt_pass along with...
2625 (pass_data_slp_vectorize): ...new pass_data instance and...
2626 (make_pass_slp_vectorize): ...new function.
2627 (pass_ipa_increase_alignment): Convert from a global struct to a
2628 subclass of simple_ipa_opt_pass along with...
2629 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
2630 (make_pass_ipa_increase_alignment): ...new function.
2631 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
2632 gimple_opt_pass along with...
2633 (pass_data_vrp): ...new pass_data instance and...
2634 (make_pass_vrp): ...new function.
2635 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
2636 subclass of simple_ipa_opt_pass along with...
2637 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
2638 (make_pass_ipa_free_lang_data): ...new function.
2639 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
2640 gimple_opt_pass along with...
2641 (pass_data_tsan): ...new pass_data instance and...
2642 (make_pass_tsan): ...new function.
2643 (pass_tsan_O0): Convert from a global struct to a subclass of
2644 gimple_opt_pass along with...
2645 (pass_data_tsan_O0): ...new pass_data instance and...
2646 (make_pass_tsan_O0): ...new function.
2647 * var-tracking.c (pass_variable_tracking): Convert from a global
2648 struct to a subclass of rtl_opt_pass along with...
2649 (pass_data_variable_tracking): ...new pass_data instance and...
2650 (make_pass_variable_tracking): ...new function.
2651 * web.c (pass_web): Convert from a global struct to a subclass of
2652 rtl_opt_pass along with...
2653 (pass_data_web): ...new pass_data instance and...
2654 (make_pass_web): ...new function.
2655 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
2656 declaration with that of...
2657 (make_pass_mode_switch_use): ...new function.
2658 (pass_resolve_sw_modes): Replace declaration with that of...
2659 (make_pass_resolve_sw_modes): ...new function.
2660 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
2661 from a global struct to a subclass of rtl_opt_pass along with...
2662 (pass_data_mode_switch_use): ...new pass_data instance and...
2663 (make_pass_mode_switch_use): ...new function.
2664 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
2665 from a global struct to a subclass of rtl_opt_pass along with...
2666 (pass_data_resolve_sw_modes): ...new pass_data instance and...
2667 (make_pass_resolve_sw_modes): ...new function.
2668 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
2669 struct to a subclass of rtl_opt_pass along with...
2670 (pass_data_insert_vzeroupper): ...new pass_data instance and...
2671 (make_pass_insert_vzeroupper): ...new function.
2672 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
2673 global struct to a subclass of rtl_opt_pass along with...
2674 (pass_data_work_around_errata): ...new pass_data instance and...
2675 (make_pass_work_around_errata): ...new function.
2676 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
2677 struct to a subclass of rtl_opt_pass along with...
2678 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
2679 (make_pass_mips_machine_reorg2): ...new function.
2681 2013-08-05 David Malcolm <dmalcolm@redhat.com>
2683 * passes.c (pass_manager::operator new): New.
2685 2013-08-05 David Malcolm <dmalcolm@redhat.com>
2687 Handwritten part of conversion of passes to C++ classes.
2689 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
2690 (toplev.o): Add dep on PASS_MANAGER_H.
2691 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
2692 of early local pases to reflect this moving from a global to a
2693 member of gcc::pass_manager.
2694 (cgraph_add_new_function): Likewise.
2695 * lto-cgraph.c (lto_output_node): Update for conversion of
2696 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
2697 * passes.c (opt_pass::clone): New.
2698 (opt_pass::gate): New.
2699 (opt_pass::execute): New.
2700 (opt_pass::opt_pass): New.
2701 (pass_manager::execute_early_local_passes): New.
2702 (pass_manager::execute_pass_mode_switching): new.
2703 (finish_optimization_passes): Convert to...
2704 (pass_manager::finish_optimization_passes): ...this.
2705 (finish_optimization_passes): Update for conversion of passes to
2707 (register_dump_files_1): Use has_gate since we cannot portably
2708 check a vtable entry against NULL.
2709 (dump_one_pass): Likewise.
2710 (ipa_write_summaries_2): Likewise.
2711 (ipa_write_optimization_summaries_1): Likewise.
2712 (ipa_read_summaries_1): Likewise.
2713 (ipa_read_optimization_summaries_1): Likewise.
2714 (execute_ipa_stmt_fixups): Likewise.
2715 (pass_manager::pass_manager): Rewrite pass-creation, invoking
2716 pass-creation functions rather than wiring up globals, and
2717 storing the results in fields of pass_manager generated using
2719 (pass_manager::dump_profile_report): Update for conversion of
2720 passes to C++ classes.
2721 (pass_manager::execute_ipa_summary_passes): Likewise.
2722 (execute_one_ipa_transform_pass): Likewise.
2723 (execute_one_pass): Use has_gate and has_execute since we cannot
2724 portably check a vtable entry against NULL.
2725 * pass_manager.h (pass_manager::finish_optimization_passes): New.
2726 (pass_manager): Use pass-instances.def to add fields for the
2727 various pass instances.
2728 * toplev.c (finalize): Update for move of
2729 finish_optimization_passes to a method of gcc::pass_manager.
2730 * toplev.h (finish_optimization_passes): Move to method of class
2732 * tree-pass.h (struct pass_data): New.
2733 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
2734 (opt_pass::gate): Convert to virtual function.
2735 (opt_pass::~opt_pass): New.
2736 (opt_pass::clone): New.
2737 (opt_pass::execute): Convert to virtual function.
2738 (opt_pass::opt_pass): New.
2739 (opt_pass::ctxt_): new.
2740 (gimple_opt_pass): Convert to subclass of opt_pass.
2741 (gimple_opt_pass::gimple_opt_pass): New.
2742 (rtl_opt_pass): Convert to subclass of opt_pass.
2743 (rtl_opt_pass::rtl_opt_pass): New.
2744 (ipa_opt_pass_d): Convert to subclass of opt_pass.
2745 (ipa_opt_pass_d::ipa_opt_pass_d): New.
2746 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
2747 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
2748 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
2749 invocation of pass_mode_switching to reflect this moving from a
2750 global to a member of gcc::pass_manager.
2751 (ix86_option_override): Rework how pass_insert_vzeroupper is
2752 added to the pass_manager to reflect autogenerated changes.
2753 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
2755 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
2757 PR rtl-optimization/57708
2758 * recog.c (peep2_find_free_register): Validate all regs in a
2761 2013-08-05 Jan Hubicka <jh@suse.cz>
2764 * cgraph.c (verify_cgraph_node): Accept local flags from other
2766 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
2767 (function_and_variable_visibility): Likewise.
2768 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
2770 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
2772 * graph.c (init_graph_slim_pretty_print): Remove.
2773 (print_graph_cfg): Do not call it. Use local pretty printer.
2774 (start_graph_dump): Likewise.
2776 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
2778 * gimple-pretty-print.c (buffer): Remove.
2779 (initialized): Likewise.
2780 (maybe_init_pretty_print): Likewise.
2781 (print_gimple_stmt): Do not call it. Use non-static local
2782 pretty_printer variable.
2783 (print_gimple_expr): Likewise.
2784 (print_gimple_seq): Likewise.
2785 (gimple_dump_bb): Likewise.
2787 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
2789 * asan.c (asan_pp): Remove.
2790 (asan_pp_initialized): Likewise.
2791 (asan_pp_initialize): Likewise.
2792 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
2793 (asan_emit_stack_protection): Tidy. Use local pretty printer.
2794 (asan_add_global): Likewise.
2796 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
2798 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
2799 * diagnostic.h (diagnostic_flush_buffer): Adjust.
2800 * pretty-print.c (pp_formatted_text_data): Likewise.
2801 (pp_indent): Rename from pp_base_indent.
2802 (pp_format): Rename from pp_base_format.
2803 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
2804 (pp_format_verbatim): Rename from pp_base_format_verbatim.
2805 (pp_flush): Rename from pp_base_flush.
2806 (pp_set_line_maximum_length): Rename from
2807 pp_base_set_line_maximum_length.
2808 (pp_clear_output_area): Rename from pp_base_clear_output_area.
2809 (pp_set_prefix): Rename from pp_base_set_prefix.
2810 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
2811 (pp_emit_prefix): Rename from pp_base_emit_prefix.
2812 (pp_append_text): Rename from pp_base_append_text.
2813 (pp_formatted_text): Rename from pp_base_formatted_text.
2814 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
2815 (pp_remaining_character_count_for_line): Rename from
2816 pp_base_remaining_character_count_for_line.
2817 (pp_newline): Rename from pp_base_newline.
2818 (pp_character): Rename from pp_base_character.
2819 (pp_string): Rename from pp_base_string.
2820 (pp_maybe_space): Rename from pp_base_maybe_space.
2821 * asan.c (asan_pp_string): Adjust.
2822 (asan_emit_stack_protection): Likewise.
2823 (asan_add_global): Likewise.
2824 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
2825 * tree-mudflap.c (mf_varname_tree): Likewise.
2826 * tree-pretty-print.c (pp_tree_identifier): Rename from
2827 pp_base_tree_identifier.
2828 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
2829 Declare as function.
2831 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
2833 * pretty-print.h (pp_bar_bar): New.
2834 (pp_ampersand_ampersand): Likewise.
2835 (pp_less_equal): Likewise.
2836 (pp_greater_equal): Likewise.
2837 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
2838 printer functions instead of pp_string or operators and punctuators.
2839 (dump_gimple_call): Likewise.
2840 (dump_gimple_omp_for): Likewise.
2841 (dump_gimple_transaction): Likewise.
2842 (dump_gimple_phi): Likewise.
2843 (pp_gimple_stmt_1): Likewise.
2844 * sched-vis.c (print_insn): Likewise.
2845 * tree-mudflap.c (mf_varname_tree): Likewise.
2846 * tree-pretty-print.c (dump_block_node): Likewise.
2847 (dump_generic_node): Likewise.
2849 2013-08-02 Jan Hubicka <jh@suse.cz>
2851 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
2853 * lto-streamer-out.c (tree_is_indexable): Results decls and
2854 parm decls are not indexable.
2855 (DFS_write_tree_body): Do not follow args and results.
2856 (hash_tree): Likewise.
2857 (output_functions): Rearrange so struct function is needed
2858 only when real body is output; be able to also ouptut abstract
2859 functions; output DECL_ARGUMENTS and DECL_RESULT.
2860 (lto_output): When not in WPA, ale store abstract functions.
2861 (write_symbol): Do not care about RESULT_DECL.
2862 (output_symbol_p): Handle correctly sbtract decls.
2863 * lto-streamer-in.c (input_function): Rearrange so struct
2864 function can be NULL at entry; allow streaming of
2865 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
2866 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
2867 sanity check during LTO.
2868 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
2869 RESULT_DECl and DECL_ARGUMENTS.
2870 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
2873 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
2875 * pretty-print.h (pp_underscore): New.
2877 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
2878 printer functions instead of pp_character.
2879 (dump_binary_rhs): Likewise.
2880 (dump_ternary_rhs): Likewise.
2881 (dump_gimple_call_args): Likewise.
2882 (pp_points_to_solution): Likewise.
2883 (dump_gimple_call): Likewise.
2884 (dump_gimple_switch): Likewise.
2885 (dump_gimple_cond): Likewise.
2886 (dump_gimple_bind): Likewise.
2887 (dump_gimple_try): Likewise.
2888 (dump_gimple_omp_for): Likewise.
2889 (dump_gimple_omp_continue): Likewise.
2890 (dump_gimple_omp_single): Likewise.
2891 (dump_gimple_omp_sections): Likewise.
2892 (dump_gimple_omp_block): Likewise.
2893 (dump_gimple_omp_critical): Likewise.
2894 (dump_gimple_transaction): Likewise.
2895 (dump_gimple_asm): Likewise.
2896 (dump_gimple_phi): Likewise.
2897 (dump_gimple_omp_parallel): Likewise.
2898 (dump_gimple_omp_task): Likewise.
2899 (dump_gimple_omp_atomic_load): Likewise.
2900 (dump_gimple_omp_atomic_store): Likewise.
2901 (dump_gimple_mem_ops): Likewise.
2902 (pp_gimple_stmt_1): Likewise.
2903 (pp_cfg_jump): Likewise.
2904 (dump_implicit_edges): Likewise.
2905 (gimple_dump_bb_for_graph): Likewise.
2906 * graph.c (draw_cfg_node): Likewise.
2907 * langhooks.c (lhd_print_error_function): Likewise.
2908 * sched-vis.c (print_exp): Likewise.
2909 (print_value): Likewise.
2910 (print_pattern): Likewise.
2911 (print_insn): Likewise.
2912 (rtl_dump_bb_for_graph): Likewise.
2913 * tree-pretty-print.c (dump_function_declaration): Likewise.
2914 (dump_array_domain): Likewise.
2915 (dump_omp_clause): Likewise.
2916 (dump_location): Likewise.
2917 (dump_generic_node): Likewise.
2918 (print_struct_decl): Likewise.
2919 * diagnostic.c (diagnostic_show_locus): Use pp_space.
2921 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
2923 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
2924 candidate table when replacing a candidate statement.
2925 (replace_rhs_if_not_dup): Likewise.
2926 (replace_one_candidate): Likewise.
2928 2013-08-02 Jan Hubicka <jh@suse.cz>
2929 Martin Liska <marxin.liska@gmail.com>
2931 * cgraphunit.c (add_new_function): Fix logic when adding from
2933 (assemble_thunk): Rename to ...
2934 (expand_thunk); .. this one; export; get it working with
2935 general functions; make produced gimple valid.
2936 * cgraph.h (expand_thunk): Declare.
2938 2013-08-02 Jan Hubicka <jh@suse.cz>
2940 * ipa-cp.c (gather_context_independent_values): Use
2941 ipa_get_param_move_cost.
2942 (get_replacement_map): Remove PARAM; move parameter folding
2944 (create_specialized_node): Update.
2945 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
2946 assert that we have gimple body; update move_cost.
2947 (count_formal_params): Assert that we have gimple body.
2948 (ipa_dump_param): New function.
2949 (ipa_alloc_node_params): Break out from ...
2950 (ipa_initialize_node_params): ... here.
2951 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
2952 (ipa_write_node_info): Stream move costs.
2953 (ipa_read_node_info): Read move costs.
2954 (ipa_update_after_lto_read): Do not recompute node params.
2955 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
2956 (ipa_get_param): Check we are not in WPA.
2957 (ipa_get_param_move_cost): New.
2958 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
2959 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
2960 parm numbers to be present.
2962 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
2964 PR rtl-optimization/58048
2965 * lra-constraints.c (process_alt_operands): Don't check asm
2966 operand on register.
2968 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
2970 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
2971 the implied StoreLoad barrier for atomic operations if before.
2973 2013-08-02 Jan Hubicka <jh@suse.cz>
2974 Martin Liska <marxin.liska@gmail.com>
2976 * cgraph.c (cgraph_function_body_availability): Do not check
2978 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
2979 symtab_node_availability): Declare.
2980 * ipa.c (can_replace_by_local_alias): New.
2981 (function_and_variable_visibility): Use it.
2982 * symtab.c (symtab_for_node_and_aliases,
2983 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
2985 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
2987 PR rtl-optimization/57963
2988 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
2989 (lra_constraints): Use them.
2991 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
2993 * config/arm/types.md (define_attr "type"): Add "load_acq"
2995 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
2997 (cortex_a53_store1): Likewise.
2999 2013-08-01 Jan Hubicka <jh@suse.cz>
3001 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
3002 partitions are not needed.
3004 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
3006 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
3007 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
3008 MAYBE_NON_Q_CLASS_P where appropriate.
3010 2013-08-01 Jan Hubicka <jh@suse.cz>
3012 * cgraph.h (release_function_body): Declare.
3013 * tree.c (free_lang_data_in_decl): Free, parameters and return values
3014 of unused delcarations.
3016 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3018 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
3019 RTL form when subtracting a constant.
3021 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3023 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
3024 Generate canonical plus rtx with negated immediate instead of minus
3026 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
3028 2013-08-01 Jan Hubicka <jh@suse.cz>
3030 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
3031 (cgraph_release_function_body): Likewise.
3032 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
3033 * cgraph.h (cgrpah_node): Rename abstract_and_needed
3034 to used_as_abstract_origin.
3035 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
3036 symbols used as abstract origins.
3037 * cgraphunit.c (analyze_functions): Update.
3038 * ipa.c (symtab_remove_unreachable_nodes): Recompute
3039 used_as_abstract_origin.
3040 * tree-inline.c (tree_function_versioning): Update
3041 used_as_abstract_origin; be ready for DECL_RESULT and
3042 DECL_ARGUMENTS to be NULL.
3044 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
3045 for abstract functions.
3046 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
3049 2013-08-01 Jan Hubicka <jh@suse.cz>
3051 * profile.c (compute_value_histograms): Fix thinko.
3053 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
3055 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
3056 aarch-common-protos.h to extra_headers.
3057 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
3058 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
3059 * config/aarch64/t-aarch64 (aarch-common.o): Define.
3061 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
3063 * config/aarch64/aarch64.md (define_attr "type"): Delete.
3064 Include "../arm/types.md". Define "type" attribute for all patterns.
3065 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
3068 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
3070 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
3071 to support power8 load fusion.
3072 (fusion_gpr_mem_load): Likewise.
3074 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
3076 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
3077 declarations for power8 load fusion.
3078 (emit_fusion_gpr_load): Likewise.
3080 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
3081 tuning for power8, turn on fusion mode by default. Turn on sign
3082 extending fusion mode if normal fusion mode is on, and we are at
3084 (fusion_gpr_load_p): New function, return true if we can fuse an
3085 addis instruction with a dependent load to a GPR.
3086 (emit_fusion_gpr_load): Emit the instructions for power8 load
3089 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
3090 (VSX load fusion peepholes): New peepholes to fuse together an
3091 addi instruction with a VSX load instruction.
3093 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
3094 peepholes to fuse an addis instruction with a load to a GPR base
3095 register. If we are supporting sign extending fusions, convert
3096 sign extending loads to zero extending loads and add an explicit
3099 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
3101 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
3102 aarch-common-protos.h to extra_headers.
3103 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
3104 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
3105 (arm_early_store_addr_dep): Likewise.
3106 (arm_no_early_alu_shift_dep): Likewise.
3107 (arm_no_early_alu_shift_value_dep): Likewise.
3108 (arm_no_early_mul_dep): Likewise.
3109 (arm_no_early_store_addr_dep): Likewise.
3110 (arm_mac_accumulator_is_mul_result): Likewise.
3111 (arm_mac_accumulator_is_result): Likewise.
3112 * config/arm/aarch-common.c: ... here. New file.
3113 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
3115 (arm_early_store_addr_dep): Likewise.
3116 (arm_no_early_alu_shift_dep): Likewise.
3117 (arm_no_early_alu_shift_value_dep): Likewise.
3118 (arm_no_early_mul_dep): Likewise.
3119 (arm_no_early_store_addr_dep): Likewise.
3120 (arm_mac_accumulator_is_mul_result): Likewise.
3121 (arm_mac_accumulator_is_result): Likewise.
3122 * config/arm/aarch-common-protos.h: ... here. New file.
3123 * config/arm/t-arm (aarch-common.o): Define.
3125 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
3127 * config/arm/arm.md: Include new file "types.md".
3128 (define_attr "type"): Move from here to ...
3129 (define_attr "mul32"): Likewise.
3130 (define_attr "mul64"): Likewise.
3131 * config/arm/types.md: ... here. New file.
3133 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
3135 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
3136 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
3138 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3140 * gen-pass-instances.awk: Fix offset of substr().
3142 2013-07-31 David Malcolm <dmalcolm@redhat.com>
3144 * Makefile.in (pass-instances.def): New.
3145 (passes.o): Replace dependency on passes.def with one on
3148 * gen-pass-instances.awk: New.
3150 * passes.c (pass_manager::pass_manager): Use pass-instances.def
3151 rather than passes.def, updating local definition of NEXT_PASS
3152 macro to add an extra NUM parameter (currently unused).
3154 2013-07-30 David Malcolm <dmalcolm@redhat.com>
3156 * Makefile.in (PASS_MANAGER_H): New.
3157 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
3158 (passes.o): Likewise.
3159 (statistics.o): Likewise.
3160 (cgraphunit.o): Likewise.
3161 (context.o): Depend on PASS_MANAGER_H.
3163 * pass_manager.h: New.
3165 * cgraphunit.c (cgraph_add_new_function): Update for moves
3166 of globals to fields of pass_manager.
3167 (analyze_function): Likewise.
3168 (expand_function): Likewise.
3169 (ipa_passes): Likewise.
3170 (compile): Likewise.
3172 * context.c (context::context): New.
3173 * context.h (context::context): New.
3174 (context::get_passes): New.
3175 (context::passes_): New.
3177 * lto-cgraph.c (input_node): Update for moves of globals to
3178 fields of pass_manager.
3180 * passes.c (all_passes): Remove, in favor of a field of the
3181 same name within the new class pass_manager.
3182 (all_small_ipa_passes): Likewise.
3183 (all_lowering_passes): Likewise.
3184 (all_regular_ipa_passes): Likewise.
3185 (all_late_ipa_passes): Likewise.
3186 (all_lto_gen_passes): Likewise.
3187 (passes_by_id): Likewise.
3188 (passes_by_id_size): Likewise.
3189 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
3190 the new class pass_manager.
3191 (set_pass_for_id): Convert to...
3192 (pass_manager::set_pass_for_id): ...method.
3193 (get_pass_for_id): Convert to...
3194 (pass_manager::get_pass_for_id): ...method.
3195 (register_one_dump_file): Move body of implementation into...
3196 (pass_manager::register_one_dump_file): ...here.
3197 (register_dump_files_1): Convert to...
3198 (pass_manager::register_dump_files_1): ...method.
3199 (register_dump_files): Convert to...
3200 (pass_manager::register_dump_files): ...method.
3201 (create_pass_tab): Update for moves of globals to fields of
3203 (dump_passes): Move body of implementation into...
3204 (pass_manager::dump_passes): ...here.
3205 (register_pass): Move body of implementation into...
3206 (pass_manager::register_pass): ...here.
3207 (init_optimization_passes): Convert into...
3208 (pass_manager::pass_manager): ...constructor for new
3209 pass_manager class, and initialize the pass_lists array.
3210 (check_profile_consistency): Update for moves of globals to
3211 fields of pass_manager.
3212 (dump_profile_report): Move body of implementation into...
3213 (pass_manager::dump_profile_report): ...here.
3214 (ipa_write_summaries_1): Update for moves of pass lists from
3215 being globals to fields of pass_manager.
3216 (ipa_write_optimization_summaries): Likewise.
3217 (ipa_read_summaries): Likewise.
3218 (ipa_read_optimization_summaries): Likewise.
3219 (execute_all_ipa_stmt_fixups): Likewise.
3221 * statistics.c (statistics_fini): Update for moves of globals to
3222 fields of pass_manager.
3224 * toplev.c (general_init): Replace call to
3225 init_optimization_passes with construction of the pass_manager
3228 * tree-pass.h (all_passes): Remove, in favor of a field of the
3229 same name within the new class pass_manager.
3230 (all_small_ipa_passes): Likewise.
3231 (all_lowering_passes): Likewise.
3232 (all_regular_ipa_passes): Likewise.
3233 (all_lto_gen_passes): Likewise.
3234 (all_late_ipa_passes): Likewise.
3235 (passes_by_id): Likewise.
3236 (passes_by_id_size): Likewise.
3237 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
3238 the new class pass_manager.
3239 (get_pass_for_id): Remove.
3241 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
3243 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
3246 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
3248 * arm.md (mulhi3): New expand pattern.
3250 2013-07-30 Jan Hubicka <jh@suse.cz>
3251 Martin Liska <marxin.liska@gmail.com>
3253 * profile.c (compute_value_histograms): Do not ICE when
3254 there is mismatch only on some counters.
3256 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3258 PR rtl-optimization/57637
3259 * function.c (move_insn_for_shrink_wrap): Also check the
3260 GEN set of the LIVE problem for the liveness analysis
3261 if it exists, otherwise give up.
3263 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
3265 PR tree-optimization/57993
3266 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
3267 replaced statement in the candidate table.
3268 (phi_add_costs): Return infinite cost when the hidden basis does
3269 not dominate all phis on which the candidate is dependent.
3270 (replace_one_candidate): Record replaced statement in the
3273 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
3275 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
3276 (ashlv2si3): New expander.
3277 (*ashlv2si3_i): New define_insn_and_split.
3278 * predicates.md (float_operation): Allow patterns with three
3281 PR rtl-optimization/58021
3282 * mode-switching.c (create_pre_exit): Always split off preceding
3283 insns if we are not at the basic block head.
3285 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
3287 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
3288 (UCLIBC_DYNAMIC_LINKER): New macro.
3289 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
3291 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
3292 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
3294 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
3295 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
3296 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
3297 for SF and DF modes. Use ieee_quad_format for TF mode.
3298 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
3299 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
3300 (mips_option_override): Handle `-mnan=legacy'.
3301 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
3302 `-mabs=2008' and `-mnan=2008'.
3303 (OPTION_DEFAULT_SPECS): Add "nan" default.
3304 (ASM_SPEC): Handle `-mnan='.
3305 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
3306 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
3307 comment accordingly.
3308 (neg<mode>2): Likewise.
3309 * config/mips/mips.opt (mabs, mnan): New options.
3310 * doc/install.texi (Configuration): Document `--with-nan=' option.
3311 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
3313 (MIPS Options): Document them.
3314 * config.gcc <mips*-*-*>: Handle `--with-nan='.
3315 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
3316 * configure: Regenerate.
3317 * config.in: Regenerate.
3319 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
3321 * config/i386/i386.md (float post-reload splitters): Do not check
3322 for subregs of SSE registers.
3324 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
3325 H.J. Lu <hongjiu.lu@intel.com>
3329 * config/i386/i386.md (post-reload splitter
3330 to avoid partial SSE reg dependency stalls): New pattern.
3332 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
3334 * config/s390/s390.md ("movcc"): Swap load and store instructions.
3336 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
3338 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
3339 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
3341 2013-07-26 Cary Coutant <ccoutant@google.com>
3343 * dwarf2out.c (die_checksum_ordered): Don't include template
3344 instantiations in signature.
3345 (is_template_parameter): New function.
3346 (is_template_instantiation): New function.
3347 (generate_skeleton_bottom_up): Don't include template instantiations
3349 (generate_skeleton): Likewise.
3350 (break_out_comdat_types): Move recursive call to break out nested
3352 (prune_unused_types_mark_generic_parms_dies): Call
3353 is_template_parameter.
3355 2013-07-26 Ian Bolton <ian.bolton@arm.com>
3357 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
3358 uses vector registers.
3359 * config/aarch64/iterators.md: Add attributes rtn and vas.
3361 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3362 Richard Earnshaw <richard.earnshaw@arm.com>
3364 * combine.c (simplify_comparison): Re-canonicalize operands
3366 * config/arm/arm.md (movcond_addsi): New splitter.
3368 2013-07-25 Sterling Augustine <saugustine@google.com>
3370 * dwarf2out.c (size_of_pubnames): Move code to...
3371 (include_pubname_in_output): ...here. New.
3372 (want_pubnames): Rearrange.
3373 (output_pubnames): Call include_pubname_in_output. Move assertion.
3375 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
3377 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
3379 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
3382 * doc/extend.texi: Remove obsolete builtins. Fix
3383 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
3385 2013-07-25 Jan Hubicka <jh@suse.cz>
3387 * cgraph.c (release_function_body): Break out from ...
3388 (cgraph_release_function_body): ... this one; also release DECL_RESULT
3390 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
3391 old_tree in the map.
3392 (create_specialized_node): Update.
3393 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
3395 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
3396 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
3397 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
3398 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
3401 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3403 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
3404 addsi3_carryin_alt2_<optab>): Correct output template.
3406 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3408 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
3409 Adjust for arm_restrict_it.
3410 Remove trailing whitespace.
3412 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
3414 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
3415 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
3417 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
3419 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
3421 PR rtl-optimization/57960
3422 * lra-constraints.c (process_alt_operands): Use the right mode
3423 when checking strict_low.
3425 2013-07-25 Jan Hubicka <jh@suse.cz>
3427 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
3428 * cgraph.c (cgraph_remove_node): Do not release function body
3429 when in cgraph streaming.
3430 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
3431 in other partitions are not considered reachable; fix handling of
3434 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3436 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
3438 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3443 * config/arm/arm.md ("*sibcall_insn): Replace use of
3444 Ss with US. Adjust output for v5 and v4t.
3445 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
3447 * config/arm/constraints.md ("Ss"): Rename to US.
3449 2013-07-25 Terry Guo <terry.guo@arm.com>
3451 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
3452 shift_add/shift_sub0/shift_sub1 RTXs.
3454 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
3455 Anton Blanchard <anton@au1.ibm.com>
3457 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
3458 (altivec_vpks<VI_char>ss): Likewise.
3459 (altivec_vpks<VI_char>us): Likewise.
3460 (altivec_vpku<VI_char>us): Likewise.
3461 (altivec_vpku<VI_char>um): Likewise.
3463 2013-07-24 David Malcolm <dmalcolm@redhat.com>
3465 Introduce context class.
3467 * Makefile.in (CONTEXT_H): New.
3468 (OBJS): Add context.o.
3469 (toplev.o): Add CONTEXT_H to dependencies.
3472 * toplev.c (general_init): Create the singleton gcc::context instance.
3478 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
3480 PR rtl-optimization/57968
3481 * mode-switching.c (create_pre_exit): Allow instructions that
3482 don't set a return register to need a non-exit mode.
3484 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
3485 Anton Blanchard <anton@au1.ibm.com>
3487 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
3488 operands to vperm for little endian.
3489 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
3490 of lvsl to create the control mask for a vperm for little endian.
3492 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3493 Anton Blanchard <anton@au1.ibm.com>
3495 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
3496 two operands for little-endian.
3498 2013-07-23 Steve Ellcey <sellcey@mips.com>
3500 * config/mips/mips.c (mips_case_values_threshold): New.
3501 (TARGET_CASE_VALUES_THRESHOLD): Define.
3503 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3504 Anton Blanchard <anton@au1.ibm.com>
3506 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
3507 selection of field for vector splat in little endian mode.
3509 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
3511 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
3512 expanders to rs6000.md.
3513 (ior<mode>3): Likewise.
3514 (and<mode>3): Likewise.
3515 (one_cmpl<mode>2): Likewise.
3516 (nor<mode>3): Likewise.
3517 (andc<mode>3): Likewise.
3518 (eqv<mode>3): Likewise.
3519 (nand<mode>3): Likewise.
3520 (orc<mode>3): Likewise.
3522 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
3525 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
3526 to split multi-word logical operations.
3527 (rs6000_split_logical_di): Likewise.
3528 (rs6000_split_logical): Likewise.
3530 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
3531 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
3532 and allow TImode operations in 32-bit.
3533 (vsx_and<mode>3_64bit): Likewise.
3534 (vsx_ior<mode>3_32bit): Likewise.
3535 (vsx_ior<mode>3_64bit): Likewise.
3536 (vsx_xor<mode>3_32bit): Likewise.
3537 (vsx_xor<mode>3_64bit): Likewise.
3538 (vsx_one_cmpl<mode>2_32bit): Likewise.
3539 (vsx_one_cmpl<mode>2_64bit): Likewise.
3540 (vsx_nor<mode>3_32bit): Likewise.
3541 (vsx_nor<mode>3_64bit): Likewise.
3542 (vsx_andc<mode>3_32bit): Likewise.
3543 (vsx_andc<mode>3_64bit): Likewise.
3544 (vsx_eqv<mode>3_32bit): Likewise.
3545 (vsx_eqv<mode>3_64bit): Likewise.
3546 (vsx_nand<mode>3_32bit): Likewise.
3547 (vsx_nand<mode>3_64bit): Likewise.
3548 (vsx_orc<mode>3_32bit): Likewise.
3549 (vsx_orc<mode>3_64bit): Likewise.
3551 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
3552 logical types in GPRs.
3554 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
3555 logical insns to rs6000.md, and allow TImode operations in
3557 (altivec_ior<mode>3): Likewise.
3558 (altivec_xor<mode>3): Likewise.
3559 (altivec_one_cmpl<mode>2): Likewise.
3560 (altivec_nor<mode>3): Likewise.
3561 (altivec_andc<mode>3): Likewise.
3563 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
3564 attributes for moving the 128-bit logical operations into
3566 (BOOL_REGS_OUTPUT): Likewise.
3567 (BOOL_REGS_OP1): Likewise.
3568 (BOOL_REGS_OP2): Likewise.
3569 (BOOL_REGS_UNARY): Likewise.
3570 (BOOL_REGS_AND_CR0): Likewise.
3571 (one_cmpl<mode>2): Add support for DI logical operations on
3572 32-bit, splitting the operations to 32-bit.
3576 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
3577 changes to combine the 32/64-bit code, allow logical operations on
3578 TI mode in 32-bit, and to use similar match_operator patterns like
3579 scalar mode uses. Combine the Altivec and VSX code for logical
3580 operations, and move it here.
3581 (ior<mode>3, 128-bit types): Likewise.
3582 (xor<mode>3, 128-bit types): Likewise.
3583 (one_cmpl<mode>3, 128-bit types): Likewise.
3584 (nor<mode>3, 128-bit types): Likewise.
3585 (andc<mode>3, 128-bit types): Likewise.
3586 (eqv<mode>3, 128-bit types): Likewise.
3587 (nand<mode>3, 128-bit types): Likewise.
3588 (orc<mode>3, 128-bit types): Likewise.
3589 (and<mode>3_internal): Likewise.
3590 (bool<mode>3_internal): Likewise.
3591 (boolc<mode>3_internal1): Likewise.
3592 (boolc<mode>3_internal2): Likewise.
3593 (boolcc<mode>3_internal1): Likewise.
3594 (boolcc<mode>3_internal2): Likewise.
3595 (eqv<mode>3_internal1): Likewise.
3596 (eqv<mode>3_internal2): Likewise.
3597 (one_cmpl1<mode>3_internal): Likewise.
3599 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
3601 * config/microblaze/microblaze.c (microblaze_expand_prologue):
3602 Rename flag_stack_usage to flag_stack_usage_info.
3604 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
3606 * config/microblaze/sync.md: New file.
3607 * config/microblaze/microblaze.md: Include sync.md
3608 * config/microblaze/microblaze.c: Add print_operand 'y'.
3609 * config/microblaze/constraints.md: Add memory_contraint
3610 'Q' which is a single register.
3612 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
3614 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
3616 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
3618 * reload.c (find_reloads): Exit loop once we find this operand
3619 cannot be reloaded somehow for this alternative.
3621 * reload.c (find_reloads): Exit loop once we find a hard register.
3623 * rtlanal.c (computed_jump_p): Exit loop once we find label
3626 * i386.c (ix86_pad_returns): Exit loop after setting replace.
3628 * cfgloopmanip.c (remove_path): Exit loop after setting
3631 * gensupport.c (subst_dup): Avoid loop if code is not
3632 MATCH_DUP nor MATCH_OP_DUP.
3634 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
3636 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
3638 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
3640 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
3641 true for SP_REGNUM if mode == ptr_mode.
3642 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
3643 with value R0_REGNUM + 31.
3645 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
3647 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
3648 pad pointer-typed argument downward.
3650 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
3652 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
3653 and __ILP32__ when the ILP32 model is in use.
3655 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
3657 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
3658 (aarch64_load_symref_appropriately): In the case of
3659 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
3660 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
3661 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
3662 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
3663 if 'mode' doesn't equal to 'ptr_mode'.
3664 (aarch64_output_mi_thunk): Add an assertion on the alignment of
3665 'vcall_offset'; change to call aarch64_emit_move differently depending
3666 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
3667 to calculate the upper bound of 'vcall_offset'.
3668 (aarch64_cannot_force_const_mem): Change to also return true if
3670 (aarch64_legitimize_reload_address): In the case of large
3671 displacements, add new local variable 'xmode' and an assertion
3672 based on it; change to use 'xmode' to generate the new rtx and
3674 (aarch64_asm_trampoline_template): Change to generate the template
3675 differently depending on TARGET_ILP32 or not; change to use
3676 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
3677 (aarch64_trampoline_size): Removed.
3678 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
3679 and replace immediate literals with it. Change to use 'ptr_mode'
3680 instead of 'DImode' and call convert_memory_address if the mode
3681 of 'fnaddr' doesn't equal to 'ptr_mode'.
3682 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
3684 (aarch64_elf_asm_destructor): Likewise.
3685 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
3686 on TARGET_ILP32 instead of aarch64_trampoline_size.
3687 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
3688 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
3689 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
3690 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
3691 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
3692 (storewb_pair<GPI:mode>_<P:mode>): ... this.
3693 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
3694 depending on the value of 'mode'.
3695 (add_losym_<mode>): New.
3696 (ldr_got_small_<mode>): New, based on ldr_got_small.
3697 (ldr_got_small): Remove.
3698 (ldr_got_small_sidi): New.
3699 * config/aarch64/iterators.md (P): New.
3700 (PTR): Change to 'ptr_mode' in the condition.
3702 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
3704 * config.gcc (aarch64*-*-*): Support --with-abi.
3705 (aarch64*-*-elf): Support --with-multilib-list.
3706 (aarch64*-*-linux*): Likewise.
3707 (supported_defaults): Add abi to aarch64*-*-*.
3708 * configure.ac: Mention AArch64 for --with-multilib-list.
3709 * configure: Re-generated.
3710 * config/aarch64/biarchilp32.h: New file.
3711 * config/aarch64/biarchlp64.h: New file.
3712 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
3714 (MULTILIB_DEFAULTS): Ditto.
3715 (DRIVER_SELF_SPECS): Ditto.
3716 (ASM_SPEC): Update to also substitute -mabi.
3717 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
3718 file whose name depends on -mabi= and -mbig-endian.
3719 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
3721 (POINTER_SIZE): New define.
3722 (POINTERS_EXTEND_UNSIGNED): Ditto.
3723 (enum aarch64_abi_type): New enumeration tag.
3724 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
3725 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
3726 (TARGET_ILP32): New define.
3727 * config/aarch64/aarch64.opt (mabi): New.
3729 (ilp32, lp64): New values for -mabi.
3730 * config/aarch64/t-aarch64 (comma): New define.
3731 (MULTILIB_OPTIONS): Ditto.
3732 (MULTILIB_DIRNAMES): Ditto.
3733 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
3734 * doc/invoke.texi: Document -mabi for AArch64.
3736 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
3738 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
3740 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
3741 Anton Blanchard <anton@au1.ibm.com>
3743 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
3744 endianness when selecting field to splat.
3746 2013-07-22 Eric Christopher <echristo@gmail.com>
3748 * dwarf2out.c (die_odr_checksum): New function to use
3749 CHECKSUM_ macros and ULEB128 for DIE tag.
3750 (generate_type_signature): Use.
3752 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
3754 * config.gcc (sparc*-*-*): Accept leon3 processor.
3755 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
3756 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
3757 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
3758 * config/sparc/sparc.opt (enum processor_type): Add leon3.
3759 (mfix-ut699): Adjust comment.
3760 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
3761 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
3762 (CPP_CPU_SPEC): Likewise.
3763 (ASM_CPU_SPEC): Likewise.
3764 * config/sparc/sparc.c (leon3_cost): New constant.
3765 (sparc_option_override): Add leon3 support.
3766 (mem_ref): New function.
3767 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
3768 (sparc_do_work_around_errata): Look into the instruction in the delay
3769 slot and adjust accordingly. Add fix for the data cache nullify issues
3770 of the UT699. Change insertion position for the NOP.
3771 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
3772 (leon3_load): New reservation.
3773 (leon_store): Bump latency to 2.
3774 (grfpu): New automaton.
3775 (grfpu_alu): New unit.
3776 (grfpu_ds): Likewise.
3777 (leon_fp_alu): Adjust.
3778 (leon_fp_mult): Delete.
3779 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
3780 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
3781 * config/sparc/sparc.md (cpu): Add leon3.
3782 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
3784 (atomic_test_and_set): Likewise.
3787 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
3789 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
3790 default for R5900 targets.
3791 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
3792 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
3793 * config/mips/mips.c (mips_option_override): Report an error for
3794 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
3795 for -march=r5900 -mhard-float.
3797 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
3799 * df-problems.c (can_move_insns_across): Exit loop once we
3800 find a non-fixed, non-global register.
3802 * ipa-pure-const.c (propagate_nothrow): Exit loop after
3805 * omega.c (omega_eliminate_red): Break after setting red_found.
3806 (omega_problem_has_red_equations): Similarly after setting found.
3807 (omega_query_variable): Similarly after setting coupled.
3809 2013-07-22 Marek Polacek <polacek@redhat.com>
3811 * gimplify.c: Don't include gimple.h twice.
3813 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3815 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
3816 instead of TARGET_THUMB1.
3817 (Pz): New constraint.
3818 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
3820 (compare_negsi_si): Likewise.
3821 (compare_addsi2_op0): Likewise.
3822 (compare_addsi2_op1): Likewise.
3823 (addsi3_carryin_<optab>): Likewise.
3824 (addsi3_carryin_alt2_<optab>): Likewise.
3825 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
3826 for arm_restrict_it.
3827 (subsi3_carryin): Likewise.
3828 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
3829 (minmax_arithsi): Disable for arm_restrict_it.
3830 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
3831 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
3832 (satsi_<SAT:code>_shift): Likewise.
3833 (arm_shiftsi3): Add alternative for 16-bit encoding.
3834 (arm32_movhf): Disable for arm_restrict_it.
3835 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
3836 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
3838 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
3840 * config/arm/arm.md (attribute "insn"): Delete.
3841 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
3842 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
3843 (not_shiftsi): Update for attribute change.
3844 (not_shiftsi_compare0): Likewise.
3845 (not_shiftsi_compare0_scratch): Likewise.
3846 (arm_one_cmplsi2): Likewise.
3847 (thumb1_one_cmplsi2): Likewise.
3848 (notsi_compare0): Likewise.
3849 (notsi_compare0_scratch): Likewise.
3850 (thumb1_movdi_insn): Likewise.
3851 (arm_movsi_insn): Likewise.
3852 (movhi_insn_arch4): Likewise.
3853 (movhi_bytes): Likewise.
3854 (arm_movqi_insn): Likewise.
3855 (thumb1_movqi_insn): Likewise.
3856 (arm32_movhf): Likewise.
3857 (thumb1_movhf): Likewise.
3858 (arm_movsf_soft_insn): Likewise.
3859 (thumb1_movsf_insn): Likewise.
3860 (thumb_movdf_insn): Likewise.
3861 (movsicc_insn): Likewise.
3862 (movsfcc_soft_insn): Likewise.
3863 (and_scc): Likewise.
3864 (cond_move): Likewise.
3865 (if_move_not): Likewise.
3866 (if_not_move): Likewise.
3867 (if_shift_move): Likewise.
3868 (if_move_shift): Likewise.
3869 (if_shift_shift): Likewise.
3870 (if_not_arith): Likewise.
3871 (if_arith_not): Likewise.
3872 (cond_move_not): Likewise.
3873 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
3874 (neon_mov<mode>): Likewise.
3875 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
3876 (thumb2_movsi_vfp): Likewise.
3877 (movsf_vfp): Likewise.
3878 (thumb2_movsf_vfp): Likewise.
3879 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
3881 (cortexa7_older_only): Likewise.
3882 (cortexa7_younger): Likewise.
3883 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
3884 (1020alu_shift_op): Likewise.
3885 (1020alu_shift_reg_op): Likewise.
3886 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
3887 (alu_shift_op): Likewise.
3888 (alu_shift_reg_op): Likewise.
3889 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
3890 (11_alu_shift_op): Likewise.
3891 (11_alu_shift_reg_op): Likewise.
3892 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
3893 (9_alu_shift_reg_op): Likewise.
3894 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
3896 (cortex_a15_alu_shift): Likewise.
3897 (cortex_a15_alu_shift_reg): Likewise.
3898 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
3899 (cortex_a5_alu_shift): Likewise.
3900 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
3902 (cortex_a53_alu_shift): Likewise.
3903 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
3905 (cortex_a7_alu_reg): Likewise.
3906 (cortex_a7_alu_shift): Likewise.
3907 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
3908 (cortex_a8_alu_shift): Likewise.
3909 (cortex_a8_alu_shift_reg): Likewise.
3910 (cortex_a8_mov): Likewise.
3911 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
3912 (cortex_a9_dp_shift): Likewise.
3913 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
3914 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
3915 (cortex_r4_mov): Likewise.
3916 (cortex_r4_alu_shift): Likewise.
3917 (cortex_r4_alu_shift_reg): Likewise.
3918 * config/arm/fa526.md (526_alu_op): Update for attribute change.
3919 (526_alu_shift_op): Likewise.
3920 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
3921 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
3922 (626te_alu_shift_op): Likewise.
3923 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
3924 (726te_alu_op): Likewise.
3925 (726te_alu_shift_op): Likewise.
3926 (726te_alu_shift_reg_op): Likewise.
3927 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
3928 (mp626_alu_shift_op): Likewise.
3929 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
3930 (pj4_alu_e1_conds): Likewise.
3931 (pj4_alu): Likewise.
3932 (pj4_alu_conds): Likewise.
3933 (pj4_shift): Likewise.
3934 (pj4_shift_conds): Likewise.
3935 (pj4_alu_shift): Likewise.
3936 (pj4_alu_shift_conds): Likewise.
3938 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3940 * config/arm/predicates.md (shiftable_operator_strict_it):
3942 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
3943 Disable cond_exec version for arm_restrict_it.
3944 (thumb2_smaxsi3): Convert to generate cond_exec.
3945 (thumb2_sminsi3): Likewise.
3946 (thumb32_umaxsi3): Likewise.
3947 (thumb2_uminsi3): Likewise.
3948 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
3949 (thumb2_neg_abssi2): Likewise.
3950 (thumb2_mov_scc): Add alternative for 16-bit encoding.
3951 (thumb2_movsicc_insn): Adjust alternatives.
3952 (thumb2_mov_negscc): Disable for arm_restrict_it.
3953 (thumb2_mov_negscc_strict_it): New pattern.
3954 (thumb2_mov_notscc_strict_it): New pattern.
3955 (thumb2_mov_notscc): Disable for arm_restrict_it.
3956 (thumb2_ior_scc): Likewise.
3957 (thumb2_ior_scc_strict_it): New pattern.
3958 (thumb2_cond_move): Adjust for arm_restrict_it.
3959 (thumb2_cond_arith): Disable for arm_restrict_it.
3960 (thumb2_cond_arith_strict_it): New pattern.
3961 (thumb2_cond_sub): Adjust for arm_restrict_it.
3962 (thumb2_movcond): Likewise.
3963 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
3964 (thumb2_zero_extendhisi2_v6): Likewise.
3965 (thumb2_zero_extendqisi2_v6): Likewise.
3966 (orsi_notsi_si): Likewise.
3967 (orsi_not_shiftsi_si): Likewise.
3969 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
3971 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
3972 instruction sequence is 1 byte shorter.
3974 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
3976 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
3977 it is not needed after split.
3979 2013-07-20 Iain Sandoe <iain@codesourcery.com>
3982 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
3983 second label for nonlocal goto receivers. Don't output pic base labels
3984 unless we're producing PIC; mark that action unreachable().
3985 (ix86_save_reg): If the function contains a nonlocal label, save the
3987 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
3988 * gcc/config/darwin.c (emitted_pic_label_num): New GTY.
3989 (update_pic_label_number_if_needed): New.
3990 (machopic_output_function_base_name): Adjust for nonlocal receiver
3992 (machopic_should_output_picbase_label): New.
3993 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
3994 (nonlocal_goto_receiver): New insn and split.
3996 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
3998 * config/aarch64/aarch64-builtins.c
3999 (aarch64_fold_builtin): Fold abs in all modes.
4000 * config/aarch64/aarch64-simd-builtins.def
4001 (abs): Enable for all modes.
4002 * config/aarch64/arm_neon.h
4003 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
4004 (vabs_f64): Add missing intrinsic.
4006 2013-07-19 Ian Bolton <ian.bolton@arm.com>
4008 * config/aarch64/arm_neon.h (vabs_s64): New function
4010 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
4013 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
4014 * config/avr/avr.md (adjust_len): Add `round'.
4015 * config/avr/avr-protos.h (avr_out_round): New prototype.
4016 (avr_out_plus): Add `out_label' argument.
4017 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
4018 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
4019 Handle the case where `insn' is just a pattern.
4020 (avr_out_bitop): Handle the case where `insn' is just a pattern.
4021 (avr_out_round): New function.
4022 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
4024 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
4026 * config/microblaze/microblaze.c (microblaze_expand_prologue):
4027 Add check for flag_stack_usage to handle -fstack-usage support
4029 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
4031 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
4032 interaction for new Power8 flags and VSX.
4034 2013-07-18 Sriraman Tallam <tmsriram@google.com>
4037 * tree-inline.c (expand_call_inline): Emit errors during
4038 early_inlining only if optimization is not turned on.
4040 2013-07-18 David Malcolm <dmalcolm@redhat.com>
4044 * passes.c (init_optimization_passes): Move the construction of
4045 the pass hierarchy into a new passes.def file.
4047 * Makefile.in (passes.o): Add dependency on passes.def.
4049 2013-07-18 David Malcolm <dmalcolm@redhat.com>
4051 * passes.c (init_optimization_passes): Introduce macros for
4052 constructing the tree of passes (INSERT_PASSES_AFTER,
4053 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
4054 TERMINATE_PASS_LIST).
4056 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
4057 Wei Mi <wmi@google.com>
4059 PR rtl-optimization/57878
4060 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
4062 (reload_pseudo_compare_func): Check nregs first for reload
4065 2013-07-18 David Malcolm <dmalcolm@redhat.com>
4067 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
4069 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
4071 * read-rtl.c (validate_const_int): Once an invalid character is
4072 seen, quit the loop.
4074 * gengtype.c (write_roots): Similarly once we find the "deletable"
4075 or "if_marked" option.
4077 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
4079 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
4080 "xtab" and "sat". Move value "clz" from here to ...
4081 (attriubte "type"): ... here.
4082 (satsi_<SAT:code>): Delete "insn" attribute.
4083 (satsi_<SAT:code>_shift): Likewise.
4084 (arm_zero_extendqisi2addsi): Likewise.
4085 (arm_extendqisi2addsi): Likewise.
4086 (clzsi2): Update for attribute changes.
4087 (rbitsi2): Likewise.
4088 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
4090 (arm_usatsihi): Likewise.
4091 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
4093 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
4095 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
4096 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
4097 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
4098 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
4099 in alphabetical order.
4100 (attribute "core_cycles"): Update for attribute changes.
4101 (arm_addsi3): Likewise.
4102 (addsi3_compare0): Likewise.
4103 (addsi3_compare0_scratch): Likewise.
4104 (addsi3_compare_op1): Likewise.
4105 (addsi3_compare_op2): Likewise.
4106 (compare_addsi2_op0): Likewise.
4107 (compare_addsi2_op1): Likewise.
4108 (addsi3_carryin_shift_<optab>): Likewise.
4109 (subsi3_carryin_shift): Likewise.
4110 (rsbsi3_carryin_shift): Likewise.
4111 (arm_subsi3_insn): Likewise.
4112 (subsi3_compare0): Likewise.
4113 (subsi3_compare): Likewise.
4114 (arm_andsi3_insn): Likewise.
4115 (thumb1_andsi3_insn): Likewise.
4116 (andsi3_compare0): Likewise.
4117 (andsi3_compare0_scratch): Likewise.
4118 (zeroextractsi_compare0_scratch
4119 (andsi_not_shiftsi_si): Likewise.
4120 (iorsi3_insn): Likewise.
4121 (iorsi3_compare0): Likewise.
4122 (iorsi3_compare0_scratch): Likewise.
4123 (arm_xorsi3): Likewise.
4124 (thumb1_xorsi3_insn): Likewise.
4125 (xorsi3_compare0): Likewise.
4126 (xorsi3_compare0_scratch): Likewise.
4127 (satsi_<SAT:code>_shift): Likewise.
4129 (arm_shiftsi3): Likewise.
4130 (shiftsi3_compare0): Likewise.
4131 (not_shiftsi): Likewise.
4132 (not_shiftsi_compare0): Likewise.
4133 (not_shiftsi_compare0_scratch): Likewise.
4134 (arm_one_cmplsi2): Likewise.
4135 (thumb_one_complsi2): Likewise.
4136 (notsi_compare0): Likewise.
4137 (notsi_compare0_scratch): Likewise.
4138 (thumb1_zero_extendhisi2): Likewise.
4139 (arm_zero_extendhisi2): Likewise.
4140 (arm_zero_extendhisi2_v6): Likewise.
4141 (arm_zero_extendhisi2addsi): Likewise.
4142 (thumb1_zero_extendqisi2): Likewise.
4143 (thumb1_zero_extendqisi2_v6): Likewise.
4144 (arm_zero_extendqisi2): Likewise.
4145 (arm_zero_extendqisi2_v6): Likewise.
4146 (arm_zero_extendqisi2addsi): Likewise.
4147 (thumb1_extendhisi2): Likewise.
4148 (arm_extendhisi2): Likewise.
4149 (arm_extendhisi2_v6): Likewise.
4150 (arm_extendqisi): Likewise.
4151 (arm_extendqisi_v6): Likewise.
4152 (arm_extendqisi2addsi): Likewise.
4153 (thumb1_extendqisi2): Likewise.
4154 (thumb1_movdi_insn): Likewise.
4155 (arm_movsi_insn): Likewise.
4156 (movsi_compare0): Likewise.
4157 (movhi_insn_arch4): Likewise.
4158 (movhi_bytes): Likewise.
4159 (arm_movqi_insn): Likewise.
4160 (thumb1_movqi_insn): Likewise.
4161 (arm32_movhf): Likewise.
4162 (thumb1_movhf): Likewise.
4163 (arm_movsf_soft_insn): Likewise.
4164 (thumb1_movsf_insn): Likewise.
4165 (movdf_soft_insn): Likewise.
4166 (thumb_movdf_insn): Likewise.
4167 (arm_cmpsi_insn): Likewise.
4168 (cmpsi_shiftsi): Likewise.
4169 (cmpsi_shiftsi_swp): Likewise.
4170 (arm_cmpsi_negshiftsi_si): Likewise.
4171 (movsicc_insn): Likewise.
4172 (movsfcc_soft_insn): Likewise.
4173 (arith_shiftsi): Likewise.
4174 (arith_shiftsi_compare0
4175 (arith_shiftsi_compare0_scratch
4176 (sub_shiftsi): Likewise.
4177 (sub_shiftsi_compare0
4178 (sub_shiftsi_compare0_scratch
4179 (and_scc): Likewise.
4180 (cond_move): Likewise.
4181 (if_plus_move): Likewise.
4182 (if_move_plus): Likewise.
4183 (if_move_not): Likewise.
4184 (if_not_move): Likewise.
4185 (if_shift_move): Likewise.
4186 (if_move_shift): Likewise.
4187 (if_shift_shift): Likewise.
4188 (if_not_arith): Likewise.
4189 (if_arith_not): Likewise.
4190 (cond_move_not): Likewise.
4191 (thumb1_ashlsi3): Set type attribute.
4192 (thumb1_ashrsi3): Likewise.
4193 (thumb1_lshrsi3): Likewise.
4194 (thumb1_rotrsi3): Likewise.
4195 (shiftsi3_compare0_scratch): Likewise.
4196 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
4197 (neon_mov<mode>): Likewise.
4198 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
4200 (thumb2_movsi_insn): Likewise.
4201 (thumb2_cmpsi_neg_shiftsi): Likewise.
4202 (thumb2_extendqisi_v6): Likewise.
4203 (thumb2_zero_extendhisi2_v6): Likewise.
4204 (thumb2_zero_extendqisi2_v6): Likewise.
4205 (thumb2_shiftsi3_short): Likewise.
4206 (thumb2_addsi3_compare0_scratch): Likewise.
4207 (orsi_not_shiftsi_si): Likewise.
4208 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
4209 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
4211 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
4212 (1020alu_shift_op): Likewise.
4213 (1020alu_shift_reg_op): Likewise.
4214 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
4215 (alu_shift_op): Likewise.
4216 (alu_shift_reg_op): Likewise.
4217 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
4218 (11_alu_shift_op): Likewise.
4219 (11_alu_shift_reg_op): Likewise.
4220 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
4221 (9_alu_shift_reg_op): Likewise.
4222 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
4224 (cortex_a15_alu_shift): Likewise.
4225 (cortex_a15_alu_shift_reg): Likewise.
4226 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
4228 (cortex_a5_alu_shift): Likewise.
4229 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
4231 (cortex_a53_alu_shift): Likewise.
4232 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
4234 (cortex_a7_alu_reg): Likewise.
4235 (cortex_a7_alu_shift): Likewise.
4236 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
4238 (cortex_a8_alu_shift): Likewise.
4239 (cortex_a8_alu_shift_reg): Likewise.
4240 (cortex_a8_mov): Likewise.
4241 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
4242 (cortex_a9_dp_shift): Likewise.
4243 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
4245 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
4247 (cortex_r4_mov): Likewise.
4248 (cortex_r4_alu_shift): Likewise.
4249 (cortex_r4_alu_shift_reg): Likewise.
4250 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
4251 (526_alu_shift_op): Likewise.
4252 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
4253 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
4254 (626te_alu_shift_op): Likewise.
4255 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
4256 (726te_alu_op): Likewise.
4257 (726te_alu_shift_op): Likewise.
4258 (726te_alu_shift_reg_op): Likewise.
4259 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
4260 (mp626_alu_shift_op): Likewise.
4261 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
4262 (pj4_alu_e1_conds): Likewise.
4263 (pj4_alu): Likewise.
4264 (pj4_alu_conds): Likewise.
4265 (pj4_shift): Likewise.
4266 (pj4_shift_conds): Likewise.
4267 (pj4_alu_shift): Likewise.
4268 (pj4_alu_shift_conds): Likewise.
4269 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
4271 (cortexa7_older_only): Likewise.
4272 (cortexa7_younger): Likewise.
4274 2013-07-18 David Malcolm <dmalcolm@redhat.com>
4276 * ipa-pure-const.c (generate_summary): Rename to...
4277 (pure_const_generate_summary): ... this.
4279 2013-07-17 Iain Sandoe <iain@codesourcery.com>
4281 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
4283 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
4286 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
4289 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4291 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
4292 enabled without -march=zEC12.
4293 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
4296 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
4298 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
4299 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
4300 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
4301 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
4302 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
4303 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
4304 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
4306 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
4307 and ISA_HAS_NMADD3_NMSUB3.
4308 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
4309 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
4310 (nmsub4<mode>, nmsub3<mode>): Likewise.
4311 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
4313 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
4315 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
4316 TARGET_MIPS5400 checking.
4318 2013-07-16 Jakub Jelinek <jakub@redhat.com>
4319 Peter Bergner <bergner@vnet.ibm.com>
4321 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
4322 registers in the comment.
4323 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
4324 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
4325 rather than FIRST_PSEUDO_REGISTERS.
4327 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
4329 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
4330 enable extra ISA flags with TARGET_HTM.
4332 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
4334 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
4337 2013-07-15 Cong Hou <congh@google.com>
4339 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
4340 in compare function for sorting.
4342 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
4344 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
4345 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
4346 * config/rs6000/rs6000.opt: Add -mhtm option.
4347 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
4348 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
4349 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
4350 __HTM__ if the HTM instructions are available.
4351 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
4352 htm_spr_reg_operand): New define_predicates.
4353 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
4354 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
4356 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
4357 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
4358 HTM builtin functions.
4359 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
4360 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
4361 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
4362 (rs6000_builtin_mask_calculate): Likewise.
4363 (rs6000_option_override_internal): Likewise.
4364 (bdesc_htm): Add new HTM builtin support.
4365 (htm_spr_num): New function.
4366 (htm_spr_regno): Likewise.
4367 (rs6000_htm_spr_icode): Likewise.
4368 (htm_expand_builtin): Likewise.
4369 (htm_init_builtins): Likewise.
4370 (rs6000_expand_builtin): Add support for HTM builtin functions.
4371 (rs6000_init_builtins): Likewise.
4372 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
4374 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
4375 (TARGET_HTM, MASK_HTM): Define macros.
4376 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
4377 (FIXED_REGISTERS): Likewise.
4378 (CALL_USED_REGISTERS): Likewise.
4379 (CALL_REALLY_USED_REGISTERS): Likewise.
4380 (REG_ALLOC_ORDER): Likewise.
4381 (enum reg_class): Likewise.
4382 (REG_CLASS_NAMES): Likewise.
4383 (REG_CLASS_CONTENTS): Likewise.
4384 (REGISTER_NAMES): Likewise.
4385 (ADDITIONAL_REGISTER_NAMES): Likewise.
4386 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
4387 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
4388 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
4389 * config/rs6000/htm.md: New file.
4390 * config/rs6000/htmintrin.h: New file.
4391 * config/rs6000/htmxlintrin.h: New file.
4393 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
4395 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
4396 Define SYMBOL_TINY_GOT, update comment.
4397 * config/aarch64/aarch64.c
4398 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
4399 (aarch64_expand_mov_immediate): Likewise.
4400 (aarch64_print_operand): Likewise.
4401 (aarch64_classify_symbol): Likewise.
4402 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
4403 (ldr_got_tiny): Define.
4405 2013-07-13 Tobias Grosser <tobias@grosser.es>
4407 PR tree-optimization/54094
4408 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
4409 scheduling dimension for the parallelism check from the polyhedral
4410 information in the AST.
4411 * graphite-dependences.c (carries_deps): Do not assume the schedule is
4414 2013-07-13 Jason Merrill <jason@redhat.com>
4416 * print-tree.c (debug_vec_tree): Use debug_raw.
4417 (debug_raw (vec<tree, va_gc> &)): New.
4418 (debug_raw (vec<tree, va_gc> *)): New.
4419 * tree.h: Declare them.
4421 2013-07-13 Bin Cheng <bin.cheng@arm.com>
4423 * ifcvt.c (ifcvt_after_combine): New static variable.
4424 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
4426 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
4427 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
4428 rest_of_handle_if_after_reload): Pass new argument for if_convert.
4430 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
4432 * config/mips/mips.c (mips_expand_call): Remove empty statement.
4434 2013-07-12 Michael Matz <matz@suse.de>
4437 * convert.c (convert_to_real): Reject non-float inner types.
4439 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
4441 * config/aarch64/aarch64-protos.h
4442 (aarch64_simd_immediate_valid_for_move): Remove.
4443 * config/aarch64/aarch64.c (simd_immediate_info): New member.
4444 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
4446 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
4448 2013-07-11 Steve Ellcey <sellcey@mips.com>
4450 * config/mips/mips.c (mips_conditional_register_usage): Do not
4451 use t[0-7] registers in MIPS16 mode when optimizing for size.
4453 2013-07-11 Sriraman Tallam <tmsriram@google.com>
4455 * config/i386/i386.c (dispatch_function_versions): Fix array
4456 indexing of function_version_info to match actual_versions.
4458 2013-07-11 Teresa Johnson <tejohnson@google.com>
4460 * vec.h (struct va_gc): Move release out-of-line.
4461 (va_gc::release): Call ggc_free on released vec.
4463 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4465 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
4466 Require GOT register as additional operand in UNSPEC.
4467 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
4468 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
4469 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
4470 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
4471 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
4472 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
4473 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
4475 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
4478 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
4479 name seen by assembler/linker rather if available.
4481 2013-07-11 Andreas Schwab <schwab@suse.de>
4483 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
4485 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
4487 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
4489 2013-07-10 Joseph Myers <joseph@codesourcery.com>
4491 * doc/tm.texi.in: Move hook documentation to ....
4492 * target.def: ... here.
4494 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
4496 * doc/tm.texi: Regenerate.
4498 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
4501 * doc/invoke.texi: Document -Wconditionally-supported.
4503 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
4506 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
4509 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
4512 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
4513 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
4514 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
4515 Remove duplicate devices.
4516 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
4517 * config/avr/t-multilib: Regenerate.
4518 * config/avr/avr-tables.opt: Regenerate.
4519 * doc/avr-mmcu.texi: Regenerate.
4521 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
4524 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
4526 2013-07-10 Graham Stott <graham.stott@btinternet.com>
4528 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
4529 the cost of MULT when optimizing for size.
4531 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4533 * config/cr16/cr16-protos.h: Don't include target.h.
4535 2013-07-09 Joseph Myers <joseph@codesourcery.com>
4537 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
4538 adjust register size for TDmode and TFmode for VSX registers.
4540 2013-07-08 Kai Tietz <ktietz@redhat.com>
4543 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
4544 hook_bool_const_tree_true.
4546 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4548 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
4549 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
4550 * config/s390/s390.md: Define FPR*_REGNUM constants.
4551 Fix FPR2_REGNUM constant (18 -> 17).
4552 ("*trunc<BFP:mode><DFP_ALL:mode>2")
4553 ("*trunc<DFP_ALL:mode><BFP:mode>2")
4554 ("trunc<BFP:mode><DFP_ALL:mode>2")
4555 ("trunc<DFP_ALL:mode><BFP:mode>2")
4556 ("*extend<BFP:mode><DFP_ALL:mode>2")
4557 ("*extend<DFP_ALL:mode><BFP:mode>2")
4558 ("extend<BFP:mode><DFP_ALL:mode>2")
4559 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
4562 2013-07-08 Graham Stott <graham.stott@btinternet.com>
4564 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
4566 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4568 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
4569 and cfun_fpr_bit_p to cfun_fpr_save_p.
4570 (s390_frame_area, s390_register_info, s390_frame_info)
4571 (s390_emit_prologue, s390_emit_epilogue)
4572 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
4574 * config/s390/s390.h: Define *_REGNUM macros for floating point
4577 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
4579 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
4581 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
4583 PR rtl-optimization/57786
4584 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
4585 and break out of the loop when it is set to false.
4587 2013-07-08 Jakub Jelinek <jakub@redhat.com>
4590 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
4591 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
4592 (const_int 63)) 0)).
4593 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
4594 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
4595 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
4597 PR rtl-optimization/57829
4598 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
4599 mask bits outside of mode are just sign-extension from mode to HWI.
4601 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
4603 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
4604 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
4605 adjust_address instead of change_address to keep info about alignment.
4606 (emit_strmov): Remove.
4607 (emit_memmov): New function.
4608 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
4609 (expand_movmem_epilogue): Likewise and return updated rtx for
4611 (expand_constant_movmem_prologue): Likewise and return updated rtx for
4612 destination and source.
4613 (decide_alignment): Refactor, handle vector_loop.
4614 (ix86_expand_movmem): Likewise.
4615 (ix86_expand_setmem): Likewise.
4616 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
4618 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
4620 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
4621 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
4623 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
4625 * config/i386/sse.md (sse_movlhps): Change alternative 3
4626 of operand 2 to "m".
4628 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
4631 * config/i386/sse.md (iptr): New mode attribute.
4632 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
4633 (<sse>_vm<plusminus_insn><mode>3): Ditto.
4634 (<sse>_vmmul<mode>3): Ditto.
4635 (<sse>_vmdiv<mode>3): Ditto.
4636 (sse_vmrcpv4sf2): Ditto.
4637 (<sse>_vmsqrt<mode>2): Ditto.
4638 (sse_vmrsqrtv4sf2): Ditto.
4639 (<sse>_vm<code><mode>3): Ditto.
4640 (avx_vmcmp<mode>3): Ditto.
4641 (<sse>_vmmaskcmp<mode>3): Ditto.
4642 (<sse>_comi): Ditto.
4643 (<sse>_ucomi): Ditto.
4644 (*xop_vmfrcz_<mode>): Ditto.
4645 (*fmai_fmadd_<mode>): Ditto.
4646 (*fmai_fmsub_<mode>): Ditto.
4647 (*fmai_fnmadd_<mode>): Ditto.
4648 (*fmai_fnmsub_<mode>): Ditto.
4649 (*fma4i_vmfmadd_<mode>): Ditto.
4650 (*fma4i_vmfmsub_<mode>): Ditto.
4651 (*fma4i_vmfnmadd_<mode>): Ditto.
4652 (*fma4i_vmfnmsub_<mode>): Ditto.
4653 (*xop_vmfrcz_<mode>): Ditto.
4654 (sse_cvtps2pi): Ditto.
4655 (sse_cvttps2pi): Ditto.
4656 (sse_cvtss2si): Ditto.
4657 (sse_cvtss2si_2): Ditto.
4658 (sse_cvtss2siq_2): Ditto.
4659 (sse_cvttss2si): Ditto.
4660 (sse_cvttss2siq): Ditto.
4661 (sse_cvtsd2si): Ditto.
4662 (sse_cvtsd2si_2): Ditto.
4663 (sse_cvtsd2siq_2): Ditto.
4664 (sse_cvttsd2si): Ditto.
4665 (sse_cvttsd2siq): Ditto.
4666 (sse_cvtsd2ss): Ditto.
4667 (sse_cvtss2sd): Ditto.
4668 (avx2_pbroadcast<mode>): Ditto.
4669 (avx2_pbroadcast<mode>_1): Ditto.
4670 (*avx_vperm_broadcast_v4sf): Ditto.
4672 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
4673 (sse_movlhps): Ditto.
4674 (sse_storehps): Ditto.
4675 (sse_loadhps): Ditto.
4676 (sse_storelps): Ditto.
4677 (sse_loadlps): Ditto.
4678 (*vec_concatv4sf): Ditto.
4679 (*vec_interleave_highv2df): Ditto.
4680 (*vec_interleave_lowv2df): Ditto.
4681 (*vec_extractv2df_1_sse): Ditto.
4682 (*vec_extractv2df_0_sse): Ditto.
4683 (sse2_storelpd): Ditto.
4684 (sse2_loadlpd): Ditto.
4685 (sse2_movsd): Ditto.
4686 (*vec_concatv4si): Ditto.
4687 (vec_concatv2di): Ditto.
4689 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
4690 for Intel asm dialect.
4691 (mmx_punpcklwd): Ditto.
4692 (mmx_punpckldq): Ditto.
4694 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
4695 for intel assembler dialect.
4697 2013-07-06 Jakub Jelinek <jakub@redhat.com>
4700 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
4701 for BUILT_IN_C{LZ,LRSB}*.
4702 * tree.h (CASE_INT_FN): Add FN##IMAX case.
4703 * tree-vrp.c (extract_range_basic): Handle
4704 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
4705 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
4706 fall thru to code calling set_value*.
4707 * builtins.c (expand_builtin): Remove *IMAX cases.
4708 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
4709 if width is bigger than 2*HWI.
4711 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
4713 PR rtl-optimization/55342
4714 * lra-int.h (lra_subreg_reload_pseudos): New.
4715 * lra.c: Add undoing optional reloads to the block diagram.
4716 (lra_subreg_reload_pseudos): New.
4717 (lra_optional_reload_pseudos): Change comments.
4718 (lra): Init and clear lra_subreg_reload_pseudos. Clear
4719 lra_optional_reload_pseudos after undo transformations.
4720 * lra-assigns.c (pseudo_prefix_title): New.
4721 (lra_setup_reg_renumber): Use it.
4722 (spill_for): Ditto. Check subreg reload pseudos too.
4723 (assign_by_spills): Consider subreg reload pseudos too.
4724 * lra-constraints.c (simplify_operand_subreg): Use
4725 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
4726 (curr_insn_transform): Recognize and do optional reloads.
4727 (undo_optional_reloads): New.
4728 (lra_undo_inheritance): Call undo_optional_reloads.
4730 2013-07-05 Thomas Quinot <quinot@adacore.com>
4732 * tree-complex.c (expand_complex_operations_1): Fix typo.
4734 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
4736 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
4737 (tune_params): New member 'const vec_costs'.
4738 * config/aarch64/aarch64.c (generic_vector_cost): New.
4739 (generic_tunings): New member 'generic_vector_cost'.
4740 (aarch64_builtin_vectorization_cost): New.
4741 (aarch64_add_stmt_cost): New.
4742 (TARGET_VECTORIZE_ADD_STMT_COST): New.
4743 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
4745 2013-07-03 Jakub Jelinek <jakub@redhat.com>
4748 * config/i386/predicates.md (vsib_address_operand): Disallow
4749 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
4751 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
4754 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
4755 expand_builtin_setjmp_receiver.
4756 (expand_label): Adjust, call expand_builtin_setjmp_receiver
4757 with NULL for the label parameter.
4758 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
4759 the frame-pointer. Adjust comments.
4760 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
4761 only if LABEL is non-NULL.
4763 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
4765 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
4766 (ARM_ABI_AAPCS64): Ditto.
4768 (ARM_DEFAULT_ABI): Ditto.
4770 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
4772 * config/aarch64/aarch64-builtins.c
4773 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
4774 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
4776 * config/aarch64/aarch64-simd.md
4777 (aarch64_ld1<VALL:mode>): New.
4778 (aarch64_st1<VALL:mode>): Likewise.
4779 * config/aarch64/arm_neon.h
4780 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
4782 2013-07-02 Sriraman Tallam <tmsriram@google.com>
4784 * config/i386/i386.c (gate_insert_vzeroupper): Check if
4786 (ix86_option_override_internal):Turn on all -mavx target flags by
4787 default as they are dependent on AVX anyway.
4789 2013-07-02 Cary Coutant <ccoutant@google.com>
4791 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
4793 (loc_checksum_ordered): Likewise.
4794 (hash_loc_operands): Remove inline keyword.
4796 2013-07-02 Jakub Jelinek <jakub@redhat.com>
4798 PR tree-optimization/57741
4799 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
4800 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
4801 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
4802 Allow REAL_CST step_exprs if flag_associative_math.
4803 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
4805 2013-07-02 Ian Bolton <ian.bolton@arm.com>
4807 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
4809 2013-07-02 Ian Bolton <ian.bolton@arm.com>
4811 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
4813 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4815 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
4817 (iorsi3_insn): Likewise.
4818 (arm_xorsi3): Likewise.
4820 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
4822 * arm.md (attribute "wtype"): Delete. Move attribute values from here
4824 (attribute "type"): ... here, and prefix with "wmmx_".
4825 (attribute "core_cycles"): Update for attribute changes.
4826 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
4827 (tbcstv4hi): Likewise.
4828 (tbcstv2si): Likewise.
4829 (iwmmxt_iordi3): Likewise.
4830 (iwmmxt_xordi3): Likewise.
4831 (iwmmxt_anddi3): Likewise.
4832 (iwmmxt_nanddi3): Likewise.
4833 (iwmmxt_arm_movdi): Likewise.
4834 (iwmmxt_movsi_insn): Likewise.
4835 (mov<mode>_internal): Likewise.
4836 (and<mode>3_iwmmxt): Likewise.
4837 (ior<mode>3_iwmmxt): Likewise.
4838 (xor<mode>3_iwmmxt): Likewise.
4839 (add<mode>3_iwmmxt): Likewise.
4840 (ssaddv8qi3): Likewise.
4841 (ssaddv4hi3): Likewise.
4842 (ssaddv2si3): Likewise.
4843 (usaddv8qi3): Likewise.
4844 (usaddv4hi3): Likewise.
4845 (usaddv2si3): Likewise.
4846 (sub<mode>3_iwmmxt): Likewise.
4847 (sssubv8qi3): Likewise.
4848 (sssubv4hi3): Likewise.
4849 (sssubv2si3): Likewise.
4850 (ussubv8qi3): Likewise.
4851 (ussubv4hi3): Likewise.
4852 (ussubv2si3): Likewise.
4853 (mulv4hi3_iwmmxt): Likewise.
4854 (smulv4hi3_highpart): Likewise.
4855 (umulv4hi3_highpart): Likewise.
4856 (iwmmxt_wmacs): Likewise.
4857 (iwmmxt_wmacsz): Likewise.
4858 (iwmmxt_wmacu): Likewise.
4859 (iwmmxt_wmacuz): Likewise.
4860 (iwmmxt_clrdi): Likewise.
4861 (iwmmxt_clrv8qi): Likewise.
4862 (iwmmxt_clr4hi): Likewise.
4863 (iwmmxt_clr2si): Likewise.
4864 (iwmmxt_uavgrndv8qi3): Likewise.
4865 (iwmmxt_uavgrndv4hi3): Likewise.
4866 (iwmmxt_uavgv8qi3): Likewise.
4867 (iwmmxt_uavgv4hi3): Likewise.
4868 (iwmmxt_tinsrb): Likewise.
4869 (iwmmxt_tinsrh): Likewise.
4870 (iwmmxt_tinsrw): Likewise.
4871 (iwmmxt_textrmub): Likewise.
4872 (iwmmxt_textrmsb): Likewise.
4873 (iwmmxt_textrmuh): Likewise.
4874 (iwmmxt_textrmsh): Likewise.
4875 (iwmmxt_textrmw): Likewise.
4876 (iwmxxt_wshufh): Likewise.
4877 (eqv8qi3): Likewise.
4878 (eqv4hi3): Likewise.
4879 (eqv2si3): Likewise.
4880 (gtuv8qi3): Likewise.
4881 (gtuv4hi3): Likewise.
4882 (gtuv2si3): Likewise.
4883 (gtv8qi3): Likewise.
4884 (gtv4hi3): Likewise.
4885 (gtv2si3): Likewise.
4886 (smax<mode>3_iwmmxt): Likewise.
4887 (umax<mode>3_iwmmxt): Likewise.
4888 (smin<mode>3_iwmmxt): Likewise.
4889 (umin<mode>3_iwmmxt): Likewise.
4890 (iwmmxt_wpackhss): Likewise.
4891 (iwmmxt_wpackwss): Likewise.
4892 (iwmmxt_wpackdss): Likewise.
4893 (iwmmxt_wpackhus): Likewise.
4894 (iwmmxt_wpackwus): Likewise.
4895 (iwmmxt_wpackdus): Likewise.
4896 (iwmmxt_wunpckihb): Likewise.
4897 (iwmmxt_wunpckihh): Likewise.
4898 (iwmmxt_wunpckihw): Likewise.
4899 (iwmmxt_wunpckilb): Likewise.
4900 (iwmmxt_wunpckilh): Likewise.
4901 (iwmmxt_wunpckilw): Likewise.
4902 (iwmmxt_wunpckehub): Likewise.
4903 (iwmmxt_wunpckehuh): Likewise.
4904 (iwmmxt_wunpckehuw): Likewise.
4905 (iwmmxt_wunpckehsb): Likewise.
4906 (iwmmxt_wunpckehsh): Likewise.
4907 (iwmmxt_wunpckehsw): Likewise.
4908 (iwmmxt_wunpckelub): Likewise.
4909 (iwmmxt_wunpckeluh): Likewise.
4910 (iwmmxt_wunpckeluw): Likewise.
4911 (iwmmxt_wunpckelsb): Likewise.
4912 (iwmmxt_wunpckelsh): Likewise.
4913 (iwmmxt_wunpckelsw): Likewise.
4914 (ror<mode>3): Likewise.
4915 (ashr<mode>3_iwmmxt): Likewise.
4916 (lshr<mode>3_iwmmxt): Likewise.
4917 (ashl<mode>3_iwmmxt): Likewise.
4918 (ror<mode>3_di): Likewise.
4919 (ashr<mode>3_di): Likewise.
4920 (lshr<mode>3_di): Likewise.
4921 (ashl<mode>3_di): Likewise.
4922 (iwmmxt_wmadds): Likewise.
4923 (iwmmxt_wmaddu): Likewise.
4924 (iwmmxt_tmia): Likewise.
4925 (iwmmxt_tmiaph): Likewise.
4926 (iwmmxt_tmiabb): Likewise.
4927 (iwmmxt_tmiatb): Likewise.
4928 (iwmmxt_tmiabt): Likewise.
4929 (iwmmxt_tmiatt): Likewise.
4930 (iwmmxt_tmovmskb): Likewise.
4931 (iwmmxt_tmovmskh): Likewise.
4932 (iwmmxt_tmovmskw): Likewise.
4933 (iwmmxt_waccb): Likewise.
4934 (iwmmxt_wacch): Likewise.
4935 (iwmmxt_waccw): Likewise.
4936 (iwmmxt_waligni): Likewise.
4937 (iwmmxt_walignr): Likewise.
4938 (iwmmxt_walignr0): Likewise.
4939 (iwmmxt_walignr1): Likewise.
4940 (iwmmxt_walignr2): Likewise.
4941 (iwmmxt_walignr3): Likewise.
4942 (iwmmxt_wsadb): Likewise.
4943 (iwmmxt_wsadh): Likewise.
4944 (iwmmxt_wsadbz): Likewise.
4945 (iwmmxt_wsadhz): Likewise.
4946 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
4947 (iwmmxt_wabsdiffb): Likewise.
4948 (iwmmxt_wabsdiffh): Likewise.
4949 (iwmmxt_wabsdiffw): Likewise.
4950 (iwmmxt_waddsubhx): Likewise
4951 (iwmmxt_wsubaddhx): Likewise.
4952 (addc<mode>3): Likewise.
4953 (iwmmxt_avg4): Likewise.
4954 (iwmmxt_avg4r): Likewise.
4955 (iwmmxt_wmaddsx): Likewise.
4956 (iwmmxt_wmaddux): Likewise.
4957 (iwmmxt_wmaddsn): Likewise.
4958 (iwmmxt_wmaddun): Likewise.
4959 (iwmmxt_wmulwsm): Likewise.
4960 (iwmmxt_wmulwum): Likewise.
4961 (iwmmxt_wmulsmr): Likewise.
4962 (iwmmxt_wmulumr): Likewise.
4963 (iwmmxt_wmulwsmr): Likewise.
4964 (iwmmxt_wmulwumr): Likewise.
4965 (iwmmxt_wmulwl): Likewise.
4966 (iwmmxt_wqmulm): Likewise.
4967 (iwmmxt_wqmulwm): Likewise.
4968 (iwmmxt_wqmulmr): Likewise.
4969 (iwmmxt_wqmulwmr): Likewise.
4970 (iwmmxt_waddbhusm): Likewise.
4971 (iwmmxt_waddbhusl): Likewise.
4972 (iwmmxt_wqmiabb): Likewise.
4973 (iwmmxt_wqmiabt): Likewise.
4974 (iwmmxt_wqmiatb): Likewise.
4975 (iwmmxt_wqmiatt): Likewise.
4976 (iwmmxt_wqmiabbn): Likewise.
4977 (iwmmxt_wqmiabtn): Likewise.
4978 (iwmmxt_wqmiatbn): Likewise.
4979 (iwmmxt_wqmiattn): Likewise.
4980 (iwmmxt_wmiabb): Likewise.
4981 (iwmmxt_wmiabt): Likewise.
4982 (iwmmxt_wmiatb): Likewise.
4983 (iwmmxt_wmiatt): Likewise.
4984 (iwmmxt_wmiabbn): Likewise.
4985 (iwmmxt_wmiabtn): Likewise.
4986 (iwmmxt_wmiatbn): Likewise.
4987 (iwmmxt_wmiattn): Likewise.
4988 (iwmmxt_wmiawbb): Likewise.
4989 (iwmmxt_wmiawbt): Likewise.
4990 (iwmmxt_wmiawtb): Likewise.
4991 (iwmmxt_wmiawtt): Likewise.
4992 (iwmmxt_wmiawbbn): Likewise.
4993 (iwmmxt_wmiawbtn): Likewise.
4994 (iwmmxt_wmiawtbn): Likewise.
4995 (iwmmxt_wmiawttn): Likewise.
4996 (iwmmxt_wmerge): Likewise.
4997 (iwmmxt_tandc<mode>3): Likewise.
4998 (iwmmxt_torc<mode>3): Likewise.
4999 (iwmmxt_torvsc<mode>3): Likewise.
5000 (iwmmxt_textrc<mode>3): Likewise.
5001 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
5002 (wmmxt_pack): Likewise.
5003 (wmmxt_mult_c1): Likewise.
5004 (wmmxt_mult_c2): Likewise.
5005 (wmmxt_alu_c1): Likewise.
5006 (wmmxt_alu_c2): Likewise.
5007 (wmmxt_alu_c3): Likewise.
5008 (wmmxt_transfer_c1): Likewise.
5009 (wmmxt_transfer_c2): Likewise.
5010 (wmmxt_transfer_c3): Likewise.
5011 (marvell_f_iwmmxt_wstr): Likewise.
5012 (marvell_f_iwmmxt_wldr): Likewise.
5014 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
5016 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
5018 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
5021 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
5022 * lra-constraints.c (need_for_split_p): Check call used hard regs
5023 living through calls.
5025 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
5026 call used regs for call insn.
5028 2013-06-28 Jakub Jelinek <jakub@redhat.com>
5031 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
5032 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
5034 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
5036 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
5038 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
5040 * lra-constraints.c (need_for_split_p): Check call used hard regs
5041 living through calls.
5043 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
5046 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
5047 to tie with any other modes. Eliminate Altivec vector mode tests,
5048 since these are a subset of ALTIVEC or VSX vector modes. Simplify
5049 code, to return 0 if testing MODE2 for a condition, if we've
5050 already tested MODE1 for the same condition.
5052 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
5054 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
5057 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
5059 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
5060 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
5062 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
5064 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
5066 (aarch64_symbolic_constant_p): Remove.
5067 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
5068 static. Fix line length and white space.
5069 (aarch64_symbolic_constant_p): Remove.
5070 * config/aarch64/predicates.md (aarch64_valid_symref):
5071 Use aarch64_classify_symbol_expression.
5073 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5075 * config/arm/constraints.md (Ts): New constraint.
5076 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
5078 (compare_scc): Use "Ts" constraint for operand 0.
5079 (ior_scc_scc): Likewise.
5080 (and_scc_scc): Likewise.
5081 (and_scc_scc_nodom): Likewise.
5082 (ior_scc_scc_cmp): Likewise for operand 7.
5083 (and_scc_scc_cmp): Likewise.
5084 * config/arm/thumb2.md (thumb2_movsi_insn):
5085 Add alternatives for 16-bit encodings.
5086 (thumb2_movhi_insn): Likewise.
5087 (thumb2_movsicc_insn): Likewise.
5088 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
5089 (thumb2_negscc): Use "Ts" constraint.
5090 Move mvn instruction outside cond_exec block.
5091 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
5092 for 16-bit encodings.
5094 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5096 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
5098 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
5099 (mulsi3subsi): Likewise.
5100 (mulsidi3adddi): Likewise.
5101 (mulsidi3_v6): Likewise.
5102 (umulsidi3_v6): Likewise.
5103 (umulsidi3adddi_v6): Likewise.
5104 (smulsi3_highpart_v6): Likewise.
5105 (umulsi3_highpart_v6): Likewise.
5106 (mulhisi3tb): Likewise.
5107 (mulhisi3bt): Likewise.
5108 (mulhisi3tt): Likewise.
5109 (maddhisi4): Likewise.
5110 (maddhisi4tb): Likewise.
5111 (maddhisi4tt): Likewise.
5112 (maddhidi4): Likewise.
5113 (maddhidi4tb): Likewise.
5114 (maddhidi4tt): Likewise.
5115 (zeroextractsi_compare0_scratch): Likewise.
5116 (insv_zero): Likewise.
5117 (insv_t2): Likewise.
5118 (anddi_notzesidi_di): Likewise.
5119 (anddi_notsesidi_di): Likewise.
5120 (andsi_notsi_si): Likewise.
5121 (iordi_zesidi_di): Likewise.
5122 (xordi_zesidi_di): Likewise.
5123 (andsi_iorsi3_notsi): Likewise.
5125 (smax_m1): Likewise.
5127 (not_shiftsi): Likewise.
5128 (unaligned_loadsi): Likewise.
5129 (unaligned_loadhis): Likewise.
5130 (unaligned_loadhiu): Likewise.
5131 (unaligned_storesi): Likewise.
5132 (unaligned_storehi): Likewise.
5133 (extv_reg): Likewise.
5134 (extzv_t2): Likewise.
5136 (udivsi3): Likewise.
5137 (arm_zero_extendhisi2addsi): Likewise.
5138 (arm_zero_extendqisi2addsi): Likewise.
5139 (compareqi_eq0): Likewise.
5140 (arm_extendhisi2_v6): Likewise.
5141 (arm_extendqisi2addsi): Likewise.
5142 (arm_movt): Likewise.
5143 (thumb2_ldrd): Likewise.
5144 (thumb2_ldrd_base): Likewise.
5145 (thumb2_ldrd_base_neg): Likewise.
5146 (thumb2_strd): Likewise.
5147 (thumb2_strd_base): Likewise.
5148 (thumb2_strd_base_neg): Likewise.
5149 (arm_negsi2): Add alternative for 16-bit encoding.
5150 (arm_one_cmplsi2): Likewise.
5152 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5154 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
5155 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
5156 (movdfcc): Likewise.
5157 * config/arm/vfp.md (*thumb2_movsf_vfp):
5158 Disable predication for arm_restrict_it.
5159 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
5160 (*thumb2_movdfcc_vfp): Likewise.
5161 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
5162 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
5163 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
5164 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
5165 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
5166 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
5167 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
5168 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
5169 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
5170 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
5171 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
5172 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
5173 Disable predication for arm_restrict_it.
5175 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
5177 * config/i386/bmiintrin.h (_bextr_u32): New.
5178 (_bextr_u64): Ditto.
5180 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
5182 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
5183 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
5184 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
5185 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
5186 * config/mips/n32-elf.h: ...this new file.
5188 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
5191 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
5192 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
5194 2013-06-27 Catherine Moore <clm@codesourcery.com>
5196 * config/mips/mips-tables.opt: Regenerate.
5197 * config/mips/mips-cpus.def: Add m14ke and m14kec.
5198 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
5199 * doc/invoke.texi: Add -m14kc.
5201 2013-06-27 Jakub Jelinek <jakub@redhat.com>
5204 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
5205 constraints of operand 1 and 2.
5208 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
5209 to match RTL canonicalization. Swap predicates and
5210 constraints of operand 1 and 2.
5212 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
5214 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
5215 Process OP_INOUT regs for splitting too.
5217 2013-06-27 Jakub Jelinek <jakub@redhat.com>
5219 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
5220 decl before the loop, initialize to NULL.
5221 (vectorizable_load): Initialize ptr_incr to NULL.
5223 2013-06-27 Martin Jambor <mjambor@suse.cz>
5226 * ipa-ref.h (ipa_maybe_record_reference): Declare.
5227 * ipa-ref.c (ipa_maybe_record_reference): New function.
5228 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
5229 * ipa-cp.c (create_specialized_node): Record potential references from
5231 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
5233 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
5235 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
5236 parameter 'mode' of type 'enum machine_mode mode'; change to pass
5237 'mode' to force_reg.
5238 (aarch64_add_offset): Update calls to aarch64_force_temporary.
5239 (aarch64_expand_mov_immediate): Likewise.
5241 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
5243 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
5244 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
5246 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5248 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
5249 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
5250 (struct machine_function): Add tbegin_p.
5251 (s390_canonicalize_comparison): Fold CC mode compares to
5252 conditional jump if possible.
5253 (s390_emit_jump): Return the emitted jump.
5254 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
5255 Handle CCRAWmode compares.
5256 (s390_option_override): Default to -mhtm if available.
5257 (s390_reg_clobbered_rtx): Handle floating point regs as well.
5258 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
5259 FPRs instead of df_regs_ever_live_p.
5260 (s390_optimize_nonescaping_tx): New function.
5261 (s390_init_frame_layout): Extend clobbered_regs array to cover
5263 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
5264 (s390_expand_tbegin): New function.
5265 (enum s390_builtin): New enum definition.
5266 (code_for_builtin): New array definition.
5267 (s390_init_builtins): New function.
5268 (s390_expand_builtin): New function.
5269 (TARGET_INIT_BUILTINS): Define.
5270 (TARGET_EXPAND_BUILTIN): Define.
5271 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
5272 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
5273 (s390_alc_comparison): Likewise.
5274 * config/s390/s390-modes.def: Add CCRAWmode.
5275 * config/s390/s390.h (processor_flags): Add PF_TX.
5276 (TARGET_CPU_HTM): Define macro.
5277 (TARGET_HTM): Define macro.
5278 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
5279 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
5280 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
5281 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
5282 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
5283 (TBEGIN_MASK, TBEGINC_MASK): New constants.
5284 ("*cc_to_int"): Move up.
5285 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
5286 constants other than 0.
5287 ("*ccraw_to_int"): New insn and splitter definition.
5288 ("tbegin", "tbegin_nofloat", "tbegin_retry")
5289 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
5290 ("tx_assist"): New expander.
5291 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
5292 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
5293 * config/s390/s390.opt: Add -mhtm option.
5294 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
5295 * config/s390/htmxlintrin.h: New file.
5296 * config/s390/htmintrin.h: New file.
5297 * config/s390/s390intrin.h: New file.
5298 * doc/extend.texi: Document htm builtins.
5299 * config.gcc: Add the new header files to extra_headers.
5301 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
5303 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
5304 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
5306 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
5307 Pat Haugen <pthaugen@us.ibm.com>
5308 Peter Bergner <bergner@vnet.ibm.com>
5310 * config/rs6000/power8.md: New.
5311 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
5312 setting for power8 entry.
5313 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
5314 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
5315 test for Power4/Power5 only.
5316 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
5318 (force_new_group): Adjust comment.
5319 * config/rs6000/rs6000.md: Include power8.md.
5321 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
5323 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
5324 * config/arm/arm-protos.h (arm_max_conditional_execute): New
5326 (tune_params): Update comment.
5327 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
5328 (arm_max_conditional_execute): New function.
5329 (thumb2_final_prescan_insn): Use max_insn_skipped and
5330 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
5332 2013-06-25 Jakub Jelinek <jakub@redhat.com>
5334 PR tree-optimization/57705
5335 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
5336 SSA_NAME step, provided that it is not defined inside the loop.
5337 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
5338 (get_initial_def_for_induction): Handle SSA_NAME IV step.
5340 2013-06-25 Martin Jambor <mjambor@suse.cz>
5343 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
5344 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
5346 (ipa_note_param_call): Initialize member_ptr flag.
5347 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
5348 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
5349 (ipa_write_indirect_edge_info): Stream member_ptr flag.
5350 (ipa_read_indirect_edge_info): Likewise.
5352 2013-06-25 Richard Biener <rguenther@suse.de>
5355 * passes.c (init_optimization_passes): Move pass_fold_builtins
5356 and pass_dce earlier with -Og.
5358 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
5360 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
5361 <BIT_FIELD_REF>: Remove trailing TAB.
5362 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
5365 2013-06-24 Martin Jambor <mjambor@suse.cz>
5367 PR tree-optimization/57358
5368 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
5369 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
5370 (ipa_analyze_params_uses): Generate pessimistic info when true.
5372 2013-06-24 Martin Jambor <mjambor@suse.cz>
5374 PR tree-optimization/57539
5375 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
5376 global.inlined_to of the new node to it. All callers changed.
5377 * ipa-inline-transform.c (clone_inlined_nodes): New variable
5378 inlining_into, pass it to cgraph_clone_node.
5379 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
5380 ipa_free_edge_args_substructures.
5381 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
5382 rdesc linked list. Do not assert rdesc edges have inlined caller.
5383 Assert we have found an rdesc in the rdesc list.
5385 2013-06-24 Richard Biener <rguenther@suse.de>
5387 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
5388 (pointer_set_lookup): Declare.
5389 (class pointer_map): New template class implementing a
5390 generic pointer to T map.
5391 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
5392 pointer_map<T>::contains, pointer_map<T>::insert,
5393 pointer_map<T>::traverse): New functions.
5394 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
5395 (pointer_set_lookup): New function.
5396 (pointer_set_contains): Use pointer_set_lookup.
5397 (pointer_set_insert): Likewise.
5398 (insert_aux): Remove.
5399 (struct pointer_map_t): Embed a pointer_set_t.
5400 (pointer_map_create): Adjust.
5401 (pointer_map_destroy): Likewise.
5402 (pointer_map_contains): Likewise.
5403 (pointer_map_insert): Likewise.
5404 (pointer_map_traverse): Likewise.
5405 * tree-streamer.h (struct streamer_tree_cache_d): Use a
5406 pointer_map<unsigned> instead of a pointer_map_t.
5407 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
5408 (streamer_tree_cache_lookup): Likewise.
5409 (streamer_tree_cache_create): Likewise.
5410 (streamer_tree_cache_delete): Likewise.
5411 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
5412 pointer_map<unsigned> instead of a pointer_map_t.
5413 (lto_init_tree_ref_encoder): Adjust.
5414 (lto_destroy_tree_ref_encoder): Likewise.
5415 * lto-section-out.c (lto_output_decl_index): Likewise.
5416 (lto_record_function_out_decl_state): Likewise.
5417 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
5419 2013-06-24 Richard Biener <rguenther@suse.de>
5421 PR tree-optimization/57488
5422 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
5424 2013-06-24 Alan Modra <amodra@gmail.com>
5426 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
5427 (gen_easy_altivec_constant): Likewise.
5428 * config/rs6000/predicates.md (easy_vector_constant_add_self,
5429 easy_vector_constant_msb): Likewise.
5431 2013-06-23 Jakub Jelinek <jakub@redhat.com>
5434 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
5435 add missing return true.
5437 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
5440 * config/sh/predicates.md (general_extend_operand): Invoke
5441 general_movsrc_operand for memory operands.
5442 (general_movsrc_operand): Allow reg+reg addressing, do not use
5443 general_operand for memory operands.
5445 2013-06-23 Sriraman Tallam <tmsriram@google.com>
5447 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
5448 when current target options does not apply.
5449 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
5450 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
5451 * config/i386/bmiintrin.h: Pass appropriate target
5452 attributes to header.
5453 * config/i386/mmintrin.h: Ditto.
5454 * config/i386/nmmintrin.h: Ditto.
5455 * config/i386/avx2intrin.h: Ditto.
5456 * config/i386/fxsrintrin.h: Ditto.
5457 * config/i386/tbmintrin.h: Ditto.
5458 * config/i386/xsaveintrin.h: Ditto.
5459 * config/i386/f16cintrin.h: Ditto.
5460 * config/i386/xtestintrin.h: Ditto.
5461 * config/i386/xsaveoptintrin.h: Ditto.
5462 * config/i386/bmi2intrin.h: Ditto.
5463 * config/i386/lzcntintrin.h: Ditto.
5464 * config/i386/smmintrin.h: Ditto.
5465 * config/i386/wmmintrin.h: Ditto.
5466 * config/i386/x86intrin.h: Remove all header include guards.
5467 * config/i386/prfchwintrin.h: Ditto.
5468 * config/i386/pmmintrin.h: Ditto.
5469 * config/i386/tmmintrin.h: Ditto.
5470 * config/i386/xmmintrin.h: Ditto.
5471 * config/i386/popcntintrin.h: Ditto.
5472 * config/i386/rdseedintrin.h: Ditto.
5473 * config/i386/ammintrin.h: Ditto.
5474 * config/i386/emmintrin.h: Ditto.
5475 * config/i386/immintrin.h: Remove all header include guards.
5476 * config/i386/fma4intrin.h: Ditto.
5477 * config/i386/lwpintrin.h: Ditto.
5478 * config/i386/xopintrin.h: Ditto.
5479 * config/i386/ia32intrin.h: Ditto.
5480 * config/i386/avxintrin.h: Ditto.
5481 * config/i386/rtmintrin.h: Ditto.
5482 * config/i386/fmaintrin.h: Ditto.
5483 * config/i386/mm3dnow.h: Ditto.
5485 2013-06-22 Sriraman Tallam <tmsriram@google.com>
5487 * common/config/i386/i386-common.c: Handle LZCNT.
5489 2013-06-22 Andi Kleen <ak@linux.intel.com>
5491 * doc/extend.texi: Use __atomic_store_n instead of
5492 __atomic_store in HLE example.
5494 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
5496 * config/sh/sh.c: Remove <cstdlib> workaround.
5498 2013-06-21 Andi Kleen <ak@linux.intel.com>
5500 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
5502 2013-06-21 Andi Kleen <ak@linux.intel.com>
5504 * doc/extend.texi: Document that __atomic_clear and
5505 __atomic_test_and_set should only be used with bool.
5507 2013-06-20 Jan Hubicka <jh@suse.cz>
5509 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
5511 * tree.c (decls_same_for_odr): New function.
5512 (same_for_edr): New function.
5513 (types_same_for_odr): New function.
5514 (get_binfo_at_offset): Use it.
5515 * tree.h (types_same_for_odr): Declare.
5517 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
5518 Jason Merrill <jason@redhat.com>
5520 * system.h: Include <cstdlib> as well as <stdlib.h>.
5522 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
5525 * config/i386/i386.c (construct_container): Report error if
5526 long double is used with disabled x87 float returns.
5528 2013-06-20 Jan Hubicka <jh@suse.cz>
5530 * lto-cgraph.c (input_symtab): Do not set cgraph state.
5532 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
5534 PR rtl-optimization/57425
5535 PR rtl-optimization/57569
5536 * alias.c (write_dependence_p): Remove parameters mem_mode and
5537 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
5538 Changed all callers.
5539 (canon_anti_dependence): Get comments and semantics in sync.
5540 Add parameter mem_canonicalized. Changed all callers.
5541 * rtl.h (canon_anti_dependence): Update prototype.
5543 2013-06-20 Richard Biener <rguenther@suse.de>
5545 * data-streamer-in.c (streamer_read_uhwi): Optimize single
5546 byte case, inline streamer_read_uchar and defer section
5549 2013-06-20 Richard Biener <rguenther@suse.de>
5551 PR tree-optimization/57584
5552 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
5553 SSA names into the expanded expression that take part in
5554 abnormal coalescing.
5556 2013-06-19 Sharad Singhai <singhai@google.com>
5558 * gcov.c (print_usage): Handle new option.
5559 (process_args): Ditto.
5560 (get_gcov_intermediate_filename): New function.
5561 (output_intermediate_file): New function.
5562 (output_gcov_file): New function
5563 (generate_results): Handle new option.
5564 (release_function): Relase demangled name.
5565 (read_graph_file): Handle demangled name.
5566 (output_lines): Ditto.
5567 * doc/gcov.texi: Document gcov intermediate format.
5569 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
5572 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
5573 (lra_emit_add): Use the functions. Add comment about Y as an
5576 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
5579 * collect2.c (collect_atexit): New.
5580 (collect_exit): Delete.
5581 (main): Register collect_atexit with atexit.
5582 (collect_wait): Change collect_exit to exit.
5584 * collect2.h (collect_exit): Delete.
5585 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
5587 2013-06-19 Wei Mi <wmi@google.com>
5589 PR rtl-optimization/57518
5590 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
5591 if regno is used in paradoxical subreg.
5592 (update_equiv_regs): Check pdx_subregs[regno] before
5593 set a reg to be equivalent with a mem.
5595 2013-06-19 Matthias Klose <doko@ubuntu.com>
5598 * file-find.h (find_a_file): Add a mode parameter.
5599 * file-find.c (find_a_file): Likewise.
5600 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
5601 with X_OK for the executables.
5602 * collect2.c (main): Call find_a_file with X_OK.
5604 2013-06-19 Steve Ellcey <sellcey@mips.com>
5607 * config/mips/mips.md (casesi_internal_mips16_<mode>):
5608 Use NEXT_INSN instead of next_real_insn.
5610 2013-06-19 Jan Hubicka <jh@suse.cz>
5612 * cgraph.h (const_value_known_p): Replace by ...
5613 (ctor_for_folding): .. this one.
5614 * cgraphunit.c (process_function_and_variable_attributes): Use it.
5615 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
5616 * expr.c (expand_expr_real_1): Likewise.
5617 (string_constant): Likewise.
5618 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
5619 * ipa.c (process_references): Likewise.
5620 (symtab_remove_unreachable_nodes): Likewise.
5621 * ipa-inline-analysis.c (param_change_prob): Likewise.
5622 * gimple-fold.c (canonicalize_constructor_val): Likewise.
5623 (get_base_constructor): Likwise.
5624 * varpool.c (varpool_remove_node): Likewise.
5625 (varpool_remove_initializer): LIkewise.
5626 (dump_varpool_node): LIkwise.
5627 (const_value_known_p): Rewrite to ...
5628 (ctor_for_folding): ... this one.
5630 2013-06-19 Jakub Jelinek <jakub@redhat.com>
5633 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
5634 PERSONALITY in $PATH derived prefixes.
5636 2013-06-19 Jeff Law <law@redhat.com>
5638 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
5641 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
5642 (simplify_bitwise_binary): Use it to simpify certain binary ops on
5645 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
5647 * config/arm/vfp.md: Move VFP instruction classification documentation
5649 * config/arm/arm.md: ... here. Update instruction classification
5652 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
5654 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
5655 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
5656 pattern. Use more efficient sequences on ARMv5 and Thumb2.
5658 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
5661 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
5662 with NEXT_INSN. Use tablejump_p to check for jump table data
5665 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
5668 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
5669 that now in C++ the value is correct per the C++ standards.
5671 2013-06-19 Richard Biener <rguenther@suse.de>
5673 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
5676 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5679 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5682 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
5683 with next_active_insn.
5685 2013-06-18 Sriraman Tallam <tmsriram@google.com>
5687 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
5688 functions are inlined during failures to flag an error.
5689 * tree-inline.c (expand_call_inline): Allow the error to be flagged
5690 in early inline pass.
5692 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
5694 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
5697 2013-06-18 Julian Brown <julian@codesourcery.com>
5699 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
5700 Permit virtual register pre-reload if !strict.
5701 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
5703 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
5705 * config/arm/neon.md (movmisalign<mode>): Use
5706 neon_perm_struct_or_reg_operand instead of
5707 neon_struct_or_register_operand.
5708 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
5709 neon_permissive_struct_operand instead of neon_struct_operand.
5710 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
5711 neon_vector_mem_operand.
5712 * config/arm/predicates.md (neon_struct_operand): Adjust call to
5713 neon_vector_mem_operand.
5714 (neon_permissive_struct_operand): New.
5715 (neon_struct_or_register_operand): Rename to...
5716 (neon_perm_struct_or_reg_operand): This. Adjust call to
5717 neon_vector_mem_operand.
5719 2013-06-18 Richard Biener <rguenther@suse.de>
5721 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
5722 * lto-streamer.h: Include pointer-set.h.
5723 (struct lto_decl_slot): Remove.
5724 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
5725 Remove next_index entry.
5726 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
5727 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
5728 (lto_init_tree_ref_encoder): Adjust.
5729 (lto_destroy_tree_ref_encoder): Likewise.
5730 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
5731 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
5732 (lto_output_decl_index): Adjust.
5733 (lto_new_out_decl_state): Likewise.
5734 (lto_record_function_out_decl_state): Likewise.
5735 * lto-streamer-out.c (copy_function): Likewise.
5737 2013-06-18 Richard Biener <rguenther@suse.de>
5739 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
5740 * cgraphunit.c: Include cfgloop.h.
5741 (init_lowered_empty_function): Initialize the loop tree.
5742 (assemble_thunk): Insert new BBs into loops.
5744 2013-06-18 Richard Biener <rguenther@suse.de>
5746 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
5747 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
5748 the map from cache entry to cache index optional.
5749 (streamer_tree_cache_replace_tree): Adjust accordingly.
5750 (streamer_tree_cache_append): Likewise.
5751 (streamer_tree_cache_delete): Likewise.
5752 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
5753 streamer cache map from cache entry to cache index.
5754 * lto-streamer-out.c (create_output_block): Adjust.
5756 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
5758 * config/arm/arm.md (attribute "insn"): Move multiplication and
5759 division attributes to...
5760 (attribute "type"): ... here. Remove mult.
5761 (attribute "mul32"): New attribute.
5762 (attribute "mul64"): Add umaal.
5763 (*arm_mulsi3): Update attributes.
5764 (*arm_mulsi3_v6): Likewise.
5765 (*thumb_mulsi3): Likewise.
5766 (*thumb_mulsi3_v6): Likewise.
5767 (*mulsi3_compare0): Likewise.
5768 (*mulsi3_compare0_v6): Likewise.
5769 (*mulsi_compare0_scratch): Likewise.
5770 (*mulsi_compare0_scratch_v6): Likewise.
5771 (*mulsi3addsi): Likewise.
5772 (*mulsi3addsi_v6): Likewise.
5773 (*mulsi3addsi_compare0): Likewise.
5774 (*mulsi3addsi_compare0_v6): Likewise.
5775 (*mulsi3addsi_compare0_scratch): Likewise.
5776 (*mulsi3addsi_compare0_scratch_v6): Likewise.
5777 (*mulsi3subsi): Likewise.
5778 (*mulsidi3adddi): Likewise.
5779 (*mulsi3addsi_v6): Likewise.
5780 (*mulsidi3adddi_v6): Likewise.
5781 (*mulsidi3_nov6): Likewise.
5782 (*mulsidi3_v6): Likewise.
5783 (*umulsidi3_nov6): Likewise.
5784 (*umulsidi3_v6): Likewise.
5785 (*umulsidi3adddi): Likewise.
5786 (*umulsidi3adddi_v6): Likewise.
5787 (*smulsi3_highpart_nov6): Likewise.
5788 (*smulsi3_highpart_v6): Likewise.
5789 (*umulsi3_highpart_nov6): Likewise.
5790 (*umulsi3_highpart_v6): Likewise.
5791 (mulhisi3): Likewise.
5792 (*mulhisi3tb): Likewise.
5793 (*mulhisi3bt): Likewise.
5794 (*mulhisi3tt): Likewise.
5795 (maddhisi4): Likewise.
5796 (*maddhisi4tb): Likewise.
5797 (*maddhisi4tt): Likewise.
5798 (maddhidi4): Likewise.
5799 (*maddhidi4tb): Likewise.
5800 (*maddhidi4tt): Likewise.
5802 (udivsi3): Likewise.
5803 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
5804 (thumb2_mulsi_short_compare0): Likewise.
5805 (thumb2_mulsi_short_compare0_scratch): Likewise.
5806 * config/arm/arm1020e.md (1020mult1): Update attribute change.
5807 (1020mult2): Likewise.
5808 (1020mult3): Likewise.
5809 (1020mult4): Likewise.
5810 (1020mult5): Likewise.
5811 (1020mult6): Likewise.
5812 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
5814 (cortex_a15_mult64): Likewise.
5815 (cortex_a15_sdiv): Likewise.
5816 (cortex_a15_udiv): Likewise.
5817 * config/arm/arm1026ejs.md (mult1): Update attribute change.
5823 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
5824 (pj4_ir_div): Likewise.
5825 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
5826 (11_mult2): Likewise.
5827 (11_mult3): Likewise.
5828 (11_mult4): Likewise.
5829 (11_mult5): Likewise.
5830 (11_mult6): Likewise.
5831 (11_mult7): Likewise.
5832 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
5833 (cortex_a8_mla): Likewise.
5834 (cortex_a8_mull): Likewise.
5835 (cortex_a8_smulwy): Likewise.
5836 (cortex_a8_smlald): Likewise.
5837 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
5838 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
5839 (cortex_r4_mul_3): Likewise.
5840 (cortex_r4_mla_4): Likewise.
5841 (cortex_r4_mla_3): Likewise.
5842 (cortex_r4_smlald): Likewise.
5843 (cortex_r4_mull): Likewise.
5844 (cortex_r4_sdiv): Likewise.
5845 (cortex_r4_udiv): Likewise.
5846 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
5847 (cortex_a7_idiv): Likewise.
5848 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
5849 (9_mult2): Likewise.
5850 (9_mult3): Likewise.
5851 (9_mult4): Likewise.
5852 (9_mult5): Likewise.
5853 (9_mult6): Likewise.
5854 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
5855 (cortex_a53_sdiv): Likewise.
5856 (cortex_a53_udiv): Likewise.
5857 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
5858 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
5859 (mp626_mult2): Likewise.
5860 (mp626_mult3): Likewise.
5861 (mp626_mult4): Likewise.
5862 * config/arm/fa526.md (526_mult1): Update attribute change.
5863 (526_mult2): Likewise.
5864 * config/arm/arm-generic.md (mult): Update attribute change.
5865 (mult_ldsched_strongarm): Likewise.
5866 (mult_ldsched): Likewise.
5867 (multi_cycle): Likewise.
5868 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
5869 * config/arm/fa606te.md (606te_mult1): Update attribute change.
5870 (606te_mult2): Likewise.
5871 (606te_mult3): Likewise.
5872 (606te_mult4): Likewise.
5873 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
5874 (cortex_a9_mac16): Likewise.
5875 (cortex_a9_multiply): Likewise.
5876 (cortex_a9_mac): Likewise.
5877 (cortex_a9_multiply_long): Likewise.
5878 * config/arm/fa626te.md (626te_mult1): Update attribute change.
5879 (626te_mult2): Likewise.
5880 (626te_mult3): Likewise.
5881 (626te_mult4): Likewise.
5883 2013-06-18 Richard Biener <rguenther@suse.de>
5886 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
5888 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5891 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
5892 with next_active_insn.
5894 2013-06-18 Alan Modra <amodra@gmail.com>
5896 * config/rs6000/rs6000.h (enum data_align): New.
5897 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
5898 (DATA_ABI_ALIGNMENT): Define.
5899 (CONSTANT_ALIGNMENT): Correct comment.
5900 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
5901 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
5903 2013-06-17 David Malcolm <dmalcolm@redhat.com>
5905 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
5906 ATTRIBUTE_UNUSED marking.
5908 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
5910 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
5911 alternative and update.
5912 (aarch64_dup_lanedi): Delete.
5913 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
5914 * config/aarch64/aarch64-simd-builtins.def: Update.
5916 2013-06-17 Richard Biener <rguenther@suse.de>
5918 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
5919 (lto_input_scc): Declare.
5920 (lto_input_tree_1): Likewise.
5921 (struct lto_stats_d): Add num_tree_bodies_output and
5922 num_pickle_refs_output.
5923 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
5924 (lto_read_tree_1): Split out from ...
5925 (lto_read_tree): ... this.
5926 (lto_input_scc): New function.
5927 (lto_input_tree_1): Split out from ...
5928 (lto_input_tree): ... this. Handle LTO_tree_scc.
5929 (lto_data_in_create): Create the streamer cache without hashes.
5930 * lto-streamer-out.c (create_output_block): Create the streamer
5931 cache with hashes when not doing WPA.
5932 (lto_write_tree_1): Split out from ...
5933 (lto_write_tree): ... this.
5934 (get_symbol_initial_value): New function.
5935 (lto_output_tree_1): Split out from ...
5936 (lto_output_tree): ... this. Write trees as series of SCCs
5937 using a DFS walk via DFS_write_tree.
5938 (struct sccs, struct scc_entry): New types.
5939 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
5940 (DFS_write_tree_body): New function.
5941 (DFS_write_tree): Likewise.
5942 (hash_tree): Likewise.
5943 (scc_entry_compare): Likewise.
5944 (hash_scc): Likewise.
5945 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
5946 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
5947 TREE_CHAIN as regular reference.
5948 (streamer_read_integer_cst): Remove.
5949 (streamer_get_pickled_tree): Adjust.
5950 * tree-streamer-out.c (streamer_write_chain): Disable streaming
5951 of DECL_EXTERNALs in BLOCK_VARS for now.
5952 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
5954 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
5955 Add hash value argument and record that if hashes are recorded
5957 (streamer_tree_cache_insert_1): Adjust.
5958 (streamer_tree_cache_insert): Likewise.
5959 (streamer_tree_cache_insert_at): Rename to ...
5960 (streamer_tree_cache_replace_tree): ... this and adjust.
5961 (streamer_tree_cache_append): Adjust.
5962 (record_common_node): Likewise.
5963 (streamer_tree_cache_create): Add argument whether to
5964 record hash values together with trees.
5965 (streamer_tree_cache_delete): Adjust.
5966 * tree-streamer.h (struct streamer_tree_cache_d): Add
5968 (streamer_read_integer_cst): Remove.
5969 (streamer_tree_cache_insert): Adjust.
5970 (streamer_tree_cache_append): Likewise.
5971 (streamer_tree_cache_insert_at): Rename to ...
5972 (streamer_tree_cache_replace_tree): ... this and adjust.
5973 (streamer_tree_cache_create): Add argument whether to record hashes.
5974 (streamer_tree_cache_get): Rename to ...
5975 (streamer_tree_cache_get_tree): ... this.
5976 (streamer_tree_cache_get_hash): New function.
5977 * tree.c (cache_integer_cst): New function.
5978 * tree.h (cache_integer_cst): Declare.
5979 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
5980 * lto-symtab.c (lto_varpool_replace_node): Only release
5981 DECL_INITIAL of non-prevailing decls.
5982 * varpool.c (varpool_remove_initializer): Do not release
5983 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
5985 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
5987 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
5988 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
5989 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
5990 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
5991 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
5992 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
5993 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
5994 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
5995 instead of TARGET_64BIT.
5996 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
5997 Require ISA_HAS_<D>DIV.
5999 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
6001 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
6002 (mips*-*-linux*): Move default with_llsc setting to where other
6004 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
6006 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
6007 Likewise. Remove default with_tune setting. Move default float
6008 setting to its own block. Handle with_llsc in the same block as above.
6010 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
6012 PR rtl-optimization/57425
6013 PR rtl-optimization/57569
6014 * alias.c (write_dependence_p): Add new parameters mem_mode,
6015 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
6016 Changed all callers.
6017 (canon_anti_dependence): New function.
6018 * cse.c (check_dependence): Use canon_anti_dependence.
6019 * cselib.c (cselib_invalidate_mem): Likewise.
6020 * rtl.h (canon_anti_dependence): Declare.
6022 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
6024 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
6025 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
6026 ".set mips3" for 64-bit targets.
6028 2013-06-15 Dehao Chen <dehao@google.com>
6030 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
6031 * gimple-low.c (gimple_check_call_matching_types): Likewise.
6032 (gimple_check_call_args): Likewise.
6033 * value-prof.c (check_ic_target): Likewise.
6034 * ipa-inline.c (early_inliner): Likewise.
6035 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
6036 * cgraph.c (cgraph_create_edge_1): Likewise.
6037 (cgraph_make_edge_direct): Likewise.
6039 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
6042 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
6043 rs6000_output_move_128bit to handle emitting quad memory
6044 operations. Set attribute length to 8 bytes.
6046 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
6048 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
6050 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
6051 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
6052 (aarch64_<su>mlsl<mode>): Likewise.
6054 2013-06-14 Mike Stump <mikestump@comcast.net>
6056 * Makefile.in (TARGET_H): Add insn-codes.h.
6058 2013-06-14 Alan Modra <amodra@gmail.com>
6062 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
6063 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
6064 bitfield expansion when EXPAND_MEMORY.
6065 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
6067 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
6069 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
6070 test for clearing quad memory on 32-bit later.
6072 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
6074 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
6075 (fold_negate_expr): Likewise.
6076 (fold_real_zero_addition_p): Handle vectors.
6077 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
6079 2013-06-14 Alan Modra <amodra@gmail.com>
6081 * varasm.c (force_const_mem): Revert 2013-06-07 change.
6083 2013-06-13 Jan Hubicka <jh@suse.cz>
6085 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
6086 Local comdats are not externally visible.
6087 * symtab.c (dump_symtab_base): Dump externally visible.
6088 (verify_symtab_base): Verify back links in the symtab hash.
6090 2013-06-13 Bin Cheng <bin.cheng@arm.com>
6092 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
6093 CONVERT_EXPR as equal nodes.
6095 2013-06-13 Bin Cheng <bin.cheng@arm.com>
6097 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
6099 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
6101 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
6102 Generalize to complex and vector.
6103 * tree.c (build_all_ones_cst): New function.
6104 * tree.h (build_all_ones_cst): Declare it.
6106 2013-06-13 Alan Modra <amodra@gmail.com>
6108 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
6109 * config/rs6000/rs6000.md (signbittf2): New insn.
6110 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
6111 (abstf2_internal, cmptf_internal2): Likewise.
6112 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
6114 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
6115 Pat Haugen <pthaugen@us.ibm.com>
6116 Peter Bergner <bergner@vnet.ibm.com>
6118 * config/rs6000/rs6000.c (emit_load_locked): Add support for
6119 power8 byte, half-word, and quad-word atomic instructions.
6120 (emit_store_conditional): Likewise.
6121 (rs6000_expand_atomic_compare_and_swap): Likewise.
6122 (rs6000_expand_atomic_op): Likewise.
6124 * config/rs6000/sync.md (larx): Add new modes for power8.
6126 (AINT): New mode iterator to include TImode as well as normal
6127 integer modes on power8.
6128 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
6129 that VSX registers are not considered. Use AINT mode iterator
6130 instead of INT1 to allow inclusion of quad word atomic operations
6132 (load_locked<mode>): Likewise.
6133 (store_conditional<mode>): Likewise.
6134 (atomic_compare_and_swap<mode>): Likewise.
6135 (atomic_exchange<mode>): Likewise.
6136 (atomic_nand<mode>): Likewise.
6137 (atomic_fetch_<fetchop_name><mode>): Likewise.
6138 (atomic_nand_fetch<mode>): Likewise.
6139 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
6141 (ATOMIC): On power8, add QImode, HImode modes.
6142 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
6143 modes that promote to SImode.
6144 (load_lockedti): Convert TImode arguments to PTImode, so that we
6145 get a guaranteed even/odd register pair.
6146 (load_lockedpti): Likewise.
6147 (store_conditionalti): Likewise.
6148 (store_conditionalpti): Likewise.
6150 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
6151 atomic load/store instructions.
6154 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
6156 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
6158 (insn_count): New attribute, with most cases extracted from...
6159 (length): ...here. Redefine most cases in terms of insn_count.
6160 (single_insn): Delete.
6161 (can_delay): Use insn_count to check for single instructions.
6162 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
6163 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
6164 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
6165 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
6166 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
6167 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
6168 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
6169 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
6170 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
6171 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
6172 rather than "length".
6173 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
6174 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
6175 Use "insn_count" rather than "length".
6176 * config/mips/mips-dsp.md
6177 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
6178 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
6181 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
6183 PR tree-optimization/57361
6184 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
6186 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
6188 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
6190 (aarch64_simd_combine<mode>): New instruction expansion.
6191 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
6193 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
6194 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
6196 2013-06-12 Jan Hubicka <jh@suse.cz>
6198 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
6199 decl has when in streaming stage.
6200 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
6201 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
6203 2013-06-12 Roland Stigge <stigge@antcom.de>
6206 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
6208 2013-06-12 Jakub Jelinek <jakub@redhat.com>
6210 PR tree-optimization/57537
6211 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
6212 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
6214 2013-06-12 Richard Biener <rguenther@suse.de>
6216 * data-streamer.h (streamer_write_char_stream): CSE
6217 obs->current_pointer.
6218 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
6219 streamer_write_char_stream manually and optimize the resulting loop.
6220 (streamer_write_hwi_stream): Likewise.
6222 2013-06-12 Jan Hubicka <jh@suse.cz>
6224 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
6225 * cgraph.h (varpool_create_empty_node): Declare.
6226 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
6228 * symtab.c (symtab_unregister_node): Be lax about missin entries
6230 (symtab_get_node): Update comment.
6231 * varpool.c (varpool_create_empty_node): Break out from ...
6232 (varpool_node_for_decl): ... here.
6233 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
6235 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
6237 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
6238 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
6239 part. Use straight-line flow at the end.
6240 <COMPONENT_REF>: Remove superfluous else.
6241 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
6243 2013-06-12 Jakub Jelinek <jakub@redhat.com>
6246 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
6247 target hook even for !TREE_PUBLIC decls. If no resolution info
6248 is available, return false for common and external decls.
6250 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
6252 * config/rl78/constraints.md (U): New constraint.
6253 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
6256 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
6259 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
6260 to allow returning address to AT_PLATFORM name.
6262 2013-06-11 Jan Hubicka <jh@suse.cz>
6264 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
6265 * cgraph.h (symtab_node_base): Add weakref flag.
6266 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
6267 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
6268 (output_weakrefs): Use weakref flag.
6269 * fold-const.c (simple_operand_p): Handle WEAK.
6270 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
6271 * ipa.c (varpool_externally_visible_p): Drop weakref.
6272 (function_and_variable_visibility): Update comment; fix weakref
6273 sanity checks; do not clear DECL_WEAK on them.
6274 * lto-cgraph.c (lto_output_node): update.
6275 (lto_output_varpool_node): Update.
6276 (input_overwrite_node): Update.
6277 (input_node): Update.
6278 (input_varpool_node): Update.
6279 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
6280 (lto_symtab_merge_symbols): Add sanity check.
6281 (lto_symtab_prevailing_decl): Do not special case weakrefs.
6282 * passes.c (rest_of_decl_compilation): Set static flag, too.
6283 * symtab.c (dump_symtab_base): Dump weakref.
6284 (verify_symtab_base): Sanity check weakrefs.
6285 (symtab_make_decl_local): Remove duplicated code.
6286 (symtab_alias_ultimate_target): Simplify.
6287 * varpool.c (varpool_create_variable_alias): Set weakref flag.
6289 2013-06-11 DJ Delorie <dj@redhat.com>
6291 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
6292 (rl78_unwind_word_mode): New.
6294 2013-06-11 David Malcolm <dmalcolm@redhat.com>
6296 * final.c (debug_prefix_maps): Make static.
6298 2013-06-11 David Malcolm <dmalcolm@redhat.com>
6300 * function.c (initial_trampoline): Remove stray copy.
6302 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
6304 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
6306 2013-06-11 Martin Jambor <mjambor@suse.cz>
6308 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
6309 within bounds at the beginning of the function.
6311 2013-06-11 Alan Modra <amodra@gmail.com>
6313 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
6315 (get_named_section): Don't NULL !DECL_P decl.
6317 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
6319 * doc/invoke.texi (core-avx2): Document.
6321 (atom): Updated with MOVBE.
6323 2013-06-11 Richard Biener <rguenther@suse.de>
6325 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
6327 2013-06-11 Anton Blanchard <anton@samba.org>
6329 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
6330 correct shift value in little-endian mode.
6332 2013-06-11 Jakub Jelinek <jakub@redhat.com>
6335 * varasm.c (get_variable_align): Move #endif to the right place.
6337 2013-06-10 Cary Coutant <ccoutant@google.com>
6339 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
6340 for hash so that hash table traversal order is deterministic.
6342 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
6343 Pat Haugen <pthaugen@us.ibm.com>
6344 Peter Bergner <bergner@vnet.ibm.com>
6346 * config/rs6000/vector.md (GPR move splitter): Do not split moves
6347 of vectors in GPRS if they are direct moves or quad word load or
6350 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
6352 (direct_move_p): Likewise.
6353 (quad_load_store_p): Likewise.
6355 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
6356 classes into bins based on the physical register type.
6357 (reg_class_to_reg_type): Likewise.
6358 (IS_STD_REG_TYPE): Likewise.
6359 (IS_FP_VECT_REG_TYPE): Likewise.
6360 (reload_fpr_gpr): Arrays to determine what insn to use if we can
6361 use direct move instructions.
6362 (reload_gpr_vsx): Likewise.
6363 (reload_vsx_gpr): Likewise.
6364 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
6365 information that is a simplification of register classes. Also
6366 precalculate direct move reload helpers.
6367 (direct_move_p): New function to return true if the operation can
6368 be done as a direct move instruciton.
6369 (quad_load_store_p): New function to return true if the operation
6370 is a quad memory operation.
6371 (rs6000_legitimize_address): If quad memory, only allow register
6372 indirect for TImode addresses.
6373 (rs6000_legitimate_address_p): Likewise.
6374 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
6375 (rs6000_reload_register_type): Likewise.
6376 (register_to_reg_type): Return register type.
6377 (rs6000_secondary_reload_simple_move): New helper function for
6378 secondary reload and secondary memory needed to identify anything
6379 that is a simple move, and does not need reloading.
6380 (rs6000_secondary_reload_direct_move): New helper function for
6381 secondary reload to identify cases that can be done with several
6382 instructions via the direct move instructions.
6383 (rs6000_secondary_reload_move): New helper function for secondary
6384 reload to identify moves between register types that can be done.
6385 (rs6000_secondary_reload): Add support for quad memory operations
6386 and for direct move.
6387 (rs6000_secondary_memory_needed): Likewise.
6388 (rs6000_debug_secondary_memory_needed): Change argument names.
6389 (rs6000_output_move_128bit): New function to return the move to
6390 use for 128-bit moves, including knowing about the various
6391 limitations of quad memory operations.
6393 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
6394 memory operations. call rs6000_output_move_128bit for the actual
6395 instruciton(s) to generate.
6396 (vsx_movti_64bit): Likewise.
6398 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
6399 (UNSPEC_P8V_MTVSRWZ): Likewise.
6400 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
6401 (UNSPEC_P8V_MTVSRD): Likewise.
6402 (UNSPEC_P8V_XXPERMDI): Likewise.
6403 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
6404 (UNSPEC_FUSION_GPR): Likewise.
6405 (FMOVE128_GPR): New iterator for direct move.
6406 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
6409 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
6410 loads and direct move instructions.
6411 (zero_extendsidi2_lfiwzx): Likewise.
6412 (extendsidi2_lfiwax): Likewise.
6413 (extendsidi2_nocell): Likewise.
6414 (floatsi<mode>2_lfiwax): Likewise.
6416 (floatunssi<mode>2_lfiwzx): Likewise.
6418 (fix_trunc<mode>_stfiwx): Likewise.
6419 (fixuns_trunc<mode>_stfiwx): Likewise.
6420 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
6421 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
6422 (parity<mode>2_cmpb): Set length/type attr.
6423 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
6424 for 'mr.' to fast_compare.
6425 (bpermd_<mode>): Change type attr to popcnt.
6426 (p8_fmrgow_<mode>): New insns for power8 direct move support.
6427 (p8_mtvsrwz_1): Likewise.
6428 (p8_mtvsrwz_2): Likewise.
6429 (reload_fpr_from_gpr<mode>): Likewise.
6430 (p8_mtvsrd_1): Likewise.
6431 (p8_mtvsrd_2): Likewise.
6432 (p8_xxpermdi_<mode>): Likewise.
6433 (reload_vsx_from_gpr<mode>): Likewise.
6434 (reload_vsx_from_gprsf): Likewise.
6435 (p8_mfvsrd_3_<mode>): LIkewise.
6436 (reload_gpr_from_vsx<mode>): Likewise.
6437 (reload_gpr_from_vsxsf): Likewise.
6438 (p8_mfvsrd_4_disf): Likewise.
6439 (multi-word GPR splits): Do not split direct moves or quad memory
6442 2013-06-10 David Malcolm <dmalcolm@redhat.com>
6444 * tree-into-ssa.c (interesting_blocks): Make static.
6446 2013-06-10 Jakub Jelinek <jakub@redhat.com>
6449 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
6450 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
6451 Use DATA_ABI_ALIGNMENT for that case instead if defined.
6452 (get_variable_align): New function.
6453 (get_variable_section, emit_bss, emit_common,
6454 assemble_variable_contents, place_block_symbol): Use
6455 get_variable_align instead of DECL_ALIGN.
6456 (assemble_noswitch_variable): Add align argument, use it
6457 instead of DECL_ALIGN.
6458 (assemble_variable): Adjust caller. Use get_variable_align
6459 instead of DECL_ALIGN.
6460 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
6462 (DATA_ABI_ALIGNMENT): Define.
6463 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
6464 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
6465 opt is false, only return the psABI mandated alignment increase.
6466 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
6467 (DATA_ABI_ALIGNMENT): ... this.
6468 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
6469 (DATA_ABI_ALIGNMENT): ... this.
6470 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
6471 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
6472 (DATA_ABI_ALIGNMENT): ... this.
6473 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
6474 * doc/tm.texi: Regenerated.
6476 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
6478 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
6479 cmp_code to construct REG_EQUAL note.
6481 2013-06-09 Jakub Jelinek <jakub@redhat.com>
6484 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
6485 that operands[2] doesn't overlap with operands[0].
6487 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
6488 Jan Hubicka <jh@suse.cz>
6490 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
6491 hack to mark symbols as used.
6493 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
6495 PR rtl-optimization/57559
6496 * lra-constraints.c (process_alt_operands): Don't discourage
6497 memory with known offset for offsetable memory constraint.
6498 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
6500 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
6502 * varasm.c (struct oc_local_state): Reorder fields.
6503 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
6504 and adjust accordingly.
6505 (output_constructor): Reorder initialization code and adjust call to
6506 output_constructor_bitfield.
6508 2013-06-07 Jan Hubicka <jh@suse.cz>
6510 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
6512 2013-06-07 David Malcolm <dmalcolm@redhat.com>
6514 * tree-object-size.c (unknown): Make const.
6516 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6518 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
6519 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
6520 for last alternative in the cpu_facility attribute.
6522 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6525 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
6526 (xordi3): Change operand 2 constraint to arm_xordi_operand.
6527 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
6528 * config/arm/constraints.md (Dg): New constraint.
6529 * config/arm/neon.md (xordi3_neon): Remove.
6530 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
6531 * config/arm/predicates.md (arm_xordi_operand): New predicate.
6533 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6535 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
6536 Clean up alternatives.
6538 2013-06-07 Alan Modra <amodra@gmail.com>
6540 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
6543 2013-06-07 Alan Modra <amodra@gmail.com>
6545 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
6547 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6549 * config/arm/constraints.md (Df): New constraint.
6550 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
6551 Correct length attribute for last two alternatives.
6553 2013-06-07 Alan Modra <amodra@gmail.com>
6555 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
6556 override user -mfp-in-toc.
6557 (offsettable_ok_by_alignment): Consider just the current access
6558 rather than the whole object, unless BLKmode. Handle
6559 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
6560 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
6561 for -mcmodel=medium.
6562 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
6563 override user -mfp-in-toc or -msum-in-toc. Default to
6564 -mno-fp-in-toc for -mcmodel=medium.
6566 2013-06-06 DJ Delorie <dj@redhat.com>
6568 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
6569 TARGET_VALID_POINTER_MODE.
6571 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
6572 Pat Haugen <pthaugen@us.ibm.com>
6573 Peter Bergner <bergner@vnet.ibm.com>
6575 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
6576 Document new power8 builtins.
6578 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
6579 condition code register, to allow 128-bit logical operations to be
6580 done in the VSX or GPR registers.
6581 (nor<mode>3): Use the canonical form for nor.
6582 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
6583 vclz*, and vpopcnt* vector instructions.
6584 (nand<mode>3): Likewise.
6585 (orc<mode>3): Likewise.
6586 (clz<mode>2): LIkewise.
6587 (popcount<mode>2): Likewise.
6589 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
6590 that only the GPRs are recognized.
6592 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
6593 support for new power8 builtins.
6595 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
6597 (xscvdpspn): Likewise.
6603 (vpopcnt): Likewise.
6604 (vpopcntb): Likewise.
6605 (vpopcnth): Likewise.
6606 (vpopcntw): Likewise.
6607 (vpopcntd): Likewise.
6612 (eqv_v16qi3): Likewise.
6613 (eqv_v8hi3): Likewise.
6614 (eqv_v4si3): Likewise.
6615 (eqv_v2di3): Likewise.
6616 (eqv_v4sf3): Likewise.
6617 (eqv_v2df3): Likewise.
6619 (nand_v16qi3): Likewise.
6620 (nand_v8hi3): Likewise.
6621 (nand_v4si3): Likewise.
6622 (nand_v2di3): Likewise.
6623 (nand_v4sf3): Likewise.
6624 (nand_v2df3): Likewise.
6626 (orc_v16qi3): Likewise.
6627 (orc_v8hi3): Likewise.
6628 (orc_v4si3): Likewise.
6629 (orc_v2di3): Likewise.
6630 (orc_v4sf3): Likewise.
6631 (orc_v2df3): Likewise.
6633 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
6634 allow power8 quad mode in 64-bit.
6635 (rs6000_builtin_vectorized_function): Add support to vectorize
6636 ISA 2.07 count leading zeros, population count builtins.
6637 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
6638 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
6639 (builtin_function_type): Add vgbbd builtin function which takes an
6641 (altivec_expand_vec_perm_const): Add support for new power8 merge
6644 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
6645 that does not include TImdoe for use with 32-bit.
6646 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
6648 (UNSPEC_VSX_CVDPSPN): Likewise.
6649 (vsx_xscvdpspn): Likewise.
6650 (vsx_xscvspdpn): Likewise.
6651 (vsx_xscvdpspn_scalar): Likewise.
6652 (vsx_xscvspdpn_directmove): Likewise.
6653 (vsx_and<mode>3): Split logical operations into 32-bit and
6654 64-bit. Add support to do logical operations on TImode as well as
6655 VSX vector types. Allow logical operations to be done in either
6656 VSX registers or in general purpose registers in 64-bit mode. Add
6657 splitters if GPRs were used. For AND, add clobber of CCmode to
6658 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
6660 (vsx_and<mode>3_32bit): Likewise.
6661 (vsx_and<mode>3_64bit): Likewise.
6662 (vsx_ior<mode>3): Likewise.
6663 (vsx_ior<mode>3_32bit): Likewise.
6664 (vsx_ior<mode>3_64bit): Likewise.
6665 (vsx_xor<mode>3): Likewise.
6666 (vsx_xor<mode>3_32bit): Likewise.
6667 (vsx_xor<mode>3_64bit): Likewise.
6668 (vsx_one_cmpl<mode>2): Likewise.
6669 (vsx_one_cmpl<mode>2_32bit): Likewise.
6670 (vsx_one_cmpl<mode>2_64bit): Likewise.
6671 (vsx_nor<mode>3): Likewise.
6672 (vsx_nor<mode>3_32bit): Likewise.
6673 (vsx_nor<mode>3_64bit): Likewise.
6674 (vsx_andc<mode>3): Likewise.
6675 (vsx_andc<mode>3_32bit): Likewise.
6676 (vsx_andc<mode>3_64bit): Likewise.
6677 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
6678 and xxlorc instructions.
6679 (vsx_eqv<mode>3_64bit): Likewise.
6680 (vsx_nand<mode>3_32bit): Likewise.
6681 (vsx_nand<mode>3_64bit): Likewise.
6682 (vsx_orc<mode>3_32bit): Likewise.
6683 (vsx_orc<mode>3_64bit): Likewise.
6685 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
6687 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
6689 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
6690 (p8_vmrgow): Likewise.
6691 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
6692 GPRs to be split under VSX.
6693 (p8v_clz<mode>2): Add power8 count leading zero support.
6694 (p8v_popcount<mode>2): Add power8 population count support.
6695 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
6698 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
6701 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
6703 (vec_nand): Likewise.
6704 (vec_vclz): Likewise.
6705 (vec_vclzb): Likewise.
6706 (vec_vclzd): Likewise.
6707 (vec_vclzh): Likewise.
6708 (vec_vclzw): Likewise.
6709 (vec_vgbbd): Likewise.
6710 (vec_vmrgew): Likewise.
6711 (vec_vmrgow): Likewise.
6712 (vec_vpopcnt): Likewise.
6713 (vec_vpopcntb): Likewise.
6714 (vec_vpopcntd): Likewise.
6715 (vec_vpopcnth): Likewise.
6716 (vec_vpopcntw): Likewise.
6718 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
6720 PR rtl-optimization/57468
6721 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
6724 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
6726 PR rtl-optimization/57459
6727 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
6728 type when setting live regs.
6730 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
6732 * config/s390/s390.opt (mlra): New option.
6733 * config/s390/s390.c (s390_decompose_address): Check displacement
6734 for all registers for LRA.
6735 (s390_secondary_reload): Don't used secondary reloads for LRA.
6736 (s390_lra_p): New function.
6737 (TARGET_LRA_P): Define.
6738 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
6739 of attribute cpu_facility to zarch for the last alternative.
6740 (*cmpmem_short): Ditto.
6742 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
6744 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
6745 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
6746 (arm_expand_prologue): Likewise.
6748 2013-06-06 Teresa Johnson <tejohnson@google.com>
6751 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
6752 as this is now done by redirect_edge_and_branch_force.
6753 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
6754 barriers, and fix interaction with splitting.
6755 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
6756 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
6757 reflect changes made in the routine.
6758 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
6759 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
6760 since this is called in cfglayout mode, and replace partition fixup
6761 with assert as that is now done by force_nonfallthru_and_redirect.
6762 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
6763 already be marked with region crossing note.
6764 (insert_section_boundary_note): Make non-static, gate on flag
6765 has_bb_partition, rewrite to also check for multiple partitions.
6766 (rest_of_handle_reorder_blocks): Remove call to
6767 insert_section_boundary_note, now done later during free_cfg.
6768 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
6769 * bb-reorder.h (insert_section_boundary_note): Declare.
6770 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
6771 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
6772 invoke insert_section_boundary_note.
6773 (try_redirect_by_replacing_jump): Remove unnecessary
6774 check for region crossing note.
6775 (fixup_partition_crossing): New function.
6776 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
6777 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
6778 in non-cfglayout mode.
6779 (force_nonfallthru_and_redirect): Fixup partition boundaries,
6780 remove old code that tried to do this. Emit barrier correctly
6781 when we are in cfglayout mode.
6782 (last_bb_in_partition): New function.
6783 (rtl_split_edge): Correctly fixup partition boundaries.
6784 (commit_one_edge_insertion): Remove old code that tried to
6785 fixup region crossing edge since this is now handled in
6786 split_block, and set up insertion point correctly since
6787 block may now end in a jump.
6788 (verify_hot_cold_block_grouping): Guard against checking when not in
6789 linearized RTL mode.
6790 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
6792 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
6793 rtl_verify_flow_info, so not called in cfglayout mode.
6794 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
6795 (fixup_reorder_chain): Remove old code that attempted to fixup region
6796 crossing note as this is now handled in force_nonfallthru_and_redirect.
6797 (duplicate_insn_chain): Don't duplicate switch section notes.
6798 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
6800 * basic-block.h (emit_barrier_after_bb): Declare.
6802 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6804 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
6805 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
6806 arm_usatsihi): Adjust alternatives for arm_restrict_it.
6808 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6810 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
6812 * config/arm/ldmstm.md: Regenerate.
6814 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6816 * config/arm/sync.md (atomic_loaddi_1):
6817 Disable predication for arm_restrict_it.
6818 (arm_load_exclusive<mode>): Likewise.
6819 (arm_load_exclusivesi): Likewise.
6820 (arm_load_exclusivedi): Likewise.
6821 (arm_load_acquire_exclusive<mode>): Likewise.
6822 (arm_load_acquire_exclusivesi): Likewise.
6823 (arm_load_acquire_exclusivedi): Likewise.
6824 (arm_store_exclusive<mode>): Likewise.
6825 (arm_store_exclusive<mode>): Likewise.
6826 (arm_store_release_exclusivedi): Likewise.
6827 (arm_store_release_exclusive<mode>): Likewise.
6829 2013-06-06 Richard Biener <rguenther@suse.de>
6831 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
6833 (lto_tag_is_tree_code_p): Adjust.
6834 (lto_tag_is_gimple_code_p): Likewise.
6835 (lto_gimple_code_to_tag): Likewise.
6836 (lto_tag_to_gimple_code): Likewise.
6837 (lto_tree_code_to_tag): Likewise.
6838 (lto_tag_to_tree_code): Likewise.
6839 * data-streamer.h (streamer_write_hwi_in_range): Use
6840 uhwi streaming to stream the normalized range.
6841 (streamer_read_hwi_in_range): Likewise.
6843 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6845 * config/arm/arm.md (enabled_for_depr_it): New attribute.
6846 (predicable_short_it): Likewise.
6847 (predicated): Likewise.
6848 (enabled): Handle above.
6849 (define_cond_exec): Set predicated attribute to yes.
6851 2013-06-05 Mike Stump <mikestump@comcast.net>
6853 * gdbinit.in (__FUNCTION__): Add.
6855 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
6857 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
6858 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
6860 2013-06-05 Jan Hubicka <jh@suse.cz>
6862 * varasm.c (mark_decl_referenced): Revert the removal until targets
6865 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
6867 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
6868 instead of mark_decl_referenced.
6870 2013-06-05 Jan Hubicka <jh@suse.cz>
6872 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
6873 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
6874 and symtab_used_from_object_file_p.
6875 (cgraph_make_node_local_1): Clear forced_by_abi.
6876 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
6877 * cgraph.h (symtab_node_base): Add forced_by_abi.
6878 (decide_is_variable_needed): Remove.
6879 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
6880 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
6881 (decide_is_symbol_needed): ... this one; handle symbols in general;
6882 always analyze virtuals; honnor forced_by_abi.
6883 (cgraph_finalize_function): Update.
6884 (varpool_finalize_decl): Update.
6885 (symbol_defined_and_needed): Remove.
6886 (analyze_functions): Update.
6887 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
6888 output_refs, input_overwrite_node): Handle forced_by_abi.
6889 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
6890 (address_taken_from_non_vtable_p): ... this one.
6891 (comdat_can_be_unshared_p_1): New function.
6892 (cgraph_comdat_can_be_unshared_p): Rename to ...
6893 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
6894 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
6895 (function_and_variable_visibility): Clear forced_by_abi as needed.
6896 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
6897 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
6898 * symtab.c (dump_symtab_base): Dump forced_by_abi.
6899 * varpool.c (decide_is_variable_needed): Remove.
6901 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6903 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
6904 (arm_option_override): Override arm_restrict_it where appropriate.
6905 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
6906 * config/arm/arm.opt (mrestrict-it): New command-line option.
6907 * doc/invoke.texi: Document -mrestrict-it.
6909 2013-06-05 David Malcolm <dmalcolm@redhat.com>
6911 * tsan.c (tsan_atomic_table): Make const.
6913 2013-06-05 Richard Biener <rguenther@suse.de>
6915 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
6916 index associated with the tree we are supposed to replace.
6917 * tree-streamer-out.c (pack_ts_base_value_fields): Output
6918 TREE_ASM_WRITTEN as zero for everything but SSA names.
6920 2013-06-05 David Malcolm <dmalcolm@redhat.com>
6922 * tree-ssa-structalias.c (call_stmt_vars): Make static.
6924 2013-06-04 Jan Hubicka <jh@suse.cz>
6926 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
6927 (input_node, input_varpool_node): Handle correctly external same
6929 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
6930 nodes at ltrans stage.
6932 2013-06-04 Jan Hubicka <jh@suse.cz>
6934 * ipa-inline.c (update_caller_keys): Fix availability test.
6935 (update_callee_keys): Likewise.
6936 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
6937 to follow ELF standard.
6939 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
6941 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
6942 (mips64r5900el-*-elf*): New configurations.
6943 * config/mips/mips-cpus.def (r5900): New processor.
6944 * config/mips/mips-tables.opt: Regenerate.
6945 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
6946 (mips_issue_rate): Handle PROCESSOR_R5900.
6947 (mips_reorg_process_insns): Force reorder mode for the R5900.
6948 * config/mips/mips.h (TARGET_MIPS5900): Define.
6949 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
6951 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
6953 * config/mips/mips.md (processor): Add r5900.
6954 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
6956 2013-06-04 Ian Bolton <ian.bolton@arm.com>
6958 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
6959 into function to generate MOVI instruction.
6960 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
6961 (aarch64_preferred_simd_mode): Turn into wrapper.
6962 (aarch64_output_scalar_simd_mov_immediate): New function.
6963 * config/aarch64/aarch64-protos.h: Add prototype for above.
6965 2013-06-04 Ian Bolton <ian.bolton@arm.com>
6967 * config/aarch64/aarch64.c (simd_immediate_info): Remove
6968 element_char member.
6969 (sizetochar): Return signed char.
6970 (aarch64_simd_valid_immediate): Remove elchar and other
6971 unnecessary variables.
6972 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
6973 Calculate element_char as required.
6974 * config/aarch64/aarch64-protos.h: Update and move prototype
6975 for aarch64_output_simd_mov_immediate.
6976 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
6979 2013-06-04 Ian Bolton <ian.bolton@arm.com>
6981 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
6982 information completed by aarch64_simd_valid_immediate.
6983 (aarch64_legitimate_constant_p): Update arguments.
6984 (aarch64_simd_valid_immediate): Work with struct rather than many
6986 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
6987 (aarch64_simd_make_constant): Update arguments.
6988 (aarch64_output_simd_mov_immediate): Work with struct rather than
6989 many pointers. Output immediate directly rather than as operand.
6990 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
6992 * config/aarch64/constraints.md (Dn): Update arguments.
6994 2013-06-04 Ian Bolton <ian.bolton@arm.com>
6996 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
6998 (aarch64_simd_immediate_valid_for_move): Remove.
6999 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
7000 (aarch64_simd_make_constant): Update call.
7001 (aarch64_output_simd_mov_immediate): Update call.
7002 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
7004 * config/aarch64/constraints.md (Dn): Update call.
7006 2013-06-04 Ian Bolton <ian.bolton@arm.com>
7008 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
7009 return type to bool for prototype.
7010 (aarch64_legitimate_constant_p): Check for true instead of not -1.
7011 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
7012 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
7014 2013-06-04 Catherine Moore <clm@codesourcery.com>
7016 * config/mips/mips.opt (meva): New.
7017 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
7018 (ASM_SPEC): Handle -meva.
7019 * doc/invoke.texi (meva): Document.
7021 2013-06-04 Alan Modra <amodra@gmail.com>
7023 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
7026 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7028 * rtl.def: Add extra fourth optional field to define_cond_exec.
7029 * gensupport.c (process_one_cond_exec): Process attributes from
7031 * doc/md.texi: Document fourth field in define_cond_exec.
7033 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
7035 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
7036 out the processing order as in store_bit_field_1.
7038 2013-06-04 Jan Hubicka <jh@suse.cz>
7041 * cgraphunit.c (cgraph_process_same_body_aliases): Create
7042 non-VAR_DECL node if it does not exist yet.
7044 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
7046 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
7047 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
7048 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
7049 target_cpu_default setting.
7051 2013-06-03 Teresa Johnson <tejohnson@google.com>
7053 * dumpfile.c (opt_info_switch_p): Change -fopt-info
7054 default to -fopt-info=optimized instead of all.
7055 * doc/invoke.texi: Ditto.
7056 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
7057 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
7058 (execute_vect_slp): Emit BB vectorization success under
7059 MSG_OPTIMIZED_LOCATIONS.
7060 * tree-vect-slp.c (vect_slp_transform_bb): Change
7061 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
7062 * tree-vect-loop.c (vect_transform_loop): Ditto.
7064 2013-06-03 Jason Merrill <jason@redhat.com>
7067 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
7068 Use TARGET_EXPR for C++.
7070 2013-06-03 Jakub Jelinek <jakub@redhat.com>
7072 PR rtl-optimization/57268
7073 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
7074 if DEBUG_INSN_P (insn).
7077 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
7079 PR rtl-optimization/57268
7080 * sched-deps.c (sched_analyze_2): Flush dependence lists if
7081 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
7083 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
7085 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
7086 (ix86_avoid_lea_for_addr): Likewise.
7087 (exact_dependency_1): Likewise.
7088 (ix86_adjust_cost): Likewise.
7089 (swap_top_of_ready_list): Fix formatting and !reload_completed check
7091 (do_reorder_for_imul): Fix typo, formatting and
7092 !reload_completed check removed.
7093 (ix86_sched_reorder): Fix typo and formatting.
7094 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
7097 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
7099 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
7101 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
7103 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
7104 <STRING_CST>: Likewise.
7105 <VECTOR_CST>: Likewise.
7107 2013-06-01 Janus Weil <janus@gcc.gnu.org>
7108 Mikael Morin <mikael@gcc.gnu.org>
7110 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
7111 * config.in: Regenerated.
7112 * configure: Regenerated.
7114 2013-06-01 Jan Hubicka <jh@suse.cz>
7117 * cgraphunit.c (compile): When weakref is not supported,
7118 set up transparent aliases before final output pass.
7119 * varasm.c (assemble_alias): Do not try to do it here.
7121 2013-06-01 Jan Hubicka <jh@suse.cz>
7124 * passes.c (for_per_function): Skip unanalyzed functions.
7126 2013-06-01 Jan Hubicka <jh@suse.cz>
7128 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
7129 (lto_symtab_merge_symbols_1): ... this one.
7130 (lto_symtab_merge_cgraph_nodes): Rename to ...
7131 (lto_symtab_merge_symbols): ... this one; simplify.
7132 * cgraph.c (same_body_aliases_done): Rename to ...
7133 (cpp_implicit_aliases_done): ... this one.
7134 (cgraph_create_function_alias): Update.
7135 (cgraph_same_body_alias): Update.
7136 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
7137 (verify_edge_corresponds_to_fndecl): Simplify.
7138 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
7139 (cgraph_node): Remove same_body_alias.
7140 (varpool_node): Remove alias_of and extra_name_alias.
7141 (same_body_aliases_done): Rename to ..
7142 (cpp_implicit_aliases_done): ... this one.
7143 (symtab_alias_ultimate_target): Add default parameter.
7144 (symtab_resolve_alias): New function.
7145 (fixup_same_cpp_alias_visibility): Declare.
7146 (cgraph_function_node): Add default parameter.
7147 (cgraph_node_asm_name): Likewise.
7148 (cgraph_function_or_thunk_node): Add default parameter; do
7149 not ICE when it is NULL.
7150 (varpool_variable_node): Likewise.
7151 * tree-emutls.c (create_emultls_var): Update.
7152 (ipa_lower_emutls): Update.
7153 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
7154 (cgraph_reset_node): Reset alias info.
7155 (cgraph_finalize_function): Update.
7156 (fixup_same_cpp_alias_visibility): Move to symtab.c.
7157 (analyze_function): Simplify.
7158 (cgraph_process_same_body_aliases): Simplify.
7159 (analyze_functions): Fixup same body aliases.
7160 (handle_alias_pairs): Simplify.
7161 (assemble_thunk): Update.
7162 (assemble_thunks_and_aliases): Update.
7163 (output_weakrefs): Rewrite.
7164 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
7165 (lto_output_varpool_node): Likewise.
7166 (compute_ltrans_boundary): Remve assert.
7167 (get_alias_symbol): New functoin.
7168 (input_node): Rewrite alias handling.
7169 (input_varpool_node): Likewise.
7170 * ipa-pure-const.c (propagate_pure_const): Fix formating.
7171 * ipa.c (process_references): Handle weakrefs correctly.
7172 (symtab_remove_unreachable_nodes): Likewise.
7173 * trans-mem.c (get_cg_data): Update.
7174 (ipa_tm_create_version_alias): Update.
7175 (ipa_tm_execute): Update.
7176 * symtab.c (dump_symtab_base): Dump aliases.
7177 (verify_symtab_base): Verify aliases.
7178 (symtab_node_availability): New function.
7179 (symtab_alias_ultimate_target): Simplify.
7180 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
7181 handle all the fixup cases.
7182 (symtab_resolve_alias): New function.
7183 * passes.c (ipa_write_summaries): Handle weakrefs.
7184 * varpool.c (varpool_analyze_node): Simplify.
7185 (assemble_aliases): Update.
7186 (varpool_create_variable_alias): Simplify.
7187 (varpool_extra_name_alias): Simplify.
7188 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
7189 (lto_symtab_merge_symbols): ... this one.
7191 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
7194 PR rtl-optimization/57268
7195 * sched-deps.c (sched_analyze_2): Flush dependence lists if
7196 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
7198 2013-06-01 Tobias Burnus <burnus@net-b.de>
7201 2013-05-31 Tobias Burnus <burnus@net-b.de>
7204 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
7207 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
7209 PR rtl-optimization/57268
7210 * sched-deps.c (sched_analyze_2): Flush dependence lists if
7211 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
7213 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
7215 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
7216 unordered comparison operators when -fno-trapping-math is in effect
7218 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
7219 and implement unordered comparison operators properly on the e500.
7221 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
7223 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
7224 for constant scalar integers.
7225 (simplify_relational_operation_1): Likewise.
7227 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
7229 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
7230 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
7233 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
7234 Igor Zamyatin <igor.zamyatin@intel.com>
7236 Silvermont (SLM) architecture performance tuning.
7237 * config/i386/i386.h (enum ix86_tune_indices): Add
7238 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
7239 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
7241 * config/i386/i386.c (initial_ix86_tune_features)
7242 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
7243 (ix86_lea_outperforms): Handle Silvermont tuning.
7244 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
7246 (ix86_use_lea_for_mov): Likewise.
7247 (ix86_avoid_lea_for_addr): Likewise.
7248 (ix86_lea_for_add_ok): Likewise.
7249 (exact_dependency_1): New function.
7250 (exact_store_load_dependency): Likewise.
7251 (ix86_adjust_cost): Handle Silvermont tuning.
7252 (do_reoder_for_imul): Likewise.
7253 (swap_top_of_ready_list): New function.
7254 (ix86_sched_reorder): Changed to handle Silvermont tuning.
7256 * config/i386/i386.md (peepholes that split memory operand in fp
7259 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
7261 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
7262 Remove un-necessary braces.
7264 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
7266 * config/aarch64/aarch64.c (aarch64_classify_symbol):
7267 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
7269 2013-05-31 Tobias Burnus <burnus@net-b.de>
7272 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
7274 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7277 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
7278 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
7279 * config/arm/neon.md (iordi3_neon): Remove.
7280 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
7281 * config/arm/predicates.md (imm_for_neon_logic_operand):
7282 Move to earlier in the file.
7283 (neon_logic_op2): Likewise.
7284 (arm_iordi_operand_neon): New predicate.
7286 2013-05-31 Richard Biener <rguenther@suse.de>
7288 PR tree-optimization/57478
7289 PR tree-optimization/57453
7290 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
7293 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
7295 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
7296 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
7298 2013-05-30 Tobias Burnus <burnus@net-b.de>
7299 Thomas Koenig <tkoenig@gcc.gnu.org>
7302 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
7303 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
7305 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
7307 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
7309 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
7311 * target.def (register_usage_leveling_p): New hook.
7312 * targhooks.c (default_register_usage_leveling_p): New.
7313 * targhooks.h (default_register_usage_leveling_p): New prototype.
7314 * lra-assigns.c (register_usage_leveling_p): Use the hook.
7315 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
7316 * doc/tm.texi: Update.
7317 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
7319 2013-05-30 Ian Bolton <ian.bolton@arm.com>
7321 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
7322 (*insv_reg<mode>): New define_insn.
7324 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
7326 PR rtl-optimization/57439
7327 * postreload.c (move2add_valid_value_p): Check that we have
7328 a zero subreg_regno_offset when accessing the register in
7331 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
7332 Igor Zamyatin <igor.zamyatin@intel.com>
7334 Silvermont (SLM) architecture pipeline model, tuning and
7336 * config.gcc: Add slm config options and target.
7338 * config/i386/slm.md: New.
7340 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
7342 * gcc/config/i386/i386-c.c (ix86_target_macros_internal): New case
7344 (ix86_target_macros_internal): Likewise.
7346 * gcc/config/i386/i386.c (slm_cost): New cost.
7347 (m_SLM): New macro flag.
7348 (initial_ix86_tune_features): Set m_SLM.
7349 (x86_accumulate_outgoing_args): Likewise.
7350 (x86_arch_always_fancy_math_387): Likewise.
7351 (processor_target_table): Add slm cost.
7352 (cpu_names): Add slm cpu name.
7353 (x86_option_override_internal): Set SLM ISA.
7354 (ix86_issue_rate): New case PROCESSOR_SLM.
7355 (ia32_multipass_dfa_lookahead): Likewise.
7356 (fold_builtin_cpu): Add slm.
7358 * config/i386/i386.h (TARGET_SLM): New target macro.
7359 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
7360 (processor_type): Add PROCESSOR_SLM.
7362 * config/i386/i386.md (cpu): Add new value "slm".
7363 (slm.md): Include slm.md.
7365 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
7366 Zhenqiang Chen <zhenqiang.chen@linaro.org>
7368 * config/arm/arm-protos.h: Add and update function protos.
7369 * config/arm/arm.c (use_simple_return_p): New added.
7370 (thumb2_expand_return): Check simple_return flag.
7371 * config/arm/arm.md: Add simple_return and conditional simple_return.
7372 * config/arm/iterators.md: Add iterator for return and simple_return.
7374 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
7376 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
7377 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
7378 (arm_emit_vfp_multi_reg_pop): Likewise.
7379 (thumb2_emit_ldrd_pop): Likewise.
7380 (arm_expand_epilogue): Add misc REG_CFA notes.
7381 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
7383 2013-05-29 Lawrence Crowl <crowl@google.com>
7385 * config/arm/t-arm: Update for below.
7387 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
7388 Change type to hash_table. Update dependent calls and types.
7390 * config/i386/t-cygming: Update for below.
7392 * config/i386/t-interix: Update for below.
7394 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
7395 Change type to hash_table. Update dependent calls and types.
7396 (i386_find_on_wrapper_list::wrappers): Likewise.
7398 * config/ia64/t-ia64: Update for below.
7400 * config/ia64/ia64.c (bundle_state_table):
7401 Change type to hash_table. Update dependent calls and types.
7403 * config/mips/mips.c (mips_reorg_process_insns::htab):
7404 Change type to hash_table. Update dependent calls and types.
7406 * config/sol2.c (solaris_comdat_htab):
7407 Change type to hash_table. Update dependent calls and types.
7409 * config/t-sol2: Update for above.
7411 2013-05-29 Teresa Johnson <tejohnson@google.com>
7413 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
7414 functions are not yet marked as defined.
7416 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
7417 Pat Haugen <pthaugen@us.ibm.com>
7418 Peter Bergner <bergner@vnet.ibm.com>
7420 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
7424 (vrotl<mode>3): Likewise.
7425 (vashl<mode>3): Likewise.
7426 (vlshr<mode>3): Likewise.
7427 (vashr<mode>3): Likewise.
7429 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
7430 support for power8 V2DI builtins.
7432 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
7433 power8 V2DI builtins.
7434 (vupkhsw): Likewise.
7435 (vupklsw): Likewise.
7436 (vaddudm): Likewise.
7441 (vpkudum): Likewise.
7442 (vpksdss): Likewise.
7443 (vpkudus): Likewise.
7444 (vpksdus): Likewise.
7449 (vsubudm): Likewise.
7450 (vcmpequd): Likewise.
7451 (vcmpgtsd): Likewise.
7452 (vcmpgtud): Likewise.
7453 (vcmpequd_p): Likewise.
7454 (vcmpgtsd_p): Likewise.
7455 (vcmpgtud_p): Likewise.
7456 (vupkhsw): Likewise.
7457 (vupklsw): Likewise.
7458 (vaddudm): Likewise.
7463 (vpksdss): Likewise.
7464 (vpksdus): Likewise.
7465 (vpkudum): Likewise.
7466 (vpkudus): Likewise.
7471 (vsubudm): Likewise.
7473 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
7474 support for power8 V2DI instructions.
7476 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
7477 power8 V2DI instructions. Combine pack and unpack insns to use an
7478 iterator for each mode. Check whether a particular mode supports
7479 Altivec instructions instead of just checking TARGET_ALTIVEC.
7480 (UNSPEC_VPKUWUM): Likewise.
7481 (UNSPEC_VPKSHSS): Likewise.
7482 (UNSPEC_VPKSWSS): Likewise.
7483 (UNSPEC_VPKUHUS): Likewise.
7484 (UNSPEC_VPKSHUS): Likewise.
7485 (UNSPEC_VPKUWUS): Likewise.
7486 (UNSPEC_VPKSWUS): Likewise.
7487 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
7488 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
7489 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
7490 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
7491 (UNSPEC_VUPKHSB): Likewise.
7492 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
7493 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
7494 (UNSPEC_VUPKHSH): Likewise.
7495 (UNSPEC_VUPKLSB): Likewise.
7496 (UNSPEC_VUPKLSH): Likewise.
7498 (VI_char): Likewise.
7499 (VI_scalar): Likewise.
7500 (VI_unit): Likewise.
7502 (VP_small): Likewise.
7503 (VP_small_lc): Likewise.
7504 (VU_char): Likewise.
7505 (add<mode>3): Likewise.
7506 (altivec_vaddcuw): Likewise.
7507 (altivec_vaddu<VI_char>s): Likewise.
7508 (altivec_vadds<VI_char>s): Likewise.
7509 (sub<mode>3): Likewise.
7510 (altivec_vsubcuw): Likewise.
7511 (altivec_vsubu<VI_char>s): Likewise.
7512 (altivec_vsubs<VI_char>s): Likewise.
7513 (altivec_vavgs<VI_char>): Likewise.
7514 (altivec_vcmpbfp): Likewise.
7515 (altivec_eq<mode>): Likewise.
7516 (altivec_gt<mode>): Likewise.
7517 (altivec_gtu<mode>): Likewise.
7518 (umax<mode>3): Likewise.
7519 (smax<mode>3): Likewise.
7520 (umin<mode>3): Likewise.
7521 (smin<mode>3): Likewise.
7522 (altivec_vpkuhum): Likewise.
7523 (altivec_vpkuwum): Likewise.
7524 (altivec_vpkshss): Likewise.
7525 (altivec_vpkswss): Likewise.
7526 (altivec_vpkuhus): Likewise.
7527 (altivec_vpkshus): Likewise.
7528 (altivec_vpkuwus): Likewise.
7529 (altivec_vpkswus): Likewise.
7530 (altivec_vpks<VI_char>ss): Likewise.
7531 (altivec_vpks<VI_char>us): Likewise.
7532 (altivec_vpku<VI_char>us): Likewise.
7533 (altivec_vpku<VI_char>um): Likewise.
7534 (altivec_vrl<VI_char>): Likewise.
7535 (altivec_vsl<VI_char>): Likewise.
7536 (altivec_vsr<VI_char>): Likewise.
7537 (altivec_vsra<VI_char>): Likewise.
7538 (altivec_vsldoi_<mode>): Likewise.
7539 (altivec_vupkhsb): Likewise.
7540 (altivec_vupkhs<VU_char>): Likewise.
7541 (altivec_vupkls<VU_char>): Likewise.
7542 (altivec_vupkhsh): Likewise.
7543 (altivec_vupklsb): Likewise.
7544 (altivec_vupklsh): Likewise.
7545 (altivec_vcmpequ<VI_char>_p): Likewise.
7546 (altivec_vcmpgts<VI_char>_p): Likewise.
7547 (altivec_vcmpgtu<VI_char>_p): Likewise.
7548 (abs<mode>2): Likewise.
7549 (vec_unpacks_hi_v16qi): Likewise.
7550 (vec_unpacks_hi_v8hi): Likewise.
7551 (vec_unpacks_lo_v16qi): Likewise.
7552 (vec_unpacks_hi_<VP_small_lc>): Likewise.
7553 (vec_unpacks_lo_v8hi): Likewise.
7554 (vec_unpacks_lo_<VP_small_lc>): Likewise.
7555 (vec_pack_trunc_v8h): Likewise.
7556 (vec_pack_trunc_v4si): Likewise.
7557 (vec_pack_trunc_<mode>): Likewise.
7559 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
7561 (vec_vmaxsd): Likewise.
7562 (vec_vmaxud): Likewise.
7563 (vec_vminsd): Likewise.
7564 (vec_vminud): Likewise.
7565 (vec_vpksdss): Likewise.
7566 (vec_vpksdus): Likewise.
7567 (vec_vpkudum): Likewise.
7568 (vec_vpkudus): Likewise.
7569 (vec_vrld): Likewise.
7570 (vec_vsld): Likewise.
7571 (vec_vsrad): Likewise.
7572 (vec_vsrd): Likewise.
7573 (vec_vsubudm): Likewise.
7574 (vec_vupkhsw): Likewise.
7575 (vec_vupklsw): Likewise.
7577 2013-05-29 Jan Hubicka <jh@suse.cz>
7579 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
7580 flags; reorder rest of fields in more consistent way.
7581 (varpool_node): Remove analyzed, finalized and alias.
7582 (cgraph_ndoe): Likewise.
7583 (symtab_alias_ultimate_target): New function.
7584 (cgraph_function_node): Move offline.
7585 (cgraph_reset_node): Declare.
7586 (cgraph_comdat_can_be_unshared_p): Remove.
7587 (varpool_remove_initializer): Declare.
7588 (varpool_first_defined_variable, varpool_next_defined_variable
7589 cgraph_first_defined_function, cgraph_next_defined_function): Update.
7590 (cgraph_function_with_gimple_body_p): Update.
7591 (varpool_all_refs_explicit_p): Update.
7592 (symtab_alias_target): New function.
7593 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
7594 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
7595 (cgraph_function_or_thunk_node): Simplify using
7596 symtab_alias_ultimate_target.
7597 (varpool_variable_node): Likewise.
7598 * cgraph.c (cgraph_create_function_alias): Update.
7599 (cgraph_add_thunk): Update.
7600 (cgraph_remove_node): Update.
7601 (dump_cgraph_node): Do not dump removed flags.
7602 (cgraph_function_body_availability): Update.
7603 (cgraph_propagate_frequency): Update.
7604 (verify_cgraph_node): Check sanity of local flag.
7605 (cgraph_function_node): Move here from cgraph.h; revamp for
7606 cgraph_function_or_thunk_node.
7607 * lto-symtab.c (lto_varpool_replace_node): Update.
7608 (lto_symtab_resolve_can_prevail_p): Update.
7609 (lto_symtab_merge_cgraph_nodes): Update.
7610 * ipa-cp.c (determine_versionability, initialize_node_lattices,
7611 propagate_constants_accross_call, devirtualization_time_bonus,
7612 ipcp_propagate_stage): Update.
7613 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
7614 * ipa-inline-transform.c (clone_inlined_nodes,
7615 preserve_function_body_p): Update.
7616 * ipa-reference.c (propagate): Update.
7617 (write_node_summary_p): Update.
7618 * toplev.c (wrapup_global_declaration_2): Update.
7619 * cgraphunit.c (cgraph_analyze_function): Rename to ...
7620 (analyze_function) ... this one.
7621 (cgraph_process_new_functions): Update.
7622 (cgraph_reset_node): Export.
7623 (cgraph_finalize_function): Update.
7624 (cgraph_add_new_function): Update.
7625 (process_function_and_variable_attributes): Update.
7626 (varpool_finalize_decl): Update.
7627 (symbol_finalized): Remove.
7628 (symbol_finalized_and_needed): Rename to ...
7629 (symbol_defined_and_needed): ... update.
7630 (cgraph_analyze_functions): Update.
7631 (handle_alias_pairs): Update.
7632 (mark_functions_to_output): Update.
7633 (assemble_thunk): Update.
7634 (output_in_order): Update.
7635 (output_weakrefs): Update.
7636 (finalize_compilation_unit): Update.
7637 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
7638 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
7639 input_node, input_varpool_node): Update.
7640 * dbxout.c (dbxout_expand_expr): Update.
7641 * cgraphclones.c (cgraph_clone_node): Update.
7642 (cgraph_copy_node_for_versioning): Update.
7643 (cgraph_materialize_clone): Update.
7644 (cgraph_materialize_all_clones): Update.
7645 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
7646 propagate_pure_const, propagate_nothrow): Update.
7647 * lto-streamer-out.c (lto_output, write_symbol): Update.
7648 * ipa-utils.c (ipa_reverse_postorder): Update.
7649 * ipa-inline.c (can_inline_edge_p): Update.
7650 (update_caller_keys, ipa_inline): Update.
7651 * dwarf2out.c (reference_to_unused,
7652 premark_types_used_by_global_vars_helper): Update.
7653 * tree-eh.c (tree_could_trap_p): Update.
7654 * ipa-split.c (consider_split, execute_split_functions): Update.
7655 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
7656 has_addr_references_p): Update; move ahead in file for better
7658 (process_references): Simplify.
7659 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
7661 (cgraph_comdat_can_be_unshared_p): Make static.
7662 (cgraph_externally_visible_p): Update.
7663 (varpool_externally_visible_p): Update.
7664 (function_and_variable_visibility): Update.
7665 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
7666 ipa_tm_mark_force_output_node): Update.
7667 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
7668 estimate_edge_devirt_benefit, inline_generate_summary,
7669 inline_write_summary): Update.
7670 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
7671 * ipa-prop.c (ipa_compute_jump_functions): Update.
7672 (ipa_print_node_params, ipa_prop_read_section,
7673 ipa_update_after_lto_read, read_replacements_section): Update.
7674 * varasm.c (mark_decl_referenced): Update.
7675 (assemble_alias, dump_tm_clone_pairs): Update.
7676 * tree-inline.c (copy_bb): Update.
7677 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
7679 * symtab.c (dump_symtab_base): Print new flags.
7680 (verify_symtab_base): Verify new flags.
7681 (symtab_alias_ultimate_target): New function.
7682 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
7683 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
7685 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
7687 * i386.c (ix86_get_function_versions_dispatcher,
7688 ix86_generate_version_dispatcher_body): Update.
7689 (fold_builtin_cpu): Use varpool_add_new_variable.
7690 * varpool.c (varpool_remove_initializer): Break out from ...
7691 (varpool_remove_node): ... this one.
7692 (dump_varpool_node, varpool_node_for_asm,
7693 cgraph_variable_initializer_availability, varpool_analyze_node,
7694 varpool_assemble_decl, varpool_remove_unreferenced_decls,
7695 varpool_finalize_named_section_flags, varpool_create_variable_alias):
7698 2013-05-29 Jan Hubicka <jh@suse.cz>
7700 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
7702 2013-05-29 Easwaran Raman <eraman@google.com>
7704 PR tree-optimization/57442
7705 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
7706 when control exits the main loop.
7708 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
7710 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
7712 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
7713 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
7714 * rx/t-rx: Add rx100 under multi library matches option for nofpu
7717 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7719 PR tree-optimization/57441
7720 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
7721 Don't limit size of incr_vec to number of candidates.
7723 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
7725 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
7726 and mips16 directories.
7727 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
7728 (MULTILIB_DIRNAMES): Ditto.
7729 (MULTILIB_EXCEPTIONS): Add new exceptions.
7730 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
7731 (MULTILIB_DIRNAMES): Ditto.
7732 (MULTILIB_EXCEPTIONS): Add new exceptions.
7734 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
7735 Marcus Shawcroft <marcus.shawcroft@arm.com>
7737 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
7738 SYMBOL_TINY_ABSOLUTE.
7739 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
7740 SYMBOL_TINY_ABSOLUTE.
7741 (aarch64_expand_mov_immediate): Likewise.
7742 (aarch64_classify_symbol): Likewise.
7743 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
7744 Permit SYMBOL_TINY_ABSOLUTE.
7745 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
7747 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
7748 Marcus Shawcroft <marcus.shawcroft@arm.com>
7750 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
7751 Refactor if/switch. Replace gcc_assert with if.
7753 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
7755 * config/i386/i386.c (initial_ix86_tune_features): Enable
7756 FP Reassociation for AMD bdver1 and bdver2.
7758 2013-05-29 Martin Jambor <mjambor@suse.cz>
7760 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
7761 and IMAGPART_EXPR do not occur within other handled_components.
7763 2013-05-29 Richard Biener <rguenther@suse.de>
7765 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
7766 access on whether the use is in the BB we currently try to
7768 (vect_bb_vectorization_profitable_p): Pass the BB we currently
7769 vectorize to vect_bb_slp_scalar_cost.
7771 2013-05-29 Richard Biener <rguenther@suse.de>
7773 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
7774 computing scalar cost offsetted by stmts that are kept live
7776 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
7777 for computation of scalar cost.
7779 2013-05-28 Steve Ellcey <sellcey@mips.com>
7781 * config/mips/mips-cpus.def (mips32r2): Change processor type.
7783 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
7785 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
7786 array notation built-in reduction functions.
7787 * doc/passes.texi (Passes): Added documentation about changes done
7789 * doc/invoke.texi (C Dialect Options): Added documentation about
7790 the -fcilkplus flag.
7791 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
7792 (BUILTINS_DEF): Depend on cilkplus.def.
7793 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
7794 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
7795 * cilkplus.def: New file.
7797 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
7799 PR rtl-optimization/57439
7800 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
7802 2013-05-28 Easwaran Raman <eraman@google.com>
7804 PR tree-optimization/57337
7805 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
7806 (find_insert_point): Correctly identify the insertion point
7807 when two statements with the same UID is compared.
7809 2013-05-28 Richard Biener <rguenther@suse.de>
7811 PR tree-optimization/56787
7812 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
7813 from the list of data references.
7814 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
7816 (vect_analyze_loop_operations): Likewise.
7817 (vect_transform_loop): Remove clobbers.
7819 2013-05-28 Martin Jambor <mjambor@suse.cz>
7821 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
7822 and REALPART_EXPRs have scalar type.
7824 2013-05-28 Richard Biener <rguenther@suse.de>
7826 PR tree-optimization/57411
7827 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
7829 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
7830 virtual operand propagation.
7832 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
7834 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
7835 destination register for bmasksi_vis.
7836 (vector_init_bshuffle): Likewise.
7837 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
7839 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
7841 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
7842 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
7843 mode if the instruction isn't available in the original mode.
7844 * config/sparc/sparc.opt (mfix-ut699): New option.
7845 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
7846 (divdf3): Turn into expander.
7847 (divdf3_nofix): New insn.
7848 (divdf3_fix): Likewise.
7849 (divsf3): Disable if -mfix-ut699.
7850 (sqrtdf2): Turn into expander.
7851 (sqrtdf2_nofix): New insn.
7852 (sqrtdf2_fix): Likewise.
7853 (sqrtsf2): Disable if -mfix-ut699.
7855 2013-05-27 Richard Biener <rguenther@suse.de>
7858 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
7859 block for the new loop.
7861 2013-05-27 Richard Biener <rguenther@suse.de>
7863 PR tree-optimization/57343
7864 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
7865 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
7866 (number_of_iterations_cond): Do not build the folded tree.
7868 2013-05-27 Richard Biener <rguenther@suse.de>
7872 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
7873 OEP_CONSTANT_ADDRESS_OF retained.
7875 PR tree-optimization/57417
7876 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
7878 (set_ssa_val_to): Compare addresses using
7879 get_addr_base_and_unit_offset.
7881 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
7883 PR rtl-optimization/56833
7884 * postreload.c (move2add_record_mode): New function.
7885 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
7886 (move2add_use_add2_insn): Use move2add_record_sym_value.
7887 (move2add_use_add3_insn): Likewise.
7888 (reload_cse_move2add): Use move2add_valid_value_p and
7889 move2add_record_mode. Invalidate call-clobbered and REG_INC
7890 affected regs by setting reg_mode to VOIDmode.
7891 (move2add_note_store): Don't pretend the inside of a SUBREG is
7892 the actual destination. Invalidate single/leading registers by
7893 setting reg_mode to VOIDmode.
7894 Use move2add_record_sym_value, move2add_valid_value_p and
7895 move2add_record_mode.
7897 2013-05-27 Richard Biener <rguenther@suse.de>
7899 PR tree-optimization/57396
7900 * tree-affine.c (double_int_constant_multiple_p): Properly
7901 return false for val == 0 and div != 0.
7903 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
7905 * config/mips/mips.h: Use #elif in preprocessor conditions.
7907 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
7910 * config/mips/constraints.md (kl): New constraint.
7911 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
7912 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
7913 constraint for operand 0. Split after CSE for MIPS16. Emit a move
7915 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
7917 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
7920 * config/mips/mips.c (mips_can_inline_p): New function.
7921 (TARGET_CAN_INLINE_P): Define.
7923 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
7925 * sched-int.h (ds_t, dw_t): Make unsigned int.
7926 Fix documentation that describes how all the ds_t bits are used.
7927 Reserve the last bit for delayed-branch scheduling.
7928 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
7929 (BITS_PER_DEP_WEAK): Fix definition and documentation.
7930 (gen_dep_weak_1): Remove prototype.
7931 * sched-deps.c (get_dep_weak_1): Make static.
7932 * target.def (speculate_insn, needs_block_p, gen_spec_check,
7933 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
7934 * doc/tm.texi: Regenerate.
7935 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
7937 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
7940 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
7942 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
7943 Sandra Loosemore <sandra@codesourcery.com>
7945 * config.gcc (powerpc-*): Allow native for with-cpu.
7947 2013-05-24 Jeff Law <law@redhat.com>
7949 PR tree-optimization/57124
7950 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
7951 conversion feeding a condition if the range has an overflow
7952 if -fstrict-overflow. Add warnings for when we do make the
7955 2013-05-24 Dehao Chen <dehao@google.com>
7957 * tree-cfg.c (locus_discrim_map): Fix the typo.
7958 (locus_discrim_hasher): Likewise.
7959 (locus_discrim_hasher::hash): Likewise.
7960 (locus_discrim_hasher::equal): Likewise.
7962 2013-05-24 Martin Jambor <mjambor@suse.cz>
7964 PR tree-optimization/57294
7965 * cgraph.h (ipa_record_stmt_references): Declare.
7966 * cgraphbuild.c (ipa_record_stmt_references): New function.
7967 (build_cgraph_edges): Use ipa_record_stmt_references.
7968 (rebuild_cgraph_edges): Likewise.
7969 (cgraph_rebuild_references): Likewise.
7970 * ipa-prop.c (ipa_modify_call_arguments): Discard references
7971 associated with the old statement and build references from the
7972 newly built statements.
7973 * ipa-ref.c (ipa_remove_stmt_references): New function.
7974 * ipa-ref.h (ipa_remove_stmt_references): Declare.
7976 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
7978 * lra-constraints.c (emit_spill_move): Use smaller mode for
7980 (check_and_process_move): Consider mem-reg moves for secondary
7982 (curr_insn_transform): Don't lose insns emitted before for
7983 secondary memory moves.
7984 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
7985 reg set up in the current insn.
7987 2013-05-24 Dehao Chen <dehao@google.com>
7989 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
7991 (locus_descrim_hasher::equal): Likewise.
7992 (build_gimple_cfg): New discriminator assignment algorithm.
7993 (make_edges): Likewise.
7994 (next_discriminator_for_locus): Likewise.
7995 (same_line_p): Likewise.
7996 (assign_discriminators): Likewise.
7997 (make_cond_expr_edges): Likewise.
7998 (make_gimple_switch_edges): Likewise.
7999 (make_goto_expr_edges): Likewise.
8000 (make_gimple_asm_edges): Likewise.
8002 2013-05-24 Ian Bolton <ian.bolton@arm.com>
8004 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
8005 X format specifier to only display bottom 16 bits.
8006 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
8007 immediate to match for operand 2, since it will be masked.
8009 2013-05-24 Richard Biener <rguenther@suse.de>
8011 PR tree-optimization/57287
8012 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
8013 all SSA names that occur in abnormal PHIs.
8015 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
8018 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
8019 that index is not negative.
8021 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
8023 PR rtl-optimization/55177
8024 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
8025 (simplify_byte_swapping_operation): New.
8026 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
8027 (simplify_relational_operation_1): Deal with BSWAP.
8029 2013-05-23 Richard Henderson <rth@redhat.com>
8032 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
8033 (ix86_reorg): Call it.
8035 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
8038 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
8039 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
8040 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
8042 2013-05-23 Christian Bruel <christian.bruel@st.com>
8045 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
8047 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
8048 Marcus Shawcroft <marcus.shawcroft@arm.com>
8050 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
8051 * config/aarch64/constraints.md (Usa): Remove.
8052 * doc/md.texi (AArch64 Usa): Remove.
8054 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
8055 Marcus Shawcroft <marcus.shawcroft@arm.com>
8057 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
8058 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
8059 * config/aarch64/predicates.md (aarch64_const_address): Remove.
8060 (aarch64_mov_operand): Use aarch64_mov_operand_p.
8062 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
8064 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
8065 instruction (AdvSIMD).
8066 * config/aarch64/aarch64-builtins.c
8067 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
8068 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
8070 2013-05-23 Martin Jambor <mjambor@suse.cz>
8073 * tree.h (contains_bitfld_component_ref_p): Declare.
8074 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
8075 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
8077 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
8078 not access a bit-field. Assert all final offsets are byte-aligned.
8080 2013-05-23 Richard Biener <rguenther@suse.de>
8082 PR tree-optimization/57380
8083 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
8084 least one invariant or re-used load.
8085 * passes.c (init_optimization_passes): Move pass_phiprop before
8088 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
8090 * config/aarch64/aarch64-simd.md
8091 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
8093 2013-05-23 Richard Biener <rguenther@suse.de>
8096 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
8097 OEP_CONSTANT_ADDRESS_OF retained.
8099 2013-05-23 Jakub Jelinek <jakub@redhat.com>
8102 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
8103 don't lower unit. Handle unit not being always BITS_PER_WORD.
8105 2013-05-23 Richard Biener <rguenther@suse.de>
8107 PR rtl-optimization/57341
8108 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
8109 instead of true_dependence.
8111 2013-05-22 David Malcolm <dmalcolm@redhat.com>
8113 * bb-reorder.c (branch_threshold): Make const.
8114 (exec_threshold): Ditto.
8116 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
8117 Pat Haugen <pthaugen@us.ibm.com>
8118 Peter Bergner <bergner@vnet.ibm.com>
8120 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
8121 documentation for the power8 crypto builtins.
8123 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
8125 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
8126 macros for defining power8 builtin functions.
8127 (BU_P8V_AV_2): Likewise.
8128 (BU_P8V_AV_P): Likewise.
8129 (BU_P8V_VSX_1): Likewise.
8130 (BU_P8V_OVERLOAD_1): Likewise.
8131 (BU_P8V_OVERLOAD_2): Likewise.
8132 (BU_CRYPTO_1): Likewise.
8133 (BU_CRYPTO_2): Likewise.
8134 (BU_CRYPTO_3): Likewise.
8135 (BU_CRYPTO_OVERLOAD_1): Likewise.
8136 (BU_CRYPTO_OVERLOAD_2): Likewise.
8137 (XSCVSPDP): Fix typo, point to the correct instruction.
8138 (VCIPHER): Add power8 crypto builtins.
8139 (VCIPHERLAST): Likewise.
8140 (VNCIPHER): Likewise.
8141 (VNCIPHERLAST): Likewise.
8142 (VPMSUMB): Likewise.
8143 (VPMSUMH): Likewise.
8144 (VPMSUMW): Likewise.
8145 (VPERMXOR_V2DI): Likewise.
8146 (VPERMXOR_V4SI: Likewise.
8147 (VPERMXOR_V8HI: Likewise.
8148 (VPERMXOR_V16QI: Likewise.
8149 (VSHASIGMAW): Likewise.
8150 (VSHASIGMAD): Likewise.
8152 (VPERMXOR): Likewise.
8153 (VSHASIGMA): Likewise.
8155 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
8156 __CRYPTO__ if the crypto instructions are available.
8157 (altivec_overloaded_builtins): Add support for overloaded power8
8160 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
8161 support for power8 crypto builtins.
8162 (builtin_function_type): Likewise.
8163 (altivec_init_builtins): Add support for builtins that take vector
8164 long long (V2DI) arguments.
8166 * config/rs6000/crypto.md: New file, define power8 crypto
8169 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
8170 Pat Haugen <pthaugen@us.ibm.com>
8171 Peter Bergner <bergner@vnet.ibm.com>
8173 * doc/invoke.texi (Option Summary): Add power8 options.
8174 (RS/6000 and PowerPC Options): Likewise.
8176 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
8177 constraints.md instead of rs6000.h. Reorder w* constraints. Add
8178 wm, wn, wr documentation.
8180 * gcc/config/rs6000/constraints.md (wm): New constraint for VSX
8181 registers if direct move instructions are enabled.
8182 (wn): New constraint for no registers.
8183 (wq): New constraint for quad word even GPR registers.
8184 (wr): New constraint if 64-bit instructions are enabled.
8185 (wv): New constraint if power8 vector instructions are enabled.
8186 (wQ): New constraint for quad word memory locations.
8188 * gcc/config/rs6000/predicates.md (const_0_to_15_operand): New
8189 constraint for 0..15 for crypto instructions.
8190 (gpc_reg_operand): If VSX allow registers in VSX registers as well
8191 as GPR and floating point registers.
8192 (int_reg_operand): New predicate to match only GPR registers.
8193 (base_reg_operand): New predicate to match base registers.
8194 (quad_int_reg_operand): New predicate to match even GPR registers
8195 for quad memory operations.
8196 (vsx_reg_or_cint_operand): New predicate to allow vector logical
8197 operations in both GPR and VSX registers.
8198 (quad_memory_operand): New predicate for quad memory operations.
8199 (reg_or_indexed_operand): New predicate for direct move support.
8201 * gcc/config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
8202 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
8203 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
8204 (POWERPC_MASKS): Add power8 options.
8205 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
8208 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
8209 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
8211 * gcc/config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
8212 (-mpower8-fusion): New power8 options.
8213 (-mpower8-fusion-sign): Likewise.
8214 (-mpower8-vector): Likewise.
8215 (-mcrypto): Likewise.
8216 (-mdirect-move): Likewise.
8217 (-mquad-memory): Likewise.
8219 * gcc/config/rs6000/rs6000.c (power8_cost): Initial definition for
8221 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
8223 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
8224 (rs6000_debug_vector_unit): Add p8_vector.
8225 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
8226 definitions. Also print fusion state.
8227 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
8228 (rs6000_builtin_mask_calculate): Add power8 builtin support.
8229 (rs6000_option_override_internal): Add support for power8.
8230 (rs6000_common_init_builtins): Add debugging for skipped builtins
8232 (rs6000_adjust_cost): Add power8 support.
8233 (rs6000_issue_rate): Likewise.
8234 (insn_must_be_first_in_group): Likewise.
8235 (insn_must_be_last_in_group): Likewise.
8236 (force_new_group): Likewise.
8237 (rs6000_register_move_cost): Likewise.
8238 (rs6000_opt_masks): Likewise.
8240 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
8241 power8 capable assembler, default to power7 options.
8242 (TARGET_DIRECT_MOVE): Likewise.
8243 (TARGET_CRYPTO): Likewise.
8244 (TARGET_P8_VECTOR): Likewise.
8245 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
8246 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
8247 (VECTOR_MEM_P8_VECTOR_P): Likewise.
8248 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
8249 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
8250 (TARGET_XSCVDPSPN): Likewise.
8251 (TARGET_XSCVSPDPN): Likewsie.
8252 (TARGET_SYNC_HI_QI): Likewise.
8253 (TARGET_SYNC_TI): Likewise.
8254 (MASK_CRYPTO): Likewise.
8255 (MASK_DIRECT_MOVE): Likewise.
8256 (MASK_P8_FUSION): Likewise.
8257 (MASK_P8_VECTOR): Likewise.
8258 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
8259 temporary used by some of the direct move instructions to get two FP
8260 temporary registers does not force creation of a stack frame.
8261 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
8262 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
8263 that any VSX registers are tieable, even if they are also an
8264 Altivec vector mode.
8265 (r6000_reg_class_enum): Add wm, wr, wv constraints.
8266 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
8267 (RS6000_BTM_CRYPTO): Likewise.
8268 (RS6000_BTM_COMMON): Likewise.
8270 * config/rs6000/rs6000.md (cpu attribute): Add power8.
8271 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
8272 (enum rs6000_vector): Add power8 vector support.
8274 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
8278 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
8279 (any_sibcall_could_use_r3): this and handle indirect calls.
8280 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
8282 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8284 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
8286 2013-05-22 Richard Biener <rguenther@suse.de>
8289 * profile.c (branch_prob): Do not split blocks that are
8290 abnormally receiving from ECF_RETURNS_TWICE functions.
8292 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
8294 * recog.c (offsettable_address_addr_space_p): Fix calculation of
8295 address mode. Move pointer mode initialization to the same place.
8297 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
8299 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
8300 while it has any effect.
8302 2013-05-21 Easwaran Raman <eraman@google.com>
8304 PR tree-optimization/57322
8305 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
8306 UID of the statement added to the BB to be 1.
8308 2013-05-21 Jakub Jelinek <jakub@redhat.com>
8310 PR tree-optimization/57331
8311 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
8312 of conversion from pointer type to integral type with integer.
8314 2013-05-21 Martin Jambor <mjambor@suse.cz>
8317 * ipa-prop.c (ipa_read_node_info): Process param_used and
8318 controlled_uses in the same order as when writing.
8320 2013-05-21 Magnus Granberg <baldrick@free.fr>
8323 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
8325 2013-05-21 Richard Biener <rguenther@suse.de>
8327 PR tree-optimization/57318
8328 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
8329 estimate stmts with side-effects as likely eliminated.
8331 2013-05-21 Richard Biener <rguenther@suse.de>
8333 PR tree-optimization/57330
8334 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
8335 preserve the call stmts fntype.
8337 2013-05-21 Richard Biener <rguenther@suse.de>
8339 PR tree-optimization/57303
8340 * tree-ssa-sink.c (statement_sink_location): Improve killing
8341 stmt detection and properly handle self-assignments.
8343 2013-05-21 Christian Bruel <christian.bruel@st.com>
8345 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
8346 spanning registers. LEAF_REG_REMAP is supported only for contiguous
8347 registers. Set register size out of the PARALLEL loop.
8349 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
8352 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
8353 (*fmasf4, *fmasf4_media): New insns.
8355 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
8357 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
8358 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
8359 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
8360 (mips_idiv_insns): Update the comments to say that the returned
8361 instruction counts are in units of BASE_INSN_LENGTH.
8362 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
8363 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
8364 using 2 rather than 4 as the length of indirect MIPS16 and
8365 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
8366 length of a NOP. Don't divide MIPS16 lengths by 2.
8367 (mips16_split_long_branches): Assume a branch is long if the
8368 length is greater than 4 rather than 8.
8369 * config/mips/mips.md (length): Give MIPS16 lengths directly,
8370 rather than multiplying them by 2. Multiply instruction counts
8371 by BASE_INSN_LENGTH rather than 4.
8372 (*jump_mips16, tls_get_tp_mips16_<mode>)
8373 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
8375 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
8377 * config/mips/mips.md (extended_mips16): Remove branch case.
8378 (length): Remove duplicated extended_mips16 test.
8380 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
8382 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
8384 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
8386 * recog.h (Recog_data): Rename to...
8387 (recog_data_d): ...this.
8388 (recog_data): Update accordingly.
8389 * recog.c (recog_data): Likewise.
8390 * reload.c (save_recog_data): Likewise.
8391 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
8392 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
8394 2013-05-17 Julian Brown <julian@codesourcery.com>
8396 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
8397 found in a REG_EQUAL note, invalidate it.
8399 2013-05-17 Easwaran Raman <eraman@google.com>
8401 * tree-ssa-reassoc.c (find_insert_point): New function.
8402 (insert_stmt_after): Likewise.
8403 (get_def_stmt): Likewise.
8404 (ensure_ops_are_available): Likewise.
8405 (not_dominated_by): Likewise.
8406 (rewrite_expr_tree): Do not move statements beyond what is
8407 necessary. Remove call to swap_ops_for_binary_stmt...
8408 (reassociate_bb): ... and move it here.
8409 (build_and_add_sum): Assign UIDs for new statements.
8410 (linearize_expr): Likewise.
8411 (do_reassoc): Renumber gimple statement UIDs.
8413 2013-05-17 Jan Hubicka <jh@suse.cz>
8415 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
8417 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
8418 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
8420 (output_weakrefs): Update.
8422 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
8423 Martin Jambor <mjambor@suse.cz>
8426 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
8427 value that corresponds to the given aggval is found in values vector.
8429 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
8431 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
8432 sse, sse2, sse3, ssse3 and sse4a flags to options.
8434 2013-05-17 David Malcolm <dmalcolm@redhat.com>
8436 * gengtype-state.c: (s_expr_writer): New class, to handle
8437 prettifying of output layout of s-expressions.
8438 (state_writer): New class, to write out gtype.state.
8439 (state_written_type_count): Move this variable into member data of
8441 (s_expr_writer::s_expr_writer): New code: constructor for new class
8442 (state_writer::state_writer(): ditto
8443 (s_expr_writer::write_new_line): New function
8444 (s_expr_writer::write_any_indent): ditto
8445 (s_expr_writer::begin_s_expr): ditto
8446 (s_expr_writer::end_s_expr): ditto
8447 (write_state_fileloc): convert to method of state_writer...
8448 (state_writer:: write_state_fileloc): ...and use methods of
8449 s_expr_writer to write indentation into the gtype.state output file
8450 to visually represent the hierarchical structure of the list
8452 (write_state_fields): ditto, renaming to...
8453 (state_writer::write_state_fields)
8454 (write_state_a_string): ditto, renaming to...
8455 (state_writer::write_state_a_string)
8456 (write_state_string_option): ditto, renaming to...
8457 (state_writer::write_state_string_option)
8458 (write_state_type_option): ditto, renaming to...
8459 (state_writer::write_state_type_option)
8460 (write_state_nested_option): ditto, renaming to...
8461 (state_writer::write_state_nested_option)
8462 (write_state_option): ditto, renaming to...
8463 (state_writer::write_state_option)
8464 (write_state_options): ditto, renaming to...
8465 (state_writer::write_state_options)
8466 (write_state_lang_bitmap): ditto, renaming to...
8467 (state_writer::write_state_lang_bitmap)
8468 (write_state_version): ditto, renaming to...
8469 (state_writer::write_state_version)
8470 (write_state_scalar_type): ditto, renaming to...
8471 (state_writer::write_state_scalar_type)
8472 (write_state_string_type): ditto, renaming to...
8473 (state_writer::write_state_string_type)
8474 (write_state_undefined_type): ditto, renaming to...
8475 (state_writer::write_state_undefined_type)
8476 (write_state_struct_union_type): ditto, renaming to...
8477 (state_writer::write_state_struct_union_type)
8478 (write_state_struct_type): ditto, renaming to...
8479 (state_writer::write_state_struct_type)
8480 (write_state_user_struct_type): ditto, renaming to...
8481 (state_writer::write_state_user_struct_type)
8482 (write_state_lang_struct_type): ditto, renaming to...
8483 (state_writer::write_state_lang_struct_type)
8484 (write_state_param_struct_type): ditto, renaming to...
8485 (state_writer::write_state_param_struct_type)
8486 (write_state_pointer_type): ditto, renaming to...
8487 (state_writer::write_state_pointer_type)
8488 (write_state_array_type): ditto, renaming to...
8489 (state_writer::write_state_array_type)
8490 (write_state_gc_used): ditto, renaming to...
8491 (state_writer::write_state_gc_used)
8492 (write_state_common_type_content): ditto, renaming to...
8493 (state_writer::write_state_common_type_content)
8494 (write_state_type): ditto, renaming to...
8495 (state_writer::write_state_type)
8496 (write_state_pair_list): ditto, renaming to...
8497 (state_writer::write_state_pair_list)
8498 (write_state_pair): ditto, renaming to...
8499 (state_writer::write_state_pair)
8500 (write_state_typedefs): ditto, renaming to...
8501 (state_writer::write_state_typedefs)
8502 (write_state_structures): ditto, renaming to...
8503 (state_writer::write_state_structures)
8504 (write_state_param_structs): ditto, renaming to...
8505 (state_writer::write_state_param_structs)
8506 (write_state_variables): ditto, renaming to...
8507 (state_writer::write_state_variables)
8508 (write_state_srcdir): ditto, renaming to...
8509 (state_writer::write_state_srcdir)
8510 (write_state_files_list): ditto, renaming to...
8511 (state_writer::write_state_files_list)
8512 (write_state_languages): ditto, renaming to...
8513 (state_writer::write_state_languages)
8514 (write_state): create a state_writer instance and use it when
8515 writing out the state file
8517 2013-05-17 Mike Stump <mikestump@comcast.net>
8519 PR rtl-optimization/57304
8520 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
8521 accessing DF_REF_REAL_LOC.
8523 2013-05-17 Jakub Jelinek <jakub@redhat.com>
8525 PR rtl-optimization/57281
8526 PR rtl-optimization/57300
8527 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
8528 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
8529 what the other splitter did if the registers are dead.
8531 2013-05-17 Richard Biener <rguenther@suse.de>
8533 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
8536 2013-05-17 Jakub Jelinek <jakub@redhat.com>
8538 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
8541 2013-05-17 Marek Polacek <polacek@redhat.com>
8543 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
8544 length when doing non-zero store of storing '\0' to '\0'.
8546 2013-05-17 Jakub Jelinek <jakub@redhat.com>
8548 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
8549 vect_external_def oprnd1 with loop_vinfo, try to emit
8550 optional cast, negation and and stmts on the loop preheader
8551 edge instead of into the pattern def seq.
8553 PR tree-optimization/57051
8554 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
8555 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
8557 2013-05-16 Nick Clifton <nickc@redhat.com>
8559 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
8560 (rl78_is_naked_func): New function.
8561 (rl78_expand_prologue): Skip prologue generation for naked functions.
8562 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
8563 * doc/extend.texi (naked): Add RL78 to the list of processors
8564 that supports this attribute.
8566 2013-05-16 Jeff Law <law@redhat.com>
8568 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
8570 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
8572 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
8573 cache parameters using detect_caches_amd also for CYRIX,
8574 NSC and TM2 signatures.
8576 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
8577 Dzianis Kahanovich <mahatma@eu.by>
8581 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
8582 VIA/Centaur processors and determine their cache parameters
8583 using detect_caches_amd.
8585 2013-05-16 Teresa Johnson <tejohnson@google.com>
8587 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
8588 (rtl_verify_edges): New function.
8589 (rtl_verify_bb_insns): Ditto.
8590 (rtl_verify_bb_pointers): Ditto.
8591 (rtl_verify_bb_insn_chain): Ditto.
8592 (rtl_verify_fallthru): Ditto.
8593 (rtl_verify_bb_layout): Ditto.
8594 (rtl_verify_flow_info_1): Outline checks into new functions.
8595 (rtl_verify_flow_info): Ditto.
8597 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
8599 * cfghooks.c (copy_bbs): Add update_dominance argument.
8600 * cfghooks.h (copy_bbs): Update prototype.
8601 * tree-cfg.c (gimple_duplicate_sese_region):
8602 Add update_dominance argument.
8603 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
8604 * tree-ssa-loop-ch.c (copy_loop_headers): Update
8605 gimple_duplicate_sese_region call.
8606 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
8607 Update copy_bbs call.
8608 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
8609 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
8611 2013-05-16 Jakub Jelinek <jakub@redhat.com>
8613 * tree-vectorizer.h (NUM_PATTERNS): Increment.
8614 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
8615 vect_recog_rotate_pattern.
8616 (vect_recog_rotate_pattern): New function.
8618 2013-05-16 Jason Merrill <jason@redhat.com>
8620 * Makefile.in (LLINKER): New variable.
8621 (mostlyclean): Remove link mutex.
8622 * configure.ac: Handle --enable-link-mutex.
8623 * lock-and-run.sh: New script.
8625 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
8628 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
8631 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8633 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
8635 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
8637 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
8638 * config/arm/arm.c (next_consecutive_mem): New function.
8639 (gen_movmem_ldrd_strd): Likewise.
8640 * config/arm/arm.md (movmemqi): Update condition and code.
8641 (unaligned_loaddi, unaligned_storedi): New patterns.
8643 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8645 * config.gcc: Obsolete *-*-solaris2.9*.
8646 * doc/install.texi (Specific, *-*-solaris2*): Document it.
8648 2013-05-16 Richard Biener <rguenther@suse.de>
8650 * passes.c (init_optimization_passes): Move pass_parallelize_loops
8651 earlier, after GRAPHITE transforms and IV canonicalization.
8653 2013-05-16 Jakub Jelinek <jakub@redhat.com>
8655 * omp-low.c (extract_omp_for_data): For collapsed loops,
8656 if at least one of the loops is known at compile time to
8657 iterate zero times, set count to 0.
8658 (expand_omp_regimplify_p): New function.
8659 (expand_omp_for_generic): For collapsed loops, if at least
8660 one of the loops isn't known to iterate at least once,
8661 add runtime check with setting count to 0.
8662 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
8663 For unsigned types if it isn't known at compile time that
8664 the loop will iterate at least once, add runtime check to bypass
8665 the whole loop if initial condition isn't true.
8667 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
8669 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
8671 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
8674 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
8675 transformations to avoid an infinite loop.
8677 2013-05-16 Marek Polacek <polacek@redhat.com>
8679 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
8681 2013-05-15 Leif Ekblad <leif@rdos.net>
8683 * config/i386/i386.c (ix86_decompose_address): Use
8684 DEFAULT_TLS_SEG_REG to access TLS segment register.
8685 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
8686 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
8687 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
8689 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
8692 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
8693 sibling calls to functions that would normally be lazily bound,
8694 unless $gp is call-clobbered.
8696 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
8698 * config/i386/i386.c (ix86_option_override_internal): Update
8699 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
8700 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
8701 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
8702 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
8704 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
8706 2013-05-15 Andreas Schwab <schwab@suse.de>
8708 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
8709 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
8712 2013-05-15 Teresa Johnson <tejohnson@google.com>
8714 * loop-unroll.c (report_unroll_peel): Check decision before
8715 emitting unroll/peel message.
8717 2013-05-15 Teresa Johnson <tejohnson@google.com>
8719 * function.h (has_bb_partition): New rtl_data flag.
8720 (bb_reorder_complete): Ditto.
8721 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
8722 instead of flag_reorder_blocks_and_partition.
8723 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
8724 with some enhancements.
8725 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
8726 * bb-reorder.c (connect_traces): Check for has_bb_partition
8727 instead of flag_reorder_blocks_and_partition.
8728 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
8729 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
8730 verify_hot_cold_block_grouping.
8731 (partition_hot_cold_basic_blocks): Set has_bb_partition.
8733 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
8736 * config/arm/predicates.md (call_insn_operand): New predicate.
8737 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
8738 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
8739 if insn is not a tail call.
8740 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
8742 * config/arm/arm.h (enum reg_class): New caller save register class.
8743 (REG_CLASS_NAMES): Likewise.
8744 (REG_CLASS_CONTENTS): Likewise.
8745 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
8748 2013-05-15 Richard Biener <rguenther@suse.de>
8750 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
8751 of MSG_OPTIMIZED_LOCATIONS.
8752 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
8753 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
8755 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
8756 of MSG_OPTIMIZED_LOCATIONS.
8757 (execute_vect_slp): Likewise.
8758 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
8759 (vect_create_cond_for_alias_checks): Likewise.
8760 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
8761 (vect_recog_widen_mult_pattern): Likewise.
8762 (vect_recog_widen_sum_pattern): Likewise.
8763 (vect_recog_over_widening_pattern): Likewise.
8764 (vect_recog_widen_shift_pattern): Likewise.
8765 (vect_recog_vector_vector_shift_pattern): Likewise.
8766 (vect_recog_divmod_pattern): Likewise.
8767 (vect_recog_mixed_size_cond_pattern): Likewise.
8768 (vect_recog_bool_pattern): Likewise.
8769 (vect_pattern_recog_1): Likewise.
8771 2013-05-15 Martin Jambor <mjambor@suse.cz>
8773 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
8774 non-functions to builtin_unreachable.
8775 * ipa-inline-transform.c (inline_call): Do not assert estimates were
8776 correct when new direct edges were discovered.
8778 2013-05-15 Martin Jambor <mjambor@suse.cz>
8780 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
8781 header, print symbol order instead of node uid, print more information
8782 about indirect edge targets.
8783 (ipa_make_edge_direct_to_target): Print symbol order instead of node
8785 (ipa_make_edge_direct_to_target): Likewise.
8786 (remove_described_reference): Likewise.
8787 (propagate_controlled_uses): Likewise.
8788 (ipa_print_node_params): Also print symbol order.
8789 (ipcp_transform_function): Print symbol order instead of node uids.
8790 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
8791 (cgraph_get_create_real_symbol_node): Likewise.
8792 * ipa-cp.c (print_lattice): Likewise.
8793 (print_all_lattices): Likewise.
8794 (determine_versionability): Likewise.
8795 (initialize_node_lattices): Likewise.
8796 (estimate_local_effects): Likewise.
8797 (update_profiling_info): Likewise.
8798 (create_specialized_node): Likewise.
8799 (perhaps_add_new_callers): Likewise.
8800 (decide_about_value): Likewise.
8801 (decide_whether_version_node): Likewise.
8802 (identify_dead_nodes): Likewise.
8803 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
8804 (dump_inline_summary): Likewise.
8805 (estimate_node_size_and_time): Likewise.
8806 (inline_analyze_function): Likewise.
8807 * ipa-inline.c (report_inline_failed_reason): Likewise.
8808 (want_early_inline_function_p): Likewise.
8809 (edge_badness): Likewise.
8810 (update_edge_key): Likewise.
8811 (inline_small_functions): Likewise. Add dumping of order to two other
8813 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
8814 instead of node uids.
8815 (propagate_pure_const): Likewise.
8816 (propagate_pure_const): Likewise.
8817 * ipa-utils.c (dump_cgraph_node_set): Likewise.
8818 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
8819 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
8821 * tree-pretty-print.c (dump_function_header): Likewise.
8822 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
8823 Print symbol order instead of node uids.
8825 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8827 * config/s390/s390.c (s390_register_move_cost): Don't impose the
8828 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
8830 2013-05-15 Richard Biener <rguenther@suse.de>
8832 PR tree-optimization/57275
8833 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
8834 return value for fail to do runtime alias checks for gather loads.
8836 2013-05-15 Jan Hubicka <jh@suse.cz>
8840 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
8841 weakrefs are not external.
8842 (lto_symtab_merge_decls): Fix thinko when dealing with
8843 non-lto_symtab decls.
8844 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
8845 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
8846 * varpool.c (dump_varpool_node): Dump more flags.
8848 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
8850 * config/i386/i386.c (processor_alias_table): Add instruction
8851 FSGSBASE for AMD bdver3 architecture.
8853 2013-05-14 Jakub Jelinek <jakub@redhat.com>
8855 * tree.c (warn_deprecated_use): Print file:line using locus color.
8856 * diagnostic.c (diagnostic_report_current_module): Print file:line
8857 and file:line:column using locus color.
8859 2013-05-14 Mike Stump <mikestump@comcast.net>
8861 * gdbinit.in: Add __null.
8863 2013-05-14 Mike Stump <mikestump@comcast.net>
8865 * recog.h: Rename struct recog_data to Recog_data.
8866 * recog.c: Likewise.
8867 * reload.c (can_reload_into): Likewise.
8868 * config/picochip/picochip.c: Likewise.
8870 2013-05-14 Mike Stump <mikestump@comcast.net>
8872 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
8874 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
8876 * resource.h (struct resources): Remove unch_memory member.
8877 (CLEAR_RESOURCE): Don't clear unch_memory.
8878 * resource.c (mark_referenced_resources): Don't set it.
8879 (mark_set_resources): Likewise.
8880 (mark_target_live_regs): Don't clear it.
8881 (init_resource_info): Likewise.
8882 * reorg.c (resource_conflicts_p): Don't compare it.
8883 (redundant_insn): Don't set it.
8885 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
8887 * emit-rtl.c (next_label): Remove unused function.
8888 (skip_consecutive_labels, link_cc0_insns): Move to ...
8889 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
8890 only place where these functions are used, and make them static.
8892 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
8894 * fold-const.c (fold_negate_expr): Handle vectors.
8895 (fold_truth_not_expr): Make it static.
8896 (fold_invert_truthvalue): New static function.
8897 (invert_truthvalue_loc): Handle vectors. Do not call
8898 fold_truth_not_expr directly.
8899 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
8900 <TRUTH_NOT_EXPR>: Do not cast to boolean.
8901 (fold_comparison): Handle vector constants.
8902 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
8903 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
8904 * tree.h (fold_truth_not_expr): Remove declaration.
8906 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
8908 * config/aarch64/aarch64-simd.md
8909 (aarch64_vcond_internal<mode>): Rename to...
8910 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
8911 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
8912 float modes. Clarify all iterator modes.
8913 (vcond<mode><mode>): Use new name for vcond expanders.
8914 (vcond<v_cmp_result><mode>): Likewise.
8915 (vcondu<mode><mode>: Likewise.
8916 * config/aarch64/iterators.md (VDQF_COND): New.
8918 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
8921 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
8922 variable for the shift amount. Check that we shift by non-negative
8925 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
8928 * config/arm/arm.h (EPILOGUE_USES): Only return true
8929 for LR_REGNUM after epilogue_completed.
8931 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
8933 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
8936 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8939 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
8940 and Solaris 11+/x86 with gld.
8941 * configure: Regenerate.
8943 2013-05-14 Jakub Jelinek <jakub@redhat.com>
8945 * expmed.c (expand_shift_1): Canonicalize rotates by
8946 constant bitsize / 2 to bitsize - 1.
8947 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
8948 case ROTATERT>: Likewise.
8951 2013-05-10 Jakub Jelinek <jakub@redhat.com>
8953 * config/i386/i386.md (rotateinv): New code attr.
8954 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
8955 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
8956 roll $31, %eax, etc.
8958 2013-05-14 Richard Biener <rguenther@suse.de>
8961 * tree-eh.c (sink_clobbers): Give up for successors with
8962 multiple predecessors and no virtual uses.
8964 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
8966 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
8967 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
8969 2013-05-14 Jakub Jelinek <jakub@redhat.com>
8972 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
8973 the case when both op0 and op1 have VOIDmode.
8975 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
8977 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
8978 in multiply-accumulate mode.
8980 2013-05-13 Guozhi Wei <carrot@google.com>
8982 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
8984 2013-05-13 Kai Tietz <ktietz@redhat.com>
8987 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
8988 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
8989 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
8990 * config/i386/i386.c (ix86_option_override_internal): Likewise.
8991 (ix86_expand_prologue): Likewise.
8992 (ix86_expand_split_stack_prologue): Likewise.
8993 (legitimate_pic_address_disp_p): Likewise.
8994 (legitimize_pic_address): Likewise.
8995 (legitimize_tls_address): Likewise.
8996 (legitimize_pe_coff_symbol): Likewise.
8997 (output_pic_addr_const): Likewise.
8998 (construct_plt_address): Likewise.
8999 (ix86_expand_call): Likewise.
9000 (x86_output_mi_thunk): Likewise.
9001 (x86_function_profiler): Likewise.
9003 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
9005 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
9006 similar switch cases.
9007 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
9008 (aarch64_simd_mov_to_<mode>low): Delete.
9009 (aarch64_simd_mov_to_<mode>high): Delete.
9010 (move_lo_quad_<mode>): Add w<-r alternative.
9011 (aarch64_simd_move_hi_quad_<mode>): Likewise.
9012 (aarch64_simd_mov_from_*): Update type attribute.
9013 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
9016 2013-05-13 Jan Hubicka <jh@suse.cz>
9018 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
9019 * config/i386/i386.c (ix86_compute_frame_layout,
9020 ix86_expand_epilogue, emit_i387_cw_initialization,
9021 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
9022 ix86_local_alignment): Fix use of size/speed predicates.
9024 2013-05-13 Jakub Jelinek <jakub@redhat.com>
9026 PR tree-optimization/45216
9027 PR tree-optimization/57157
9028 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
9029 the (-Y) & (B - 1) variant if OP is |.
9030 * expmed.c (expand_shift_1): For rotations by const0_rtx just
9031 return shifted. Use (-op1) & (prec - 1) as other_amount
9032 instead of prec - op1.
9034 2013-05-13 Martin Jambor <mjambor@suse.cz>
9037 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
9038 (ipa_constant_data): New type.
9039 (ipa_jump_func): Use ipa_constant_data to hold information about
9040 constant jump functions.
9041 (ipa_get_jf_constant): Adjust to jump function type changes.
9042 (ipa_get_jf_constant_rdesc): New function.
9043 (ipa_param_descriptor): New field controlled_uses.
9044 (ipa_get_controlled_uses): New function.
9045 (ipa_set_controlled_uses): Likewise.
9046 * ipa-ref.h (ipa_find_reference): Declare.
9047 * ipa-prop.c (ipa_cst_ref_desc): New type.
9048 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
9050 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
9051 New parameter cs. Adjust all callers.
9052 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
9053 (remove_described_reference): New function.
9054 (jfunc_rdesc_usable): Likewise.
9055 (try_make_edge_direct_simple_call): Decrement controlled use count,
9056 attempt to remove reference if it hits zero.
9057 (combine_controlled_uses_counters): New function.
9058 (propagate_controlled_uses): Likewise.
9059 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
9060 (ipa_edge_duplication_hook): Duplicate reference descriptions.
9061 (ipa_print_node_params): Print described use counter.
9062 (ipa_write_jump_function): Adjust to jump function type changes.
9063 (ipa_read_jump_function): New parameter CS, pass it to
9064 ipa_set_jf_constant. Adjust caller.
9065 (ipa_write_node_info): Stream controlled use count
9066 (ipa_read_node_info): Likewise.
9067 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
9069 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
9070 count. Remove cloning-added reference if it reaches zero.
9071 * ipa-ref.c (ipa_find_reference): New function.
9073 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
9075 * config/i386/i386.c (processor_target_table): Modified default
9076 alignment values for AMD BD and BT architectures.
9078 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
9080 * tree-vect-generic.c (uniform_vector_p): Move ...
9081 * tree.c (uniform_vector_p): ... here.
9082 * tree.h (uniform_vector_p): Declare it.
9083 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
9086 2013-05-13 Jakub Jelinek <jakub@redhat.com>
9088 PR tree-optimization/57230
9089 * tree-ssa-strlen.c (handle_char_store): Record length for
9090 array store from STRING_CST.
9092 PR tree-optimization/57230
9093 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
9096 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
9098 * config/epiphany/epiphany.c (epiphany_init): Check size of
9099 NUM_MODES_FOR_MODE_SWITCHING.
9100 (epiphany_expand_prologue):
9101 Remove CONFIG_REGNUM initial value handling code.
9102 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
9103 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
9104 (emit_set_fp_mode, epiphany_mode_after): Likewise.
9105 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
9106 Don't return 1 for FP_MODE_NONE.
9107 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
9108 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
9109 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
9110 * config/epiphany/epiphany.md (save_config): New pattern.
9112 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
9114 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
9116 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
9118 * config/i386/i386.md (memory): Handle sseishft1.
9119 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
9120 (*vec_extractv2di_1): Ditto.
9122 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
9124 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
9127 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
9129 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
9130 Add mthumb/march=armv7-a multilib.
9131 Add mthumb/march=armv7-r multilib.
9132 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
9134 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
9136 * config/v850/t-rtems: Add more multilibs.
9138 2013-05-10 Richard Biener <rguenther@suse.de>
9140 PR tree-optimization/57214
9141 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
9142 not propagate from SSA names that occur in abnormal PHI nodes.
9144 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
9146 * stor-layout.c (element_precision): New function.
9147 * machmode.h (element_precision): Declare it.
9148 * tree.c (build_minus_one_cst): New function.
9149 (element_precision): Likewise.
9150 * tree.h (build_minus_one_cst): Declare new function.
9151 (element_precision): Likewise.
9152 * fold-const.c (operand_equal_p): Use element_precision.
9153 (fold_binary_loc): Handle vector types.
9154 * convert.c (convert_to_integer): Use element_precision.
9155 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
9158 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
9160 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
9161 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
9162 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
9163 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
9164 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
9165 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
9166 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
9167 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
9168 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
9169 (Uuw8): New constraints.
9170 (Usb4): Move into alphabetical order.
9171 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
9172 (sd8_operand, ub8_operand, uw8_operand): New predicates.
9173 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
9174 previously unnamed patterns.
9175 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
9176 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
9177 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
9178 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
9179 of set_attr_alternative/if_then_else. Use extended_mips16 instead
9180 of specific lengths.
9182 2013-05-10 Jakub Jelinek <jakub@redhat.com>
9184 * config/i386/i386.md (rotateinv): New code attr.
9185 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
9186 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
9187 roll $31, %eax, etc.
9189 PR tree-optimization/45216
9190 PR tree-optimization/57157
9191 * tree-ssa-forwprop.c (simplify_rotate): New function.
9192 (ssa_forward_propagate_and_combine): Call it.
9194 2013-05-10 Richard Biener <rguenther@suse.de>
9196 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
9197 disable peeling when we version for aliasing.
9198 (vector_alignment_reachable_p): Honor explicit user alignment.
9199 (vect_supportable_dr_alignment): Likewise.
9200 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
9201 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
9202 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
9203 then peeling to arrange for the cost-model check to come first.
9205 2013-05-10 Alan Modra <amodra@gmail.com>
9207 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
9208 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
9209 * configure: Regenerate.
9211 2013-05-10 Alan Modra <amodra@gmail.com>
9214 * varasm.c (default_elf_select_section): Move !DECL_P check..
9215 (get_named_section): ..to here before calling get_section_name.
9217 (default_section_type_flags): Add DECL_P check.
9218 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
9219 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
9221 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
9223 * config/epiphany/epiphany.c (epiphany_expand_prologue):
9224 When using gen_stack_adjust_str with a register offset, add a
9225 REG_FRAME_RELATED_EXPR note.
9227 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
9229 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
9230 (*vec_extractv4si_zext_mem): Ditto.
9231 (*vec_extractv2di): Add 0->x and x->x alternatives.
9232 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
9233 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
9235 2013-05-09 Jason Merrill <jason@redhat.com>
9237 N3639 C++1y VLA support
9238 * gimplify.c (gimplify_vla_decl): Don't touch an existing
9241 * tree.c (build_constructor_va): New.
9242 * tree.h: Declare it.
9244 2013-05-09 Martin Jambor <mjambor@suse.cz>
9247 * gimple-fold.c (canonicalize_constructor_val): Call
9248 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
9250 2013-05-09 Jan Hubicka <jh@suse.cz>
9251 Richard Biener <rguenther@suse.de>
9254 * symtab.c (symtab_make_decl_local): Do not add private names.
9256 2013-05-09 Jan Hubicka <jh@suse.cz>
9259 * symtab.c (insert_to_assembler_name_hash): Handle clones.
9260 (unlink_from_assembler_name_hash): Likewise.
9261 (symtab_prevail_in_asm_name_hash, symtab_register_node,
9262 symtab_unregister_node, symtab_initialize_asm_name_hash,
9263 change_decl_assembler_name): Update.
9265 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
9267 * config/aarch64/aarch64.md: New movtf split.
9268 (*movtf_aarch64): Update.
9269 (aarch64_movdi_tilow): Handle TF modes and rename to
9270 aarch64_movdi_<mode>low.
9271 (aarch64_movdi_tihigh): Handle TF modes and rename to
9272 aarch64_movdi_<mode>high
9273 (aarch64_movtihigh_di): Handle TF modes and rename to
9274 aarch64_mov<mode>high_di
9275 (aarch64_movtilow_di): Handle TF modes and rename to
9276 aarch64_mov<mode>low_di
9277 (aarch64_movtilow_tilow): Remove spurious whitespace.
9278 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
9280 (aarch64_print_operand): Update.
9282 2013-05-09 Alan Modra <amodra@gmail.com>
9284 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
9286 * configure: Regenerate.
9288 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
9290 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
9291 splitter preparation statements.
9292 * config/i386/sse.md (*vec_extract* splitters): Ditto.
9293 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
9296 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9298 * gimple-ssa-strength-reduction.c (count_candidates): Change
9299 return value to int.
9300 (analyze_candidates_and_replace): Change type of length to int.
9302 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
9304 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
9305 (*vec_extract<mode>): Use VI12_128 mode iterator.
9306 (*vec_extract<mode>_mem): Ditto.
9307 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
9310 2013-05-08 Diego Novillo <dnovillo@google.com>
9315 2012-08-17 Diego Novillo <dnovillo@google.com>
9318 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
9319 * config.in: Regenerate.
9320 * configure: Regenerate.
9321 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
9323 2013-05-08 Jan Hubicka <jh@suse.cz>
9326 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
9327 * cgraph.h (symtab_node_base): Add unique_name.
9328 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
9329 input_overwrite_node, input_varpool_node): Stream unique_name.
9330 * cgraphclones.c (cgraph_create_virtual_clone,
9331 cgraph_function_versioning): Set unique_name.
9332 * ipa.c (function_and_variable_visibility): Set unique_name.
9334 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9336 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
9337 (alloc_cand_and_find_basis): Restrict conditional candidate
9338 processing to CAND_MULTs.
9340 2013-05-08 Jan Hubicka <jh@suse.cz>
9343 lto-symtab.c (lto_symtab_symbol_p): New function.
9344 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
9345 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
9346 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
9347 Skip static symbols.
9349 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
9351 PR tree-optimization/57200
9352 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
9353 Only call inform if the preceding warning_at returns true.
9355 2013-05-07 Han Shen <shenhan@google.com>
9357 * cfgexpand.c (record_or_union_type_has_array_p): New function.
9358 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
9359 * common.opt (fstack-protector-strong): New option.
9360 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
9361 * doc/invoke.texi (Optimization Options): Document
9362 "-fstack-protector-strong".
9363 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
9365 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
9367 * config/mips/mips.c (mips_machine_reorg2): Return 0.
9369 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
9371 * ira.c (update_equiv_regs): Add insn having equiv memory even if
9372 it is not lhs of the insn.
9373 (setup_reg_equiv): Remove insn having equiv memory which it is not
9375 * lra-constraints.c (process_address): Try to improve generation
9376 code for address base + disp.
9377 (lra_constraints): Make correct the code for checking insn setting
9378 up backward equivalence. Remove insn only if it is in the init
9380 * lra-eliminations.c (update_reg_eliminate): Change return value.
9381 (lra_eliminate): Use the result.
9383 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
9385 * config/i386/sse.md (ssescalarnummask): New mode attribute.
9386 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
9387 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
9388 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
9389 register target operands.
9390 (*vec_extractv8hi_sse2): New pattern.
9391 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
9392 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
9393 (*vec_extract<mode>_mem): New insn and split pattern.
9395 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
9397 * config/arm/arm.c (arm_asan_shadow_offset): New function.
9398 (TARGET_ASAN_SHADOW_OFFSET): Define.
9399 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
9400 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
9402 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9404 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
9405 (incr_vec_index): Return -1 if increment not found.
9406 (create_add_on_incoming_edge): Assert if increment not found.
9407 (record_increment): Limit number of increments recorded.
9408 (all_phi_incrs_profitable): Return false if an increment not found.
9409 (replace_profitable_candidates): Don't process increments that were
9411 (analyze_candidates_and_replace): Limit size of incr_vec.
9413 2013-05-07 Richard Biener <rguenther@suse.de>
9415 * calls.c (special_function_p): setjmp-like functions are leaf.
9416 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
9417 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
9419 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
9421 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
9422 (aarch64_simd_mov<mode>): New expander.
9423 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
9424 (aarch64_simd_mov_to_<mode>high): Likewise.
9425 (aarch64_simd_mov_from_<mode>low): Likewise.
9426 (aarch64_simd_mov_from_<mode>high): Likewise.
9427 (aarch64_dup_lane<mode>): Update.
9428 (aarch64_dup_lanedi): New instruction pattern.
9429 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
9430 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
9432 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9434 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
9435 (replace_mult_candidate): Remove unnecessary argument; remove
9436 unnecessary parameter from call to introduce_cast_before_cand.
9437 (replace_unconditional_candidate): Remove unnecessary parameter
9438 from call to replace_mult_candidate.
9439 (replace_conditional_candidate): Likewise.
9440 (insert_initializers): Use make_temp_ssa_name.
9441 (introduce_cast_before_cand): Remove unnecessary argument; use
9443 (replace_one_candidate): Remove unnecessary argument; remove
9444 unnecessary parameter from calls to introduce_cast_before_cand.
9445 (replace_profitable_candidates): Remove unnecessary parameters
9446 from calls to replace_one_candidate.
9448 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9450 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
9451 phi def as possibly hiding a basis for a CAND_ADD whose operands
9452 have been commuted in the analysis.
9453 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
9455 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
9457 * config/aarch64/aarch64.md
9458 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
9459 shift value between 0-4.
9461 2013-05-07 Richard Biener <rguenther@suse.de>
9463 * double-int.h (rshift): New overload.
9464 * double-int.c (rshift): New function.
9465 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
9466 (create_reference_ops_from_ref): Remove.
9467 (vn_reference_insert): Use shared ops for constructing the
9468 reference and copy it.
9470 2013-05-07 Richard Biener <rguenther@suse.de>
9473 * tree-eh.c (sink_clobbers): Properly propagate
9474 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
9476 2013-05-07 Jakub Jelinek <jakub@redhat.com>
9478 PR tree-optimization/57149
9479 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
9480 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
9481 collect_phi_def_edges, execute_late_warn_uninitialized): Use
9482 uninit_undefined_value_p instead of ssa_undefined_value_p.
9485 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
9486 for modifier == EXPAND_INITIALIZER.
9488 2013-05-07 Anton Blanchard <anton@samba.org>
9490 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
9491 for powerpc64 little endian.
9492 * configure: Regenerate.
9494 2013-05-06 Graham Stott <grahams@btinternet.com>
9496 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
9497 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
9498 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
9499 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
9501 2013-05-06 Graham Stott <grahams@btinternet.com>
9503 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
9504 codes which allow non-lvalues.
9506 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
9508 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
9509 components are all 1s.
9510 (integer_minus_onep): New function.
9511 * tree.h (integer_minus_onep): Declare it.
9512 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
9513 integer_minus_onep instead of integer_all_onesp.
9515 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
9518 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
9519 variations of these patterns.
9521 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
9523 * config/i386/i386.md (isa): Add x64_sse4 member.
9524 (enabled): Handle x64_sse4.
9525 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
9526 instruction for 64bit SSE4_1 targets. Update insn attributes.
9527 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
9528 instruction for SSE4_1 targets. Update insn attributes.
9529 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
9530 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
9531 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
9533 (*vec_extractv4si): Rename from *sse4_1_pextrd.
9534 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
9535 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
9537 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
9540 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
9542 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
9544 * final.c (do_assembler_dialects): Don't handle curly braces and
9545 vertical bar escaped by % as dialect delimiters.
9546 (output_asm_insn): Print curly braces and vertical bar if escaped
9547 by % and ASSEMBLER_DIALECT defined.
9548 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
9549 * doc/tm.texi: Regenerated.
9551 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
9553 * config/mips/mips.c: Include tree-pass.h.
9554 (mips_reorg): Split in pre- and post-dbr_schedule parts.
9555 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
9556 (pass_mips_machine_reorg2): New machine specific pass.
9557 (insert_pass_mips_machine_reorg2): New pass plugin definition.
9558 (mips_option_override): Register the new pass.
9559 * rtl.h (cleanup_barriers): Remove prototype.
9560 (dbr_schedule): Likewise.
9561 * jump.c (cleanup_barriers): Make static.
9562 * reorg.c (dbr_schedule): Likewise.
9564 2013-05-06 Richard Biener <rguenther@suse.de>
9566 PR tree-optimization/57185
9567 * tree-parloops.c (add_field_for_reduction): Handle anonymous
9570 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
9573 * config/i386/i386.c (add_parameter_dependencies): Add dependence
9574 between "first_arg" and "insn", not "last" and "insn".
9576 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9578 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
9579 (find_candidates_in_block): Re-enable slsr_process_phi.
9580 (create_phi_basis): Fix double counting of candidate adjustment.
9582 2013-05-06 Richard Biener <rguenther@suse.de>
9585 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
9586 the edge is also fallthru, preserve it and just clear the
9588 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
9589 also complex, preserve that and just clear the fallthru flag.
9590 * tree-inline.c (update_ssa_across_abnormal_edges): Also
9591 update virtual operands.
9593 2013-05-06 Alan Modra <amodra@gmail.com>
9595 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
9596 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
9597 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
9598 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
9599 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
9600 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
9602 2013-05-06 Alan Modra <amodra@gmail.com>
9604 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
9605 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
9606 (DEFAULT_ASM_ENDIAN): Define.
9607 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
9608 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
9609 Update -K PIC clause from sysv4.h.
9610 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
9611 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
9613 2013-05-06 Alan Modra <amodra@gmail.com>
9615 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
9616 twice for little-endian.
9617 (ashrdi3_no_power, ashrdi3): Support little-endian.
9619 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
9622 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
9623 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
9625 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
9626 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
9627 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
9630 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
9631 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
9633 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
9634 * config/pa/pa.opt: Make mbig-switch a no-op.
9635 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
9636 (CASE_VECTOR_MODE): Always return SImode.
9637 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
9638 for the !TARGET_BIG_SWITCH case.
9639 * config/pa/pa-linux.h: Likewise.
9640 * config/pa/pa-openbsd.h: Likewise.
9641 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
9642 * config/pa/pa.md (short_jump): Remove define_insn.
9643 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
9644 (casesi0): Remove define_insn.
9645 (type): Remove btable_branch.
9646 (pa_combine_type): Likewise.
9647 (in_nullified_branch_delay): Likewise.
9648 (in_call_delay): Likewise.
9649 (define_delay): Likewise.
9650 (define_insn_reservation "Z3"): Likewise.
9651 (define_insn_reservation "Z4"): Likewise.
9652 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
9653 (pa_adjust_insn_length): Remove adjustment for btable branches.
9654 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
9657 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
9659 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
9660 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
9661 Add m->r,x alternatives.
9662 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
9663 splitters using SWI48x mode iterator.
9664 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
9665 TARGET_64BIT. Add m->x alternative.
9666 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
9667 Add o->x alternative. Enable for TARGET_SSE.
9668 (sse_storeq): Remove expander.
9669 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
9670 with memory input operand.
9671 (*vec_extractv2di_1 splitter): New.
9672 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
9673 * config/i386/i386.md (ssevecmodelower): New mode attribute.
9675 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
9677 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
9678 (INT_LOWPART): Delete.
9679 (extract_MB): Adjust.
9680 (extract_ME): Adjust.
9681 (print_operand): Adjust.
9683 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
9685 * config/rs6000/predicates.md (reg_or_add_cint_operand,
9686 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
9687 (reg_or_logical_cint_operand, easy_fp_constant,
9688 logical_const_operand): Delete "CONST_DOUBLE" case.
9689 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
9690 "HOST_BITS_PER_WIDE_INT == 64" test.
9691 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
9692 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
9693 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
9694 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
9696 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
9697 CONST_DOUBLE DImode/VOIDmode case.
9698 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
9699 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
9700 CONST_DOUBLE VOIDmode case.
9701 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
9702 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
9703 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
9704 Delete CONST_DOUBLE case.
9705 (splitters for mov FMOVE64 const_double): Delete
9706 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
9707 "HOST_BITS_PER_WIDE_INT >= 64" test.
9708 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
9710 (mov DI const_double): Delete.
9712 2013-05-04 Jakub Jelinek <jakub@redhat.com>
9714 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
9715 on op shows all bits zero in mode of a lowpart subreg, return zero.
9717 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
9720 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
9721 to save TFmode registers and DImode to save TImode registers for
9722 caller save operations.
9723 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
9724 mark being partially clobbered since they only use the first
9727 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
9728 and TDmode only use the upper 64-bits of each VSX register.
9730 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9732 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
9733 (find_candidates_in_block): Disable slsr_process_phi.
9735 2013-05-03 Guozhi Wei <carrot@google.com>
9737 * coverage.c (coverage_obj_init): Move the construction of gcov
9739 (build_init_ctor): ... here.
9741 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9743 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
9744 (slsr_cand_d): Redefine def_phi.
9745 (stride_status, phi_adjust_status, count_phis_status): New enums.
9746 (find_phi_def): New.
9747 (find_basis_for_base_expr): New.
9748 (find_basis_for_candidate): Handle hidden bases.
9749 (alloc_cand_and_find_basis): Handle phi candidates.
9750 (slsr_process_phi): New.
9751 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
9752 (create_mul_imm_cand): Likewise.
9753 (create_add_ssa_cand): Exclude phi base candidates.
9754 (create_add_imm_cand): Likewise.
9755 (slsr_process_cast): Likewise.
9756 (slsr_process_copy): Likewise.
9757 (find_candidates_in_block): Handle phi candidates.
9758 (dump_candidate): Likewise.
9759 (unconditional_cands): Delete.
9760 (unconditional_cands_with_known_stride_p): Delete.
9761 (phi_dependent_cand_p): New.
9762 (cand_increment): Handle phi-dependent candidates.
9763 (replace_dependent): Delete.
9764 (replace_mult_candidate): New.
9765 (replace_unconditional_candidate): New.
9766 (incr_vec_index): Move to avoid forward reference.
9767 (create_add_on_incoming_edge): New.
9768 (create_phi_basis): New.
9769 (replace_dependents): Delete.
9770 (replace_conditional_candidate): New.
9771 (phi_add_costs): New.
9772 (replace_uncond_cands_and_profitable_phis): New.
9773 (record_increment): Handle phi adjustments.
9774 (record_phi_increments): New.
9775 (record_increments): Handle phi adjustments.
9776 (phi_incr_cost): New.
9777 (lowest_cost_path): Handle phis.
9778 (total_savings): Likewise.
9779 (analyze_increments): Likewise.
9780 (ncd_with_phi): New.
9781 (ncd_of_cand_and_phis): New.
9782 (nearest_common_dominator_for_cands): Handle phi increments.
9783 (all_phi_incrs_profitable): New.
9784 (replace_profitable_candidates): Handle phi-dependent candidates.
9785 (analyze_candidates_and_replace): Likewise.
9787 2013-05-03 Teresa Johnson <tejohnson@google.com>
9790 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
9791 do not exceed REG_BR_PROB_BASE.
9793 2013-05-03 Jeff Law <law@redhat.com>
9795 PR tree-optimization/57144
9796 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
9797 operand of the condition will bit into the new type when eliminating
9798 a cast feeding a condition.
9800 2013-05-03 Jakub Jelinek <jakub@redhat.com>
9802 PR rtl-optimization/57130
9803 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
9804 of COMPARE as in_code to the recursive call if needed.
9806 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
9808 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
9809 (enabled): Handle new members.
9810 * config/i386/sse.md (*vec_concatv2si): Merge from
9811 *vec_concatv2si_sse2 and vec_concatv2si_sse.
9812 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
9814 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
9816 PR tree-optimization/57027
9817 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
9818 for fnms opportunity, check we got the prerequisite kind
9819 of tree / gimple before using accessor functions.
9821 2013-05-03 Richard Biener <rguenther@suse.de>
9823 * double-int.h (lshift): New overload without precision
9825 (operator *=, operator +=, operator -=): Move ...
9826 * double-int.c (operator *=, operator +=, operator -=): ... here
9827 and implement more efficiently.
9828 (mul_double_with_sign): Remove.
9829 (lshift_double): Adjust to take unsinged shift argument, push
9830 dispatching code to callers.
9831 (mul_double_wide_with_sign): Add early out for callers that
9832 are not interested in high parts or overflow.
9833 (lshift): New function.
9834 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
9836 (lrotate, rrotate): Use logical shifts.
9837 * expr.c (get_inner_reference): Use lshift.
9838 * fixed-value.c (do_fixed_divide): Likewise.
9839 * tree-dfa.c (get_ref_base_and_extent): Likewise.
9840 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
9841 (indirect_refs_may_alias_p): Likewise.
9842 (stmt_kills_ref_p_1): Likewise.
9844 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
9846 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
9848 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
9850 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
9851 scalar form of FABD instruction.
9853 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
9855 * lra-constraints.c (process_alt_operands): Add checking alt
9856 number to choose the best alternative.
9858 2013-05-02 Richard Biener <rguenther@suse.de>
9860 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
9861 bitmap and its handling.
9862 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
9864 2013-05-02 Richard Biener <rguenther@suse.de>
9867 * tree-inline.c (copy_loops): Properly handle removed loops.
9868 (copy_cfg_body): Mark destination loops for fixup if source
9871 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
9874 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
9875 generating simple_return for naked functions.
9877 2013-05-02 Martin Jambor <mjambor@suse.cz>
9880 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
9881 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
9883 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
9884 ipa_agg_replacement_value structures.
9885 (known_aggs_to_agg_replacement_list): Likewise.
9886 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
9887 (read_agg_replacement_chain): Likewise.
9888 (ipcp_transform_function): Also check that by_ref flags match.
9890 2013-05-02 Richard Biener <rguenther@suse.de>
9892 * graphds.h (struct graph): Add obstack member.
9893 * graphds.c (new_graph): Initialize obstack and allocate
9895 (add_edge): Allocate edge from the obstack.
9896 (free_graph): Free the obstack instead of all edges and vertices.
9898 2013-05-02 Teresa Johnson <tejohnson@google.com>
9900 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
9902 * cfg.c (update_bb_profile_for_threading): Ditto.
9903 * tree-inline.c (copy_bb): Ditto.
9904 (copy_edges_for_bb): Ditto.
9905 (initialize_cfun): Ditto.
9906 (copy_cfg_body): Ditto.
9907 (expand_call_inline): Ditto.
9908 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
9909 (estimate_node_size_and_time): Ditto.
9910 (inline_merge_summary): Ditto.
9911 * cgraphclones.c (cgraph_clone_edge): Ditto.
9912 (cgraph_clone_node): Ditto.
9913 * sched-rgn.c (compute_dom_prob_ps): Ditto.
9914 (compute_trg_info): Ditto.
9916 2013-05-02 Ian Bolton <ian.bolton@arm.com>
9918 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
9919 S reg when fp attribute set.
9920 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
9922 2013-05-02 Ian Bolton <ian.bolton@arm.com>
9924 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
9926 (*and_one_cmplsi3_compare0_uxtw): Likewise.
9927 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
9928 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
9930 2013-05-02 Richard Biener <rguenther@suse.de>
9932 * tree-scalar-evolution.c (scev_info_hasher): Remove.
9933 (struct instantiate_cache_entry): New type.
9934 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
9935 (struct instantiate_cache_type): New type.
9936 (set_instantiated_value, get_instantiated_value): Remove.
9937 (get_instantiated_value_entry): New function.
9938 (instantiate_scev_name): Use the new cache and adjust.
9939 (instantiate_scev_poly): Adjust.
9940 (instantiate_scev_binary): Likewise.
9941 (instantiate_array_ref): Likewise.
9942 (instantiate_scev_convert): Likewise.
9943 (instantiate_scev_not): Likewise.
9944 (instantiate_scev_3): Likewise.
9945 (instantiate_scev_2): Likewise.
9946 (instantiate_scev_r): Likewise.
9947 (instantiate_scev): Likewise.
9948 (resolve_mixers): Likewise.
9950 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
9953 * lra-constraints.c (best_small_class_operands_num): Remove.
9954 (process_alt_operands): Remove small_class_operands_num. Take
9955 small classes operands into losers and only if the operand is not
9956 matched. Modify debugging output.
9957 (curr_insn_transform): Remove best_small_class_operands_num.
9960 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
9962 * config/aarch64/aarch64-builtins.c
9963 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
9964 * config/aarch64/aarch64-simd-builtins.def
9965 (reduc_splus_): Add new modes.
9966 (reduc_uplus_): New.
9967 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
9968 (reduc_uplus_v4sf): Likewise.
9969 (reduc_splus_v4sf): Likewise.
9970 (aarch64_addv<mode>): Likewise.
9971 (reduc_uplus_<mode>): Likewise.
9972 (reduc_splus_<mode>): Likewise.
9973 (aarch64_addvv2di): Likewise.
9974 (reduc_uplus_v2di): Likewise.
9975 (reduc_splus_v2di): Likewise.
9976 (aarch64_addvv2si): Likewise.
9977 (reduc_uplus_v2si): Likewise.
9978 (reduc_splus_v2si): Likewise.
9979 (reduc_<sur>plus_<mode>): New.
9980 (reduc_<sur>plus_v2di): Likewise.
9981 (reduc_<sur>plus_v2si): Likewise.
9982 (reduc_<sur>plus_v4sf): Likewise.
9983 (aarch64_addpv4sf): Likewise.
9984 * config/aarch64/arm_neon.h
9985 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
9986 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
9987 add UNSPEC_SADDV, UNSPEC_UADDV.
9989 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
9991 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
9993 * config/aarch64/arm_neon.h
9994 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
9996 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
9998 * config/aarch64/aarch64-builtins
9999 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
10001 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10003 * config/aarch64/aarch64-simd-builtins.def
10004 (reduc_smax_): New.
10005 (reduc_smin_): Likewise.
10006 (reduc_umax_): Likewise.
10007 (reduc_umin_): Likewise.
10008 (reduc_smax_nan_): Likewise.
10009 (reduc_smin_nan_): Likewise.
10012 (smax): Update for V2SF, V4SF and V2DF modes.
10015 (smin_nan): Likewise.
10016 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
10017 (<su><maxmin><mode>3): ...This, refactor.
10018 (s<maxmin><mode>3): New.
10019 (<maxmin_uns><mode>3): Likewise.
10020 (reduc_<maxmin_uns>_<mode>): Refactor.
10021 (reduc_<maxmin_uns>_v4sf): Likewise.
10022 (reduc_<maxmin_uns>_v2si): Likewise.
10023 (aarch64_<fmaxmin><mode>: Remove.
10024 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
10026 (vmin<q>_f<32,64>): Likewise.
10027 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
10029 (su): Add mappings for smax, smin, umax, umin.
10031 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
10032 (FMAXMIN): Rename as...
10033 (FMAXMIN_UNS): ...This.
10035 (fmaxminv): Likewise.
10036 (fmaxmin): Likewise.
10038 (maxmin_uns_op): Likewise.
10040 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10042 * config/aarch64/arm_neon.h
10043 (vac<ge, gt><sd>_f<32, 64>): Rename to...
10044 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
10045 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
10047 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10049 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
10050 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
10052 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10054 * config/aarch64/aarch64-simd.md
10055 (vcond<mode>_internal): Handle special cases for constant masks.
10056 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
10057 (vcondu<mode><mode>): Likewise.
10058 (vcond<v_cmp_result><mode>): New.
10060 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10062 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
10063 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
10064 * config/aarch64/aarch64-simd-builtins.def
10065 (cmeq): Update to BUILTIN_VALLDI.
10070 * config/aarch64/arm_neon.h
10071 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
10072 to builtins or C as appropriate.
10074 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10076 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
10078 (cmhi): Rename to...
10080 * config/aarch64/aarch64-simd.md
10081 (simd_mode): Add SF.
10082 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
10083 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
10084 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
10085 (cstore<mode>_neg): ...This.
10086 * config/aarch64/iterators.md
10088 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
10089 (COMPARISONS): New.
10090 (UCOMPARISONS): Likewise.
10091 (optab): Add missing comparisons.
10097 (VCMP_S): Likewise.
10098 (VCMP_U): Likewise.
10099 (V_cmp_result): Add DF, SF modes.
10100 (v_cmp_result): Likewise.
10102 (vmtype): Likewise.
10103 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
10105 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
10107 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
10108 define_insn to define_insn_and_split.
10109 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
10110 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
10111 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
10112 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
10113 (thumb2_negscc): Likewise.
10115 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
10117 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
10119 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
10121 * config/arm/thumb2.md: Remove trailing whitespaces.
10123 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
10125 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
10126 Use gen_int_mode rather than GEN_INT.
10128 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
10130 * value-prof.c (stream_in_histogram_value): Remove the strayed
10133 2013-04-30 Richard Biener <rguenther@suse.de>
10135 PR middle-end/57122
10136 * cfghooks.c (split_edge): Properly check for the loop latch edge.
10138 2013-04-30 Richard Biener <rguenther@suse.de>
10140 PR middle-end/57107
10141 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
10143 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
10145 PR rtl-optimization/56957
10146 PR rtl-optimization/57105
10147 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
10148 variable. Use just INSN_UID for determining whether an insn
10149 should be only disconnected from the insn stream.
10150 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
10152 2013-04-30 Jakub Jelinek <jakub@redhat.com>
10154 PR tree-optimization/57104
10155 * tsan.c (instrument_expr): Don't instrument accesses to
10156 DECL_HARD_REGISTER VAR_DECLs.
10158 2013-04-30 Richard Biener <rguenther@suse.de>
10160 * function.h (loops_for_fn): New inline function.
10161 (set_loops_for_fn): Likewise.
10162 * cfgloop.h (place_new_loop): Add struct function parameter.
10163 (get_loop): Likewise.
10164 (get_loops): Likewise.
10165 (number_of_loops): Likewise.
10166 (fel_next): Adjust.
10167 (fel_init): Likewise.
10168 * cfg.c (get_loop_copy): Adjust.
10169 * cfgloop.c (flow_loops_dump): Likewise.
10170 (record_loop_exits): Likewise.
10171 (verify_loop_structure): Likewise.
10172 * cfgloopanal.c (mark_irreducible_loops): Likewise.
10173 (estimate_reg_pressure_cost): Likewise.
10174 (mark_loop_exit_edges): Likewise.
10175 * cfgloopmanip.c (place_new_loop): Likewise.
10176 (add_loop): Likewise.
10177 (duplicate_loop): Likewise.
10178 * graph.c (draw_cfg_nodes): Likewise.
10179 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
10180 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
10181 (extract_affine_chrec): Likewise.
10182 (build_scop_iteration_domain): Likewise.
10183 * graphite.c (graphite_initialize): Likewise.
10184 * ira-build.c (create_loop_tree_nodes): Likewise.
10185 (more_one_region_p): Likewise.
10186 (rebuild_regno_allocno_maps): Likewise.
10187 (mark_loops_for_removal): Likewise.
10188 (mark_all_loops_for_removal): Likewise.
10189 (remove_unnecessary_regions): Likewise.
10190 (ira_build): Likewise.
10191 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
10192 * loop-init.c (fix_loop_structure): Likewise.
10193 (gate_rtl_move_loop_invariants): Likewise.
10194 (gate_rtl_unswitch): Likewise.
10195 (gate_rtl_unroll_and_peel_loops): Likewise.
10196 (rtl_doloop): Likewise.
10197 * lto-streamer-in.c (input_cfg): Likewise.
10198 * lto-streamer-out.c (output_cfg): Likewise.
10199 * modulo-sched.c (sms_schedule): Likewise.
10200 * predict.c (tree_estimate_probability): Likewise.
10201 (tree_estimate_probability_driver): Likewise.
10202 (estimate_loops): Likewise.
10203 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
10204 (move_sese_region_to_fn): Likewise.
10205 (debug_loop_num): Likewise.
10206 * tree-chrec.c (chrec_evaluate): Likewise.
10207 (hide_evolution_in_other_loops_than_loop): Likewise.
10208 (chrec_component_in_loop_num): Likewise.
10209 (reset_evolution_in_loop): Likewise.
10210 (evolution_function_is_invariant_rec_p): Likewise.
10211 * tree-if-conv.c (main_tree_if_conversion): Likewise.
10212 * tree-inline.c (copy_loops): Likewise.
10213 (copy_cfg_body): Likewise.
10214 (tree_function_versioning): Likewise.
10215 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
10216 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
10218 (add_to_evolution_1): Likewise.
10219 (scev_const_prop): Likewise.
10220 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
10221 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
10222 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
10223 (tree_ssa_lim_initialize): Likewise.
10224 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
10225 (verify_loop_closed_ssa): Likewise.
10226 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
10227 (tree_ssa_loop_im): Likewise.
10228 (tree_ssa_loop_unswitch): Likewise.
10229 (tree_vectorize): Likewise.
10230 (check_data_deps): Likewise.
10231 (tree_ssa_loop_ivcanon): Likewise.
10232 (tree_ssa_loop_bounds): Likewise.
10233 (tree_complete_unroll): Likewise.
10234 (tree_complete_unroll_inner): Likewise.
10235 (tree_parallelize_loops): Likewise.
10236 (tree_ssa_loop_prefetch): Likewise.
10237 (tree_ssa_loop_ivopts): Likewise.
10238 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
10239 * tree-vectorizer.c (vectorize_loops): Likewise.
10241 2013-04-29 Mike Frysinger <vapier@gentoo.org>
10243 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
10244 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
10245 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
10246 with EABI_LINK_SPEC.
10248 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
10251 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
10254 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
10257 * lra-constraints.c (process_alt_operands): Discourage a bit more
10258 using memory for pseudos. Print cost dump for alternatives.
10259 Modify cost values for conflicts with early clobbers.
10260 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
10262 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
10265 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
10267 2013-04-29 Ian Bolton <ian.bolton@arm.com>
10269 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
10270 from/to S register.
10271 (movdi_aarch64): Support LDR/STR from/to D register.
10273 2013-04-29 Ian Bolton <ian.bolton@arm.com>
10275 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
10276 or higher by default.
10278 2013-04-29 Richard Biener <rguenther@suse.de>
10280 PR middle-end/57075
10281 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
10282 even if not adding abnormal edges for calls that can make
10285 2013-04-29 Richard Biener <rguenther@suse.de>
10287 PR middle-end/57103
10288 * tree-cfg.c (move_stmt_op): Fix condition under which to update
10290 (move_stmt_r): Remove redundant checking.
10292 2013-04-29 Teresa Johnson <tejohnson@google.com>
10295 * basic-block.h (apply_scale): New function.
10296 (apply_probability): Use apply_scale.
10297 * gimple-streamer-in.c (input_bb): Ditto.
10298 * lto-streamer-in.c (input_cfg): Ditto.
10299 * lto-cgraph.c (merge_profile_summaries): Ditto.
10300 * tree-optimize.c (execute_fixup_cfg): Ditto.
10301 * tree-inline.c (copy_bb): Update comment to use apply_scale.
10302 (copy_edges_for_bb): Ditto.
10303 (copy_cfg_body): Ditto.
10305 2013-04-29 Tom de Vries <tom@codesourcery.com>
10307 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
10308 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
10309 (tail_merge_optimize): Handle current_loops == NULL.
10311 2013-04-26 Jeff Law <law@redhat.com>
10313 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
10314 (simplify_cond_using_ranges): Generalize code to simplify
10315 COND_EXPRs where one argument is a constant and the other
10316 is an SSA_NAME created by an integral type conversion.
10318 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10320 * config/arm/arm.md (store_minmaxsi): Use only when
10321 optimize_insn_for_size_p.
10323 2013-04-29 Christian Bruel <christian.bruel@st.com>
10326 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
10328 2013-04-29 Richard Biener <rguenther@suse.de>
10330 PR middle-end/57089
10331 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
10332 loop tree make sure to schedule a fixup for the child as well.
10333 (expand_omp_for_generic): Properly add loops.
10334 (expand_omp_for_static_nochunk): Likewise.
10335 (expand_omp_for_static_chunk): Likewise.
10336 (expand_omp_for): For the degenerate case fixup loops.
10337 (expand_omp_sections): Fix default bb placement in loops.
10338 (expand_omp_atomic_pipeline): Properly add loops.
10340 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10342 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
10344 2013-04-29 Tom de Vries <tom@codesourcery.com>
10346 * tree-ssa-tail-merge.c: Update header comment.
10348 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
10350 * config/aarch64/arm_neon.h
10351 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
10352 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
10353 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
10354 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
10355 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
10356 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
10357 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
10358 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
10360 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
10362 * config/aarch64/aarch64-simd.md
10363 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
10364 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
10365 fix_trunc, fixuns_trunc.
10366 (ftrunc<VDQF:mode>2): New.
10367 * config/aarch64/iterators.md (optab): Add fix, fixuns.
10368 (fix_trunc_optab): New.
10370 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
10372 * config/aarch64/aarch64-builtins.c
10373 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
10374 iceilf, lround, iroundf.
10376 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
10379 * config/i386/i386.h (enum ix86_tune_indices)
10380 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
10381 New, split from X86_TUNE_INTER_UNIT_MOVES.
10382 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
10383 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
10384 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
10385 (TARGET_INTER_UNIT_MOVES): Remove.
10386 * config/i386/i386.c (initial_ix86_tune_features): Update.
10387 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
10388 (ix86_expand_convert_uns_didf_sse): Use
10389 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
10390 (ix86_expand_vector_init_one_nonzero): Ditto.
10391 (ix86_expand_vector_init_interleave): Ditto.
10392 (inline_secondary_memory_needed): Return true for moves from SSE class
10393 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
10394 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
10395 * config/i386/constraints.md (Yi, Ym): Depend on
10396 TARGET_INTER_UNIT_MOVES_TO_VEC.
10397 (Yj, Yn): New constraints.
10398 * config/i386/i386.md (*movdi_internal): Change constraints of
10399 operand 1 from Yi to Yj and from Ym to Yn.
10400 (*movsi_internal): Ditto.
10401 (*movdf_internal): Ditto.
10402 (*movsf_internal): Ditto.
10403 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
10404 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
10405 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
10406 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
10407 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
10408 * config/i386/sse.md (movdi_to_sse): Ditto.
10409 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
10410 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
10411 TARGET_INTER_UNIT_MOVES.
10412 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
10413 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
10414 instead of TARGET_INTER_UNIT_MOVES.
10415 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
10416 operand 1 from Yi to Yj and from Ym to Yn.
10418 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
10420 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
10421 (float_truncate_hi_): Likewise.
10422 (float_extend_lo_): Likewise.
10423 (float_truncate_lo_): Likewise.
10424 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
10425 (aarch64_float_extend_lo_v2df): Likewise.
10426 (vec_unpacks_hi_v4sf): Likewise.
10427 (aarch64_float_truncate_lo_v2sf): Likewise.
10428 (aarch64_float_truncate_hi_v4sf): Likewise.
10429 (vec_pack_trunc_v2df): Likewise.
10430 (vec_pack_trunc_df): Likewise.
10432 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
10434 * config/aarch64/aarch64-builtins.c
10435 (aarch64_fold_builtin): Fold float conversions.
10436 * config/aarch64/aarch64-simd-builtins.def
10437 (floatv2si, floatv4si, floatv2di): New.
10438 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
10439 * config/aarch64/aarch64-simd.md
10440 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
10441 * config/aarch64/iterators.md (FLOATUORS): New.
10442 (optab): Add float, floatuns.
10443 (su_optab): Likewise.
10445 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
10447 * config/aarch64/aarch64-builtins.c
10448 (aarch64_builtin_vectorized_function): Use new names for
10450 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
10451 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
10452 (fcvtzu): Split as...
10453 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
10454 (fcvtas): Split as...
10455 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
10456 (fcvtau): Split as...
10457 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
10458 (fcvtps): Split as...
10459 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
10460 (fcvtpu): Split as...
10461 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
10462 (fcvtms): Split as...
10463 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
10464 (fcvtmu): Split as...
10465 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
10466 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
10467 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
10468 (lfrintnusf, lfrintnudf): Likewise.
10469 * config/aarch64/aarch64-simd.md
10470 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
10472 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
10473 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
10474 (fcvt_pattern): Likewise.
10476 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
10478 * config/aarch64/aarch64-simd.md
10479 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
10480 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
10482 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
10484 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
10485 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
10486 (vrnd<a,m,n,p>_f32): Implement using builtins.
10487 (vrnd<i,x><q>_f<32, 64>): New.
10489 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
10491 * config/aarch64/aarch64-builtins.c
10492 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
10493 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
10494 (frintz): Rename to...
10496 (frintp): Rename to...
10498 (frintm): Rename to...
10500 (frinti): Rename to...
10501 (nearbyint): ...this.
10502 (frintx): Rename to...
10504 (frinta): Rename to...
10506 * config/aarch64/aarch64-simd.md
10507 (aarch64_frint<frint_suffix><mode>): Delete.
10508 (<frint_pattern><mode>2): Convert to insn.
10509 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
10510 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
10511 (frint_pattern): Likewise.
10512 (frint_suffix): Likewise.
10514 2013-04-29 Richard Biener <rguenther@suse.de>
10516 PR tree-optimization/57081
10517 * loop-init.c: Include tree-flow.h.
10518 (loop_optimizer_finalize): Free number of iteration estimates.
10519 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
10521 2013-04-29 Jakub Jelinek <jakub@redhat.com>
10523 PR tree-optimization/57083
10524 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
10525 non-singleton shift count range, zero extend low_bound for uns case.
10527 * config/i386/predicates.md (general_vector_operand): New predicate.
10528 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
10529 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
10530 if they aren't nonimmediate operands. If their original values
10531 satisfy const_vector_equal_evenodd_p, don't shift them.
10532 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
10533 predicates. For the SSE4.1 case force operands[{1,2}] into registers
10534 if not nonimmediate_operand.
10535 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
10536 instead of register_operand.
10537 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
10539 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
10541 * stor-layout.c (finalize_size_functions): Allocate a structure and
10542 reset cfun before dumping the functions.
10544 2013-04-27 Jakub Jelinek <jakub@redhat.com>
10546 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
10549 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
10550 use xop_pmacsdqh if uns_p.
10551 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
10552 the immediate rotate count.
10554 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
10556 * rtl.h (struct rtx_def): Add comment for field jump.
10557 (LRA_SUBREG_P): New macro.
10558 * recog.c (register_operand): Check LRA_SUBREG_P.
10559 * lra.c (lra): Add note at the end of RTL code. Align non-empty
10561 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
10562 (lra_final_code_change): Skip subreg change for operators.
10563 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
10564 if there are no operand changes.
10565 * lra-constraints.c (curr_insn_set): New.
10566 (match_reload): Set LRA_SUBREG_P.
10567 (emit_spill_move): Ditto.
10568 (check_and_process_move): Use curr_insn_set. Process only single
10569 set insns. Don't initialize sec_mem_p and change_p.
10570 (simplify_operand_subreg): Use LRA_SUBREG_P.
10571 (reg_in_class_p): New function.
10572 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
10573 of #ifdef. Add code to remove cycling.
10574 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
10575 non-null disp. Reload inner instead of disp when base and index
10576 are null. Try to put lo_sum into register.
10577 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
10578 (check_and_process_move): Move code for move cost check to
10579 simple_move_p. Remove equiv_substitution.
10580 (simple_move_p): New function.
10581 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
10582 curr_insn_set. Call check_and_process_move only for single set
10583 insns. Use the new function. Move call of check_and_process_move
10584 after operand equiv substitution and address process.
10586 2013-04-26 Jakub Jelinek <jakub@redhat.com>
10589 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
10590 with nonlocal goto receivers or returns twice calls, ignore
10591 unininitialized values from abnormal edges to nl goto receiver
10592 or returns twice call.
10594 2013-04-26 Jakub Jelinek <jakub@redhat.com>
10596 PR tree-optimization/57051
10597 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
10598 and VEC_RSHIFT_EXPR if shift count is a multiple of element
10601 2013-04-26 Richard Biener <rguenther@suse.de>
10603 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
10604 (expand_omp_taskreg): Likewise. Mark loops for fixup.
10605 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
10606 (fixup_loop_arrays_after_move): New function.
10607 (move_sese_region_to_fn): Properly outline the loop tree parts
10608 of the SESE region.
10610 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
10612 * config/i386/i386.md (type, unit): Fix long lines.
10614 2013-04-26 Richard Biener <rguenther@suse.de>
10616 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
10617 (lto-streamer-out.o): Likewise.
10618 * cfgloop.c (init_loops_structure): Export, add struct function
10619 argument and adjust.
10620 (flow_loops_find): Adjust.
10621 * cfgloop.h (enum loop_estimation): Add EST_LAST.
10622 (init_loops_structure): Declare.
10623 * lto-streamer-in.c: Include cfgloop.h.
10624 (input_cfg): Input the loop tree.
10625 * lto-streamer-out.c: Include cfgloop.h.
10626 (output_cfg): Output the loop tree.
10627 (output_struct_function_base): Do not drop PROP_loops.
10629 2013-03-26 Richard Biener <rguenther@suse.de>
10631 * tree-cfg.c (execute_build_cfg): Build the loop tree.
10632 (pass_build_cfg): Provide PROP_loops.
10633 (move_sese_region_to_fn): Remove loops that are outlined into fn
10635 * tree-inline.c: Include cfgloop.h.
10636 (initialize_cfun): Do not drop PROP_loops.
10637 (copy_loops): New function.
10638 (copy_cfg_body): Copy loop structure.
10639 (tree_function_versioning): Initialize destination loop tree.
10640 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
10641 (pass_parallelize_loops): Do IL verification.
10642 * loop-init.c (loop_optimizer_init): Fixup loops if required.
10643 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
10644 the CFG make sure we fixup loops as well.
10645 * tree-ssa-tail-merge.c: Include cfgloop.h.
10646 (replace_block_by): When merging loop latches mark loops for fixup.
10647 * lto-streamer-out.c (output_struct_function_base): Drop
10648 PROP_loops for now.
10649 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
10650 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
10651 * ipa-split.c: Include cfgloop.h.
10652 (split_function): Add the new return block to the loop tree root.
10653 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
10654 whether we have removed the forwarder block.
10655 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
10656 * cfgloop.h (place_new_loop): Declare.
10657 * cfgloopmanip.c (place_new_loop): Export.
10658 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
10659 (tree-switch-conversion.o): Likewise.
10660 (tree-complex.o): Likewise.
10661 (tree-inline.o): Likewise.
10662 (tree-ssa-tailmerge.o): Likewise.
10663 (ipa-split.o): Likewise.
10664 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
10665 (tree-ssa-copy.o): Likewise.
10666 * tree-switch-conversion.c: Include cfgloop.h
10667 (process_switch): If we emit a bit-test cascade, schedule loops
10669 * tree-complex.c: Include cfgloop.h.
10670 (expand_complex_div_wide): Properly add new basic-blocks to loops.
10671 * asan.c: Include cfgloop.h.
10672 (create_cond_insert_point): Properly add new basic-blocks to
10673 loops, schedule loop fixup.
10674 * cfgloop.c (verify_loop_structure): Check that looks are not
10676 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
10678 (expand_omp_for_generic): Likewise.
10679 (expand_omp_sections): Likewise.
10680 (expand_omp_atomic_pipeline): Schedule loops for fixup.
10681 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
10682 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
10683 is initialized, not when loops are present.
10684 * tree-parloops.c (parallelize_loops): Remove checking here.
10685 * passes.c (init_optimization_passes): Schedule a copy-propagation
10686 pass before complete unrolling of inner loops.
10688 2013-04-26 Jakub Jelinek <jakub@redhat.com>
10690 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
10691 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
10692 (colorize_init): Add argument to _WIN32 version.
10693 * toplev.c: Include diagnostic-color.h.
10694 (process_options): Default to -fdiagnostics-color=auto if
10695 GCC_COLORS env var is in the environment.
10696 * common.opt (fdiagnostics-color=): Add Var and Init.
10697 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
10698 env var is in the environment, the default is auto rather than never.
10700 * diagnostic.h (file_name_as_prefix): Add context argument.
10701 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
10702 the string as locus.
10703 * langhooks.c (lhd_print_error_function): Adjust caller.
10705 2013-04-25 Lawrence Crowl <crowl@google.com>
10707 * var-tracking.c (shared_hash_def::htab):
10708 Change type to hash_table. Update dependent calls and types.
10710 2013-04-25 Lawrence Crowl <crowl@google.com>
10712 * Makefile.in: Update as needed below.
10714 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
10715 Move declaration to after the type's method definitons.
10717 * attribs.c (htab_t scoped_attributes::attribute_hash):
10718 Change type to hash_table. Update dependent calls and types.
10720 * bitmap.c (htab_t bitmap_desc_hash):
10721 Change type to hash_table. Update dependent calls and types.
10723 * cselib.c (htab_t cselib_hash_table):
10724 Change type to hash_table. Update dependent calls and types.
10726 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
10727 (hash_string_slot_node): Move implementation into lto-streamer.h
10728 struct string_slot_hasher.
10729 (eq_string_slot_node): Likewise.
10731 * data-streamer-out.c: Update output_block::string_hash_table
10732 dependent calls and types.
10734 * dwarf2cfi.c (htab_t trace_index):
10735 Change type to hash_table. Update dependent calls and types.
10737 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
10738 Change type to hash_table. Update dependent calls and types.
10739 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
10740 (htab_t optimize_external_refs::map): Likewise.
10741 (htab_t output_comp_unit::extern_map): Likewise.
10742 (htab_t output_comdat_type_unit::extern_map): Likewise.
10743 (htab_t output_macinfo::macinfo_htab): Likewise.
10744 (htab_t optimize_location_lists::htab): Likewise.
10745 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
10747 * except.c (htab_t ehspec_hash_type):
10748 Change type to hash_table. Update dependent calls and types.
10749 (assign_filter_values::ttypes): Likewise.
10750 (assign_filter_values::ehspec): Likewise.
10751 (sjlj_assign_call_site_values::ar_hash): Likewise.
10752 (convert_to_eh_region_ranges::ar_hash): Likewise.
10754 * gcse.c (htab_t pre_ldst_table):
10755 Change type to hash_table. Update dependent calls and types.
10757 * ggc-common.c (htab_t saving_htab):
10758 Change type to hash_table. Update dependent calls and types.
10759 (htab_t loc_hash): Likewise.
10760 (htab_t ptr_hash): Likewise.
10761 (call_count): Rename ggc_call_count.
10762 (call_alloc): Rename ggc_call_alloc.
10763 (loc_descriptor): Rename make_loc_descriptor.
10764 (add_statistics): Rename ggc_add_statistics.
10766 * ggc-common.c (saving_htab):
10767 Change type to hash_table. Update dependent calls and types.
10769 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
10770 (push_gimplify_context): Likewise.
10771 (pop_gimplify_context): Likewise.
10772 (struct gimple_temp_hash_elt): Added.
10773 (struct gimplify_hasher): Likewise.
10774 (struct gimplify_ctx.temp_htab):
10775 Change type to hash_table. Update dependent calls and types.
10777 * gimple-fold.c: Include gimplify-ctx.h.
10779 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
10780 Change type to hash_table. Update dependent calls and types.
10781 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
10782 avoid potential global name collision.
10784 * gimplify.c: Include gimplify-ctx.h.
10785 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
10786 (htab_t gimplify_ctx::temp_htab):
10787 Update dependent calls and types for new type hash_table.
10788 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
10789 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
10791 * gimplify-ctx.h: New.
10792 (struct gimple_temp_hash_elt): Move from gimplify.c.
10793 (class gimplify_hasher): New.
10794 (struct gimplify_ctx): Move from gimple.h.
10795 (htab_t gimplify_ctx::temp_htab):
10796 Change type to hash_table. Update dependent calls and types.
10798 * graphite-clast-to-gimple.c: Include graphite-htab.h.
10799 (htab_t ivs_params::newivs_index):
10800 Change type to hash_table. Update dependent calls and types.
10801 (htab_t ivs_params::params_index): Likewise.
10802 (htab_t print_generated_program::params_index): Likewise.
10803 (htab_t gloog::newivs_index): Likewise.
10804 (htab_t gloog::params_index): Likewise.
10806 * graphite.c: Include graphite-htab.h.
10807 4htab_t graphite_transform_loops::bb_pbb_mapping):
10808 Change type to hash_table. Update dependent calls and types.
10810 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
10811 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
10812 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
10814 * graphite-dependences.c: Include graphite-htab.h.
10815 (loop_is_parallel_p): Change hash table type of parameter.
10817 * graphite-htab.h: New.
10818 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
10819 (extern find_pbb_via_hash): Move from graphite-poly.h.
10820 (extern loop_is_parallel_p): Move from graphite-poly.h.
10821 (extern get_loop_body_pbbs): Move from graphite-poly.h.
10823 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
10824 (extern loop_is_parallel_p): Move to graphite-htab.h.
10825 (extern get_loop_body_pbbs): Move to graphite-htab.h.
10827 * haifa-sched.c (htab_t delay_htab):
10828 Change type to hash_table. Update dependent calls and types.
10829 (htab_t delay_htab_i2): Likewise.
10831 * ira-color.c (htab_t allocno_hard_regs_htab):
10832 Change type to hash_table. Update dependent calls and types.
10834 * ira-costs.c (htab_t cost_classes_htab):
10835 Change type to hash_table. Update dependent calls and types.
10837 * loop-invariant.c (htab_t merge_identical_invariants::eq):
10838 Change type to hash_table. Update dependent calls and types.
10840 * loop-iv.c (htab_t bivs):
10841 Change type to hash_table. Update dependent calls and types.
10843 * loop-unroll.c (htab_t opt_info::insns_to_split):
10844 Change type to hash_table. Update dependent calls and types.
10845 (htab_t opt_info::insns_with_var_to_expand): Likewise.
10847 * lto-streamer.h (struct string_slot): Move from data-streamer.h
10848 (struct string_slot_hasher): New.
10849 (htab_t output_block::string_hash_table):
10850 Change type to hash_table. Update dependent calls and types.
10852 * lto-streamer-in.c (freeing_string_slot_hasher): New.
10853 (htab_t file_name_hash_table):
10854 Change type to hash_table. Update dependent calls and types.
10856 * lto-streamer-out.c: Update output_block::string_hash_table dependent
10859 * lto-streamer.c (htab_t tree_htab):
10860 Change type to hash_table. Update dependent calls and types.
10862 * omp-low.c: Include gimplify-ctx.h.
10864 * passes.c (htab_t name_to_pass_map):
10865 Change type to hash_table. Update dependent calls and types.
10866 (pass_traverse): Rename to passes_pass_traverse.
10868 * plugin.c (htab_t event_tab):
10869 Change type to hash_table. Update dependent calls and types.
10871 * postreload-gcse.c (htab_t expr_table):
10872 Change type to hash_table. Update dependent calls and types.
10873 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
10875 * sese.c (debug_rename_map_1): Make extern.
10876 (htab_t copy_bb_and_scalar_dependences::rename_map):
10877 Change type to hash_table. Update dependent calls and types.
10879 * sese.h (extern debug_rename_map): Move to .c file.
10881 * store-motion.c (htab_t store_motion_mems_table):
10882 Change type to hash_table. Update dependent calls and types.
10884 * trans-mem.c (htab_t tm_new_mem_hash):
10885 Change type to hash_table. Update dependent calls and types.
10887 * tree-browser.c (htab_t TB_up_ht):
10888 Change type to hash_table. Update dependent calls and types.
10890 * tree-cfg.c (htab_t discriminator_per_locus):
10891 Change type to hash_table. Update dependent calls and types.
10893 * tree-complex.c: Include tree-hasher.h
10894 (htab_t complex_variable_components):
10895 Change type to hash_table. Update dependent calls and types.
10897 * tree-eh.c (htab_t finally_tree):
10898 Change type to hash_table. Update dependent calls and types.
10900 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
10901 struct int_tree_hasher.
10902 (extern int_tree_map_eq): Likewise.
10903 (uid_decl_map_hash): Removed.
10904 (extern decl_tree_map_eq): Likewise.
10906 * tree-hasher.h: New.
10907 (struct int_tree_hasher): New.
10908 (typedef int_tree_htab_type): New.
10910 * tree-inline.c: Include gimplify-ctx.h.
10912 * tree-mudflap.c: Include gimplify-ctx.h.
10914 * tree-parloops.c: Include tree-hasher.h.
10915 (htab_t eliminate_local_variables_stmt::decl_address):
10916 Change type to hash_table. Update dependent calls and types.
10917 (htab_t separate_decls_in_region::decl_copies): Likewise.
10919 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
10920 Change type to hash_table. Update dependent calls and types.
10922 * tree-sra.c (candidates):
10923 Change type to hash_table. Update dependent calls and types.
10925 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
10927 (int_tree_map_hash): Likewise.
10929 * tree-ssa-dom.c (htab_t avail_exprs):
10930 Change type to hash_table. Update dependent calls and types.
10932 * tree-ssa-live.c (var_map_base_init::tree_to_index):
10933 Change type to hash_table. Update dependent calls and types.
10935 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
10936 Change type to hash_table. Update dependent calls and types.
10938 * tree-ssa-phiopt.c (seen_ssa_names):
10939 Change type to hash_table. Update dependent calls and types.
10941 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
10942 Change type to hash_table. Update dependent calls and types.
10944 * tree-ssa-uncprop.c (equiv):
10945 Change type to hash_table. Update dependent calls and types.
10947 2013-04-25 Jakub Jelinek <jakub@redhat.com>
10949 PR rtl-optimization/57003
10950 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
10951 call note_stores with kill_clobbered_value callback again after
10952 killing regs_invalidated_by_call.
10954 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
10956 * config/aarch64/aarch64-simd.md
10957 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
10958 (aarch64_simd_bsl<mode>): Likewise.
10959 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
10961 2013-04-25 Marek Polacek <polacek@redhat.com>
10963 PR tree-optimization/57066
10964 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
10966 2013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
10968 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
10970 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
10972 * config/aarch64/aarch64-builtins.c
10973 (aarch64_fold_builtin): New.
10974 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
10975 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
10976 * config/aarch64/aarch64-simd-builtins.def (abs): New.
10977 * config/aarch64/arm_neon.h
10978 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
10980 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
10981 Tejas Belagod <tejas.belagod@arm.com>
10983 * config/aarch64/aarch64-builtins.c
10984 (aarch64_gimple_fold_builtin): New.
10985 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
10986 * config/aarch64/aarch64-simd-builtins.def (addv): New.
10987 * config/aarch64/aarch64-simd.md (addpv4sf): New.
10988 (addvv4sf): Update.
10989 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
10991 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
10993 * config/aarch64/aarch64.md
10994 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
10996 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
10998 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
10999 (*ngcsi_uxtw): New pattern.
11001 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11002 Julian Brown <julian@codesourcery.com>
11004 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
11005 (TB_DREG): Add T_V4HF.
11006 (v4hf_UP): New macro.
11007 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
11008 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
11009 Handle initialisation of V4HF. Adjust initialisation of reinterpret
11011 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
11012 (arm_vector_mode_supported_p): Handle V4HF.
11013 (arm_mangle_map): Handle V4HFmode.
11014 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
11015 * config/arm/arm_neon_builtins.def: Add entries for
11016 vcvtv4hfv4sf, vcvtv4sfv4hf.
11017 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
11018 (neon_vcvtv4hfv4sf): Likewise.
11019 * config/arm/neon-gen.ml: Handle half-precision floating point
11021 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
11022 * config/arm/arm_neon.h: Regenerate.
11023 * config/arm/neon.ml (type elts): Add F16.
11024 (type vectype): Add T_float16x4, T_floatHF.
11025 (type vecmode): Add V4HF.
11026 (type features): Add Requires_FP_bit feature.
11027 (elt_width): Handle F16.
11028 (elt_class): Likewise.
11029 (elt_of_class_width): Likewise.
11030 (mode_of_elt): Refactor.
11031 (type_for_elt): Handle F16, fix error messages.
11032 (vectype_size): Handle T_float16x4.
11033 (vcvt_sh): New function.
11034 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
11035 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
11036 (string_of_mode): Handle V4HF.
11037 * doc/arm-neon-intrinsics.texi: Regenerate.
11039 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
11041 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
11042 format specifier in 'X' case.
11044 2013-04-25 Alan Modra <amodra@gmail.com>
11047 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
11048 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
11049 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
11050 Repeat for many other rotate/shift and mask patterns using subregs.
11051 Name lshiftrt insns.
11052 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
11053 on WORDS_BIG_ENDIAN.
11055 2013-04-25 Alan Modra <amodra@gmail.com>
11057 * config.gcc: Support little-endian powerpc-linux targets.
11058 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
11059 (LINK_OS_LINUX_SPEC): Define.
11060 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
11061 Preserve MASK_LITTLE_ENDIAN.
11062 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
11063 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
11064 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
11065 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
11066 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
11067 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
11068 Correct fp word order for little-endian. Don't shift toc entries
11069 smaller than a word for little-endian.
11070 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
11071 (bswapdi2 splits): Correct low-part subreg for little-endian.
11072 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
11073 low/high where such is correct only for be.
11074 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
11075 little-endian for -mcall-aixdesc.
11077 2013-04-25 Alan Modra <amodra@gmail.com>
11079 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
11080 replace_equiv_address_nv.
11082 2013-04-25 Alan Modra <amodra@gmail.com>
11084 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
11086 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
11089 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
11090 * rtl.h (struct rtx_def): ...
11092 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
11094 PR rtl-optimizations/57046
11095 * lra-constraints (split_reg): Set up lra_risky_transformations_p
11096 for multi-reg splits.
11098 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
11100 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
11102 2013-04-24 Sterling Augustine <saugustine@google.com>
11104 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
11105 (comp_dir_string, debug_str_dwo_section): New.
11106 (DEBUG_STR_DWO_SECTION): Rename to ...
11107 (DEBUG_DWO_STR_SECTION): ... this.
11108 (DEBUG_NORM_STR_SECTION): Delete.
11109 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
11110 (DEBUG_STR_DWO_SECTION_FLAGS): New.
11111 (find_AT_string): Move most logic to ...
11112 (find_AT_string_in_table): ... here. New.
11113 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
11114 add_skeleton_AT_string. Delete logic.
11115 (output_skeleton_debug_sections): Remove call to
11116 add_top_level_skeleton_die_attrs.
11117 (add_comp_dir_attribute): Move logic to comp_dir_string.
11118 (dwarf2out_init): Initialize debug_str_dwo_section.
11119 (output_indirect_string): Call find_string_form.
11120 (output_indirect_strings): Rewrite.
11121 (prune_unused_types): Empty skeleton_debug_str_hash.
11122 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
11123 (dwarf2out_finish): Call output_indirect_strings.
11125 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
11127 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
11129 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
11131 * rtl.h (struct rtx_def): Add comment for field jump.
11132 (LRA_SUBREG_P): New macro.
11133 * recog.c (register_operand): Check LRA_SUBREG_P.
11134 * lra.c (lra): Add note at the end of RTL code. Align non-empty
11136 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
11137 (lra_final_code_change): Skip subreg change for operators.
11138 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
11139 if there are no operand changes.
11140 * lra-constraints.c (curr_insn_set): New.
11141 (match_reload): Set LRA_SUBREG_P.
11142 (emit_spill_move): Ditto.
11143 (check_and_process_move): Use curr_insn_set. Process only single
11144 set insns. Don't initialize sec_mem_p and change_p.
11145 (simplify_operand_subreg): Use LRA_SUBREG_P.
11146 (reg_in_class_p): New function.
11147 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
11148 of #ifdef. Add code to remove cycling.
11149 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
11150 non-null disp. Reload inner instead of disp when base and index
11151 are null. Try to put lo_sum into register.
11152 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
11153 (check_and_process_move): Move code for move cost check to
11154 simple_move_p. Remove equiv_substitution.
11155 (simple_move_p): New function.
11156 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
11157 curr_insn_set. Call check_and_process_move only for single set
11158 insns. Use the new function. Move call of check_and_process_move
11159 after operand equiv substitution and address process.
11161 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
11163 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
11164 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
11165 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
11167 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
11169 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
11171 2013-04-24 Marek Polacek <polacek@redhat.com>
11173 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
11174 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
11175 (select_loops_exit_conditions): Likewise.
11176 (number_of_iterations_for_all_loops): Likewise.
11177 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
11178 (scev_analysis): Likewise.
11180 2013-04-02 Catherine Moore <clm@codesourcery.com>
11181 Chao-ying Fu <fu@mips.com>
11183 * config/mips/micromips.md (jraddiusp): New pattern.
11184 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
11185 instruction if possible.
11187 2013-04-24 Alan Modra <amodra@gmail.com>
11189 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
11191 2013-04-24 Julian Brown <julian@codesourcery.com>
11192 Chung-Lin Tang <cltang@codesourcery.com>
11194 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
11195 dependency behavior in enumeration type DIE generation. Add TODO note
11196 to comments about future DW_FORM_sdata/udata re-work of related code.
11198 2013-04-23 Lawrence Crowl <crowl@google.com>
11200 * Makefile.in: Update as needed below.
11202 * hash-table.h (class hash_table):
11203 Correct many methods with parameter types compare_type to the correct
11204 value_type. (Correct code was unlikely to notice the change.)
11205 (hash_table::elements_with_deleted) New.
11206 (class hashtable::iterator): New.
11207 (hashtable::begin()): New.
11208 (hashtable::end()): New.
11209 (FOR_EACH_HASH_TABLE_ELEMENT): New.
11211 * statistics.c (statistics_hashes):
11212 Change type to hash_table. Update dependent calls and types.
11214 * tree-into-ssa.c (var_infos):
11215 Change type to hash_table. Update dependent calls and types.
11217 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
11218 Change type to hash_table. Update dependent calls and types.
11220 * tree-ssa-loop-im.c (struct mem_ref.refs):
11221 Change type to hash_table. Update dependent calls and types.
11223 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
11224 Change type to hash_table. Update dependent calls and types.
11226 * tree-ssa-sccvn.c (vn_tables_s::nary):
11227 Change type to hash_table. Update dependent calls and types.
11228 (vn_tables_s::phis): Likewise.
11229 (vn_tables_s::references): Likewise.
11231 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
11232 (vn_reference_eq): Update parameter and return types.
11234 * tree-ssa-structalias.c (pointer_equiv_class_table):
11235 Change type to hash_table. Update dependent calls and types.
11236 (location_equiv_class_table): Likewise.
11238 * tree-vect-data-refs.c: Consequential changes for making
11239 peeling a hash_table.
11241 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
11242 (destroy_loop_vec_info): Dependent hash_table update.
11244 * tree-vectorizer.h (peeling_htab):
11245 Change type to hash_table. Update dependent calls and types.
11247 2013-04-23 Shiva Chen <shiva0217@gmail.com>
11249 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
11250 to check the register content is equal or not.
11251 * lra-constraints.c (match_reload): Use lra_assign_reg_val
11252 to assign register content record.
11253 * lra-eliminations.c (update_reg_eliminate): Use
11254 lra_update_reg_val_offset to update register content offset.
11255 * lra-int.h (struct lra_reg): Add offset member.
11256 (lra_reg_val_equal_p): New static inline function.
11257 (lra_update_reg_val_offset): New static inline function.
11258 (lra_assign_reg_val): New static inline function.
11259 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
11260 to assign register content record.
11261 (initialize_lra_reg_info_element): Initial offset to zero.
11263 2013-04-23 Catherine Moore <clm@codesourcery.com>
11265 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
11266 operands. Record compression.
11268 2013-04-23 Xinliang David Li <davidxl@google.com>
11270 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
11272 2013-04-23 Richard Biener <rguenther@suse.de>
11274 PR middle-end/57036
11275 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
11276 parameter, only add abnormal goto edges from the copied body
11277 if the call could perform abnormal gotos.
11278 (copy_cfg_body): Adjust.
11280 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
11282 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
11284 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
11286 * coretypes.h (gimple_stmt_iterator): Add struct to make
11289 2013-04-23 Richard Biener <rguenther@suse.de>
11291 PR tree-optimization/57026
11292 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
11293 from SSA names occuring in abnormal PHI nodes.
11295 2013-04-22 Andi Kleen <ak@linux.intel.com>
11297 * lto/lto.c (print_lto_report_1): Fix LTO report names.
11299 2013-04-22 Andi Kleen <ak@linux.intel.com>
11301 * lto/lto.c (print_lto_report_1): Declare early.
11302 (read_cgraph_and_symbols): Call print_lto_report_1 early.
11304 2013-04-22 Andi Kleen <ak@linux.intel.com>
11306 * common.opt (-flto-report-wpa): Add.
11307 * doc/invoke.texi (-flto-report-wpa): Add.
11308 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
11311 2013-04-22 Xinliang David Li <davidxl@google.com>
11313 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
11314 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
11315 * Makefile.in: New dependency
11317 David Daney <ddaney.cavm@gmail.com>
11319 * configure.ac (gcc_cv_as_micromips_support): Use the
11320 --fatal-warnings option.
11321 * configure: Regenerate.
11323 2013-04-22 Marek Polacek <polacek@redhat.com>
11326 * tsan.c (instrument_expr): Don't instrument expression
11327 in case its size is zero.
11329 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
11333 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
11335 * config/alpha/alpha.c (TARGET_LRA_P): New define.
11337 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
11339 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
11340 (gimple_stmt_iterator): New typedef.
11341 * gimple.h (gimple_stmt_iterator): Rename to...
11342 (gimple_stmt_iterator_d): ... This.
11343 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
11344 trees be valid for GIMPLE and GENERIC.
11345 (TARGET_GIMPLE_FOLD_BUILTIN): New.
11346 * gimple-fold.c (gimple_fold_call): Call target hook
11347 gimple_fold_builtin.
11348 * hooks.c (hook_bool_gsiptr_false): New.
11349 * hooks.h (hook_bool_gsiptr_false): New.
11350 * target.def (fold_stmt): New.
11351 * doc/tm.texi: Regenerate.
11353 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
11356 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
11357 a set sp if no stack realignment.
11359 2013-04-22 Nick Clifton <nickc@redhat.com>
11361 * config.gcc (tilegx-linux): Extend extra_objs rather than
11363 (tilepro-linux): Likewise.
11365 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
11367 * config/aarch64/aarch64-builtins.c
11369 (CF0, CF1, CF2, CF3, CF4, CF10): New.
11370 (VAR<1-12>): Add MAP parameter.
11371 (BUILTIN_*): Likewise.
11372 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
11373 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
11374 (aarch64_ushl_n<mode>): Likewise.
11375 (aarch64_sshr_n<mode>): Likewise.
11376 (aarch64_ushr_n<mode>): Likewise.
11377 (aarch64_<maxmin><mode>): Likewise.
11378 (aarch64_sqrt<mode>): Likewise.
11379 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
11380 (vshr<q>_n_*): Likewise.
11382 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
11384 * config/aarch64/aarch64-builtins.c
11385 (aarch64_simd_builtin_type_mode): Handle SF types.
11387 (BUILTIN_GPF): Define.
11388 (aarch64_init_simd_builtins): Handle SF types.
11389 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
11390 (frecps): Likewise.
11391 (frecpx): Likewise.
11392 * config/aarch64/aarch64-simd.md
11393 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
11394 (aarch64_frecpe<mode>): New.
11395 (aarch64_frecps<mode>): Likewise.
11396 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
11397 (v8type): Add frecp<esx>.
11398 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
11399 (aarch64_frecps<mode>): Likewise.
11400 * config/aarch64/iterators.md (FRECP): New.
11401 (frecp_suffix): Likewise.
11402 * config/aarch64/arm_neon.h
11403 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
11405 2013-04-22 Christian Bruel <christian.bruel@st.com>
11408 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
11409 (REG_CLASS_NAMES): Idem.
11410 (REG_CLASS_CONTENTS): Idem.
11411 (REGCLASS_HAS_FP_REG): Idem.
11412 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
11413 (sh_conditional_register_usage): Idem.
11415 2013-04-21 Jeff Law <law@redhat.com>
11417 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
11418 (ssa_forward_propagate_and_combine): Use it.
11420 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
11422 * lra.c: Update the flow chart diagram.
11424 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
11426 PR rtl-optimization/56847
11427 * lra-constraints.c (process_alt_operands): Discourage alternative
11428 with non-matche doffsettable memory constraint fro memory with
11431 2013-04-19 Richard Biener <rguenther@suse.de>
11433 PR tree-optimization/56982
11434 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
11436 * gimplify.c (gimplify_call_expr): Notice special calls.
11437 (gimplify_modify_expr): Likewise.
11438 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
11439 abnormal control flow receivers.
11440 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
11441 in the same way as cfun->has_nonlocal_labels.
11442 (gimple_purge_dead_abnormal_call_edges): Likewise.
11443 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
11444 receivers start a basic-block.
11446 2013-04-19 Richard Biener <rguenther@suse.de>
11448 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
11450 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
11451 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
11452 (SLP_TREE_LOAD_PERMUTATION): Add.
11453 (vect_transform_slp_perm_load): Adjust prototype.
11454 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
11455 (vect_free_slp_instance): Likewise.
11456 (vect_create_new_slp_node): Likewise.
11457 (vect_supported_slp_permutation_p): Remove.
11458 (vect_slp_rearrange_stmts): Adjust.
11459 (vect_supported_load_permutation_p): Likewise. Inline
11460 vect_supported_slp_permutation_p here.
11461 (vect_analyze_slp_instance): Compute load permutations per
11462 slp node instead of per instance.
11463 (vect_get_slp_defs): Adjust.
11464 (vect_transform_slp_perm_load): Likewise.
11465 (vect_schedule_slp_instance): Remove redundant code.
11466 (vect_schedule_slp): Remove hack for PR56270, add it ...
11467 * tree-vect-stmts.c (vectorizable_load): ... here, do not
11468 CSE loads for SLP. Adjust.
11470 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
11472 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
11473 spelling in two comments.
11475 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
11478 * config/arm/arm.c (load_multiple_sequence): Require SP
11479 as base register for loads if SP is in the register list.
11481 2013-04-19 Martin Jambor <mjambor@suse.cz>
11483 PR tree-optimization/56718
11484 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
11485 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
11486 and made public. Adjusted all callers.
11487 (ipa_intraprocedural_devirtualization): New function.
11488 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
11489 (ipa_intraprocedural_devirtualization): Likewise.
11490 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
11492 2013-04-19 Richard Biener <rguenther@suse.de>
11494 PR tree-optimization/57000
11495 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
11497 2013-04-19 Terry Guo <terry.guo@arm.com>
11499 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
11501 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
11502 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
11503 (cortex_m4_fmacs): Use new reservations.
11504 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
11506 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
11508 PR rtl-optimization/56999
11509 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
11511 (lra_coalesce): Remove split_origin_bitmap and related code.
11512 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
11513 ranges if necessary.
11515 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
11517 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
11519 (ix86_expand_call): Remove clobbered_registers array and use
11520 x86_64_ms_sysv_extra_clobbered_registers instead.
11521 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
11523 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
11525 * config/i386/i386.md (*call_rex64_ms_sysv): Use
11526 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
11527 (*call_value_rex64_ms_sysv): Ditto.
11529 2013-04-18 Cary Coutant <ccoutant@google.com>
11531 * dwarf2out.c (output_pubnames): Check die_perennial_p of
11532 parent instead of die_mark.
11534 2013-04-18 Diego Novillo <dnovillo@google.com>
11536 * gimple.c (create_gimple_tmp): New.
11537 (get_expr_type): New.
11538 (build_assign): New.
11539 (build_type_cast): New.
11540 * gimple.h (enum ssa_mode): Define.
11541 (gimple_seq_set_location): New.
11542 * asan.c (build_check_stmt): Change some gimple_build_* calls
11543 to use build_assign and build_type_cast.
11545 2013-04-18 Richard Biener <rguenther@suse.de>
11547 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
11548 handle negative step. Remove redundant checks.
11549 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
11550 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
11551 for negative step and grouped loads fail to vectorize.
11553 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
11555 * emit-rtl.c (reset_insn_used_flags): New function.
11556 (reset_all_used_flags): Use it.
11557 (verify_insn_sharing): New function.
11558 (verify_rtl_sharing): Fix verification for SEQUENCEs.
11560 2013-04-18 Jakub Jelinek <jakub@redhat.com>
11562 PR tree-optimization/56984
11563 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
11564 and (x >> M) >= N don't register any assertion if N << M is the
11567 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
11569 * lower-subreg.c (resolve_simple_move): If called self-recursive,
11570 do not delete_insn insns that have not yet been emitted, only
11571 unlink them with remove_insn.
11572 * df-scan.c (df_insn_delete): Revert r197492.
11574 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
11576 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
11577 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
11579 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
11581 * config/arm/arm.md (movsicc_insn): Convert define_insn into
11582 define_insn_and_split.
11583 (and_scc,ior_scc,negscc): Likewise.
11584 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
11586 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
11588 * config/arm/arm.c (use_return_insn): Return 0 for targets that
11589 can benefit from using a sequence of LDRD instructions in epilogue
11590 instead of a single LDM instruction.
11592 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
11595 * doc/extend.texi: Fix typo.
11597 2013-04-17 Richard Biener <rguenther@suse.de>
11599 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
11600 (vect_build_slp_tree): ... here.
11601 (vect_build_slp_tree_1): Compute which stmts of the SLP group
11602 match. Remove special-casing of mismatched complex loads.
11603 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
11604 re-try the match with swapped commutative operands.
11605 (vect_supported_load_permutation_p): Remove special-casing of
11606 mismatched complex loads.
11607 (vect_analyze_slp_instance): Adjust.
11609 2013-04-17 Richard Biener <rguenther@suse.de>
11611 PR rtl-optimization/56921
11612 * cfgloop.h (struct loop): Add simple_loop_desc member.
11613 (struct niter_desc): Mark with GTY(()).
11614 (simple_loop_desc): Do not use aux field but simple_loop_desc.
11615 * loop-iv.c (get_simple_loop_desc): Likewise.
11616 (free_simple_loop_desc): Likewise.
11619 2013-04-16 Richard Biener <rguenther@suse.de>
11621 PR rtl-optimization/56921
11622 * loop-init.c (pass_rtl_move_loop_invariants): Add
11623 TODO_do_not_ggc_collect to todo_flags_finish.
11624 (pass_rtl_unswitch): Same.
11625 (pass_rtl_unroll_and_peel_loops): Same.
11626 (pass_rtl_doloop): Same.
11628 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
11630 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
11631 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
11632 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
11634 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
11635 * tree-streamer.c (record_common_node): Adjust reference in comment.
11637 2013-04-17 Terry Guo <terry.guo@arm.com>
11639 * config/arm/cortex-m4.md: Add a new bypass.
11641 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
11643 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
11645 (*subs_<optab><mode>_multp2): New pattern.
11646 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
11647 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
11649 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
11651 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
11652 (*subs_mul_imm_<mode>): New pattern.
11654 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
11657 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
11658 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
11659 (vsx_movti_32bit): Change j->wa to O->wa.
11661 2013-04-16 Richard Biener <rguenther@suse.de>
11663 PR rtl-optimization/56921
11664 * loop-init.c (pass_rtl_move_loop_invariants): Add
11665 TODO_do_not_ggc_collect to todo_flags_finish.
11666 (pass_rtl_unswitch): Same.
11667 (pass_rtl_unroll_and_peel_loops): Same.
11668 (pass_rtl_doloop): Same.
11670 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
11672 * config/arm/arm.c (emit_multi_reg_push): New declaration
11673 for an existing function.
11674 (arm_emit_strd_push): New function.
11675 (arm_expand_prologue): Used here.
11676 (arm_emit_ldrd_pop): New function.
11677 (arm_expand_epilogue): Used here.
11678 (arm_get_frame_offsets): Update condition.
11679 (arm_emit_multi_reg_pop): Add a special case for load of a single
11680 register with writeback.
11682 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
11684 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
11687 2013-04-16 Richard Biener <rguenther@suse.de>
11689 PR tree-optimization/56756
11690 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
11691 (first_mem_ref_loc): New.
11692 (execute_sm): Place the load temporarily before a previous
11693 access instead of in the latch edge to ensure its SSA dependencies
11694 are defined at points dominating the load.
11696 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
11698 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
11699 correct fix by moving header and footer insn to the footer of
11700 the merged basic block. Clear BB_END of the merged-away block.
11702 PR middle-end/43631
11703 * emit-rtl.c (make_note_raw): New function.
11704 (link_insn_into_chain): New static inline function.
11705 (add_insn): Use it.
11706 (add_insn_before, add_insn_after): Factor insn chain linking code...
11707 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
11708 using link_insn_into_chain.
11709 (note_outside_basic_block_p): New helper function for emit_note_after
11710 and emit_note_before.
11711 (emit_note_after): Use nobb variant of add_insn_after if the note
11712 should not be contained in a basic block.
11713 (emit_note_before): Use nobb variant of add_insn_before if the note
11714 should not be contained in a basic block.
11715 (emit_note_copy): Use make_note_raw.
11716 (emit_note): Likewise.
11717 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
11718 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
11719 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
11720 the moved barrier the tail of the basic block it follows.
11721 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
11723 2013-04-15 Jakub Jelinek <jakub@redhat.com>
11725 PR tree-optimization/56962
11726 * gimple-ssa-strength-reduction.c (record_increment): Only set
11727 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
11728 either rhs1 or rhs2 is equal to c->base_expr.
11730 2013-04-15 Richard Biener <rguenther@suse.de>
11732 PR tree-optimization/56933
11733 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
11735 (GROUP_READ_WRITE_DEPENDENCE): Remove.
11736 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
11737 * tree-vect-data-refs.c (vect_analyze_group_access): Move
11738 dependence check ...
11739 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
11741 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
11742 GROUP_READ_WRITE_DEPENDENCE.
11744 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11746 * emit-rtl.c (reset_all_used_flags): New function.
11747 (verify_rtl_sharing): Call reset_all_used_flags before and after
11748 performing the checks.
11750 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11752 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
11753 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
11754 * config/arm/constraints.md (De): New constraint.
11755 * config/arm/neon.md (anddi3_neon): Delete.
11756 (neon_vand<mode>): Expand to standard anddi3 pattern.
11757 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
11758 Move earlier in the file.
11759 (neon_inv_logic_op2): Likewise.
11760 (arm_anddi_operand_neon): New predicate.
11762 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
11764 * configure.ac (gcc_cv_ld_as_needed): Set
11765 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
11766 Use -z ignore, -z record on *-*-solaris2*.
11767 (HAVE_LD_AS_NEEDED): Update comment.
11768 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
11769 * configure: Regenerate.
11770 * config.in: Regenerate.
11771 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
11772 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
11773 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
11774 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
11775 equivalents. Fix markup.
11776 * doc/tm.texi: Regenerate.
11778 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
11780 * config/i386/i386.opt: New option mstack-protector-guard=.
11781 * config/i386/i386-opts.h: Add enum stack_protector_guard.
11782 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
11783 TARGET_SSP_TLS_GUARD.
11784 * config/i386/i386.c (ix86_option_override_internal): Set
11785 ix86_stack_protector_guard.
11786 * config/i386/i386.md (stack_protect_set): Enable for
11787 TARGET_SSP_TLS_GUARD only.
11788 (stack_protect_set_<mode>): Ditto.
11789 (stack_protect_test): Ditto.
11790 (stack_protect_test_<mode>): Ditto.
11791 * doc/invoke.texi (i386 Option): Document.
11793 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
11796 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
11797 (S_MODES): Set H_MODE bit.
11798 (SF_MODES): Set only S_MODE and SF_MODE bits.
11799 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
11800 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
11801 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
11802 <MODE_FLOAT>: Likewise.
11804 2013-04-15 Joey Ye <joey.ye@arm.com>
11806 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
11808 2013-04-15 Joey Ye <joey.ye@arm.com>
11810 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
11812 (thumb_far_jump_used_p): Count instruction size and set
11815 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
11817 * reorg.c (fill_simple_delay_slots): Reindent block of code.
11818 * resource.c (mark_target_live_regs): Reformat conditional block.
11820 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
11822 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
11823 notes, they are emitted only just before final.
11824 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
11826 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
11828 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
11829 * cfgrtl.c (delete_insn): Call it here instead.
11830 * lra-spills.c (lra_final_code_change): Use delete_insn.
11831 * haifa-sched.c (sched_remove_insn): Likewise.
11832 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
11833 returning to the nop pool.
11834 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
11835 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
11837 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
11839 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
11840 * doc/tm.texi: Regenerated.
11842 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
11844 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
11847 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
11849 * df-core.c (df_find_def): Compare register numbers.
11850 (df_find_use): Likewise.
11852 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
11855 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
11856 lra_in_progress for return.
11858 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
11860 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
11861 define_insn into define_insn_and_split and emit movsicc patterns.
11863 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
11865 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
11867 2013-04-12 Richard Biener <rguenther@suse.de>
11869 * tree-pass.h (TODO_do_not_ggc_collect): New.
11870 * passes.c (execute_one_ipa_transform_pass): Honor
11871 TODO_do_not_ggc_collect.
11872 (execute_one_pass): Likewise.
11875 2013-04-10 Richard Biener <rguenther@suse.de>
11877 * passes.c (init_optimization_passes): Remove reload pass.
11878 * ira.c (do_reload): Merge into ...
11880 (rest_of_handle_reload): Remove.
11881 (pass_reload): Likewise.
11882 * config/i386/i386.c (ix86_option_override): Refer to ira instead
11883 of reload for vzeroupper pass placement.
11885 2013-04-12 Jakub Jelinek <jakub@redhat.com>
11887 PR tree-optimization/56918
11888 PR tree-optimization/56920
11889 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
11890 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
11891 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
11892 use wide_mul_with_sign method.
11894 2013-04-12 Richard Biener <rguenther@suse.de>
11896 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
11897 not be considered a gimple constant.
11899 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
11901 * fold-const.c (const_binop): Handle vector shifts by a scalar.
11902 (fold_binary_loc): Call const_binop also for mixed vector-scalar
11905 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
11906 Jakub Jelinek <jakub@redhat.com>
11908 * opts.c: Include diagnostic-color.h.
11909 (common_handle_option): Handle OPT_fdiagnostics_color_.
11910 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
11911 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
11912 (diagnostic-color.o): New.
11913 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
11914 (diagnostic_color_rule): New enum.
11915 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
11916 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
11917 the location string.
11918 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
11919 either NULL, or color kind.
11920 * diagnostic-color.c: New file.
11921 * diagnostic-color.h: New file.
11922 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
11924 * doc/invoke.texi (-fdiagnostics-color): Document.
11925 * pretty-print.h (pp_show_color): Define.
11926 (struct pretty_print_info): Add show_color field.
11927 * diagnostic.c: Include diagnostic-color.h.
11928 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
11929 macros. Colorize error:, warning: etc. strings and also the location
11931 (diagnostic_show_locus): Colorize the caret line.
11932 * pretty-print.c: Include diagnostic-color.h.
11933 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
11934 inside of %< %> quotes or quoted through q format modifier.
11936 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11938 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
11940 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
11942 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
11943 code in CC_NZ mode.
11944 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
11947 2013-04-11 Marek Polacek <polacek@redhat.com>
11949 PR tree-optimization/48184
11950 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
11952 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
11954 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
11955 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
11956 (skip_simple_arithmetic): Tidy up.
11957 * tree.h (skip_simple_constant_arithmetic): Declare.
11959 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
11961 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
11963 2013-04-11 Richard Biener <rguenther@suse.de>
11965 * tree-vect-loop.c (get_initial_def_for_induction): Properly
11966 generate vector constants.
11968 2013-04-11 Richard Biener <rguenther@suse.de>
11970 PR tree-optimization/56878
11971 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
11972 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
11974 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
11975 Prefer to align the DR with the most invariant base address.
11977 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
11979 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
11982 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
11984 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
11985 floating-point vector comparisons against 0.
11987 2013-04-11 Jakub Jelinek <jakub@redhat.com>
11989 PR tree-optimization/56899
11990 * fold-const.c (extract_muldiv_1): Apply distributive law
11991 only if TYPE_OVERFLOW_WRAPS (ctype).
11993 2013-04-11 Bin Cheng <bin.cheng@arm.com>
11996 * ira-costs.c (scan_one_insn): Check whether the source rtx of
11997 loading has side effect.
11999 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
12001 * config/sparc/sparc.c: Include tree-pass.h.
12002 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
12003 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
12004 head of file. Change return type. Split off gate function.
12005 (sparc_gate_work_around_errata): New function.
12006 (pass_work_around_errata): New pass definition.
12007 (insert_pass_work_around_errata) New pass insert definition to
12008 insert pass_work_around_errata just after delayed-branch scheduling.
12009 (sparc_option_override): Insert the pass.
12010 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
12012 2013-04-10 David S. Miller <davem@davemloft.net>
12014 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
12015 or -mcpu=hypersparc.
12017 * target.def (cstore_mode): New hook.
12018 * target.h: Include insn-codes.h
12019 * targhooks.c: Likewise.
12020 (default_cstore_mode): New function.
12021 * targhooks.h: Declare it.
12022 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
12023 * doc/tm.texi: Rebuild.
12024 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
12025 target hook, rather than inspecting the insn_data.
12026 * config/sparc/sparc.c (sparc_cstore_mode): New function.
12027 (TARGET_CSTORE_MODE): Redefine.
12028 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
12030 * config/sparc/predicates.md (cstore_result_operand): New special
12032 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
12033 Use it for operand 0.
12034 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
12035 (*snesi_special): Likewise.
12036 (*snesi_zero): Likewise.
12037 (*seqsi_zero): Likewise.
12038 (*sltu_insn): Likewise.
12039 (*sgeu_insn): Likewise.
12040 (*seqdi_special): Make operand 0 and comparison operation be of
12042 (*snedi_special): Likewise.
12043 (*snedi_special_vis3): Likewise.
12044 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
12045 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
12046 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
12047 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
12048 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
12049 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
12050 (*sltu_extend_sp64): Likewise.
12051 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
12052 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
12053 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
12054 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
12055 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
12057 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
12059 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
12060 (aarch64_start_file): Use the new function.
12062 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
12063 Jason Merrill <jason@redhat.com>
12065 * common.opt: Add -gdwarf.
12066 * opts.c (common_handle_option): Handle it.
12067 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
12069 2013-04-10 Richard Biener <rguenther@suse.de>
12071 * passes.c (execute_todo): Do not call ggc_collect conditional here.
12072 (execute_one_ipa_transform_pass): But unconditionally here.
12073 (execute_one_pass): And here.
12074 (init_optimization_passes): Remove reload pass.
12075 * tree-pass.h (TODO_ggc_collect): Remove.
12076 (pass_reload): Likewise.
12077 * ira.c (do_reload): Merge into ...
12079 (rest_of_handle_reload): Remove.
12080 (pass_reload): Likewise.
12081 * config/i386/i386.c (ix86_option_override): Refer to ira instead
12082 of reload for vzeroupper pass placement.
12083 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
12084 and todo_flags_finish of all passes.
12086 2013-04-10 Richard Biener <rguenther@suse.de>
12088 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
12089 first_const_oprnd field, rename first_def_type to first_op_type.
12090 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
12091 (vect_get_and_check_slp_defs): Always use the type of the
12092 operand. Allow mixed vect_external_def, vect_constant_def types.
12093 (vect_get_constant_vectors): Handle mixed vect_external_def,
12094 vect_constant_def types.
12096 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
12098 PR tree-optimization/55524
12099 * tree-ssa-math-opts.c
12100 (convert_mult_to_fma): Don't use an fms construct
12101 when we don't have an fms operation, but fnma, and it looks
12102 likely that we'll be able to use the latter.
12104 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
12106 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
12108 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
12109 inline fail caused by overwritable functions.
12111 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
12113 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
12114 unnecessary bits in the constant power of two case.
12116 2013-04-10 Richard Biener <rguenther@suse.de>
12118 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
12119 broken code swapping operands.
12120 (vect_build_slp_tree): Do not compute load permutations here.
12121 (vect_analyze_slp_instance): Compute load permutations here,
12122 after building the SLP tree.
12124 2013-04-09 Christian Bruel <christian.bruel@st.com>
12126 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
12127 of next/prev_real_insn.
12129 2013-04-09 Jan Hubicka <jh@suse.cz>
12131 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
12132 Drop aliased parameter.
12133 (function_and_variable_visibility): Do not handle alias pairs.
12134 * cgraph.c (varpool_externally_visible_p): Update prototype.
12135 * varpool.c (varpool_add_new_variable): Update.
12137 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12139 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
12141 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
12143 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
12145 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
12147 2013-04-09 Marek Polacek <polacek@redhat.com>
12149 PR tree-optimization/48762
12150 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
12152 2013-04-09 Richard Biener <rguenther@suse.de>
12154 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
12156 (vect_build_slp_tree): Likewise.
12157 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
12158 calculating the cost of a SLP instance.
12159 (vect_analyze_slp_instance): Use it from here, after building
12162 2013-04-09 Jakub Jelinek <jakub@redhat.com>
12164 PR middle-end/56883
12165 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
12166 expand_omp_for_static_chunk): Use simple_p = true in
12167 force_gimple_operand_gsi calls when assigning to addressable decls.
12169 2013-04-09 Jeff Law <law@redhat.com>
12171 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
12172 when the boolean was created by converting a wider object which
12173 had a boolean range.
12175 2013-04-09 Richard Biener <rguenther@suse.de>
12177 * tree-vectorizer.h (slp_void_p): Remove.
12178 (slp_tree): Typedef before _slp_tree declaration.
12179 (struct _slp_tree): Use a vector of slp_tree as children.
12180 (vect_get_place_in_interleaving_chain): Remove.
12181 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
12183 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
12185 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
12186 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
12187 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
12188 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
12189 Use slp_node instead of slp_void_p and adjust.
12191 2013-04-09 Richard Biener <rguenther@suse.de>
12193 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
12194 work that is not necessary.
12196 2013-04-09 Jakub Jelinek <jakub@redhat.com>
12198 PR tree-optimization/56854
12199 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
12200 forward into clobber stmts if it would change MEM_REF lhs into
12203 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
12205 * tree.c (type_hash_lookup, type_hash_add): Make static.
12206 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
12208 2013-04-09 Richard Biener <rguenther@suse.de>
12210 * tree.h (unsave_expr_now): Remove.
12211 * tree-inline.c (mark_local_for_remap_r): Remove.
12212 (unsave_expr_1): Likewise.
12213 (unsave_r): Likewise.
12214 (unsave_expr_now): Likewise.
12215 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
12216 (propagate_tree_value): Likewise.
12218 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
12220 * doc/rtl.texi (sequence): Rewrite documentation to match the
12221 current use of SEQUENCE rtl objects.
12222 * rtl.def (SEQUENCE): Likewise.
12224 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
12225 Update documentation.
12226 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
12227 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
12229 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
12231 2013-04-08 Teresa Johnson <tejohnson@google.com>
12233 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
12234 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
12236 (estimate_edge_size_and_time): Add comment to suggest using rounding
12238 (estimate_node_size_and_time): Ditto.
12239 (remap_edge_change_prob): Use helper rounding divide methods.
12240 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
12241 (gimple_mod_pow2_value_transform): Ditto.
12242 (gimple_mod_subtract_transform): Ditto.
12243 (gimple_ic_transform): Ditto.
12244 (gimple_stringops_transform): Ditto.
12245 * stmt.c (conditional_probability): Ditto.
12246 (emit_case_dispatch_table): Ditto.
12247 * lto-cgraph.c (merge_profile_summaries): Ditto.
12248 * tree-optimize.c (execute_fixup_cfg): Ditto.
12249 * cfgcleanup.c (try_forward_edges): Ditto.
12250 * cfgloopmanip.c (scale_loop_profile): Ditto.
12252 (duplicate_loop_to_header_edge): Ditto.
12253 (lv_adjust_loop_entry_edge): Ditto.
12254 * tree-vect-loop.c (vect_transform_loop): Ditto.
12255 * profile.c (compute_branch_probabilities): Ditto.
12256 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
12257 * lto-streamer-in.c (input_cfg): Ditto.
12258 * gimple-streamer-in.c (input_bb): Ditto.
12259 * ipa-cp.c (update_profiling_info): Ditto.
12260 (update_specialized_profile): Ditto.
12261 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
12262 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
12264 * sched-rgn.c (compute_dom_prob_ps): Ditto.
12265 (compute_trg_info): Ditto.
12266 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
12267 (purge_dead_edges): Ditto.
12268 * loop-unswitch.c (unswitch_loop): Ditto.
12269 * cgraphclones.c (cgraph_clone_edge): Ditto.
12270 (cgraph_clone_node): Ditto.
12271 * tree-inline.c (copy_bb): Ditto.
12272 (copy_edges_for_bb): Ditto.
12273 (initialize_cfun): Ditto.
12274 (copy_cfg_body): Ditto.
12275 (expand_call_inline): Ditto.
12277 2013-04-08 Kai Tietz <ktietz@redhat.com>
12279 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
12280 TARGET_CYGWIN64 by TARGET_64BIT.
12282 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
12284 * config/epiphany/epiphany.md (GPR_1): New constant.
12285 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
12286 * config/epiphany/epiphany.c (gen_compare_reg):
12287 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
12288 is already in place.
12289 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
12290 Don't require being called during rtl expansion; If y operlaps r0,
12292 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
12293 (epiphany_expand_epilogue): Likewise.
12295 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
12296 Don't use CC_FPmode for ORDERED / UNORDERED.
12297 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
12299 * config/epiphany/constraints.md (CnL): New constraint.
12300 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
12301 * config/epiphany/predicates.md (add_operand): Allow 1024.
12303 * config/epiphany/epiphany.md (logical_op): New code iterator.
12304 (op_mnc): New code attribute.
12305 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
12306 (mov_f+1, mov_f+2): New peephole2 patterns.
12308 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
12309 (cstoresi4): Also allow re-use of zero result when doing a NE
12310 comparison to a non-zero operand.
12311 Use (clobber (scratch)) for first insn if the gpr output is not needed.
12313 * config/epiphany/epiphany.md (<insn_opname>v2si3):
12314 Use gen_addsi3_i / gen_subsi3_i.
12316 2013-04-08 Jakub Jelinek <jakub@redhat.com>
12320 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
12321 contain anything but clobbers, at most one __builtin_stack_restore,
12322 optionally debug stmts and final resx, and if it has at least one
12323 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
12324 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
12325 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
12326 which isn't defaut definition, remove them.
12327 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
12328 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
12329 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
12330 with MEM_REF LHS with SSA_NAME address.
12332 2013-04-08 Jeff Law <law@redhat.com>
12334 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
12336 2013-04-08 Richard Biener <rguenther@suse.de>
12338 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
12340 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
12341 determined vector type.
12342 (vect_analyze_data_refs): Likewise.
12343 (vect_get_new_vect_var): Adjust.
12344 (vect_create_destination_var): Preserve SSA name versions.
12345 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
12346 not dump anything here.
12348 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
12350 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
12351 Add member lr_slot_known.
12352 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
12354 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
12355 Remove code that sets lr_slot_offset according to what a previous
12356 version of epiphany_emit_save_restore used to do.
12357 (epiphany_emit_save_restore): When doing an lr save or restore,
12358 set/verify lr_slot_known and lr_slot_offset.
12360 2013-04-08 Xinyu Qi <xyqi@marvell.com>
12363 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
12366 2013-04-08 Richard Biener <rguenther@suse.de>
12368 * alias.c (find_base_term): Fix thinko in previous change.
12370 2013-04-08 Jakub Jelinek <jakub@redhat.com>
12372 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
12373 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
12374 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
12375 if possible to compute val.
12376 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
12377 For QImode integers don't require anything about precision. Use
12378 const_with_all_bytes_same to find out if the constant doesn't have
12379 repeated bytes in it.
12381 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12383 * config/s390/s390.c (s390_expand_insv): Only accept insertions
12386 2013-04-08 Marek Polacek <polacek@redhat.com>
12388 PR rtl-optimization/48182
12389 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
12392 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
12395 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
12396 nuses, make sure we have a label.
12398 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12401 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
12402 (rs6000_emit_swdiv_low_precision): Remove.
12403 (rs6000_emit_swdiv): Rewrite to handle between one and four
12404 iterations of Newton-Raphson generally; modify required number of
12405 iterations for some cases.
12406 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
12408 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
12410 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
12411 set-but-unused variable.
12413 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
12414 basic blocks of released function bodies garbage-collectable.
12416 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
12417 (struct rtl_opt_pass): Add TODO_df_finish.
12419 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
12421 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
12423 * config/arm/constraints.md (q): New constraint.
12424 * config/arm/ldrdstrd.md: New file.
12425 * config/arm/arm.md (ldrdstrd.md) New include.
12426 (arm_movdi): Use "q" instead of "r" constraint
12427 for double-word memory access.
12428 (movdf_soft_insn): Likewise.
12429 * config/arm/vfp.md (movdi_vfp): Likewise.
12430 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
12431 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
12432 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
12433 (mem_ok_for_ldrd_strd): Likewise.
12434 (output_move_double): Update assertion.
12436 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
12438 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
12440 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
12442 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
12443 define_insn_and_split.
12444 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
12446 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
12448 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
12449 define_insn_and_split.
12450 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
12451 (shiftsi3_compare): New pattern.
12452 (rrx): New pattern.
12453 * config/arm/unspecs.md (UNSPEC_RRX): New.
12455 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
12457 * config/arm/arm.md (negdi_extendsidi): New pattern.
12458 (negdi_zero_extendsidi): Likewise.
12460 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
12462 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
12463 define_insn_and_split.
12464 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
12465 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
12467 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
12469 * config/arm/arm.md (arm_subdi3): Convert define_insn into
12470 define_insn_and_split.
12471 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
12472 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
12474 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
12476 * config/arm/arm.md (subsi3_carryin): New pattern.
12477 (subsi3_carryin_const): Likewise.
12478 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
12479 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
12481 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
12483 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
12485 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
12487 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
12488 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
12490 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12492 * config/arm/arm.c (arm_expand_builtin): Change fcode
12493 type to unsigned int.
12495 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
12497 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
12499 2013-04-04 Ian Lance Taylor <iant@google.com>
12501 * doc/standards.texi (Standards): The Go frontend supports the Go 1
12504 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
12506 PR middle-end/56729
12507 * df-scan.c (df_insn_delete): Disable failing assert.
12509 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12511 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
12512 New function prototype.
12513 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
12514 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
12515 (arm_builtin_vectorized_function): New function.
12517 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12519 * config/arm/arm_neon_builtins.def: New file.
12520 * config/arm/arm.c (neon_builtin_data): Move contents to
12521 arm_neon_builtins.def.
12522 (enum arm_builtins): Include neon builtin definitions.
12523 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
12524 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
12526 2013-04-04 Marek Polacek <polacek@redhat.com>
12528 PR tree-optimization/48186
12529 * predict.c (maybe_hot_frequency_p): Return false if
12530 HOT_BB_FREQUENCY_FRACTION is 0.
12531 (cgraph_maybe_hot_edge_p): Likewise.
12533 2013-04-04 Richard Biener <rguenther@suse.de>
12535 PR tree-optimization/56826
12536 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
12539 2013-04-04 Richard Biener <rguenther@suse.de>
12541 PR tree-optimization/56213
12542 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
12543 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
12545 2013-04-04 Richard Biener <rguenther@suse.de>
12547 PR tree-optimization/56837
12548 * tree-loop-distribution.c (classify_partition): For non-zero
12549 values require that the value has the same precision as its
12550 mode to be useful as memset value.
12552 2013-04-03 Nick Clifton <nickc@redhat.com>
12554 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
12555 (fmssf4): Use fmsf.s on E3V5 architectures.
12556 (fnmasf4): Use fnmaf.s on E3V5 architectures.
12557 (fnmssf4): Use fnmsf.s on E3V5 architectures.
12559 2013-04-03 Jeff Law <law@redhat.com>
12561 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
12562 (lra-eliminations.o): Likewise.
12564 2013-04-03 Teresa Johnson <tejohnson@google.com>
12566 * gcov-io.c (compute_working_sets): Moved most of body of old
12567 compute_working_sets here from profile.c.
12568 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
12569 (gcov_working_set_t): Moved typedef here from basic-block.h
12570 (compute_working_set): Declare.
12571 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
12572 (get_working_sets): Renamed from compute_working_set,
12573 replace most of body with call to new compute_working_sets.
12574 (get_exec_counts): Replace call to compute_working_sets
12575 to get_working_sets.
12576 * profile.h (get_working_sets): Renamed from compute_working_set.
12577 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
12578 to get_working_sets.
12579 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
12580 * gcov-dump.c (dump_working_sets): New function.
12582 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
12584 * hwint.c (sext_hwi, zext_hwi): New functions.
12585 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
12586 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
12587 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
12588 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
12589 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
12590 (sext_hwi, zext_hwi): New functions.
12592 2013-04-03 Jeff Law <law@redhat.com>
12594 PR tree-optimization/56799
12595 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
12596 back test for widening conversion erroneously dropped in prior change.
12598 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12601 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
12602 instead of next_real_insn.
12604 2013-04-03 Marek Polacek <polacek@redhat.com>
12607 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
12609 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12612 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
12614 (thumb1_output_casesi): Likewise.
12615 (thumb2_output_casesi): Likewise.
12617 2013-04-03 Richard Biener <rguenther@suse.de>
12619 PR tree-optimization/56817
12620 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
12622 (tree_unroll_loops_completely_1): ... new function to manually
12623 walk the loop tree, properly defering outer loops of unrolled
12624 loops to later iterations.
12626 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
12628 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
12629 (vectorizable_load): Likewise.
12630 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
12631 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
12633 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
12635 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
12638 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
12640 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
12642 2013-04-03 Bin Cheng <bin.cheng@arm.com>
12644 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
12646 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
12648 PR tree-optimization/56790
12649 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
12652 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
12654 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
12656 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
12657 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
12660 2013-04-03 Jakub Jelinek <jakub@redhat.com>
12663 * tree.h (force_folding_builtin_constant_p): New decl.
12664 * builtins.c (force_folding_builtin_constant_p): New variable.
12665 (fold_builtin_constant_p): Fold immediately also if
12666 force_folding_builtin_constant_p.
12668 2013-04-03 Richard Biener <rguenther@suse.de>
12670 PR tree-optimization/56812
12671 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
12672 DRs of the same interleaving chain are independent.
12674 2013-04-02 Jason Merrill <jason@redhat.com>
12676 * gdbinit.in (pbb): Use debug fn.
12678 2013-04-02 Lawrence Crowl <crowl@google.com>
12680 * sese.h (struct ivtype_map_elt_s): Remove unused.
12681 (extern debug_ivtype_map): Remove unused.
12682 (extern eq_ivtype_map_elts): Remove unused.
12683 * sese.c (debug_ivtype_map): Removed unused.
12684 (debug_ivtype_map_1): Removed unused.
12685 (debug_ivtype_elt): Remove unused.
12686 (eq_ivtype_map_elts): Remove unused.
12689 2013-04-02 Kai Tietz <ktietz@redhat.com>
12692 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
12693 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
12694 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
12696 (legitimize_pe_coff_symbol): Likewise.
12697 (is_imported_p): New helper-function.
12698 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
12699 for Windows x64 targets.
12700 (ix86_expand_prologue): Optimize for pe-coff targets.
12701 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
12702 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
12703 medium/large code-model.
12704 (legitimize_pic_address): Likewise.
12705 (legitimize_tls_address): Likewise.
12706 (ix86_expand_call): Likewise.
12707 (x86_output_mi_thunk): Likewise.
12708 (get_dllimport_decl): Add new beimport argument.
12709 (construct_plt_address): Don't assert for x64 pe-coff targets.
12710 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
12712 (SYMBOL_FLAG_STUBVAR): New macro.
12713 (SYMBOL_REF_STUBVAR_P): Likewise.
12714 * config/i386/winnt.c (stub_list): New structure.
12715 (stub_head): New local variable.
12716 (i386_pe_record_stub): New function.
12717 (i386_pe_file_end): Emit refptr-stubs.
12719 2013-04-02 Jakub Jelinek <jakub@redhat.com>
12721 PR rtl-optimization/56745
12722 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
12723 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
12726 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
12727 and both of them are MEM_REFs, just compare first argument for
12728 equality and attempt to deal even with differing offsets.
12731 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
12732 of gimple_clobber_p to be MEM_REF.
12733 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
12734 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
12735 after gimplification.
12736 * asan.c (get_mem_ref_of_assignment): Don't instrument
12737 gimple_clobber_p stmts.
12738 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
12739 gimple_clobber_p stmt if they have MEM_REF lhs and
12740 are dead because of another gimple_clobber_p stmt.
12741 * tree-ssa-live.c (clear_unused_block_pointer): Treat
12742 gimple_clobber_p stmts like debug stmts.
12743 (remove_unused_locals): Remove clobbers with MEM_REF lhs
12744 that refer to unused VAR_DECLs or uninitialized values.
12745 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
12746 gimple_clobber_p stmts if they refer to removed parameters.
12747 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
12750 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
12752 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
12753 using SWI48 mode attribute.
12755 2013-04-02 Wei Mi <wmi@google.com>
12757 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
12758 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
12759 *<rotate_insn><mode>3_mask in i386.md.
12761 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
12763 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
12765 2013-04-02 Richard Biener <rguenther@suse.de>
12767 PR tree-optimization/56778
12768 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
12769 Runtime alias tests are not supported for gather loads.
12770 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
12771 stmts referenced from SSA operands before updating SSA form.
12773 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
12774 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
12776 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
12777 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
12778 * config/arm/cortex-a53.md: New file.
12779 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
12780 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
12781 * config/arm/arm.c (arm_issue_rate): Likewise.
12782 * config/arm/arm-tune.md: Regenerate
12783 * config/arm/arm-tables.opt: Regenerate.
12784 * config/arm/arm-cores.def: Add cortex-a53.
12786 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
12788 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
12791 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
12793 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
12794 scalar load/store operations using B/H registers.
12795 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
12797 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
12799 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
12801 * config/aarch64/aarch64.c
12802 (aarch64_simd_scalar_immediate_valid_for_move): New.
12803 * config/aarch64/aarch64-protos.h
12804 (aarch64_simd_scalar_immediate_valid_for_move): New.
12805 * config/aarch64/constraints.md (Dh, Dq): New.
12806 * config/aarch64/iterators.md (hq): New.
12808 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
12810 * reorg.c (get_branch_condition): Deal with conditional returns.
12811 (fill_simple_delay_slots): Remove dead code dealing with jumps.
12813 2013-04-01 Wei Mi <wmi@google.com>
12815 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
12816 Truncate operand 2 using %b asm operand modifier.
12817 (*<shift_insn><mode>3_mask): Ditto.
12818 (*<rotate_insn><mode>3_mask): Ditto.
12820 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
12822 PR middle-end/56798
12823 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
12825 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
12827 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
12829 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
12831 2013-03-30 Lawrence Crowl <crowl@google.com>
12833 * dse.c (clear_alias_sets): Remove never set.
12834 (disqualified_clear_alias_sets): Remove never set.
12835 (clear_alias_mode_pool): Remove never set.
12836 (dse_step0): Remove condition that is never true.
12837 (canon_address): Remove condition that is never true.
12838 (dse_step7): Remove condition that is never true.
12839 (rest_of_handle_dse): Remove condition that is never true.
12840 (rest_of_handle_dse::did_global): Remove never read from above.
12841 (dse_step2_spill): Remove never called from above.
12842 (dse_step5_spill): Remove never called from above.
12844 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
12846 * doc/md.texi (Standard Names) <casesi>: Update documentation for
12847 JUMP_TABLE_DATA changes.
12848 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
12849 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
12850 (Insns) <jump_table_data>: New entry.
12851 * doc/tm.texi: Regenerate.
12853 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
12855 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
12856 for table jump at the end of a basic block using tablejump_p.
12857 * targhooks.c (default_invalid_within_doloop): Likewise.
12858 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
12859 target hook implementation that is identical to the default hook.
12860 (rs6000_invalid_within_doloop): Remove.
12862 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
12863 unused variable from tablejump_p call.
12865 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
12866 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
12867 (INSN_DELETED_P): Likewise.
12868 (emit_jump_table_data): New prototype.
12869 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
12870 after 4th as unused.
12871 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
12872 * sched-vis.c (print_insn): Likewise.
12873 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
12874 insn for compatibility with back ends that use next_active_insn to
12875 identify jump table data.
12876 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
12877 (remove_insn): Likewise.
12878 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
12880 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
12881 (emit_jump_table_data): New function.
12883 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
12884 basic block, a JUMP_TABLE_DATA never is.
12885 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
12886 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
12887 off from code handling real insns.
12888 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
12889 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
12890 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
12891 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
12892 is not a NONDEBUG_INSN_P.
12893 * ira-costs.c (scan_one_insn): Likewise.
12894 * jump.c (mark_all_labels): Likewise.
12895 (mark_jump_label_1): Likewise.
12896 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
12897 * lra.c (get_insn_freq): Expect all insns reaching here to be in
12899 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
12900 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
12901 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
12902 JUMP_TABLE_DATA_P insns.
12903 (calculate_elim_costs_all_insns): Likewise.
12904 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
12905 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
12906 (delete_output_reload): Code style fixups.
12907 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
12908 insn flags on this non-insn.
12909 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
12910 as scheduling barriers, for pre-change compatibility.
12911 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
12912 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
12914 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
12915 redundant JUMP_TABLE_DATA_P test.
12916 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
12917 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
12918 (frv_for_each_packet): Likewise.
12919 * config/i386/i386.c (min_insn_size): Likewise.
12920 (ix86_avoid_jump_mispredicts): Likewise.
12921 * config/m32r/m32r.c (m32r_is_insn): Likewise.
12922 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
12923 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
12924 (mips16_insn_length): Robustify.
12925 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
12926 (mips16_split_long_branches): Likewise.
12927 * config/pa/pa.c (pa_combine_instructions): Likewise.
12928 * config/rs6000/rs6000.c (get_next_active_insn): Treat
12929 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
12930 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
12931 as contributing to pool range lengths.
12932 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
12933 Remove redundant JUMP_TABLE_DATA_P test.
12934 (sh_loop_align): Likewise.
12935 (split_branches): Likewise.
12936 (sh_insn_length_adjustment): Likewise.
12937 * config/spu/spu.c (get_branch_target): Likewise.
12939 2013-03-29 Jan Hubicka <jh@suse.cz>
12941 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
12942 gcov streaming; stream hot bb threshold to ltrans.
12943 * predict.c (get_hot_bb_threshold): Break out from ....
12944 (maybe_hot_count_p): ... here.
12945 (set_hot_bb_threshold): New function.
12946 * lto-section-in.c (lto_section_name): Add profile.
12947 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
12948 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
12949 and data-streamer.h
12950 (histogram_entry): New structure.
12951 (histogram, histogram_pool): New global vars.
12952 (histogram_hash): New structure.
12953 (histogram_hash::hash): New method.
12954 (histogram_hash::equal): Likewise.
12955 (account_time_size): New function.
12956 (cmp_counts): New function.
12957 (dump_histogram): New function.
12958 (ipa_profile_generate_summary): New function.
12959 (ipa_profile_write_summary): New function.
12960 (ipa_profile_read_summary): New function.
12961 (ipa_profile): Decide on threshold.
12962 (pass_ipa_profile): Add ipa_profile_write_summary and
12963 ipa_profile_read_summary.
12964 * Makefile.in (ipa.o): Update dependencies.
12965 * lto-streamer.h (LTO_section_ipa_profile): New section.
12967 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
12969 * tree.h (VAR_P): New.
12971 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
12974 * doc/invoke.texi ([-fwhole-program]): Fix typo.
12976 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
12978 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
12979 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
12980 (control_flow_insn_p): Likewise.
12981 * cfgrtl.c (duplicate_insn_chain): Likewise.
12982 * final.c (get_attr_length_1): Likewise.
12983 (shorten_branches): Likewise.
12984 (final_scan_insn): Likewise.
12985 * function.c (instantiate_virtual_regs): Likewise.
12986 * gcse.c (insert_insn_end_basic_block): Likewise.
12987 * ira-costs.c (scan_one_insn): Likewise.
12988 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
12989 * lra.c (check_rtl): Likewise.
12990 * reload1.c (elimination_costs_in_insn): Likewise.
12991 * reorg.c (follow_jumps): Likewise.
12993 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
12994 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
12995 (thumb_far_jump_used_p): Likewise.
12996 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
12997 (workaround_speculation): Likewise.
12998 (add_sched_insns_for_speculation): Likewise.
12999 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
13000 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
13001 (frv_for_each_packet): Likewise.
13002 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
13003 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
13004 (final_emit_insn_group_barriers): Likewise.
13005 * config/m32r/m32r.c (m32r_is_insn): Likewise.
13006 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
13007 (mips16_insn_length): Likewise.
13008 * config/pa/pa.c (pa_reorg): Likewise.
13009 (pa_combine_instructions): Likewise.
13010 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
13011 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
13012 (sh_reorg): Likewise.
13013 (split_branches): Likewise.
13014 * config/spu/spu.c (get_branch_target): Likewise.
13016 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
13019 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
13021 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
13022 Fix declaration name.
13024 2013-03-28 Lawrence Crowl <crowl@google.com>
13026 * graphds.h (struct graph.indicies): Remove unused.
13027 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
13028 (SCOP_ORIGINAL_PDDRS): Remove unused.
13029 * sese.h (extern insert_loop_close_phis): Removed unused.
13030 (extern insert_guard_phis): Removed unused.
13031 (extern ivtype_map_elt_info): Removed unused.
13032 (new_ivtype_map_elt): Removed unused.
13033 * sese.c (ivtype_map_elt_info): Removed unused.
13035 2013-03-28 Lawrence Crowl <crowl@google.com>
13037 * Makefile.in: Add several missing include dependences.
13039 (test-dump.o): New. This object is not added to any executable,
13040 but is present for ad-hoc testing.
13042 (debug (const bitmap_head_def &)): New.
13043 (debug (const bitmap_head_def *)): New.
13045 (extern debug (const bitmap_head_def &)): New.
13046 (extern debug (const bitmap_head_def *)): New.
13048 (debug (edge_def &)): New.
13049 (debug (edge_def *)): New.
13051 (debug (basic_block_def &)): New.
13052 (debug (basic_block_def *)): New.
13054 (dump_node (const_tree, int, FILE *)): Correct source file.
13056 (debug (die_struct &)): New.
13057 (debug (die_struct *)): New.
13059 (extern debug (die_struct &)): New.
13060 (extern debug (die_struct *)): New.
13061 * gimple-pretty-print.c
13062 (debug (gimple_statement_d &)): New.
13063 (debug (gimple_statement_d *)): New.
13064 * gimple-pretty-print.h
13065 (extern debug (gimple_statement_d &)): New.
13066 (extern debug (gimple_statement_d *)): New.
13068 (debug (ira_allocno_copy &)): New.
13069 (debug (ira_allocno_copy *)): New.
13070 (debug (ira_allocno &)): New.
13071 (debug (ira_allocno *)): New.
13073 (extern debug (ira_allocno_copy &)): New.
13074 (extern debug (ira_allocno_copy *)): New.
13075 (extern debug (ira_allocno &)): New.
13076 (extern debug (ira_allocno *)): New.
13078 (debug (live_range &)): New.
13079 (debug (live_range *)): New.
13081 (debug (lra_live_range &)): New.
13082 (debug (lra_live_range *)): New.
13084 (debug (lra_live_range &)): New.
13085 (debug (lra_live_range *)): New.
13087 (debug (omega_pb_d &)): New.
13088 (debug (omega_pb_d *)): New.
13090 (extern debug (omega_pb_d &)): New.
13091 (extern debug (omega_pb_d *)): New.
13093 (debug (const rtx_def &)): New.
13094 (debug (const rtx_def *)): New.
13096 (debug_tree (tree): Move within file.
13097 (debug_raw (const tree_node &)): New.
13098 (debug_raw (const tree_node *)): New.
13099 (dump_tree_via_hooks (const tree_node *, int)): New.
13100 (debug (const tree_node &)): New.
13101 (debug (const tree_node *)): New.
13102 (debug_verbose (const tree_node &)): New.
13103 (debug_verbose (const tree_node *)): New.
13104 (debug_head (const tree_node &)): New.
13105 (debug_head (const tree_node *)): New.
13106 (debug_body (const tree_node &)): New.
13107 (debug_body (const tree_node *)): New.
13108 (debug_vec_tree (tree): Move and reimplement in terms of dump.
13109 (debug (vec<tree, va_gc> &)): New.
13110 (debug (vec<tree, va_gc> *)): New.
13112 (extern debug (const rtx_def &)): New.
13113 (extern debug (const rtx_def *)): New.
13115 (debug_raw (simple_bitmap_def &)): New.
13116 (debug_raw (simple_bitmap_def *)): New.
13117 (debug (simple_bitmap_def &)): New.
13118 (debug (simple_bitmap_def *)): New.
13120 (extern debug (simple_bitmap_def &)): New.
13121 (extern debug (simple_bitmap_def *)): New.
13122 (extern debug_raw (simple_bitmap_def &)): New.
13123 (extern debug_raw (simple_bitmap_def *)): New.
13125 (debug (vinsn_def &)): New.
13126 (debug (vinsn_def *)): New.
13127 (debug_verbose (vinsn_def &)): New.
13128 (debug_verbose (vinsn_def *)): New.
13129 (debug (expr_def &)): New.
13130 (debug (expr_def *)): New.
13131 (debug_verbose (expr_def &)): New.
13132 (debug_verbose (expr_def *)): New.
13133 (debug (vec<rtx> &)): New.
13134 (debug (vec<rtx> *)): New.
13136 (extern debug (vinsn_def &)): New.
13137 (extern debug (vinsn_def *)): New.
13138 (extern debug_verbose (vinsn_def &)): New.
13139 (extern debug_verbose (vinsn_def *)): New.
13140 (extern debug (expr_def &)): New.
13141 (extern debug (expr_def *)): New.
13142 (extern debug_verbose (expr_def &)): New.
13143 (extern debug_verbose (expr_def *)): New.
13144 (extern debug (vec<rtx> &)): New.
13145 (extern debug (vec<rtx> *)): New.
13147 (_list_iter_cond_expr): Make inline instead of static.
13149 (debug (sreal &)): New.
13150 (debug (sreal *)): New.
13152 (extern debug (sreal &)): New.
13153 (extern debug (sreal *)): New.
13155 (extern debug_raw (const tree_node &)): New.
13156 (extern debug_raw (const tree_node *)): New.
13157 (extern debug (const tree_node &)): New.
13158 (extern debug (const tree_node *)): New.
13159 (extern debug_verbose (const tree_node &)): New.
13160 (extern debug_verbose (const tree_node *)): New.
13161 (extern debug_head (const tree_node &)): New.
13162 (extern debug_head (const tree_node *)): New.
13163 (extern debug_body (const tree_node &)): New.
13164 (extern debug_body (const tree_node *)): New.
13165 (extern debug (vec<tree, va_gc> &)): New.
13166 (extern debug (vec<tree, va_gc> *)): New.
13168 (debug (struct loop &)): New.
13169 (debug (struct loop *)): New.
13170 (debug_verbose (struct loop &)): New.
13171 (debug_verbose (struct loop *)): New.
13172 * tree-dump.c: Add header dependence.
13174 (extern debug (struct loop &)): New.
13175 (extern debug (struct loop *)): New.
13176 (extern debug_verbose (struct loop &)): New.
13177 (extern debug_verbose (struct loop *)): New.
13179 (debug (data_reference &)): New.
13180 (debug (data_reference *)): New.
13181 (debug (vec<data_reference_p> &)): New.
13182 (debug (vec<data_reference_p> *)): New.
13183 (debug (vec<ddr_p> &)): New.
13184 (debug (vec<ddr_p> *)): New.
13186 (extern debug (data_reference &)): New.
13187 (extern debug (data_reference *)): New.
13188 (extern debug (vec<data_reference_p> &)): New.
13189 (extern debug (vec<data_reference_p> *)): New.
13190 (extern debug (vec<ddr_p> &)): New.
13191 (extern debug (vec<ddr_p> *)): New.
13193 (debug (pt_solution &)): New.
13194 (debug (pt_solution *)): New.
13196 (extern debug (pt_solution &)): New.
13197 (extern debug (pt_solution *)): New.
13199 (debug (_var_map &)): New.
13200 (debug (_var_map *)): New.
13201 (debug (tree_live_info_d &)): New.
13202 (debug (tree_live_info_d *)): New.
13204 (extern debug (_var_map &)): New.
13205 (extern debug (_var_map *)): New.
13206 (extern debug (tree_live_info_d &)): New.
13207 (extern debug (tree_live_info_d *)): New.
13209 2013-03-28 Jan Hubicka <jh@suse.cz>
13211 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
13213 2013-03-28 Ian Bolton <ian.bolton@arm.com>
13215 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
13216 record only when desired or required.
13218 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
13220 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
13221 *vec_extractv2di_1_rex64. Use x64 isa attribute.
13223 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
13225 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
13226 (*andsi3_compare0_uxtw): New pattern.
13227 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
13228 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
13230 2013-03-28 Jan Hubicka <jh@suse.cz>
13232 * data-streamer-in.c (streamer_read_gcov_count): New function.
13233 * gimple-streamer-out.c: Include value-prof.h.
13234 (output_gimple_stmt): Output histogram.
13235 (output_bb): Use streamer_write_gcov_count.
13236 * value-prof.c: Include data-streamer.h
13237 (dump_histogram_value): Add HIST_TYPE_MAX.
13238 (stream_out_histogram_value): New function.
13239 (stream_in_histogram_value): New function.
13240 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
13241 (stream_out_histogram_value, stream_in_histogram_value): Declare.
13242 * data-streamer-out.c (streamer_write_gcov_count): New function.
13243 (streamer_write_gcov_count_stream): New function.
13244 * lto-cgraph.c (lto_output_edge): Update counter streaming.
13245 (lto_output_node): Likewise.
13246 (input_node, input_edge): Likewise.
13247 * lto-streamer-out.c (output_cfg): Update streaming.
13248 * lto-streamer-in.c (input_cfg): Likewise.
13249 * data-streamer.h (streamer_write_gcov_count,
13250 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
13251 * gimple-streamer-in.c: Include value-prof.h
13252 (input_gimple_stmt): Input histograms.
13253 (input_bb): Update profile streaming.
13255 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
13257 * genmodes.c (emit_max_int): New function.
13258 (emit_insn_modes_h): Added call to emit_max_function.
13259 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
13261 * machmode.def: Fixed comment.
13263 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
13265 * combine.c (try_combine): Removed useless assert.
13266 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
13268 2013-03-28 Marek Polacek <polacek@redhat.com>
13269 Richard Biener <rguenther@suse.de>
13271 PR tree-optimization/56695
13272 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
13273 build signed result of a vector comparison.
13274 * tree-cfg.c (verify_gimple_comparison): Check that a result
13275 of a vector comparison has signed type.
13277 2013-03-28 Richard Biener <rguenther@suse.de>
13279 PR tree-optimization/37021
13280 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
13281 do not restrict gaps between groups.
13282 * tree-vect-stmts.c (vectorizable_load): Properly account for
13283 a gap between groups.
13285 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
13287 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
13288 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
13291 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
13293 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
13294 * doc/extend.texi (Named Address Spaces): Ditto.
13295 (Variable Attributes): Ditto.
13297 2013-03-27 Kai Tietz <ktietz@redhat.com>
13299 * config.build: Add support for cygwin x64 target.
13300 * config.gcc: Likewise.
13301 * config.host: Likewise.
13302 * configure.ac: Likewise
13303 * configure: Regenerated.
13305 2013-03-27 Kai Tietz <ktietz@redhat.com>
13307 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
13308 * config/i386/t-cygwin-w64: New file.
13309 * config/i386/cygwin-w64.h: New file.
13310 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
13311 and add support for x64-cygwin target.
13312 (CPP_SPEC): Likewise.
13313 (CXX_WRAP_SPEC_LIST): Undefine before define.
13314 (LIBGCJ_SONAME): Use 15 as version.
13316 2013-03-27 Richard Biener <rguenther@suse.de>
13318 PR tree-optimization/56716
13319 * tree-ssa-structalias.c (perform_var_substitution): Adjust
13320 dumping for ref nodes.
13322 2013-03-27 Martin Jambor <mjambor@suse.cz>
13324 PR tree-optimization/55334
13325 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
13326 restricted pointers to arrays.
13328 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
13330 * Makefile.in (.SUFFIXES): Add .cc.
13331 (.c.o): Apply same recipe for implicit rule .cc.o.
13333 2013-03-27 Richard Biener <rguenther@suse.de>
13335 PR tree-optimization/37021
13336 * tree-vect-data-refs.c (vect_check_strided_load): Allow
13337 REALPART/IMAGPART_EXPRs around the supported refs.
13338 * tree-ssa-structalias.c (find_func_aliases): Assume that
13339 floating-point values are not used to transfer pointers.
13341 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
13343 * gcc/target.def (TARGET_HAS_IFUNC_P): New target hook.
13344 * gcc/doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
13345 * gcc/doc/tm.texi: Regenerate.
13346 * gcc/targhooks.h (default_has_ifunc_p): New.
13347 * gcc/targhooks.c (default_has_ifunc_p): Ditto.
13348 * gcc/config/linux-protos.h: New file.
13349 * gcc/config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of
13350 this hook for linux which disables support of indirect functions in
13352 * gcc/config/linux-android.c: New file.
13353 * gcc/config/t-linux-android.c: Ditto.
13354 * gcc/config.gcc: Added new object file linux-android.o.
13355 * gcc/config/i386/i386.c (ix86_get_function_versions_dispatcher):
13356 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
13357 * gcc/varasm.c (do_assemble_alias): Likewise.
13358 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
13359 doesn't support indirect functions.
13360 * configure: Regenerate.
13362 2013-03-27 Bin Cheng <bin.cheng@arm.com>
13365 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
13366 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
13369 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13371 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
13373 2013-03-27 Terry Guo <terry.guo@arm.com>
13375 * config/arm/arm-cores.def: Added core cortex-r7.
13376 * config/arm/arm-tune.md: Regenerated.
13377 * config/arm/arm-tables.opt: Regenerated.
13378 * doc/invoke.texi: Added entry for core cortex-r7.
13380 2013-03-27 Walter Lee <walt@tilera.com>
13382 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
13383 double-decrement of next_scratch_regno.
13385 2013-03-27 Walter Lee <walt@tilera.com>
13387 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
13389 (insn_v1mulus): Ditto.
13390 (insn_v2muls): Ditto.
13392 2013-03-27 Walter Lee <walt@tilera.com>
13394 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
13395 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
13397 2013-03-27 Walter Lee <walt@tilera.com>
13399 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
13400 (*sibcall_value): Ditto.
13402 2013-03-27 Walter Lee <walt@tilera.com>
13404 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
13405 (insn_mnz_v8qi): ... this ...
13406 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
13408 (insn_v<n>mnz): Replaced by ...
13409 (insn_v1mnz): ... this ...
13410 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
13412 (insn_mz_<mode>): Replaced by ...
13413 (insn_mz_v8qi): ... this ...
13414 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
13416 (insn_v<n>mz): Replaced by ...
13417 (insn_v1mz): ... this ...
13418 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
13421 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
13423 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
13425 2013-03-26 Roland McGrath <mcgrathr@google.com>
13427 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
13428 than fprintf with a non-constant, non-format string.
13430 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
13432 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
13433 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
13434 operand 0 predicate.
13435 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
13436 attribute. Use general_x64nomem_operand as operand 1 predicate.
13437 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
13438 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
13439 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
13440 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
13441 (mov<mode>_insv_1): Remove expander. Merge insn with
13442 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
13443 Use general_x64nomem_operand as operand 1 predicate.
13444 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
13445 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
13446 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
13447 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
13448 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
13449 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
13450 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
13451 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
13452 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
13453 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
13454 (general_x64nomem_operand): Ditto.
13456 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
13458 * config/rtems.opt: Add -pthread option.
13460 2013-03-26 Richard Biener <rguenther@suse.de>
13462 * alias.c (find_base_term): Avoid redundant and not used recursion.
13463 (base_alias_check): Get the initial base term from the caller.
13464 (true_dependence_1): Compute and pass base terms to base_alias_check.
13465 (write_dependence_p): Likewise.
13466 (may_alias_p): Likewise.
13468 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
13470 * config/aarch64/aarch64.c (aarch64_classify_address): Support
13471 PC-relative load in SI modes and above only.
13473 2013-03-26 Xinyu Qi <xyqi@marvell.com>
13475 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
13476 * config/arm/iwmmxt.md (WCGR0): Update.
13477 (WCGR1, WCGR2, WCGR3): Likewise.
13479 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
13481 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
13482 Use x64 and nox64 isa attributes.
13484 2013-03-26 Richard Biener <rguenther@suse.de>
13486 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
13487 alignment computations and rely on get_object_alignment_1
13488 for the !TYPE_P case.
13489 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
13491 2013-03-26 Walter Lee <walt@tilera.com>
13493 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
13494 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
13496 2013-03-25 Jeff Law <law@redhat.com>
13498 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
13499 check for INTEGRAL_TYPE_P that was missing due to checking in
13500 wrong version of prior patch.
13502 2013-03-25 Walter Lee <walt@tilera.com>
13504 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
13505 TILEGX_INSN_SHUFFLEBYTES1.
13506 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
13508 (tilegx_builtins): Ditto.
13509 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
13511 2013-03-25 Walter Lee <walt@tilera.com>
13513 * config/tilegx/tilegx.md (floatsisf2): New pattern.
13514 (floatunssisf2): New pattern.
13515 (floatsidf2): New pattern.
13516 (floatunssidf2): New pattern.
13518 2013-03-25 Walter Lee <walt@tilera.com>
13520 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
13521 tests for constraint J, K, N, P.
13523 2013-03-25 Walter Lee <walt@tilera.com>
13525 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
13526 Use indirect/pcrel encoding.
13527 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
13530 2013-03-25 Steve Ellcey <sellcey@mips.com>
13532 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
13533 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
13534 * config/mips/mips.c (mips_option_override): Set IMADD default.
13535 * config/mips/mips.h (PTF_AVOID_IMADD): New.
13536 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
13537 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
13538 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
13539 * doc/invoke.texi (-mimadd/-mno-imadd): New.
13541 2013-03-25 Jeff Law <law@redhat.com>
13543 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
13544 slightly to avoid creating and folding useless trees. Simplify
13545 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
13547 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
13549 * config/i386/i386.md (*zero_extendsidi2): Merge with
13550 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
13551 * config/i386/predicates.md (x86_64_zext_operand): Rename from
13552 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
13553 targets. Clarify comment.
13555 2013-03-25 Martin Jambor <mjambor@suse.cz>
13557 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
13558 pass-through jump functions differently.
13559 (ipa_read_jump_function): Likewise. Also use setter functions to set
13562 2013-03-25 Martin Jambor <mjambor@suse.cz>
13564 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
13565 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
13567 (ipa_get_indirect_edge_target): New function.
13568 (devirtualization_time_bonus): New parameter known_aggs, pass it to
13569 ipa_get_indirect_edge_target. Update all callers.
13570 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
13571 ipa_get_indirect_edge_target_1 instead of calling
13572 ipa_get_indirect_edge_target.
13573 (create_specialized_node): Pass aggvlas to
13574 ipcp_discover_new_direct_edges.
13576 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13578 * config/arm/arm.md (f_sels, f_seld): New types.
13579 (*cmov<mode>): New pattern.
13580 * config/arm/predicates.md (arm_vsel_comparison_operator): New
13583 2013-03-25 Kai Tietz <ktietz@redhat.com>
13585 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
13586 POSIX-printf for mingw-hosted builds.
13588 2013-03-25 Richard Biener <rguenther@suse.de>
13590 PR middle-end/56694
13591 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
13592 must-not-throw stmt location.
13594 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13596 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
13597 Emit load-acquire versions when acq is true.
13598 (arm_emit_store_exclusive): Add rel parameter.
13599 Emit store-release versions when rel is true.
13600 (arm_split_compare_and_swap): Use acquire-release instructions
13602 of barriers when appropriate.
13603 (arm_split_atomic_op): Likewise.
13604 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
13605 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
13606 (VUNSPEC_SLX): Likewise.
13607 (VUNSPEC_LDA): Likewise.
13608 (VUNSPEC_STL): Likewise.
13609 * config/arm/sync.md (atomic_load<mode>): New pattern.
13610 (atomic_store<mode>): Likewise.
13611 (arm_load_acquire_exclusive<mode>): Likewise.
13612 (arm_load_acquire_exclusivesi): Likewise.
13613 (arm_load_acquire_exclusivedi): Likewise.
13614 (arm_store_release_exclusive<mode>): Likewise.
13616 2013-03-25 Catherine Moore <clm@codesourcery.com>
13618 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
13619 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
13620 * config/mip/predicates.md (lwsp_swsp_operand,
13621 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
13622 sb16_operand, db4_operand, db7_operand, ib3_operand,
13623 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
13624 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
13625 andi16_operand): New predicates.
13626 * config/mips/mips.md (compression): New attribute.
13627 (enabled): New attribute.
13628 (length): Consider compression in computing length.
13629 (shift_compression): New code attribute.
13630 (*add<mode>3): New operands. Record compression.
13631 (sub<mode>3): Likewise.
13632 (one_cmpl<mode>2): Likewise.
13633 (*and<mode>3): Likewise.
13634 (*ior<mode>3): Likewise.
13635 (unnamed pattern for xor): Likewise.
13636 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
13637 (*<optab><mode>3): Likewise.
13638 (*mov<mode>_internal: Likewise.
13639 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
13640 (mips_unsigned_immediate_p): New.
13641 (umips_lwsp_swsp_address_p): New.
13642 (m16_based_address_p): New.
13643 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
13644 (mips_unsigned_immediate_p): New prototype.
13645 (lwsp_swsp_address_p): New prototype.
13646 (m16_based_address_p): New prototype.
13647 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
13648 (mips_signed_immediate_p): New function.
13649 (m16_based_address_p): New function.
13650 (lwsp_swsp_address_p): New function.
13651 (mips_print_operand_punctuation): Recognize short delay slot insns
13652 for microMIPS.add<mode>3"
13654 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13657 * config/arm/iterators.md (v_cmp_result): New mode attribute.
13658 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
13660 2013-03-25 Richard Biener <rguenther@suse.de>
13662 PR tree-optimization/56689
13663 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
13666 2013-03-25 Richard Biener <rguenther@suse.de>
13668 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
13670 (memory_references): Likewise.
13671 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
13672 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
13673 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
13674 (gather_mem_refs_in_loops): Fold into ...
13675 (analyze_memory_references): ... this. Move initialization
13676 to tree_ssa_lim_initialize.
13677 (fill_always_executed_in): Rename to ...
13678 (fill_always_executed_in_1): ... this.
13679 (fill_always_executed_in): Move contains_call computation to
13680 this new function from ...
13681 (tree_ssa_lim_initialize): ... here.
13682 (tree_ssa_lim): Call fill_always_executed_in.
13684 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
13686 * postreload.c (reload_combine): Fix code detecting returns.
13688 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
13690 * function.c (emit_use_return_register_into_block): On cc0 targets,
13691 do not emit the sequence between cc0 setter and user.
13693 2013-03-25 Kai Tietz <ktietz@redhat.com>
13695 * config/i386/predicates.md (local_symbolic_operand): Interpret
13696 dll-imported symbols as none-local.
13698 2013-03-25 Richard Biener <rguenther@suse.de>
13700 * tree-ssa-loop-im.c (struct depend): Remove.
13701 (struct lim_aux_data): Make depends a vec of gimples.
13702 (free_lim_aux_data): Adjust.
13703 (add_dependency): Likewise.
13704 (set_level): Likewise.
13706 2013-03-25 Richard Biener <rguenther@suse.de>
13708 PR middle-end/56434
13709 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
13710 the pointer returned by calls with ECF_MALLOC set.
13712 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
13714 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
13716 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
13718 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
13719 using MMXMODE mode iterator.
13720 (*move<mode>_internal): Merge with *movv2sf_internal and
13721 *movv2sf_internal_rex64 using MMXMODE mode iterator.
13723 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
13725 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
13726 (record_last_mem_set_info): Likewise.
13728 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
13729 of XNEWVEC followed by memset.
13730 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
13732 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
13734 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
13735 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
13736 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
13737 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
13738 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
13739 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
13740 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
13741 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
13742 BARRIER_P instead of GET_CODE.
13744 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
13746 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
13747 inaccuracy in the probing code.
13749 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
13750 (ctrapdi4): Likewise.
13752 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
13754 * calls.c (expand_call): Add missing guard to code handling return
13755 of non-BLKmode structures in MSB.
13756 * function.c (expand_function_end): Likewise.
13758 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
13760 * combine.c (try_combine): Adjust comment. Do not add the set of
13761 insn #0 if the destination indirectly is set or dies in insn #2.
13762 Tidy up code to distribute a new note.
13764 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
13766 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
13767 also for alternatives 16 and 17.
13769 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
13771 * config/i386/sse.md (*mov<mode>_internal): Merge with
13772 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
13773 Emit insn template depending on type attribute. Use
13774 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
13775 movd instead of movq mnemonic for interunit moves. Rewrite mode
13776 attribute calculation. Remove unit attribute calculation.
13777 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
13778 Set prefix_data16 attribute for DImode ssemov types.
13779 Use Ym instead of y for SSE-MMX conversion alternatives.
13780 Reorder operand constraints.
13782 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
13784 * df.h (df_insn_delete): Adjust prototype.
13785 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
13786 and let it decide whether mark the basic block dirty.
13787 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
13788 * df-scan.c (df_insn_info_delete): New helper function, split
13789 off from df_insn_delete.
13790 (df_scan_free_bb_info): Use it.
13791 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
13793 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
13794 that the insn is actually an insn and it has a non-NULL basic block.
13795 Do not mark basic block dirty if only deleting a DEBUG_INSN.
13797 2013-03-22 Richard Biener <rguenther@suse.de>
13799 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
13801 (mem_ref_alloc): Do not allocate them.
13802 (refs_independent_p): Do not query or maintain a cache.
13804 2013-03-22 Richard Biener <rguenther@suse.de>
13806 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
13807 (gather_mem_refs_in_loops): Do not compute it.
13808 (analyze_memory_references): Do not allocate it.
13809 (tree_ssa_lim_finalize): Do not free it.
13810 (for_all_locs_in_loop): Do not query all_refs_in_loop.
13812 2013-03-22 Richard Biener <rguenther@suse.de>
13814 * is-a.h (as_a): Use gcc_checking_assert.
13816 2013-03-22 Ian Bolton <ian.bolton@arm.com>
13818 * config/aarch64/aarch64.c (aarch64_print_operand): New
13819 format specifier for printing a constant in hex.
13820 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
13821 format specifier for printing second operand.
13823 2013-03-22 Richard Biener <rguenther@suse.de>
13825 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
13827 (gather_mem_refs_in_loops): Perform store accumulation here.
13828 (create_vop_ref_mapping_loop): Remove.
13829 (create_vop_ref_mapping): Likewise.
13830 (analyze_memory_references): Initialize refs_stored_in_loop.
13831 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
13832 (record_indep_loop): Remove.
13833 (record_dep_loop): New function.
13834 (ref_indep_loop_p_1): Adjust to only walk over references
13835 in the loop, not its subloops.
13836 (ref_indep_loop_p): Rename to ...
13837 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
13838 maintaining a more fine-grained cache.
13839 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
13840 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
13842 2013-03-22 Richard Biener <rguenther@suse.de>
13844 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
13845 (struct mem_ref): Make accesses_in_loop a vec of a vec of
13846 aggregate mem_ref_loc.
13847 (free_mem_ref_locs): Inline into ...
13848 (memref_free): ... this and adjust.
13849 (mem_ref_alloc): Adjust.
13850 (mem_ref_locs_alloc): Remove.
13851 (record_mem_ref_loc): Adjust.
13852 (get_all_locs_in_loop): Rewrite into ...
13853 (for_all_locs_in_loop): ... this iterator.
13854 (rewrite_mem_ref_loc): New functor.
13855 (rewrite_mem_refs): Use for_all_locs_in_loop.
13856 (sm_set_flag_if_changed): New functor.
13857 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
13858 (ref_always_accessed): New functor.
13859 (ref_always_accessed_p): Use for_all_locs_in_loop.
13861 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
13863 * tree-pass.h (PROP_gimple_lvec): New.
13864 * passes.c (dump_properties): Handle PROP_gimple_lvec.
13865 (init_optimization_passes): Move pass_lower_vector.
13866 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
13868 (pass_lower_vector): Provide PROP_gimple_lvec.
13869 (pass_lower_vector_ssa): Likewise.
13870 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
13872 2013-03-21 Mark Wielaard <mjw@redhat.com>
13874 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
13876 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
13878 * config/i386/i386.md (*movdi_internal): Disparage slightly
13879 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
13880 conversion alternatives.
13882 2013-03-21 Jakub Jelinek <jakub@redhat.com>
13884 PR middle-end/48087
13885 * diagnostic.def (DK_WERROR): New kind.
13886 * diagnostic.h (werrorcount): Define.
13887 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
13888 promoted to DK_ERROR, increment DK_WERROR counter instead of
13890 * toplev.c (toplev_main): Call print_ignored_options even if
13891 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
13892 even if just werrorcount is non-zero.
13895 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
13897 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
13898 (string_cst_pool_decl): New function.
13899 (optimize_one_addr_into_implicit_ptr): New function.
13900 (resolve_addr_in_expr): Optimize DWARF location expression
13901 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
13902 which doesn't live in memory, but has DW_AT_location or
13903 DW_AT_const_value, or refers to a string literal, into
13904 DW_OP_GNU_implicit_pointer.
13905 (optimize_location_into_implicit_ptr): New function.
13906 (resolve_addr): If removing DW_AT_location of a variable because
13907 it was DW_OP_addr of address of the variable, but the variable doesn't
13908 live in memory, try to emit const value attribute for the initializer.
13910 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
13912 * tree.h (VECTOR_TYPE_P): New macro.
13913 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
13914 TYPE_MODE): Use it.
13915 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
13916 VEC_COND_EXPR cannot be lvalues.
13917 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
13919 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
13921 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
13922 Restrict the transformation to equal modes.
13924 2013-03-21 Richard Biener <rguenther@suse.de>
13926 PR tree-optimization/39326
13927 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
13928 (MEM_ANALYZABLE): Adjust.
13929 (record_mem_ref_loc): Move bitmap ops ...
13930 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
13931 unanalyzable refs, do not record locations for it.
13932 (analyze_memory_references): Allocate ref zero as shared
13934 (refs_independent_p): Do not test for unanalyzed mems here.
13935 (ref_indep_loop_p_1): Special-case disambiguation against
13936 the unanalyzed ref.
13937 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
13939 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
13941 * config/arm/arm-protos.h (tune_params): Add
13942 prefer_neon_for_64bits field.
13943 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
13944 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
13945 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
13946 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
13947 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
13948 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
13949 (arm_option_override): Handle -mneon-for-64bits new option.
13950 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
13951 (prefer_neon_for_64bits): Declare new variable.
13952 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
13953 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
13954 (arch_enabled): Handle new arch types. Remove support for onlya8
13956 (one_cmpldi2): Use new arch names.
13957 (zero_extend<mode>di2, extend<mode>di2): Ditto.
13958 * config/arm/arm.opt (mneon-for-64bits): Add option.
13959 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
13960 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
13961 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
13963 * doc/invoke.texi (-mneon-for-64bits): Document.
13965 2013-03-21 Richard Biener <rguenther@suse.de>
13967 PR tree-optimization/39326
13968 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
13969 (sort_bbs_in_loop_postorder_cmp): New function.
13970 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
13972 2013-03-21 Richard Biener <rguenther@suse.de>
13974 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
13975 (vect_insert_into_interleaving_chain): Likewise.
13976 (vect_drs_dependent_in_basic_block): Inline ...
13977 (vect_slp_analyze_data_ref_dependence): ... here. New function,
13979 (vect_analyze_data_ref_dependence): ... here. Simplify.
13980 (vect_check_interleaving): Simplify.
13981 (vect_analyze_data_ref_dependences): Likewise. Split out ...
13982 (vect_slp_analyze_data_ref_dependences): ... this new function.
13983 (dr_group_sort_cmp): New function.
13984 (vect_analyze_data_ref_accesses): Compute data-reference groups
13985 here instead of in vect_analyze_data_ref_dependence. Use
13986 a more efficient algorithm.
13987 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
13988 vect_slp_analyze_data_ref_dependences. Call
13989 vect_analyze_data_ref_accesses earlier.
13990 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
13991 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
13992 (vect_slp_analyze_data_ref_dependences): New prototype.
13994 2013-03-21 Richard Biener <rguenther@suse.de>
13996 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
13997 ref is stored in the loop.
13998 (find_refs_for_sm): Walk only over all stores.
13999 (store_motion_loop): Allocate from lim_bitmap_obstack.
14000 (store_motion): Likewise.
14002 2013-03-21 Richard Biener <rguenther@suse.de>
14004 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
14005 Update virtual SSA form.
14007 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
14009 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
14010 * configure: Regenerate.
14011 * config.in: Regenerate.
14012 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
14013 if !HAVE_LD_EH_FRAME_CIEV3.
14015 2013-03-21 Richard Biener <rguenther@suse.de>
14017 * tree-cfg.c (verify_expr_no_block): New function.
14018 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
14019 nor DECL_VALUE_EXPR have locations with associated blocks.
14020 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
14021 (clear_unused_block_pointer): Remove code dealing with
14022 blocks in DECL_DEBUG_EXPR locations.
14024 2013-03-21 Richard Biener <rguenther@suse.de>
14026 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
14027 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
14028 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
14029 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
14030 instead of DECL_DEBUG_EXPR_IS_FROM.
14031 * gimplify.c (gimplify_modify_expr): Likewise.
14032 * tree-cfg.c (verify_expr_location_1): Likewise.
14033 * tree-complex.c (create_one_component_var): Likewise.
14034 * tree-sra.c (create_access_replacement): Likewise.
14035 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
14036 (clear_unused_block_pointer): Likewise.
14037 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
14038 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
14039 * var-tracking.c (var_debug_decl): Likewise.
14040 (track_expr_p): Likewise.
14041 * tree-inline.c (add_local_variables): Likewise. Set
14042 DECL_HAS_DEBUG_EXPR_P after copying it.
14043 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
14044 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
14046 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
14049 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
14050 * configure: Regenerate.
14051 * config.in: Regenerate.
14052 * config/i386/i386.md (*movdf_internal): Use
14053 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
14054 movd instead of movq mnemonic for interunit moves.
14055 (*movdi_internal): Ditto.
14057 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
14059 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
14060 (abd<mode>_3): New pattern.
14061 (aba<mode>_3): New pattern.
14062 (fabd<mode>_3): New pattern.
14064 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
14066 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
14067 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
14068 occurrence of REGISTER_PREFIX as its empty string.
14070 2013-03-20 Jeff Law <law@redhat.com>
14072 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
14073 addititional equivalences for equality comparisons between an SSA_NAME
14074 and a constant where the SSA_NAME was set from a widening conversion.
14076 2013-03-20 Walter Lee <walt@tilera.com>
14078 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
14080 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
14082 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
14083 depending on type attribute.
14084 (*movti_internal): Ditto.
14085 (*movtf_internal): Ditto.
14086 (*movxf_internal): Ditto.
14087 (*movdf_internal): Ditto.
14088 (*movsf_internal): Ditto.
14090 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
14092 * config/i386/i386.md (*movti_internal): Set prefix attribute to
14093 maybe_vex for sselog1 and ssemov types.
14094 (*movdi_internal): Reorder operand constraints.
14095 (*movsi_internal): Ditto. Set prefix attribute to
14096 maybe_vex for sselog1 and ssemov types.
14097 (*movtf_internal): Set prefix attribute to maybe_vex
14098 for sselog1 and ssemov types.
14099 (*movdf_internal): Ditto. Set prefix_data16 attribute for
14100 DImode ssemov types. Reorder operand constraints.
14101 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
14102 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
14103 attribute for SImode ssemov types. Reorder operand constraints.
14105 2013-03-20 Martin Jambor <mjambor@suse.cz>
14107 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
14108 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
14110 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
14112 * config/rs6000/predicates.md (indexed_address, update_address_mem
14113 update_indexed_address_mem): New predicates.
14114 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
14115 attribute for load/store instructions.
14116 * config/rs6000/dfp.md (movsd_store): Likewise.
14117 (movsd_load): Likewise.
14118 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
14119 (unnamed HI->DI extend define_insn): Likewise.
14120 (unnamed SI->DI extend define_insn): Likewise.
14121 (unnamed QI->SI extend define_insn): Likewise.
14122 (unnamed QI->HI extend define_insn): Likewise.
14123 (unnamed HI->SI extend define_insn): Likewise.
14124 (unnamed HI->SI extend define_insn): Likewise.
14125 (extendsfdf2_fpr): Likewise.
14126 (movsi_internal1): Likewise.
14127 (movsi_internal1_single): Likewise.
14128 (movhi_internal): Likewise.
14129 (movqi_internal): Likewise.
14130 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
14131 attribute for load/store instructions.
14132 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
14134 (mov<mode>_softfloat): Likewise.
14135 (mov<mode>_hardfloat32): Likewise.
14136 (mov<mode>_hardfloat64): Likewise.
14137 (mov<mode>_softfloat64): Likewise.
14138 (movdi_internal32): Likewise.
14139 (movdi_internal64): Likewise.
14140 (probe_stack_<mode>): Likewise.
14142 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
14144 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
14145 floating point, and decimal floating point to reload iterator.
14147 * config/rs6000/constraints.md (wl constraint): New constraints to
14148 return FLOAT_REGS if certain options are used to reduce the number
14149 of separate patterns that exist in the file.
14150 (wx constraint): Likewise.
14151 (wz constraint): Likewise.
14153 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
14154 -mdebug=reg, print wg, wl, wx, and wz constraints.
14155 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
14156 Initialize the reload functions for 64-bit binary/decimal floating
14158 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
14159 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
14160 create the buffer on the stack to overcome not having a 32-bit
14162 (rs6000_emit_move): Likewise.
14163 (rs6000_secondary_memory_needed_rtx): Likewise.
14164 (rs6000_alloc_sdmode_stack_slot): Likewise.
14165 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
14166 via xxlxor, just like DFmode 0.0.
14168 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
14169 define as 1 if we are running on a power7 or newer.
14170 (enum r6000_reg_class_enum): Add new constraints.
14172 * config/rs6000/dfp.md (movsd): Delete, combine with binary
14173 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
14174 with other moves by using conditional constraits (wg). Use LFIWZX
14175 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
14176 (movsd splitter): Likewise.
14177 (movsd_hardfloat): Likewise.
14178 (movsd_softfloat): Likewise.
14180 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
14181 binary and decimal floating point moves.
14182 (fmove_ok): New attributes to combine binary and decimal floating
14183 point moves, and to combine power6x (mfpgpr) moves along normal
14185 (real_value_to_target): Likewise.
14186 (f32_lr): Likewise.
14187 (f32_lm): Likewise.
14188 (f32_li): Likewise.
14189 (f32_sr): Likewise.
14190 (f32_sm): Likewise.
14191 (f32_si): Likewise.
14192 (movsf): Combine binary and decimal floating point moves. Combine
14193 power6x (mfpgpr) moves with other moves by using conditional
14194 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
14195 (mov<mode> for SFmode/SDmode); Likewise.
14196 (SFmode/SDmode splitters): Likewise.
14197 (movsf_hardfloat): Likewise.
14198 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
14199 (movsf_softfloat): Likewise.
14200 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
14202 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
14203 wx and wz constraints.
14205 * config/rs6000/constraints.md (wg constraint): New constraint to
14206 return FLOAT_REGS if -mmfpgpr (power6x) was used.
14208 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
14211 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
14212 -mdebug=reg, print wg, wl, wx, and wz constraints.
14213 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
14214 Initialize the reload functions for 64-bit binary/decimal floating
14216 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
14217 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
14218 create the buffer on the stack to overcome not having a 32-bit
14220 (rs6000_emit_move): Likewise.
14221 (rs6000_secondary_memory_needed_rtx): Likewise.
14222 (rs6000_alloc_sdmode_stack_slot): Likewise.
14223 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
14224 via xxlxor, just like DFmode 0.0.
14226 * config/rs6000/dfp.md (movdd): Delete, combine with binary
14227 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
14228 with other moves by using conditional constraits (wg). Use LFIWZX
14229 and STFIWX for loading SDmode on power7.
14230 (movdd splitters): Likewise.
14231 (movdd_hardfloat32): Likewise.
14232 (movdd_softfloat32): Likewise.
14233 (movdd_hardfloat64_mfpgpr): Likewise.
14234 (movdd_hardfloat64): Likewise.
14235 (movdd_softfloat64): Likewise.
14237 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
14238 64-bit binary and decimal floating point moves.
14239 (FMOVE64X): Likewise.
14240 (movdf): Combine 64-bit binary and decimal floating point moves.
14241 Combine power6x (mfpgpr) moves with other moves by using
14242 conditional constraits (wg).
14243 (mov<mode> for DFmode/DDmode): Likewise.
14244 (DFmode/DDmode splitters): Likewise.
14245 (movdf_hardfloat32): Likewise.
14246 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
14247 (movdf_softfloat32): Likewise.
14248 (movdf_hardfloat64_mfpgpr): Likewise.
14249 (movdf_hardfloat64): Likewise.
14250 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
14251 (movdf_softfloat64): Likewise.
14252 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
14253 (reload_<mode>_load): Move to later in the file so they aren't in
14254 the middle of the floating point move insns.
14255 (reload_<mode>_store): Likewise.
14257 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
14260 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
14261 constraint if -mdebug=reg.
14262 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
14263 Enable using dd reload support if needed.
14265 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
14266 binary and decimal floating point moves in rs6000.md.
14267 (movtd_internal): Likewise.
14269 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
14270 decimal floating point moves.
14272 (movtf_internal): Likewise.
14273 (mov<mode>_internal, TDmode/TFmode): Likewise.
14274 (movtf_softfloat): Likewise.
14275 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
14277 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
14278 movdi_internal64, using wg constraint for move direct operations.
14279 (movdi_internal64): Likewise.
14281 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
14282 MODES_TIEABLE_P for selected modes. Print the numerical value of
14283 the various virtual registers. Use GPR/FPR first/last values,
14284 instead of hard coding the register numbers. Print which modes
14285 have reload functions registered.
14286 (rs6000_option_override_internal): If -mdebug=reg, trace the options
14287 settings before/after setting cpu, target and subtarget settings.
14288 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
14289 and for secondary reload failures in rs6000_secondary_reload_inner.
14290 (rs6000_secondary_reload_fail): Likewise.
14291 (rs6000_secondary_reload_inner): Likewise.
14293 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
14294 macros for first/last GPR and FPR registers.
14295 (LAST_GPR_REGNO): Likewise.
14296 (FIRST_FPR_REGNO): Likewise.
14297 (LAST_FPR_REGNO): Likewise.
14299 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
14300 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
14301 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
14302 (vcond<mode><mode>): Likewise.
14303 (vcondu<mode><mode>): Likewise.
14304 (vector_gtu<mode>): Likewise.
14305 (vector_gte<mode>): Likewise.
14306 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
14307 to prevent the compiler from converting DImode operations to TImode.
14308 (ior<mode>3): Likewise.
14309 (and<mode>3): Likewise.
14310 (one_cmpl<mode>2): Likewise.
14311 (nor<mode>3): Likewise.
14312 (andc<mode>3): Likewise.
14314 * config/rs6000/constraints.md (wt constraint): New constraint
14315 that returns VSX_REGS if TImode is allowed in VSX registers.
14317 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
14318 constant under VSX.
14320 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
14321 similar to TImode, but it is restricted to being in the GPRs.
14323 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
14324 TImode to occupy a single VSX register.
14326 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
14327 -mvsx-timode for power7/power8.
14328 (power7 cpu): Likewise.
14329 (power8 cpu): Likewise.
14331 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
14332 sure that TFmode/TDmode take up two registers if they are ever
14333 allowed in the upper VSX registers.
14334 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
14336 (rs6000_init_hard_regno_mode_ok): Likewise.
14337 (rs6000_debug_reg_global): Add debugging for PTImode and wt
14338 constraint. Print if LRA is turned on.
14339 (rs6000_option_override_internal): Give an error if -mvsx-timode
14340 and VSX is not enabled.
14341 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
14342 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
14343 to reg+offset addressing. Use PTImode when checking offset
14344 addresses for validity.
14345 (reg_offset_addressing_ok_p): Likewise.
14346 (rs6000_legitimate_offset_address_p): Likewise.
14347 (rs6000_legitimize_address): Likewise.
14348 (rs6000_legitimize_reload_address): Likewise.
14349 (rs6000_legitimate_address_p): Likewise.
14350 (rs6000_eliminate_indexed_memrefs): Likewise.
14351 (rs6000_emit_move): Likewise.
14352 (rs6000_secondary_reload): Likewise.
14353 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
14354 reloads to fpr registers to continue to use reg+offset addressing,
14355 but 64-bit reloads to altivec registers need reg+reg addressing.
14356 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
14357 it. Treat LO_SUM like a PLUS operation.
14358 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
14359 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
14360 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
14361 registers to share a register with a smaller sized type, since VSX
14362 puts scalars in the upper 64-bits.
14363 (print_operand): Add support for PTImode.
14364 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
14365 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
14366 registers, but don't have arithmetic support.
14367 (rs6000_memory_move_cost): Add test for VSX.
14368 (rs6000_opt_masks): Add -mvsx-timode.
14370 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
14373 (VSr): Use wt constraint for TImode.
14374 (VSv): Drop TImode support.
14375 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
14376 (vsx_movti_64bit): Likewise.
14377 (vsx_movti_32bit): Likewise.
14378 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
14379 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
14380 one '?' on the appropriate output constraint. Do not allow TImode
14381 logical operations on 32-bit systems.
14382 (vsx_ior<mode>3): Likewise.
14383 (vsx_xor<mode>3): Likewise.
14384 (vsx_one_cmpl<mode>2): Likewise.
14385 (vsx_nor<mode>3): Likewise.
14386 (vsx_andc<mode>3): Likewise.
14387 (vsx_concat_<mode>): Likewise.
14388 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
14390 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
14391 OPTION_MASK_VSX_TIMODE.
14392 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
14393 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
14395 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
14396 (TI2 iterator): New iterator for TImode, PTImode.
14397 (wd mode attribute): Add values for vector types.
14398 (movti_string): Replace TI move operations with operations for TImode
14399 and PTImode. Add support for TImode being allowed in VSX registers.
14400 (mov<mode>_string, TImode/PTImode): Likewise.
14401 (movti_ppc64): Likewise.
14402 (mov<mode>_ppc64, TImode/PTImode): Likewise.
14403 (TI mode splitters): Likewise.
14405 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
14408 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
14410 PR tree-optimization/56355
14411 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
14412 Also handle integers with undefined overflow.
14414 2013-03-20 Catherine Moore <clm@codesourcery.com>
14415 Maciej W. Rozycki <macro@codesourcery.com>
14416 Tom de Vries <tom@codesourcery.com>
14417 Nathan Sidwell <nathan@codesourcery.com>
14418 Iain Sandoe <iain@codesourcery.com>
14419 Nathan Froyd <froydnj@codesourcery.com>
14420 Chao-ying Fu <fu@mips.com>
14422 * doc/extend.texi: (micromips, nomicromips, nocompression):
14423 Document new function attributes.
14424 * doc/invoke.texi (minterlink-compressed, mmicromips,
14425 m14k, m14ke, m14kec): Document new options.
14426 (minterlink-mips16): Update documentation.
14427 * doc/md.texi (ZC, ZD): Document new constraints.
14428 * configure.ac (gcc_cv_as_micromips): Check if linker
14429 supports the .set micromips directive.
14430 * configure: Regenerate.
14431 * config.in: Regenerate.
14432 * config/mips/mips-tables.opt: Regenerate.
14433 * config/mips/micromips.md: New file.
14434 * constraints.md (ZC, ZD): New constraints.
14435 * config/mips/predicates.md (movep_src_register): New predicate.
14436 (movep_src_operand): New predicate.
14437 (non_volatile_mem_operand): New predicate.
14438 * config/mips/mips.md (multimem): New type.
14439 (length): Differentiate between 17-bit and 18-bit branch offsets.
14440 (MOVEP1, MOVEP2): New mode iterator.
14441 (mov_<load>l): Use ZC constraint.
14442 (mov_<load>r): Likewise.
14443 (mov_<store>l): Likewise.
14444 (mov_<store>r): Likewise.
14445 (*branch_equality<mode>_inverted): Add microMIPS support.
14446 (*branch_equality<mode>): Likewise.
14447 (*jump_absolute): Likewise.
14448 (indirect_jump_<mode>): Likewise.
14449 (tablejump_<mode>): Likewise.
14450 (<optab>_internal): Likewise.
14451 (sibcall_internal): Likewise.
14452 (sibcall_value_internal): Likewise.
14453 (prefetch): Use constraint ZD.
14454 * config/mips/mips.opt (minterlink-compressed): New option.
14455 (minterlink-mips16): Now an alias for minterlink-compressed.
14456 (mmicromips): New option.
14457 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
14458 (compare_and_swap_12): Likewise.
14459 (sync_add<mode>): Likewise.
14460 (sync_<optab>_12): Likewise.
14461 (sync_old_<optab>_12): Likewise.
14462 (sync_new_<optab>_12): Likewise.
14463 (sync_nand_12): Likewise.
14464 (sync_old_nand_12): Likewise.
14465 (sync_new_nand_12): Likewise.
14466 (sync_sub<mode>): Likewise.
14467 (sync_old_add<mode>): Likewise.
14468 (sync_old_sub<mode>): Likewise.
14469 (sync_new_add<mode>): Likewise.
14470 (sync_new_sub<mode>): Likewise.
14471 (sync_<optab><mode>): Likewise.
14472 (sync_old_<optab><mode>): Likewise.
14473 (sync_new_<optab><mode>): Likewise.
14474 (sync_nand<mode>): Likewise.
14475 (sync_old_nand<mode>): Likewise.
14476 (sync_new_nand<mode>): Likewise.
14477 (sync_lock_test_and_set<mode>): Likewise.
14478 (test_and_set_12): Likewise.
14479 (atomic_compare_and_swap<mode>): Likewise.
14480 (atomic_exchange<mode>_llsc): Likewise.
14481 (atomic_fetch_add<mode>_llsc): Likewise.
14482 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
14483 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
14484 (umips_save_restore_pattern_p): Likewise.
14485 (umips_load_store_pair_p): Likewise.
14486 (umips_output_load_store_pair): Likewise.
14487 (umips_movep_target_p): Likewise.
14488 (umips_12bit_offset_address_p): Likewise.
14489 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
14490 (mips_base_mips16): Rename this...
14491 (mips_base_compression_flags): ...to this. Update all uses.
14492 (mips_attribute_table): Add micromips, nomicromips and nocompression.
14493 (mips_mips16_decl_p): Delete.
14494 (mips_nomips16_decl_p): Delete.
14495 (mips_get_compress_on_flags): New function.
14496 (mips_get_compress_off_flags): New function.
14497 (mips_get_compress_mode): New function.
14498 (mips_get_compress_on_name): New function.
14499 (mips_get_compress_off_name): New function.
14500 (mips_insert_attributes): Support multiple compression types.
14501 (mips_merge_decl_attributes): Likewise.
14502 (umips_12bit_offset_address_p): New function.
14503 (mips_start_function_definition): Emit .set micromips directive.
14504 (mips_call_may_need_jalx_p): New function.
14505 (mips_function_ok_for_sibcall): Add microMIPS support.
14506 (mips_print_operand_punctuation): Support short delay slots and
14508 (umips_swm_mask, umips_swm_encoding): New.
14509 (umips_build_save_restore): New function.
14510 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
14511 (was_mips16_p): Remove.
14512 (old_compression_mode): New.
14513 (mips_set_compression_mode): New function.
14514 (mips_set_current_function): Add microMIPS support.
14515 (mips_option_override): Likewise.
14516 (umips_save_restore_pattern_p): New function.
14517 (umips_output_save_restore): New function.
14518 (umips_load_store_pair_p_1): New function.
14519 (umips_load_store_pair_p): New function.
14520 (umips_output_load_store_pair_1): New function.
14521 (umips_output_load_store_pair): New function.
14522 (umips_movep_target_p) New function.
14523 (mips_prepare_pch_save): Add microMIPS support.
14524 * config/mips/mips.h (TARGET_COMPRESSION): New.
14525 (TARGET_CPU_CPP_BUILTINS): Update macro
14526 to use new compression flags and to support microMIPS.
14527 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
14528 (MIPS_ARCH_FLOAT_SPEC): Likewise.
14529 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
14530 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
14531 (ASM_SPEC): Support mmicromips and mno-micromips.
14532 (M16STORE_REG_P): New macro.
14533 (MIPS_CALL): Support TARGET_MICROMIPS.
14534 (MICROMIPS_J): New macro.
14535 (mips_base_mips16): Rename this...
14536 (mips_base_compression_flags): ...to this.
14537 (UMIPS_12BIT_OFFSET_P): New macro.
14538 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
14539 (MULTILIB_DIRNAMES): Likewise.
14540 2013-03-20 Richard Biener <rguenther@suse.de>
14542 PR tree-optimization/56661
14543 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
14544 the result does not have to be distinct.
14546 2013-03-20 Richard Biener <rguenther@suse.de>
14548 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
14551 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14552 Steven Bosscher <steven@gcc.gnu.org>
14554 PR rtl-optimization/56605
14555 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
14557 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
14560 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
14561 that require movd instead of movq.
14563 2013-03-20 Richard Biener <rguenther@suse.de>
14565 * tree-ssa-structalias.c (struct variable_info): Add pointer
14566 to the first field of an aggregate with sub-vars. Make
14567 this and the pointer to the next subfield its ID.
14568 (vi_next): New function.
14569 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
14570 storedanything_id, integer_id): Increment by one.
14571 (new_var_info, get_call_vi, lookup_call_clobber_vi,
14572 get_call_clobber_vi): Adjust.
14573 (solution_set_expand): Simplify and speedup.
14574 (solution_set_add): Inline into ...
14575 (set_union_with_increment): ... this. Adjust accordingly.
14576 (do_sd_constraint): Likewise.
14577 (do_ds_constraint): Likewise.
14578 (do_complex_constraint): Simplify.
14579 (build_pred_graph): Adjust.
14580 (solve_graph): Likewise. Simplify and speedup.
14581 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
14582 get_constraint_for_component_ref, get_constraint_for_1,
14583 first_vi_for_offset, first_or_preceding_vi_for_offset,
14584 create_function_info_for, create_variable_info_for_1,
14585 create_variable_info_for, intra_create_variable_infos): Adjust.
14586 (init_base_vars): Push NULL for ID zero.
14587 (compute_points_to_sets): Adjust.
14589 2013-03-20 Richard Biener <rguenther@suse.de>
14591 * cfgloop.c (verify_loop_structure): Streamline and avoid
14592 ICEing on corrupt loop tree.
14593 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
14596 2013-03-20 Richard Biener <rguenther@suse.de>
14598 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
14599 check whether an SSA update is needed.
14601 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
14603 * config/mips/constraints.md (T): Rename to...
14607 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
14608 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
14610 2013-03-19 Ian Bolton <ian.bolton@arm.com>
14612 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
14613 (*subsi3_carryin_uxtw): Likewise.
14615 2013-03-19 Ian Bolton <ian.bolton@arm.com>
14617 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
14618 (*rorsi3_insn_uxtw): Likewise.
14620 2013-03-19 Ian Bolton <ian.bolton@arm.com>
14622 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
14623 (*extrsi5_insn_uxtw): Likewise.
14625 2013-03-19 Richard Biener <rguenther@suse.de>
14627 PR tree-optimization/56273
14628 * passes.c (init_optimization_passes): Move second VRP after DOM.
14630 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
14632 * config/i386/i386.md (*movti_internal): Merge from
14633 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
14634 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
14635 nox64 isa attributes.
14637 2013-03-18 Richard Biener <rguenther@suse.de>
14639 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
14641 (merge_node_constraints): Likewise.
14642 (build_succ_graph): Likewise.
14643 (valid_graph_edge): Inline into single caller.
14644 (unify_nodes): Likewise. Use bitmap_set_bit return value
14646 (scc_visit): Fix formatting and variable use.
14647 (do_sd_constraint): Use gcc_checking_assert.
14648 (do_ds_constraint): Likewise.
14649 (do_complex_constraint): Likewise.
14650 (condense_visit): Likewise. Cleanup.
14651 (dump_pred_graph): New function.
14652 (perform_var_substitution): Dump the pred-graph before
14653 variable substitution.
14654 (find_equivalent_node): Use gcc_checking_assert.
14655 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
14657 2013-03-18 Richard Biener <rguenther@suse.de>
14659 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
14660 Remove cond_expr_stmt_list argument and do not gimplify the
14662 (vect_loop_versioning): Adjust.
14663 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
14664 Cleanup to use less temporaries.
14665 (vect_create_data_ref_ptr): Cleanup.
14667 2013-03-18 Jakub Jelinek <jakub@redhat.com>
14669 PR tree-optimization/56635
14670 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
14671 require types_compatible_p types.
14673 2013-03-18 Nick Clifton <nickc@redhat.com>
14675 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
14676 spurious backslash.
14678 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
14679 Add missing line to comment describing function.
14681 2013-03-18 Richard Biener <rguenther@suse.de>
14683 PR tree-optimization/56210
14684 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
14685 Handle string / character search functions.
14686 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
14688 2013-03-18 Richard Biener <rguenther@suse.de>
14690 PR middle-end/56483
14691 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
14692 and implement properly.
14693 * gimple.h (gimple_cond_single_var_p): Remove.
14695 2013-03-18 Richard Biener <rguenther@suse.de>
14697 * tree-data-ref.h (find_data_references_in_loop): Declare.
14698 * tree-data-ref.c (get_references_in_stmt): Use a stack
14699 vector pre-allocated in the callers.
14700 (find_data_references_in_stmt): Adjust.
14701 (graphite_find_data_references_in_stmt): Likewise.
14702 (create_rdg_vertices): Likewise.
14703 (find_data_references_in_loop): Export.
14704 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
14705 Compute dependences here...
14706 (vect_analyze_data_refs): ...not here. When we encounter
14707 a non-vectorizable data reference in basic-block vectorization
14708 truncate the data reference vector. Do not bother to
14709 fixup data-dependence information for gather loads.
14710 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
14711 of data references, as reported.
14713 2013-03-18 Richard Biener <rguenther@suse.de>
14715 PR tree-optimization/3713
14716 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
14717 has_constants and expr.
14718 (stmt_has_constants): Properly valueize SSA names when deciding
14719 whether the stmt has constants.
14721 2013-03-18 Richard Biener <rguenther@suse.de>
14723 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
14724 whole function when there is nothing to do.
14725 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
14726 * tree-vectorizer.c (vectorize_loops): Update virtual and
14727 loop-closed SSA once.
14728 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
14730 2013-03-18 Richard Biener <rguenther@suse.de>
14732 PR middle-end/56113
14733 * domwalk.c (bb_postorder): New global static.
14734 (cmp_bb_postorder): New function.
14735 (walk_dominator_tree): Replace scheme imposing an order for
14736 visiting dominator sons by one sorting them at the time they
14737 are pushed on the stack.
14739 2013-03-18 Richard Biener <rguenther@suse.de>
14741 PR tree-optimization/39326
14742 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
14743 (struct mem_ref): Replace mem member with ao_ref typed member.
14744 (MEM_ANALYZABLE): Adjust.
14745 (memref_eq): Likewise.
14746 (mem_ref_alloc): Likewise.
14747 (gather_mem_refs_stmt): Likewise.
14748 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
14749 (execute_sm_if_changed_flag_set): Adjust.
14750 (execute_sm): Likewise.
14751 (ref_always_accessed_p): Likewise.
14752 (refs_independent_p): Likewise.
14753 (can_sm_ref_p): Likewise.
14755 2013-03-18 Jakub Jelinek <jakub@redhat.com>
14758 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
14759 return 1 even for !unsignedp.
14761 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
14763 * config/i386/i386.md (isa): Add x64 and nox64.
14764 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
14765 (*pushtf): Enable *roF alternative for x64 isa only.
14766 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
14767 mode attribute of integer alternatives to DImode for TARGET_64BIT.
14768 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
14769 (*movtf_internal): Merge from *movtf_internal_rex64 and
14770 *movtf_internal_sse. Use x64 and nox64 isa attributes.
14771 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
14772 nox64 isa attributes.
14773 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
14774 nox64 isa attributes.
14775 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
14777 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
14779 * config/alpha/alpha.c (TARGET_LRA_P): New define.
14781 2013-03-17 Jakub Jelinek <jakub@redhat.com>
14784 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
14785 class names. Remove trailing comma after "ALL_REGS".
14787 2013-03-16 Jan Hubicka <jh@suse.cz>
14789 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
14790 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
14791 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
14792 of cgraph_get_create_node.
14793 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
14795 2013-03-16 Jason Merrill <jason@redhat.com>
14798 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
14799 with DW_AT_default_value.
14801 2013-03-16 Jakub Jelinek <jakub@redhat.com>
14803 * BASE-VER: Set to 4.9.0.
14805 2013-03-14 Andi Kleen <ak@linux.intel.com>
14808 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
14809 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
14810 Document _x* TSX intrinsics.
14812 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
14813 David Holsgrove <david.holsgrove@xilinx.com>
14815 * configure.ac: Add MicroBlaze TLS support detection.
14816 * configure: Regenerate.
14817 * config/microblaze/microblaze-protos.h
14818 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
14819 symbol_mentioned_p, label_mentioned_p): Add prototypes.
14820 * config/microblaze/microblaze.c (microblaze_address_type): Add
14821 ADDRESS_TLS and tls_reloc address types.
14822 (microblaze_address_info): Add tls_reloc.
14823 (TARGET_HAVE_TLS): Define.
14824 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
14825 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
14826 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
14827 load_tls_operand, microblaze_call_tls_get_addr,
14828 microblaze_legitimize_tls_address): New functions.
14829 (microblaze_classify_unspec): Handle UNSPEC_TLS.
14830 (get_base_reg): Use microblaze_tls_symbol_p.
14831 (microblaze_classify_address): Handle TLS.
14832 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
14833 label_mentioned_p and microblaze_tls_referenced_p.
14834 (microblaze_legitimize_address): Handle TLS.
14835 (microblaze_address_insns): Handle ADDRESS_TLS.
14836 (pic_address_needs_scratch): Handle TLS.
14837 (print_operand_address): Handle TLS.
14838 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
14839 (microblaze_expand_move): Handle TLS.
14840 (microblaze_legitimate_constant_p): Check
14841 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
14842 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
14843 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
14844 (PIC_OFFSET_TABLE_REGNUM): Set.
14845 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
14846 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
14847 (addsi3, movsi_internal2, movdf_internal): Update constraints
14848 * config/microblaze/predicates.md (arith_plus_operand): Define
14849 (move_operand): Redefine as move_src_operand,
14850 check microblaze_tls_referenced_p.
14852 2013-03-14 Ian Bolton <ian.bolton@arm.com>
14854 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
14855 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
14857 2013-03-14 Ian Bolton <ian.bolton@arm.com>
14859 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
14862 2013-03-14 Jakub Jelinek <jakub@redhat.com>
14864 PR tree-optimization/53265
14865 * common.opt (Waggressive-loop-optimizations): New option.
14866 * tree-ssa-loop-niter.c: Include tree-pass.h.
14867 (do_warn_aggressive_loop_optimizations): New function.
14868 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
14869 if number_of_latch_executions returned constant.
14870 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
14871 early. If number_of_latch_executions returned constant, set
14872 nb_iterations_upper_bound back to it.
14873 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
14875 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
14876 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
14878 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
14879 (MULTILIB_OSDIRNAMES): Set.
14880 * genmultilib: If defaultosdirname doesn't start with :: , set
14881 defaultosdirname2 instead, clear it and emit two . multilib_raw
14882 entries instead of just one.
14884 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
14886 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
14887 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
14888 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
14889 (SUBTARGET_OVERRIDE_OPTIONS): New.
14891 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
14894 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
14895 (musermode): Convert to Var(TARGET_USERMODE).
14896 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
14897 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
14898 * config/sh/sh.c (sh_option_override): Use
14899 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
14900 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
14902 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
14904 (udivsi3_i4_single, divsi3_i4_single): Use
14905 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
14907 2013-03-13 Dave Korn <dave.korn.cygwin@....>
14909 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
14912 2013-03-13 Richard Biener <rguenther@suse.de>
14914 PR tree-optimization/56608
14915 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
14916 calls when vectorizing basic-blocks.
14918 2013-03-13 Jakub Jelinek <jakub@redhat.com>
14921 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
14924 2013-03-12 Jakub Jelinek <jakub@redhat.com>
14926 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
14928 2013-03-11 Jan Hubicka <jh@suse.cz>
14931 * lto-streamer-out.c (output_symbol_p): Skip references from
14932 constructors of external variables.
14934 2013-03-11 Jan Hubicka <jh@suse.cz>
14936 PR middle-end/56571
14937 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
14939 * emit-rtl.c (verify_rtx_sharing): Likewise.
14940 (copy_insn_1): Likewise.
14941 * rtl.c (copy_rtx): Likewise.
14943 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
14946 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
14947 output_operand_lossage message.
14949 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
14952 * arm.c (shift_op): Validate RTL pattern on the fly.
14953 (arm_print_operand, case 'S'): Don't use shift_operator to validate
14956 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
14959 * config/pa/pa.md (call_value): Check for calls to powf and direct to
14960 new call patterns that clobber %fr12.
14961 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
14962 split and postreload patterns.
14963 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
14964 registers %fr12 and %fr12R as call used.
14966 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
14968 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
14969 (canon_address, record_store, replace_read, check_mem_read_rtx,
14970 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
14971 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
14972 rest_of_handle_dse): Likewise.
14974 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
14976 PR middle-end/56524
14977 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
14979 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
14980 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
14981 (save_optabs_if_changed): Replace with...
14982 (init_tree_optimization_optabs): ...this.
14983 * optabs.c (save_optabs_if_changed): Rename to...
14984 (init_tree_optimization_optabs): ...this. Take the optimization node
14985 as argument. Do nothing if the base optabs are already correct.
14986 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
14987 to recompute optabs.
14988 * function.h (function): Remove optabs field.
14989 * function.c (invoke_set_current_function_hook): Call
14990 init_tree_optimization_optabs. Use the result to initialize
14993 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
14995 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
14996 if GTMA_HAS_NO_INSTRUMENTATION.
14997 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
14998 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
14999 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
15000 * gimple-pretty-print.c (dump_gimple_transaction): Handle
15001 GTMA_HAS_NO_INSTRUMENTATION.
15003 2013-03-08 Jakub Jelinek <jakub@redhat.com>
15005 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
15008 2013-03-08 Marek Polacek <polacek@redhat.com>
15009 Jakub Jelinek <jakub@redhat.com>
15011 PR tree-optimization/56478
15012 * predict.c (is_comparison_with_loop_invariant_p): Change the
15013 type of loop_step to tree.
15014 (predict_loops): Adjust.
15015 (predict_iv_comparison): Perform the computations on double_ints.
15017 2013-03-08 Richard Biener <rguenther@suse.de>
15019 PR tree-optimization/56570
15020 * tree-cfg.c (verify_expr_location_1): Verify locations for
15022 * tree-sra.c (create_access_replacement): Strip locations
15023 from DECL_DEBUG_EXPRs.
15025 2013-03-08 Richard Biener <rguenther@suse.de>
15027 * tree-inline.c (expand_call_inline): Do not associate
15028 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
15029 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
15031 2013-03-08 Richard Biener <rguenther@suse.de>
15033 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
15034 or block changes with -Og. Fix for location / block encoding
15035 changes and PHI arguments with locations.
15037 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
15039 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
15041 (struct output_info): Likewise.
15042 (register_overhead): Remove bad gcc_assert.
15043 (bitmap_find_bit): If there is only a single bitmap element, do not
15044 count a miss as a search.
15045 (print_statistics): Update for counter type changes.
15046 (dump_bitmap_statistics): Likewise. Print headers such that they
15047 are properly lined up with the printed counters.
15049 2013-03-07 Jakub Jelinek <jakub@redhat.com>
15051 PR tree-optimization/56559
15052 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
15053 check that it has only a single use.
15055 2013-03-07 Richard Biener <rguenther@suse.de>
15057 * doc/invoke.texi (fwhole-program): Discourage use in combination
15060 2013-03-06 Jakub Jelinek <jakub@redhat.com>
15062 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
15064 PR tree-optimization/56539
15065 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
15066 instead of GSI_CONTINUE_LINKING as last argument to
15067 force_gimple_operand_gsi. Adjust function comment.
15069 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
15072 PR middle-end/56548
15073 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
15074 promoted mode, convert the result back to the original mode.
15076 2013-03-06 Richard Biener <rguenther@suse.de>
15078 PR middle-end/56294
15079 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
15080 (insert_updated_phi_nodes_compare_uids): New function.
15081 (update_ssa): Sort symbols_to_rename after UID before
15082 traversing it to insert PHI nodes.
15084 2013-03-06 Richard Biener <rguenther@suse.de>
15086 PR middle-end/50494
15087 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
15088 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
15091 2013-02-13 Richard Biener <rguenther@suse.de>
15094 * varasm.c (output_constant_def_1): Get the decl representing
15095 the constant as argument.
15096 (output_constant_def): Wrap output_constant_def_1.
15097 (make_decl_rtl): Use output_constant_def_1 with the decl
15098 representing the constant.
15099 (build_constant_desc): Optionally re-use a decl already
15100 representing the constant.
15101 (tree_output_constant_def): Adjust.
15103 2013-03-06 Joey Ye <joey.ye@arm.com>
15106 * gcc.c (convert_white_space): New function.
15107 (main): Handles white space in function name.
15109 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
15112 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
15113 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
15114 to SH_DIV_CALL_TABLE for TARGET_SH2.
15115 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
15117 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
15118 call-table options.
15120 2013-03-05 Sterling Augustine <saugustine@google.com>
15121 Cary Coutant <ccoutant@google.com>
15124 * dwarf2out.c (resolve_addr): Don't call
15125 remove_loc_list_addr_table_entries a second time for the same
15128 2013-03-05 Jakub Jelinek <jakub@redhat.com>
15131 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
15132 (avoid_complex_debug_insns): New function.
15133 (expand_debug_locations): Call it.
15135 PR rtl-optimization/56484
15136 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
15137 lifetimes of hard registers on small register class machines.
15139 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
15141 * config/microblaze/microblaze-protos.h: Rename
15142 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
15143 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
15145 (microblaze_fast_interrupt_function_p): New function.
15146 (microblaze_is_interrupt_handler): Rename to
15147 microblaze_is_interrupt_variant and add fast_interrupt check.
15148 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
15149 (save_restore_insns): Likewise.
15150 (compute_frame_size): Likewise.
15151 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
15152 (microblaze_globalize_label): Likewise.
15153 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
15154 * config/microblaze/microblaze.md: Use wrapper
15155 microblaze_is_interrupt_variant.
15157 2013-03-05 Kai Tietz <ktietz@redhat.com>
15159 * sdbout.c (sdbout_one_type): Switch to current function's section
15160 supporting cold/hot.
15162 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
15164 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
15167 2013-03-05 Jakub Jelinek <jakub@redhat.com>
15169 PR middle-end/56461
15170 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
15171 if VALGRIND_GET_VBITS is defined, temporarily make object
15172 memory all defined, and restore previous valgrind addressability
15173 and definability afterwards. Free this_object at the end.
15175 PR middle-end/56461
15176 * lra.c (lra): Call lra_clear_live_ranges if live_p,
15177 right before calling lra_create_live_ranges, also call it
15178 when clearing live_p. Only call lra_clear_live_ranges
15179 at the end if live_p.
15181 PR middle-end/56461
15182 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
15184 2013-03-05 Richard Biener <rguenther@suse.de>
15186 PR tree-optimization/56521
15187 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
15190 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
15193 * except.h (remove_unreachable_eh_regions): New prototype.
15194 * except.c (remove_eh_handler_splicer): New function, split out
15195 of remove_eh_handler.
15196 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
15197 warning about running it on many EH regions one at a time.
15198 (remove_unreachable_eh_regions_worker): New function, walk the
15199 EH tree in depth-first order and remove non-marked regions.
15200 (remove_unreachable_eh_regions): New function.
15201 * tree-eh.c (mark_reachable_handlers): New function, split out
15202 from remove_unreachable_handlers.
15203 (remove_unreachable_handlers): Use mark_reachable_handlers and
15204 remove_unreachable_eh_regions.
15205 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
15206 and remove_unreachable_eh_regions.
15208 2013-03-05 Richard Biener <rguenther@suse.de>
15210 PR middle-end/56525
15211 * loop-init.c (fix_loop_structure): Remove loops in two stages,
15212 not freeing them until the end.
15214 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15216 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
15218 2013-03-05 Richard Biener <rguenther@suse.de>
15220 PR tree-optimization/56270
15221 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
15222 of loads after scheduling an SLP instance.
15224 2013-03-05 Jakub Jelinek <jakub@redhat.com>
15226 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
15228 (check_gcc_parallelize): Run guality.exp as a separate job from
15229 vect.exp with unsorted.exp and $(dg_target_exps) separately from
15230 struct-layout-1.exp with stackalign.exp.
15232 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
15234 PR middle-end/56461
15235 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
15236 load_index sbitmap even if some bit in it isn't set.
15238 PR middle-end/56461
15239 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
15240 (discover_iteration_bound_by_body_walk): Change queues to
15241 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
15242 spelling in comment. Call safe_push on queues[bound_index] directly.
15243 Release queues[queue_index] in every iteration unconditionally.
15244 Release bounds vector.
15246 PR middle-end/56461
15247 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
15248 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
15249 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
15252 2013-03-05 Richard Biener <rguenther@suse.de>
15255 * tree-inline.c (remap_blocks_to_null): New function.
15256 (expand_call_inline): When expanding a call stmt without
15257 an associated block inline remap all callee blocks to NULL.
15259 2013-03-05 Jakub Jelinek <jakub@redhat.com>
15261 PR rtl-optimization/56494
15262 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
15263 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
15264 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
15266 PR middle-end/56461
15267 * sel-sched-ir.c (free_sched_pools): Release
15268 succs_info_pool.stack[succs_info_pool.max_top] vectors too
15269 if succs_info_pool.max_top isn't -1.
15272 * opts.c (opts_obstack, opts_concat): Moved to...
15273 * opts-common.c (opts_obstack, opts_concat): ... here.
15275 2013-03-04 Jakub Jelinek <jakub@redhat.com>
15277 PR middle-end/56461
15278 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
15280 2013-03-04 Martin Jambor <mjambor@suse.cz>
15282 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
15283 all appropriate places.
15285 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
15287 PR tree-optimization/56424
15288 * ipa-split.c (split_function): Do not set the RSO flag if result is
15289 not by reference and its type is a register type.
15291 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
15293 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
15294 (microblaze_legitimate_pic_operand): Likewise
15295 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
15296 new function microblaze_legitimate_pic_operand
15297 * config/microblaze/microblaze-protos.h
15298 (microblaze_legitimate_pic_operand): Declare.
15300 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
15302 * config/microblaze/predicates.md (call_insn_simple_operand):
15303 New predicate for supported rtx code types.
15304 * config/microblaze/microblaze.md (call_internal1): Use
15305 call_insn_simple_operand predicate.
15307 2013-03-04 Jakub Jelinek <jakub@redhat.com>
15309 PR middle-end/56461
15310 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
15311 partitions.ordered_remove.
15313 PR middle-end/56461
15314 * tree-vect-stmts.c (vectorizable_conversion): Don't call
15315 vec_oprnds0.create (1) for modifier == NONE.
15317 PR middle-end/56461
15318 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
15319 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
15320 vec_oprnds1 right before pushing anything to it for
15323 PR middle-end/56461
15324 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
15325 set nbbs to 0 instead of having separate code path.
15326 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
15327 instead of false as last argument if returning NULL.
15329 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
15331 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
15332 the attribute is now called "target" instead of "option".
15333 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
15334 * doc/tm.texi.in (Target Attributes): Likewise document the correct
15335 attribute/pragma name for TARGET_OPTION_VALID_P and
15336 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
15337 * doc/tm.texi: Regenerated.
15339 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
15341 * config/microblaze/microblaze.c:
15342 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
15343 * config/microblaze/microblaze.h: Add -mxl-reorder to
15345 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
15346 instructions emitted if TARGET_REORDER.
15347 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
15348 or 0 for -m/-mno case, but initialises as 2 to detect default use case
15351 2013-03-01 Xinliang David Li <davidxl@google.com>
15353 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
15356 2013-03-01 Jakub Jelinek <jakub@redhat.com>
15358 PR middle-end/56461
15359 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
15360 vector even when returning true. Fix up function comment formatting.
15362 PR middle-end/56461
15363 * ira-build.c (ira_loop_nodes_count): New variable.
15364 (create_loop_tree_nodes): Initialize it.
15365 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
15367 PR middle-end/56461
15368 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
15369 method on dr_chain and result_chain.
15370 * tree-vect-stmts.c (vectorizable_store): Only call
15371 result_chain.create if j == 0.
15373 PR middle-end/56461
15374 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
15375 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
15376 before overwriting it.
15378 2013-03-01 Tobias Burnus <burnus@net-b.de>
15380 * doc/extended.texi (C Extensions): Change order in @menu
15382 (Other MIPS Built-in Functions): Move last MIPS entry before
15383 "picoChip Built-in Functions".
15384 (SH Built-in Functions): Move after RX Built-in Functions.
15385 * doc/gcc.texi (Introduction): Change order in @menu
15387 * doc/md.texi (Constraints): Ditto.
15388 * gty.texi (Type Information): Ditto.
15389 (User-provided marking routines for template types): Make
15391 * doc/invoke.texi (AArch64 Options): Move before
15392 "Adapteva Epiphany Options".
15394 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
15395 Jakub Jelinek <jakub@redhat.com>
15398 * asan.c (gate_asan): Lookup no_sanitize_address instead of
15399 no_address_safety_analysis attribute.
15400 * doc/extend.texi (no_address_safety_attribute): Rename to
15401 no_sanitize_address attribute, mention no_address_safety_analysis
15402 attribute as deprecated alias.
15404 2013-02-28 Jakub Jelinek <jakub@redhat.com>
15406 PR middle-end/56461
15407 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
15408 type to vec<vec<tree> > *.
15409 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
15410 to be vec<tree> instead of vec<tree> *, set vec_defs
15411 to vNULL and call vec_defs.create (number_of_vects), adjust other
15413 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
15414 vectorizable_condition): Adjust vect_get_slp_defs callers.
15416 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
15418 * config/aarch64/aarch64.c
15419 (aarch64_float_const_representable): Remove unused variable.
15421 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
15423 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
15425 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
15427 * config/aarch64/aarch64-builtins.c
15428 (aarch64_init_simd_builtins): Make static.
15430 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
15432 * config/aarch64/aarch64.c
15433 (aarch64_simd_make_constant): Make static.
15435 2013-02-28 Martin Jambor <mjambor@suse.cz>
15437 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
15438 with no initialization to the RHS of debug statements.
15440 2013-02-28 Martin Jambor <mjambor@suse.cz>
15442 PR tree-optimization/56294
15443 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
15445 (get_access_replacement): Do not call create_access_replacement.
15446 Assert a replacement exists.
15447 (get_repl_default_def_ssa_name): Create the replacement declaration
15450 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
15452 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
15453 final_end_function.
15455 2013-02-28 Marek Polacek <polacek@redhat.com>
15457 PR rtl-optimization/56466
15458 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
15459 if we're changing a loop.
15460 (peel_loops_completely): Likewise.
15462 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
15465 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
15467 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
15470 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
15471 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
15472 INTX_FTYPE_FX, FX_FTYPE_INTX.
15473 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
15475 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
15477 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
15478 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
15479 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
15480 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
15481 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
15482 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
15483 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
15484 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
15485 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
15486 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
15487 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
15488 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
15489 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
15490 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
15491 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
15492 (avrxmega6): Increase max flash segments from 5 to 6.
15493 * config/avr/t-multilib: Regenerate.
15494 * config/avr/avr-tables.opt: Regenerate.
15495 * doc/avr-mmcu.texi: Regenerate.
15497 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
15499 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
15500 (avr_device_to_arch): Rename to avr_device_to_ld.
15501 (avr_device_to_as): New prototype.
15502 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
15503 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
15504 * config/avr/driver-avr.c (avr_device_to_as): New.
15505 (avr_device_to_arch): Rename to avr_device_to_ld.
15507 2013-02-27 Jakub Jelinek <jakub@redhat.com>
15509 PR middle-end/56461
15510 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
15511 method on dr_chain and result_chain.
15513 PR middle-end/56461
15514 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
15515 pointer_set_destroy on not_executed_last_iteration.
15517 PR middle-end/56461
15518 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
15520 PR middle-end/56461
15521 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
15522 FOR_EACH_DEFINED_FUNCTION when freeing state.
15524 PR middle-end/56461
15525 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
15527 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
15530 PR middle-end/56461
15531 * ipa-cp.c (decide_whether_version_node): Call vec_free on
15532 known_aggs[i].items and release known_aggs vector.
15534 PR middle-end/56461
15535 * ipa-reference.c (propagate): Free node_info even for alias nodes.
15537 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
15539 * config/microblaze/microblaze.c (microblaze_emit_compare):
15540 Use xor for EQ/NE comparisions.
15541 * config/microblaze/microblaze.md (cstoresf4): Add constraints
15542 (cbranchsf4): Adjust operator to comparison_operator.
15544 2013-02-27 Jakub Jelinek <jakub@redhat.com>
15546 PR middle-end/56461
15547 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
15549 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
15550 vec_safe_push, always update *slot.
15551 (redirect_edge_var_map_clear): Use vec_free.
15552 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
15553 (free_var_map_entry): Use vec_free.
15554 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
15555 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
15557 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
15559 PR middle-end/45472
15560 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
15561 when the may_trap_p bit of the exprs being merged differs.
15562 Reorder tests for speculativeness in the logical and operator.
15564 2013-02-27 Jakub Jelinek <jakub@redhat.com>
15566 * incpath.c (add_standard_paths): Use reconcat instead of concat
15567 where appropriate and avoid leaking memory.
15569 * opts.h: Include obstack.h.
15570 (opts_concat): New prototype.
15571 (opts_obstack): New declaration.
15572 * opts.c (opts_concat): New function.
15573 (opts_obstack): New variable.
15574 (init_options_struct): Call gcc_init_obstack on opts_obstack.
15575 (finish_options): Use opts_concat instead of concat
15576 and XOBNEWVEC instead of XNEWVEC.
15577 * opts-common.c (generate_canonical_option, decode_cmdline_option,
15578 generate_option): Likewise.
15579 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
15580 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
15583 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
15584 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
15586 2013-02-26 Jakub Jelinek <jakub@redhat.com>
15588 PR middle-end/56461
15589 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
15591 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
15593 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
15594 (arm_block_move_unaligned_straight): Likewise.
15595 (arm_adjust_block_mem): Likewise.
15597 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
15600 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
15601 temp, cond and label.
15602 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
15605 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
15606 * config/c6x/c6x.h (dbx_register_map): Update declaration.
15609 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
15610 of prologue/epilogue functions.
15613 * config/tilegx/tilegx.c (tilegx_expand_prologue):
15614 Remove unused variable cfa_offset.
15615 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
15618 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
15619 type promotion to unsigned.
15622 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
15623 for HOST_WIDE_INT of 32 bit / same size as int.
15624 (arm_block_move_unaligned_straight): Likewise.
15625 (arm_adjust_block_mem): Likewise.
15628 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
15631 2013-02-26 Marek Polacek <polacek@redhat.com>
15633 PR tree-optimization/56426
15634 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
15636 2013-02-26 Richard Biener <rguenther@suse.de>
15639 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
15640 unused variable loops.
15642 2013-02-26 Jakub Jelinek <jakub@redhat.com>
15644 PR tree-optimization/56448
15645 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
15646 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
15647 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
15648 later operands of the references, or even first operand for
15649 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
15651 PR tree-optimization/56443
15652 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
15653 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
15654 to type_for_mode langhook.
15656 2013-02-25 Matt Turner <mattst88@gmail.com>
15658 * doc/invoke.texi: Document r4700.
15660 2013-02-25 Richard Biener <rguenther@suse.de>
15662 PR tree-optimization/56175
15663 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
15665 (simplify_bitwise_binary): ... here. Also guard the conversion
15666 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
15668 2013-02-25 Catherine Moore <clm@codesourcery.com>
15671 2013-02-24 Catherine Moore <clm@codesourcery.com>
15672 Maciej W. Rozycki <macro@codesourcery.com>
15673 Tom de Vries <tom@codesourcery.com>
15674 Nathan Sidwell <nathan@codesourcery.com>
15675 Iain Sandoe <iain@codesourcery.com>
15676 Nathan Froyd <froydnj@codesourcery.com>
15677 Chao-ying Fu <fu@mips.com>
15679 * doc/extend.texi: (micromips, nomicromips, nocompression):
15680 Document new function attributes.
15681 * doc/invoke.texi (minterlink-compressed, mmicromips,
15682 m14k, m14ke, m14kec): Document new options.
15683 (minterlink-mips16): Update documentation.
15684 * doc/md.texi (ZC, ZD): Document new constraints.
15685 * configure.ac (gcc_cv_as_micromips): Check if linker
15686 supports the .set micromips directive.
15687 * configure: Regenerate.
15688 * config.in: Regenerate.
15689 * config/mips/mips-tables.opt: Regenerate.
15690 * config/mips/micromips.md: New file.
15691 * constraints.md (ZC, AD): New constraints.
15692 * config/mips/predicates.md (movep_src_register): New predicate.
15693 (movep_src_operand): New predicate.
15694 (non_volatile_mem_operand): New predicate.
15695 * config/mips/mips.md (multimem): New type.
15696 (length): Differentiate between 17-bit and 18-bit branch offsets.
15697 (MOVEP1, MOVEP2): New mode iterator.
15698 (mov_<load>l): Use ZC constraint.
15699 (mov_<load>r): Likewise.
15700 (mov_<store>l): Likewise.
15701 (mov_<store>r): Likewise.
15702 (*branch_equality<mode>_inverted): Add microMIPS support.
15703 (*branch_equality<mode>): Likewise.
15704 (*jump_absolute): Likewise.
15705 (indirect_jump_<mode>): Likewise.
15706 (tablejump_<mode>): Likewise.
15707 (<optab>_internal): Likewise.
15708 (sibcall_internal): Likewise.
15709 (sibcall_value_internal): Likewise.
15710 (prefetch): Use constraint ZD.
15711 * config/mips/mips.opt (minterlink-compressed): New option.
15712 (minterlink-mips16): Now an alias for minterlink-compressed.
15713 (mmicromips): New option.
15714 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
15715 (compare_and_swap_12): Likewise.
15716 (sync_add<mode>): Likewise.
15717 (sync_<optab>_12): Likewise.
15718 (sync_old_<optab>_12): Likewise.
15719 (sync_new_<optab>_12): Likewise.
15720 (sync_nand_12): Likewise.
15721 (sync_old_nand_12): Likewise.
15722 (sync_new_nand_12): Likewise.
15723 (sync_sub<mode>): Likewise.
15724 (sync_old_add<mode>): Likewise.
15725 (sync_old_sub<mode>): Likewise.
15726 (sync_new_add<mode>): Likewise.
15727 (sync_new_sub<mode>): Likewise.
15728 (sync_<optab><mode>): Likewise.
15729 (sync_old_<optab><mode>): Likewise.
15730 (sync_new_<optab><mode>): Likewise.
15731 (sync_nand<mode>): Likewise.
15732 (sync_old_nand<mode>): Likewise.
15733 (sync_new_nand<mode>): Likewise.
15734 (sync_lock_test_and_set<mode>): Likewise.
15735 (test_and_set_12): Likewise.
15736 (atomic_compare_and_swap<mode>): Likewise.
15737 (atomic_exchange<mode>_llsc): Likewise.
15738 (atomic_fetch_add<mode>_llsc): Likewise.
15739 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
15740 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
15741 (umips_save_restore_pattern_p): Likewise.
15742 (umips_load_store_pair_p): Likewise.
15743 (umips_output_load_store_pair): Likewise.
15744 (umips_movep_target_p): Likewise.
15745 (umips_12bit_offset_address_p): Likewise.
15746 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
15747 (mips_base_mips16): Rename this...
15748 (mips_base_compression_flags): ...to this. Update all uses.
15749 (mips_attribute_table): Add micromips, nomicromips and nocompression.
15750 (mips_mips16_decl_p): Delete.
15751 (mips_nomips16_decl_p): Delete.
15752 (mips_get_compress_on_flags): New function.
15753 (mips_get_compress_off_flags): New function.
15754 (mips_get_compress_mode): New function.
15755 (mips_get_compress_on_name): New function.
15756 (mips_get_compress_off_name): New function.
15757 (mips_insert_attributes): Support multiple compression types.
15758 (mips_merge_decl_attributes): Likewise.
15759 (umips_12bit_offset_address_p): New function.
15760 (mips_start_function_definition): Emit .set micromips directive.
15761 (mips_call_may_need_jalx_p): New function.
15762 (mips_function_ok_for_sibcall): Add microMIPS support.
15763 (mips_print_operand_punctuation): Support short delay slots and
15765 (umips_swm_mask, umips_swm_encoding): New.
15766 (umips_build_save_restore): New function.
15767 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
15768 (was_mips16_p): Remove.
15769 (old_compression_mode): New.
15770 (mips_set_compression_mode): New function.
15771 (mips_set_current_function): Add microMIPS support.
15772 (mips_option_override): Likewise.
15773 (umips_save_restore_pattern_p): New function.
15774 (umips_output_save_restore): New function.
15775 (umips_load_store_pair_p_1): New function.
15776 (umips_load_store_pair_p): New function.
15777 (umips_output_load_store_pair_1): New function.
15778 (umips_output_load_store_pair): New function.
15779 (umips_movep_target_p) New function.
15780 (mips_prepare_pch_save): Add microMIPS support.
15781 * config/mips/mips.h (TARGET_COMPRESSION): New.
15782 (TARGET_CPU_CPP_BUILTINS): Update macro
15783 to use new compression flags and to support microMIPS.
15784 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
15785 (MIPS_ARCH_FLOAT_SPEC): Likewise.
15786 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
15787 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
15788 (ASM_SPEC): Support mmicromips and mno-micromips.
15789 (M16STORE_REG_P): New macro.
15790 (MIPS_CALL): Support TARGET_MICROMIPS.
15791 (MICROMIPS_J): New macro.
15792 (mips_base_mips16): Rename this...
15793 (mips_base_compression_flags): ...to this.
15794 (UMIPS_12BIT_OFFSET_P): New macro.
15795 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
15796 (MULTILIB_DIRNAMES): Likewise.
15798 2013-02-25 Tom de Vries <tom@codesourcery.com>
15800 PR rtl-optimization/56131
15801 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
15802 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
15803 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
15805 2013-02-25 Tobias Burnus <burnus@net-b.de>
15807 * doc/invoke.texi (-fsanitize=): Move from optimization
15808 to debugging options.
15810 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
15812 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
15814 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
15815 Alexander Monakov <amonakov@ispras.ru>
15817 PR middle-end/56077
15818 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
15819 flush pending lists also on non-jumps. Adjust comment.
15821 2013-02-24 Catherine Moore <clm@codesourcery.com>
15822 Maciej W. Rozycki <macro@codesourcery.com>
15823 Tom de Vries <tom@codesourcery.com>
15824 Nathan Sidwell <nathan@codesourcery.com>
15825 Iain Sandoe <iain@codesourcery.com>
15826 Nathan Froyd <froydnj@codesourcery.com>
15827 Chao-ying Fu <fu@mips.com>
15829 * doc/extend.texi: (micromips, nomicromips, nocompression):
15830 Document new function attributes.
15831 * doc/invoke.texi (minterlink-compressed, mmicromips,
15832 m14k, m14ke, m14kec): Document new options.
15833 (minterlink-mips16): Update documentation.
15834 * doc/md.texi (ZC, ZD): Document new constraints.
15835 * configure.ac (gcc_cv_as_micromips): Check if linker
15836 supports the .set micromips directive.
15837 * configure: Regenerate.
15838 * config.in: Regenerate.
15839 * config/mips/mips-tables.opt: Regenerate.
15840 * config/mips/micromips.md: New file.
15841 * constraints.md (ZC, AD): New constraints.
15842 * config/mips/predicates.md (movep_src_register): New predicate.
15843 (movep_src_operand): New predicate.
15844 (non_volatile_mem_operand): New predicate.
15845 * config/mips/mips.md (multimem): New type.
15846 (length): Differentiate between 17-bit and 18-bit branch offsets.
15847 (MOVEP1, MOVEP2): New mode iterator.
15848 (mov_<load>l): Use ZC constraint.
15849 (mov_<load>r): Likewise.
15850 (mov_<store>l): Likewise.
15851 (mov_<store>r): Likewise.
15852 (*branch_equality<mode>_inverted): Add microMIPS support.
15853 (*branch_equality<mode>): Likewise.
15854 (*jump_absolute): Likewise.
15855 (indirect_jump_<mode>): Likewise.
15856 (tablejump_<mode>): Likewise.
15857 (<optab>_internal): Likewise.
15858 (sibcall_internal): Likewise.
15859 (sibcall_value_internal): Likewise.
15860 (prefetch): Use constraint ZD.
15861 * config/mips/mips.opt (minterlink-compressed): New option.
15862 (minterlink-mips16): Now an alias for minterlink-compressed.
15863 (mmicromips): New option.
15864 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
15865 (compare_and_swap_12): Likewise.
15866 (sync_add<mode>): Likewise.
15867 (sync_<optab>_12): Likewise.
15868 (sync_old_<optab>_12): Likewise.
15869 (sync_new_<optab>_12): Likewise.
15870 (sync_nand_12): Likewise.
15871 (sync_old_nand_12): Likewise.
15872 (sync_new_nand_12): Likewise.
15873 (sync_sub<mode>): Likewise.
15874 (sync_old_add<mode>): Likewise.
15875 (sync_old_sub<mode>): Likewise.
15876 (sync_new_add<mode>): Likewise.
15877 (sync_new_sub<mode>): Likewise.
15878 (sync_<optab><mode>): Likewise.
15879 (sync_old_<optab><mode>): Likewise.
15880 (sync_new_<optab><mode>): Likewise.
15881 (sync_nand<mode>): Likewise.
15882 (sync_old_nand<mode>): Likewise.
15883 (sync_new_nand<mode>): Likewise.
15884 (sync_lock_test_and_set<mode>): Likewise.
15885 (test_and_set_12): Likewise.
15886 (atomic_compare_and_swap<mode>): Likewise.
15887 (atomic_exchange<mode>_llsc): Likewise.
15888 (atomic_fetch_add<mode>_llsc): Likewise.
15889 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
15890 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
15891 (umips_save_restore_pattern_p): Likewise.
15892 (umips_load_store_pair_p): Likewise.
15893 (umips_output_load_store_pair): Likewise.
15894 (umips_movep_target_p): Likewise.
15895 (umips_12bit_offset_address_p): Likewise.
15896 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
15897 (mips_base_mips16): Rename this...
15898 (mips_base_compression_flags): ...to this. Update all uses.
15899 (mips_attribute_table): Add micromips, nomicromips and nocompression.
15900 (mips_mips16_decl_p): Delete.
15901 (mips_nomips16_decl_p): Delete.
15902 (mips_get_compress_on_flags): New function.
15903 (mips_get_compress_off_flags): New function.
15904 (mips_get_compress_mode): New function.
15905 (mips_get_compress_on_name): New function.
15906 (mips_get_compress_off_name): New function.
15907 (mips_insert_attributes): Support multiple compression types.
15908 (mips_merge_decl_attributes): Likewise.
15909 (umips_12bit_offset_address_p): New function.
15910 (mips_start_function_definition): Emit .set micromips directive.
15911 (mips_call_may_need_jalx_p): New function.
15912 (mips_function_ok_for_sibcall): Add microMIPS support.
15913 (mips_print_operand_punctuation): Support short delay slots and
15915 (umips_swm_mask, umips_swm_encoding): New.
15916 (umips_build_save_restore): New function.
15917 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
15918 (was_mips16_p): Remove.
15919 (old_compression_mode): New.
15920 (mips_set_compression_mode): New function.
15921 (mips_set_current_function): Add microMIPS support.
15922 (mips_option_override): Likewise.
15923 (umips_save_restore_pattern_p): New function.
15924 (umips_output_save_restore): New function.
15925 (umips_load_store_pair_p_1): New function.
15926 (umips_load_store_pair_p): New function.
15927 (umips_output_load_store_pair_1): New function.
15928 (umips_output_load_store_pair): New function.
15929 (umips_movep_target_p) New function.
15930 (mips_prepare_pch_save): Add microMIPS support.
15931 * config/mips/mips.h (TARGET_COMPRESSION): New.
15932 (TARGET_CPU_CPP_BUILTINS): Update macro
15933 to use new compression flags and to support microMIPS.
15934 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
15935 (MIPS_ARCH_FLOAT_SPEC): Likewise.
15936 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
15937 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
15938 (ASM_SPEC): Support mmicromips and mno-micromips.
15939 (M16STORE_REG_P): New macro.
15940 (MIPS_CALL): Support TARGET_MICROMIPS.
15941 (MICROMIPS_J): New macro.
15942 (mips_base_mips16): Rename this...
15943 (mips_base_compression_flags): ...to this.
15944 (UMIPS_12BIT_OFFSET_P): New macro.
15945 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
15946 (MULTILIB_DIRNAMES): Likewise.
15948 2013-02-24 Jakub Jelinek <jakub@redhat.com>
15951 * target-globals.c (save_target_globals): For init_reg_sets and
15952 target_reinit remporarily set this_fn_optabs to this_target_optabs.
15954 2013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
15956 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
15957 * config/aarch64/t-aarch64
15958 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
15960 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
15962 PR inline-asm/56148
15963 * lra-constraints.c (process_alt_operands): Reload operand
15964 conflicting with earlier clobber only if no more other conflicting
15967 2013-02-22 Jakub Jelinek <jakub@redhat.com>
15970 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
15971 if not linking a shared library.
15973 2013-02-22 Seth LaForge <sethml@google.com>
15975 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
15977 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
15979 * config/arm/arm.md (split for extendsidi): Update condition.
15980 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
15981 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
15982 (qhs_zextenddi_cstr): Likewise.
15984 2013-02-21 Jakub Jelinek <jakub@redhat.com>
15986 PR middle-end/56420
15987 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
15988 avoid signed wrapping.
15989 (expand_mult): Handle properly multiplication by
15990 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
15991 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
15992 in the compiler if coeff is HOST_WIDE_INT_MIN.
15993 (expand_divmod): Don't make ext_op1 static, change it's type to
15994 uhwi. Avoid undefined behavior in -INTVAL (op1).
15996 PR rtl-optimization/50339
15997 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
15999 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
16000 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
16001 into splitting_ashiftrt field.
16002 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
16004 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
16007 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
16009 PR middle-end/56108
16010 * trans-mem.c (execute_tm_mark): Do not expand transactions that
16011 are sure to go irrevocable.
16013 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
16015 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
16016 scalars are valid operands.
16018 2013-02-21 Martin Jambor <mjambor@suse.cz>
16020 PR tree-optimization/56310
16021 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
16022 only matching indices and non-negative final offsets.
16023 (intersect_aggregates_with_edge): Pass src_idx to
16024 agg_replacements_to_vector. Pass src_idx insstead of index to
16025 intersect_with_agg_replacements.
16027 2013-02-21 Martin Jambor <mjambor@suse.cz>
16029 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
16030 instead of hard-wired defaults.
16032 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
16034 * doc/invoke.texi (MIPS Options): Update documentation of the
16035 floating-point multiply-accumulate instruction restrictions.
16037 2013-02-21 Kostya Serebryany <kcc@google.com>
16039 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
16040 asan_shadow_offset on x86_64 linux.
16042 2013-02-21 Richard Biener <rguenther@suse.de>
16044 PR tree-optimization/56415
16046 2013-02-11 Richard Biener <rguenther@suse.de>
16048 PR tree-optimization/56273
16049 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
16052 2013-02-21 Jakub Jelinek <jakub@redhat.com>
16055 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
16058 PR inline-asm/56405
16059 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
16060 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
16062 2013-02-20 Jan Hubicka <jh@suse.cz>
16064 PR tree-optimization/56265
16065 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
16066 when target is referenced for first time.
16068 2013-02-20 Richard Biener <rguenther@suse.de>
16070 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
16071 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
16072 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
16073 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
16074 not return anything.
16075 (rename_ssa_copies): Do not remove unused locals.
16076 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
16077 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
16078 * passes.c (execute_function_todo): Do not schedule unused locals
16079 removal if cleanup_tree_cfg did something.
16080 * tree-ssa-live.c (remove_unused_locals): Dump statistics
16081 about the number of removed locals.
16083 2013-02-20 Richard Biener <rguenther@suse.de>
16085 PR tree-optimization/56398
16086 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
16088 2013-02-20 Martin Jambor <mjambor@suse.cz>
16090 PR tree-optimization/55334
16091 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
16092 restricted pointers to arrays.
16094 2013-02-20 Richard Biener <rguenther@suse.de>
16095 Jakub Jelinek <jakub@redhat.com>
16097 PR tree-optimization/56396
16098 * tree-ssa-ccp.c (n_const_val): New static variable.
16099 (get_value): Return NULL for SSA names we don't have a lattice
16101 (ccp_initialize): Initialize n_const_val.
16102 * tree-ssa-copy.c (n_copy_of): New static variable.
16103 (init_copy_prop): Initialize n_copy_of.
16104 (get_value): Return NULL_TREE for SSA names we don't have a
16107 2013-02-20 Martin Jambor <mjambor@suse.cz>
16109 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
16111 2013-02-20 Richard Biener <rguenther@suse.de>
16113 * genpreds.c (write_lookup_constraint): Do not compare first
16114 letter of the constraint again.
16116 2013-02-20 Richard Biener <rguenther@suse.de>
16118 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
16120 (get_use_iv_cost): Terminate hashtable walk when coming across
16123 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
16125 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
16126 reassociation for avx2 targets.
16128 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
16130 * config/microblaze/microblaze.c: microblaze_has_clz = 0
16131 Add version check for v8.10.a to enable microblaze_has_clz
16132 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
16133 version and TARGET_PATTERN_COMPARE check
16134 * config/microblaze/microblaze.md: New clzsi2 instruction
16136 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
16138 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
16139 function before branching.
16141 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
16143 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
16145 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
16147 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
16149 PR middle-end/55889
16150 * sel-sched.c: Include ira.h.
16151 (implicit_clobber_conflict_p): New function.
16152 (moveup_expr): Use it.
16153 * Makefile.in (sel-sched.o): Depend on ira.h.
16155 2013-02-19 Richard Biener <rguenther@suse.de>
16157 PR tree-optimization/56384
16158 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
16159 (vn_hash_type): Split out from ...
16160 (vn_hash_constant_with_type): ... here.
16161 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
16162 (vn_phi_eq): Compare types from vn_phi_s structure.
16163 (vn_phi_lookup): Populate vn_phi_s type.
16164 (vn_phi_insert): Likewise.
16166 2013-02-19 Jakub Jelinek <jakub@redhat.com>
16168 PR tree-optimization/56350
16169 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
16170 if haven't found reduction or nested cycle operand, rather than
16171 asserting we must find it.
16173 PR tree-optimization/56381
16174 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
16177 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
16178 Jakub Jelinek <jakub@redhat.com>
16181 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
16182 (swap_optab_enable): Same.
16183 (init_all_optabs): Use argument instead of global.
16184 * tree.h (struct tree_optimization_option): New field target_optabs.
16185 * expr.h (init_all_optabs): Add argument to prototype.
16186 (TREE_OPTIMIZATION_OPTABS): New.
16187 (save_optabs_if_changed): Protoize.
16188 * optabs.h: Declare this_fn_optabs.
16189 * optabs.c (save_optabs_if_changed): New.
16190 Declare this_fn_optabs.
16191 (init_optabs): Add argument to init_all_optabs() call.
16192 * function.c (invoke_set_current_function_hook): Handle per
16194 * function.h (struct function): New field optabs.
16195 * config/mips/mips.c (mips_set_mips16_mode): Handle when
16196 optimization_current_node has changed.
16197 * target-globals.h (save_target_globals_default_opts): Protoize.
16198 * target-globals.c (save_target_globals_default_opts): New.
16200 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
16203 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
16204 registers %fr12 and %fr12R as call used.
16207 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
16208 and HImode, require all displacements to be an integer multiple of
16210 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
16211 only allow QImode and HImode when reload is in progress and strict is
16212 true. Likewise for symbolic addresses. Use base14_operand to check
16213 displacements in REG+BASE addresses.
16215 2013-02-18 Richard Biener <rguenther@suse.de>
16217 PR tree-optimization/56366
16218 * tree-vect-loop.c (get_initial_def_for_induction): Properly
16219 handle sign-conversion of outer-loop initial induction value.
16221 2013-02-18 Richard Biener <rguenther@suse.de>
16223 PR middle-end/56349
16224 * cfghooks.c (merge_blocks): If we merge a latch into another
16225 block adjust references to it.
16226 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
16227 (verify_loop_structure): Verify that a recorded latch is in fact
16230 2013-02-18 Richard Biener <rguenther@suse.de>
16232 PR tree-optimization/56321
16233 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
16234 order SSA name release and virtual operand unlinking.
16236 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
16238 * config/microblaze/microblaze.md (save_stack_block): Define.
16239 (restore_stack_block): Likewise.
16241 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
16243 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
16244 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
16245 * config/microblaze/microblaze.c (microblaze_option_override):
16246 Bail out early for PIC modes when target does not support PIC.
16248 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
16250 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
16251 Replace with a microblaze version.
16252 (microblaze_trampoline_init): Adapt for microblaze.
16253 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
16256 2013-02-16 Jakub Jelinek <jakub@redhat.com>
16257 Dodji Seketeli <dodji@redhat.com>
16260 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
16261 (instrument_mem_region_access): Do not forget to always put
16262 instrumentation of the of 'base' and 'base + len' in a "if (len !=
16263 0) statement, even for cases where either 'base' or 'base + len'
16264 are not instrumented -- because they have been previously
16265 instrumented. Simplify the logic by putting all the statements
16266 instrument 'base + len' inside a sequence, and then insert that
16267 sequence right before the current insertion point. Then, to
16268 instrument 'base + len', just get an iterator on that statement.
16269 And do not forget to update the pointer to iterator the function
16270 received as argument.
16272 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
16274 PR rtl-optimization/56348
16275 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
16277 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
16279 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
16280 (clean_graph_dump_file): Pass base to start_graph_dump.
16282 2013-02-14 Richard Henderson <rth@redhat.com>
16285 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
16287 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
16289 * collect2-aix.h: Define F_LOADONLY.
16291 2013-02-14 Richard Biener <rguenther@suse.de>
16294 * varasm.c (output_constant_def_1): Get the decl representing
16295 the constant as argument.
16296 (output_constant_def): Wrap output_constant_def_1.
16297 (make_decl_rtl): Use output_constant_def_1 with the decl
16298 representing the constant.
16299 (build_constant_desc): Optionally re-use a decl already
16300 representing the constant.
16301 (tree_output_constant_def): Adjust.
16303 2013-02-14 Dodji Seketeli <dodji@redhat.com>
16306 * asan.c (instrument_builtin_call): Really put the length of the
16307 second source argument into src1_len.
16309 2013-02-13 Jakub Jelinek <jakub@redhat.com>
16311 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
16312 argument. If it is false, don't create edge from then_bb to
16314 (insert_if_then_before_iter): Pass true to it.
16315 (build_check_stmt): Pass false to it.
16316 (transform_statements): Flush hash table only on extended basic
16317 block boundaries, rather than at the beginning of every bb.
16318 Don't flush hash table on nonfreeing_call_p calls.
16319 * tree-flow.h (nonfreeing_call_p): New prototype.
16320 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
16322 2013-02-13 David S. Miller <davem@davemloft.net>
16324 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
16326 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
16329 * ira.c (max_regno_before_ira): Move from ...
16331 (fix_reg_equiv_init): Use max_regno_before_ira instead of
16334 2013-02-13 Jakub Jelinek <jakub@redhat.com>
16336 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
16338 2013-02-13 Richard Biener <rguenther@suse.de>
16341 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
16342 globals in MEM_REFs.
16344 2013-02-13 Richard Biener <rguenther@suse.de>
16346 * loop-init.c (loop_optimizer_init): Clear loop state when
16347 re-initializing preserved loops.
16348 * loop-unswitch.c (unswitch_single_loop): Return whether
16349 we unswitched the loop. Do not verify loop state here.
16350 (unswitch_loops): When we unswitched a loop discover new loops.
16352 2013-02-13 Kostya Serebryany <kcc@google.com>
16354 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
16356 * sanitizer.def: Rename __asan_init to __asan_init_v1.
16358 2013-02-12 Dodji Seketeli <dodji@redhat.com>
16360 Avoid instrumenting duplicated memory access in the same basic block
16361 * Makefile.in (asan.o): Add new dependency on hash-table.h
16362 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
16363 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
16364 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
16365 (free_mem_ref_resources, has_mem_ref_been_instrumented)
16366 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
16367 (get_mem_ref_of_assignment): New functions.
16368 (get_mem_refs_of_builtin_call): Extract from
16369 instrument_builtin_call and tweak a little bit to make it fit with
16371 (instrument_builtin_call): Use the new
16372 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
16373 of is_gimple_builtin_call.
16374 (instrument_derefs, instrument_mem_region_access): Insert the
16375 instrumented memory reference into the hash table.
16376 (maybe_instrument_assignment): Renamed instrument_assignment into
16377 this, and change it to advance the iterator when instrumentation
16378 actually happened and return true in that case. This makes it
16379 homogeneous with maybe_instrument_assignment, and thus give a
16380 chance to callers to be more 'regular'.
16381 (transform_statements): Clear the memory reference hash table
16382 whenever we enter a new BB, when we cross a function call, or when
16383 we are done transforming statements. Use
16384 maybe_instrument_assignment instead of instrumentation. No more
16385 need to special case maybe_instrument_assignment and advance the
16386 iterator after calling it; it's now handled just like
16387 maybe_instrument_call. Update comment.
16389 2013-02-13 Richard Biener <rguenther@suse.de>
16391 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
16392 Fix loop discovery code.
16394 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
16396 PR inline-asm/56148
16397 * lra-constraints.c (process_alt_operands): Match early clobber
16398 operand with itself. Check conflicts with earlyclobber only if
16399 the operand is not reloaded. Prefer to reload conflicting operand
16400 if earlyclobber and matching operands are the same.
16402 2013-02-12 Richard Biener <rguenther@suse.de>
16405 * lto-streamer-out.c (write_symbol): Do not output symbols
16406 for hard register variables.
16408 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
16411 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
16412 (umulsidi3_insn, mulsidi3_insn): New insns.
16414 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
16416 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
16417 (struct tune_params): Add vec_costs field.
16418 * config/arm/arm.c (arm_builtin_vectorization_cost)
16419 (arm_add_stmt_cost): New functions.
16420 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
16421 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
16422 (arm_default_vec_cost): New struct of type cpu_vec_costs.
16423 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
16424 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
16425 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
16426 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
16428 2013-02-12 Richard Biener <rguenther@suse.de>
16431 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
16432 decls again if possible.
16434 2013-02-12 Richard Biener <rguenther@suse.de>
16436 PR middle-end/56288
16437 * tree-ssa.c (verify_ssa_name): Fix check, move
16438 SSA_NAME_IN_FREE_LIST check up.
16440 2013-02-12 Jakub Jelinek <jakub@redhat.com>
16441 Steven Bosscher <steven@gcc.gnu.org>
16443 PR rtl-optimization/56151
16444 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
16445 equal to op0 or op1, and last_insn pattern is CODE operation
16446 with MEM dest and one of the operands matches that MEM.
16448 2013-02-11 Sriraman Tallam <tmsriramgoogle.com>
16450 * doc/extend.texi: Document Function Multiversioning and "default"
16451 parameter string to target attribute.
16452 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
16453 target attribute parameter is "default".
16454 (ix86_compare_version_priority): Remove checks for target attribute.
16455 (ix86_mangle_function_version_assembler_name): Change error to sorry.
16456 Remove check for target attribute equal to NULL. Add assert.
16457 (ix86_generate_version_dispatcher_body): Change error to sorry.
16459 2013-02-11 Iain Sandoe <iain@codesourcery.com>
16460 Jack Howarth <howarth@bromo.med.uc.edu>
16461 Patrick Marlier <patrick.marlier@gmail.com>
16464 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
16465 define ENDFILE_SPEC as TM_DESTRUCTOR.
16466 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
16468 2013-02-11 Alexander Potapenko <glider@google.com>
16469 Jack Howarth <howarth@bromo.med.uc.edu>
16470 Jakub Jelinek <jakub@redhat.com>
16473 * config/darwin.c (cdtor_record): Rename ctor_record.
16474 (sort_cdtor_records): Rename sort_ctor_records.
16475 (finalize_dtors): New routine to sort destructors by
16476 priority before use in assemble_integer.
16477 (machopic_asm_out_destructor): Use finalize_dtors if needed.
16479 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
16481 PR rtl-optimization/56275
16482 * simplify-rtx.c (avoid_constant_pool_reference): Check that
16483 offset is non-negative and less than cmode size before
16484 calling simplify_subreg.
16486 2013-02-11 Richard Biener <rguenther@suse.de>
16488 PR tree-optimization/56264
16489 * cfgloop.h (fix_loop_structure): Adjust prototype.
16490 * loop-init.c (fix_loop_structure): Return the number of
16491 newly discovered loops.
16492 * tree-cfgcleanup.c (repair_loop_structures): When new loops
16493 are discovered, do a full loop-closed SSA rewrite.
16495 2013-02-11 Richard Biener <rguenther@suse.de>
16497 PR tree-optimization/56273
16498 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
16500 (check_array_ref): Fix missing newline in dumps.
16501 (search_for_addr_array): Likewise.
16503 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
16505 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
16507 2013-02-09 Jakub Jelinek <jakub@redhat.com>
16510 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
16512 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
16514 PR rtl-optimization/56246
16515 * lra-constraints.c (simplify_operand_subreg): Try to reuse
16517 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
16518 constraints are satisfied.
16520 2013-02-08 Jeff Law <law@redhat.com>
16523 * emit-rtl.c (reg_is_parm_p): New function.
16524 * regs.h (reg_is_parm_p): New prototype.
16525 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
16526 callee-clobbered registers.
16528 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
16531 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
16532 If there is no implicit builtin declaration, just return NULL.
16534 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
16536 * config/i386/sse.md (FMAMODEM): New mode iterator.
16537 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
16538 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
16540 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
16542 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
16543 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
16544 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
16546 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
16548 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
16549 (microblaze*-*-elf): Likewise.
16550 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
16552 * config/microblaze/microblaze-c.c: Add builtin defines for
16553 _LITTLE_ENDIAN and _BIG_ENDIAN.
16554 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
16555 add to TARGET_DEFAULT flags.
16556 Expand ASM_SPEC and LINK_SPEC.
16557 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
16558 * config/microblaze/microblaze.md: Update extendsidi2 and
16559 movdi_internal instructions to use low-order / high-order reg
16561 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
16562 options and inversemask / mask of LITTLE_ENDIAN.
16563 * config/microblaze/t-microblaze: Expand multilib options to
16564 include mlittle-endian (le) and update exceptions patterns.
16566 2013-02-08 Jakub Jelinek <jakub@redhat.com>
16568 PR rtl-optimization/56195
16569 * lra-constraints.c (get_reload_reg): Don't reuse regs
16570 if they have smaller mode than requested, if they have
16571 wider mode than requested, try to return a SUBREG.
16573 PR tree-optimization/56250
16574 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
16575 if type is unsigned and code isn't MULT_EXPR.
16577 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
16579 PR tree-optimization/56064
16580 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
16581 bits according to mode.
16582 * fixed-value.h (fixed_from_double_int)
16583 (const_fixed_from_double_int): Adjust comments.
16585 2013-02-08 Richard Biener <rguenther@suse.de>
16588 * lto-streamer.h (struct data_in): Remove current_file, current_line
16589 and current_col members.
16590 * lto-streamer-out.c (lto_output_location): Stream changed bits
16591 en-block for efficiency.
16592 * lto-streamer-in.c (clear_line_info): Remove.
16593 (lto_input_location): Cache current file, line and column
16594 globally via local statics. Read changed bits en-block.
16595 (input_function): Do not call clear_line_info.
16596 (lto_read_body): Likewise.
16597 (lto_input_toplevel_asms): Likewise.
16599 2013-02-08 Michael Matz <matz@suse.de>
16601 PR tree-optimization/52448
16602 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
16603 (nt_call_phase): New static.
16604 (add_or_mark_expr): Only mark accesses with newer phase than any
16606 (nonfreeing_call_p): New.
16607 (nt_init_block): Update nt_call_phase, mark blocks as visited.
16608 (nt_fini_block): Keep blocks marked as visited.
16609 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
16611 2013-02-08 Richard Biener <rguenther@suse.de>
16613 * ira.c (ira): Free broken dominator information.
16615 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
16617 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
16619 2013-02-08 Marek Polacek <polacek@redhat.com>
16621 * cfgloop.c (verify_loop_structure): Add more checking of headers.
16623 2013-02-08 Richard Biener <rguenther@suse.de>
16625 PR middle-end/56181
16626 * cfgloop.h (flow_loops_find): Adjust.
16627 (bb_loop_header_p): Declare.
16628 * cfgloop.c (bb_loop_header_p): New function split out from ...
16629 (flow_loops_find): ... here. Adjust function signature,
16630 support incremental loop structure update.
16631 (verify_loop_structure): Cleanup. Verify a loop is a loop.
16632 * cfgloopmanip.c (fix_loop_structure): Move ...
16633 * loop-init.c (fix_loop_structure): ... here.
16634 (apply_loop_flags): Split out from ...
16635 (loop_optimizer_init): ... here.
16636 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
16637 in incremental mode, only remove dead loops here.
16639 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
16642 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
16643 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
16644 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
16645 (*round<mode>3.libgcc): New insns for fixed-modes.
16646 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
16647 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
16648 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
16649 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
16650 implementations. Define to __builtin_avr_absFX,
16651 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
16652 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
16653 __builtin_avr_countlsFX, respectively.
16654 * config/avr/avr-c.c (target.h): Include it.
16655 (enum avr_builtin_id): New enum.
16656 (avr_resolve_overloaded_builtin): New static function.
16657 (avr_register_target_pragmas): Use it to set
16658 targetm.resolve_overloaded_builtin.
16659 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
16660 tree nodes used by DEF_BUILTIN.
16661 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
16662 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
16663 <AVR_BUILTIN_xxBITS>: Same.
16665 2013-02-08 Richard Biener <rguenther@suse.de>
16667 * cfgloop.c (verify_loop_structure): Properly handle
16668 a loop exiting to another loop header.
16669 * ira-int.h (ira_loops): Remove.
16670 * ira.c (ira_loops): Remove.
16671 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
16672 (do_reload): Use loop_optimizer_finalize.
16673 * ira-build.c (create_loop_tree_nodes): Use get_loops and
16674 number_of_loops to access the loop tree.
16675 (more_one_region_p): Likewise.
16676 (finish_loop_tree_nodes): Likewise.
16677 (rebuild_regno_allocno_maps): Likewise.
16678 (mark_loops_for_removal): Likewise.
16679 (mark_all_loops_for_removal): Likewise.
16680 (remove_unnecessary_regions): Likewise.
16681 (ira_build): Likewise.
16682 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
16684 2013-02-08 Richard Biener <rguenther@suse.de>
16686 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
16687 * ipa-pure-const.c (analyze_function): Avoid calling
16688 mark_irreducible_loops twice.
16689 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
16691 2013-02-07 David S. Miller <davem@davemloft.net>
16693 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
16695 * var-tracking.c (vt_add_function_parameter): Test the presence of
16696 HAVE_window_save properly and do not remap argument registers when
16697 we have a leaf function.
16699 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
16702 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
16704 * config/i386/i386.c (ix86_print_operand): Ditto.
16706 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
16708 * lra-constraints.c (process_alt_operands): Fix recently added comment.
16710 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
16712 PR rtl-optimization/56225
16713 * lra-constraints.c (process_alt_operands): Check that reload hard
16714 reg can hold value for strict_low_part.
16716 2013-02-07 Jakub Jelinek <jakub@redhat.com>
16719 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
16720 dwarf2out_end_function.
16721 (in_first_function_p, maybe_at_text_label_p,
16722 first_loclabel_num_not_at_text_label): New variables.
16723 (dwarf2out_var_location): In the first function find out
16724 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
16725 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
16728 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
16730 PR rtl-optimization/56178
16731 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
16732 SUBREG of a register. Tidy up related block of code.
16733 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
16734 note if the source is a register or a SUBREG of a register.
16736 2013-02-07 Jakub Jelinek <jakub@redhat.com>
16739 * config/rs6000/rs6000.md (ptrm): New mode attr.
16740 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
16741 call_value_indirect_aix<pttrsize>,
16742 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
16745 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
16747 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
16748 if -bnortl. Convert to strcmp and strncmp.
16750 2013-02-07 Alan Modra <amodra@gmail.com>
16753 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
16754 addresses won't wrap when offsetting.
16755 (rs6000_secondary_reload): Provide secondary reloads needed for
16756 wrapping LO_SUM addresses.
16758 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
16760 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
16761 MACH, just __MACH__.
16763 2013-02-06 Richard Biener <rguenther@suse.de>
16765 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
16766 instead of calling fix_loop_structure.
16768 2013-02-06 Jakub Jelinek <jakub@redhat.com>
16770 PR middle-end/56217
16771 * omp-low.c (use_pointer_for_field): Return false if
16772 lower_send_shared_vars doesn't generate any copy-out code.
16774 2013-02-06 Tom de Vries <tom@codesourcery.com>
16776 PR rtl-optimization/56131
16777 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
16778 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
16779 of the label is NULL. Add comment.
16781 2013-02-05 Jakub Jelinek <jakub@redhat.com>
16783 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
16786 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
16787 (STATIC_LIBTSAN_LIBS): Likewise.
16788 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
16789 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
16790 is defined, don't add anything else beyond that.
16791 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
16792 (LINK_COMMAND_SPEC): Use them.
16794 PR tree-optimization/56205
16795 * tree-stdarg.c (check_all_va_list_escapes): Return true if
16796 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
16797 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
16799 2013-02-05 Richard Biener <rguenther@suse.de>
16801 PR tree-optimization/53342
16802 PR tree-optimization/53185
16803 * tree-vectorizer.h (vect_check_strided_load): Remove.
16804 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
16805 not disallow peeling for vectorized strided loads.
16806 (vect_check_strided_load): Make static and simplify.
16807 (vect_analyze_data_refs): Adjust.
16808 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
16809 correctly when vectorizing strided loads.
16811 2013-02-05 Richard Biener <rguenther@suse.de>
16813 * doc/install.texi: Refer to ISL, not PPL.
16815 2013-02-05 Jan Hubicka <jh@suse.cz>
16817 PR tree-optimization/55789
16818 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
16820 2013-02-05 Jan Hubicka <jh@suse.cz>
16822 PR tree-optimization/55789
16823 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
16824 the dead call anyway.
16826 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
16829 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
16831 2013-02-04 Alexander Potapenko <glider@google.com>
16832 Jack Howarth <howarth@bromo.med.uc.edu>
16833 Jakub Jelinek <jakub@redhat.com>
16836 * config/darwin.c (sort_ctor_records): Stabilized qsort
16837 on constructor priority by using original position.
16838 (finalize_ctors): New routine to sort constructors by
16839 priority before use in assemble_integer.
16840 (machopic_asm_out_constructor): Use finalize_ctors if needed.
16842 2013-02-04 Jakub Jelinek <jakub@redhat.com>
16845 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
16846 about visibility on artificial decls.
16847 * config/sol2.c (solaris_assemble_visibility): Likewise.
16849 2013-02-04 Kai Tietz <ktietz@redhat.com>
16852 * config/i386/i386.c (function_value_ms_64): Add additional valtype
16853 argument and improve checking of return-argument types for 16-byte
16855 (ix86_function_value_1): Add additional valtype argument on call
16856 of function_value_64.
16857 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
16858 handling infunction_value_64 function.
16860 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
16862 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
16864 2013-02-04 Richard Biener <rguenther@suse.de>
16866 PR tree-optimization/56188
16867 * tree-ssa-structalias.c (label_visit): Consider case with
16868 initially non-empty points-to set.
16869 (perform_var_substitution): Dump node mapping and clean up.
16871 2013-02-04 Richard Guenther <rguenther@suse.de>
16874 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
16875 node prevail as last resort.
16876 (lto_symtab_merge_decls): Remove guard on LTRANS here.
16877 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
16879 2013-02-04 Richard Biener <rguenther@suse.de>
16881 PR tree-optimization/56113
16882 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
16884 (equiv_class_lookup_or_add): ... this.
16885 (label_visit): Adjust and fix error in previous patch.
16886 (perform_var_substitution): Adjust.
16888 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
16890 * config/sh/divtab.c: Fix formatting and comments throughout the file.
16891 * config/sh/sh4-300.md: Likewise.
16892 * config/sh/sh4a.md: Likewise.
16893 * config/sh/constraints.md: Likewise.
16894 * config/sh/sh.md: Likewise.
16895 * config/sh/netbsd-elf.h: Likewise.
16896 * config/sh/predicates.md: Likewise.
16897 * config/sh/sh-protos.h: Likewise.
16898 * config/sh/ushmedia.h: Likewise.
16899 * config/sh/linux.h: Likewise.
16900 * config/sh/sh.c: Likewise.
16901 * config/sh/superh.h: Likewise.
16902 * config/sh/elf.h: Likewise.
16903 * config/sh/sh4.md: Likewise.
16904 * config/sh/sh.h: Likewise.
16906 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
16908 * config/pa/constraints.md: Adjust unused letters. Change "T"
16909 constraint to match_test floating_point_store_memory_operand().
16910 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
16911 (base14_operand): New.
16912 (floating_point_store_memory_operand): New.
16913 (integer_store_memory_operand): Revise to use base14_operand and
16914 reg_plus_base_memory_operand.
16915 (move_dest_operand): Allow symbolic_memory_operands.
16916 (symbolic_memory_operand): Check for LO_SOM.
16917 (symbolic_operand): Change default case to break.
16918 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
16919 CONST_DOUBLE values to be reloaded by putting them into memory when
16920 the destination is a floating point register.
16921 (movdf): Remove code to handle CONST_DOUBLE.
16923 (reload_indf_r1): New.
16924 (reload_insf_r1): New.
16925 Consistently use "Q" and "T" constraints with integer and floating
16926 point move instructions, respectively.
16927 (movdi): Remove FAIL.
16928 Change predicate for source operand unamed DImode move from
16929 general_operand to move_src_operand.
16930 (umulsidi3): Change predicate for destination operand to
16932 Likewise for similar unamed patterns.
16933 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
16934 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
16935 (hppa_legitimize_address): Simplify mask calculation.
16936 (pa_emit_move_sequence): Revised handling of secondary reloads from
16937 REG+D addresses for floating point loads and stores. Directly handle
16938 loading CONST0_RTX (mode) to a floating point register.
16939 (pa_secondary_reload): Handle reloading DF and SFmode constant values
16940 to floating point registers. Don't restrict secondary reloads to
16941 floating point registers to integer modes. Revise some comments and
16943 (TARGET_LEGITIMATE_ADDRESS_P): Define.
16944 (pa_legitimate_address_p): New.
16945 (pa_legitimize_reload_address): New.
16946 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
16947 (STRICT_REG_OK_FOR_BASE_P): New.
16948 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
16949 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
16951 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
16952 Andrew Dixie <andrewd@gentrack.com>
16954 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
16957 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
16959 * expmed.c (extract_bit_field_1): Pass the full width of the
16960 structure to get_best_reg_extraction_insn.
16962 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
16965 * configure.ac (use_cxa_atexit): Add AIX.
16966 * configure: Regenerate.
16968 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
16970 2013-02-01 Jakub Jelinek <jakub@redhat.com>
16973 * final.c (need_profile_function): New variable.
16974 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
16975 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
16976 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
16977 notes, targetm.asm_out.function_prologue doesn't emit anything,
16978 HAVE_prologue and profiler should be emitted before prologue,
16979 set need_profile_function instead of emitting it.
16980 (final_scan_insn): If need_profile_function, emit
16981 profile_function on the first NOTE_INSN_BASIC_BLOCK or
16982 NOTE_INSN_FUNCTION_BEG note.
16984 2013-02-01 Richard Henderson <rth@redhat.com>
16986 * config/rs6000/rs6000.md (smulditi3): New.
16989 * config/alpha/alpha.md (umulditi3): New.
16991 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
16993 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
16994 (ASM_OUTPUT_ALIGNED_LOCAL): New.
16996 2013-02-01 Richard Biener <rguenther@suse.de>
16998 PR tree-optimization/56113
16999 * tree-ssa-structalias.c (label_visit): Reduce work for
17000 single-predecessor nodes.
17002 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
17004 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
17005 range isn't testing for zero.
17007 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
17009 PR middle-end/56113
17010 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
17012 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
17013 Nick Clifton <nickc@redhat.com>
17015 * config/v850/constraints.md (Q): Define as a memory constraint.
17016 * config/v850/predicates.md (label_ref_operand): New predicate.
17017 (e3v5_shift_operand): New predicate.
17018 (ior_operator): New predicate.
17019 * config/v850/t-v850: Add e3v5 multilib.
17020 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
17021 (v850_gen_movdi): Prototype.
17022 * config/v850/v850.c: Add support for e3v5 architecture.
17023 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
17025 (construct_save_jarl): Add e3v5 long JARL support.
17026 (v850_adjust_insn_length): New function. Adjust length of call
17027 insns when using e3v5 instructions.
17028 (v850_gen_movdi): New function: Generate instructions to move a
17030 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
17031 (CPP_SPEC): Define __v850e3v5__ as appropriate.
17032 (TARGET_USE_FPU): Enable for e3v5.
17033 (CONST_OK_FOR_W): New macro.
17034 (ADJUST_INSN_LENGTH): Define.
17035 * config/v850/v850.md (UNSPEC_LOOP): Define.
17036 (attr cpu): Add v850e3v5.
17037 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
17038 (movdi): New pattern.
17039 (movdi_internal): New pattern.
17040 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
17041 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
17042 (cstoresf4): Likewise.
17043 (cstoredf4): Likewise.
17044 (insv): New pattern.
17045 (rotlso3_a): New pattern.
17046 (rotlsi3_b): New pattern
17047 (rotlsi3_v850e3v5): New pattern.
17048 (doloop_begin): New pattern.
17049 (fix_loop_counter): New pattern.
17050 (doloop_end): New pattern.
17051 (branch_normal): Add e3v5 long branch support.
17052 (branch_invert): Likewise.
17053 (branch_z_normal): Likewise.
17054 (branch_z_invert): Likewise.
17055 (branch_nz_normal): Likewise.
17056 (branch_nz_invert): Likewise.
17057 (call_internal_short): Add e3v5 register-indirect JARL support.
17058 (call_internal_long): Likewise.
17059 (call_value_internal_short): Likewise.
17060 (call_value_internal_long): Likewise.
17061 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
17062 (mloop): New option.
17063 * config.gcc: Add support for configuring v840e3v5 target.
17064 * doc/invoke.texi: Document new v850 specific command line options.
17066 2013-01-31 Paul Koning <ni1d@arrl.net>
17070 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
17071 children if parent is a class.
17072 (prune_unused_types_prune): Don't add DW_AT_declaration.
17074 2013-01-31 Richard Biener <rguenther@suse.de>
17076 PR tree-optimization/56157
17077 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
17078 match up operand with SLP child.
17080 2013-01-31 Jason Merrill <jason@redhat.com>
17083 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
17084 parameters the first time.
17085 (gen_scheduled_generic_parms_dies): Check completeness here.
17087 2013-01-31 Richard Biener <rguenther@suse.de>
17089 PR middle-end/53073
17090 * common.opt (faggressive-loop-optimizations): New flag,
17091 enabled by default.
17092 * doc/invoke.texi (faggressive-loop-optimizations): Document.
17093 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
17094 infer_loop_bounds_from_undefined by it.
17096 2013-01-31 Richard Biener <rguenther@suse.de>
17098 PR tree-optimization/56150
17099 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
17100 visit virtual operands.
17101 (find_uses_to_rename_bb): Likewise.
17103 2013-01-31 Richard Biener <rguenther@suse.de>
17105 PR tree-optimization/56150
17106 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
17107 mixed store non-store stmts.
17109 2013-01-30 Jakub Jelinek <jakub@redhat.com>
17112 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
17113 LIBASAN_EARLY_SPEC is defined.
17114 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
17115 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
17117 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
17120 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
17121 invalid args instead of ICEing on it.
17122 (ix86_valid_target_attribute_tree): Return error_mark_node if
17123 ix86_valid_target_attribute_inner_p failed.
17124 (ix86_valid_target_attribute_p): Return false only if
17125 ix86_valid_target_attribute_tree returned error_mark_node. Allow
17126 target("default") attribute.
17127 (sorted_attr_string): Change argument from const char * to tree,
17128 merge in all target attribute arguments rather than just one.
17129 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
17130 instead of free. Avoid using strcat.
17131 (ix86_mangle_function_version_assembler_name): Mangle
17132 target("default") as if no target attribute is present. Adjust
17133 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
17134 instead of xmalloc and XDELETEVEC instead of free.
17135 (ix86_function_versions): Don't return true if one of the decls
17136 doesn't have target attribute. If they don't and one of the decls
17137 is DECL_FUNCTION_VERSIONED, report an error. Adjust
17138 sorted_attr_string caller. Use XDELETEVEC instead of free.
17139 (ix86_supports_function_versions): Remove.
17140 (make_name): Fix up formatting.
17141 (make_dispatcher_decl): Remove resolver_name and its initialization.
17142 Avoid leaking memory.
17143 (is_function_default_version): Return true if there is
17144 target("default") attribute rather than no target attribute at all.
17145 (make_resolver_func): Avoid leaking memory.
17146 (ix86_generate_version_dispatcher_body): Likewise.
17147 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
17148 * target.def (supports_function_versions): Remove.
17149 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
17150 * doc/tm.texi: Regenerated.
17152 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
17154 PR rtl-optimization/56144
17155 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
17156 for values with side effects.
17158 2013-01-30 Richard Biener <rguenther@suse.de>
17160 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
17161 (sparseset_pop): Likewise.
17162 * cfganal.c (compute_idf): Likewise. Increase work-stack size
17163 to be able to use quick_push in the worker loop.
17165 2013-01-30 Marek Polacek <polacek@redhat.com>
17167 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
17169 2013-01-30 Richard Biener <rguenther@suse.de>
17172 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
17174 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
17176 PR tree-optimization/56064
17177 * fixed-value.c (fixed_from_double_int): New function.
17178 * fixed-value.h (fixed_from_double_int): New prototype.
17179 (const_fixed_from_double_int): New static inline function.
17180 * fold-const.c (native_interpret_fixed): New static function.
17181 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
17182 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
17183 (native_encode_fixed): New static function.
17184 (native_encode_expr) <FIXED_CST>: Use it.
17185 (native_interpret_int): Move double_int worker code to...
17186 * double-int.c (double_int::from_buffer): ...this new static method.
17187 * double-int.h (double_int::from_buffer): Prototype it.
17189 2013-01-30 Richard Biener <rguenther@suse.de>
17191 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
17192 New pointer-map and obstack.
17193 (init_alias_vars): Allocate pointer-map and obstack.
17194 (delete_points_to_sets): Free them.
17195 (find_what_var_points_to): Cache result.
17196 (find_what_p_points_to): Adjust for changed interface of
17197 find_what_var_points_to.
17198 (compute_points_to_sets): Likewise.
17199 (ipa_pta_execute): Likewise.
17201 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
17203 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
17204 * configure: Regenerate.
17205 * config.in: Regenerate.
17206 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
17207 #nobits/#progbits if supported.
17209 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
17212 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
17213 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
17214 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
17216 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
17218 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
17219 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
17221 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
17223 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
17224 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
17226 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
17228 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
17230 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
17231 * config/arm/cortex-a7.md: New bypasses using
17232 arm_mac_accumulator_is_result.
17234 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
17236 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
17237 (cortex_a7_neon_mla): Likewise.
17238 (cortex_a7_fpfmad): New reservation.
17239 (cortex_a7_fpmacs): Use ffmas and update required units.
17240 (cortex_a7_fpmuld): Update required units and latency.
17241 (cortex_a7_fpmacd): Likewise.
17242 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
17243 (cortex_a7_neon). Likewise.
17244 (bypass) Update participating units.
17246 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
17248 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
17249 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
17251 * config/arm/vfp11.md (vfp_farith): Use ffmas.
17252 (vfp_fmul): Use ffmad.
17253 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
17254 (cortex_r4_fmacd): Use ffmad.
17255 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
17256 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
17257 (cortex_a9_fmacd): Use ffmad.
17258 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
17259 (cortex_a8_vfp_macd): Use ffmad.
17260 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
17261 (cortex_a5_fpmacd): Use ffmad.
17262 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
17263 (cortex_a15_vfp_macd): Use ffmad.
17264 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
17266 2013-01-29 Jason Merrill <jason@redhat.com>
17269 * varasm.c (default_assemble_visibility): Don't warn about
17270 visibility on artificial decls.
17272 2013-01-29 Richard Biener <rguenther@suse.de>
17274 PR tree-optimization/56113
17275 * tree-ssa-structalias.c (equiv_class_lookup): Also return
17277 (label_visit): Free duplicate bitmaps and record the leader instead.
17278 (perform_var_substitution): Adjust.
17280 2013-01-29 Richard Biener <rguenther@suse.de>
17282 PR tree-optimization/55270
17283 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
17284 the CFG, schedule loops for fixup.
17286 2013-01-29 Nick Clifton <nickc@redhat.com>
17288 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
17291 2013-01-28 Leif Ekblad <leif@rdos.net>
17293 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
17294 * config/i386/i386.h (TARGET_RDOS): New macro.
17295 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
17296 * config/i386/i386.c (ix86_option_override_internal): For 64bit
17297 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
17298 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
17299 DEFAULT_LARGE_SECTION_THRESHOLD.
17300 * config/i386/i386.md (R14_REG, R15_REG): New constants.
17301 * config/i386/rdos.h: New file.
17302 * config/i386/rdos64.h: New file.
17304 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
17307 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
17310 2013-01-28 Jakub Jelinek <jakub@redhat.com>
17312 PR rtl-optimization/56117
17313 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
17314 call cselib_lookup_from_insn on the MEM before calling
17315 add_insn_mem_dependence.
17317 2013-01-28 Richard Biener <rguenther@suse.de>
17319 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
17320 to a stmt that didn't have one.
17321 (copy_phis_for_bb): Likewise for PHI arguments.
17322 (copy_debug_stmt): Likewise for debug stmts.
17324 2013-01-28 Richard Biener <rguenther@suse.de>
17326 PR tree-optimization/56034
17327 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
17328 (partition_builtin_p): Adjust.
17329 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
17330 it is the last partition.
17331 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
17332 up the vertex for the definition.
17333 (classify_partition): Classify whether a partition is a
17334 PKIND_REDUCTION, thus has uses outside of the loop.
17335 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
17336 Merge all PKIND_REDUCTION partitions into the last partition.
17337 (tree_loop_distribution): Seed partitions from reductions as well.
17339 2013-01-28 Jakub Jelinek <jakub@redhat.com>
17341 PR tree-optimization/56125
17342 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
17343 pow(x,c) into sqrt(x) * powi(x, n/2) or
17344 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
17345 optimizing for size.
17346 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
17347 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
17350 PR tree-optimization/56094
17351 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
17352 to UNKNOWN_LOCATION while gimplifying expr.
17354 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
17357 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
17358 operand 0 in movabs insn template for -masm=intel asm alternative.
17359 (*movabs<mode>_2): Ditto for operand 1.
17361 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
17364 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
17367 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
17369 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
17372 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17374 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
17375 Undef to avoid warning.
17377 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
17379 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
17380 * configure: Regenerate.
17382 2013-01-25 Jakub Jelinek <jakub@redhat.com>
17384 PR tree-optimization/56098
17385 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
17386 for stmts with volatile ops.
17387 (cond_store_replacement): Don't optimize if assign has volatile ops.
17388 (cond_if_else_store_replacement_1): Don't optimize if either
17389 then_assign or else_assign have volatile ops.
17390 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
17393 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
17395 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
17397 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
17399 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
17400 missing ':' in asm example.
17402 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
17404 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
17405 entries into lane and laneq entries.
17406 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
17407 Remove AdvSIMD scalar modes.
17408 (aarch64_sq<r>dmulh_laneq<mode>): New.
17409 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
17411 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
17412 builtin implementations to relfect changes in RTL in aarch64-simd.md.
17413 * config/aarch64/iterators.md (VCOND): New.
17416 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
17419 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
17420 Add NULL LIBNAME argument to existing definitions.
17421 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
17422 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
17423 * config/avr/avr.c (DEF_BUILTIN): Same.
17424 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
17425 (avr_expand_builtin): Expand to a vanilla call if a libgcc
17426 implementation is available (DECL_ASSEMBLER_NAME is set).
17427 (avr_fold_absfx): New static function.
17428 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
17429 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
17430 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
17431 AVR_BUILTIN_ABSLLK.
17432 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
17433 (abshk, absk, abslk, absllk): Provide as static inline functions.
17435 2013-01-25 Marek Polacek <polacek@redhat.com>
17437 PR tree-optimization/56035
17438 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
17440 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
17442 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
17443 (*movtf_internal_rex64): Add (!o,C) alternative
17444 (*movxf_internal_rex64): Ditto.
17445 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
17447 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
17449 * doc/invoke.texi: fix typo.
17450 * doc/objc.texi: fix typo.
17452 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
17454 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
17455 for the first two alternatives.
17457 2013-01-24 Diego Novillo <dnovillo@google.com>
17459 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
17460 (ggc-zone.o): Remove.
17461 * configure.ac: Remove option --with-gc.
17462 * configure: Re-generate.
17463 * doc/install.texi: Remove documentation for --with-gc.
17464 * gengtype.c (write_enum_defn): Remove. Update all users.
17465 (write_Types_process_field): Remove generation of gt_e_* argument.
17466 (output_type_enum): Remove. Update all users.
17467 (write_enum_defn): Remove. Update all users.
17468 (enum alloc_zone): Remove. Update all users.
17469 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
17470 * ggc-common.c (ggc_splay_alloc): Remove first argument.
17471 Update all callers.
17472 (struct ptr_data): Remove field TYPE. Update all users.
17473 (gt_pch_note_object): Remove argument TYPE. Update all users.
17474 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
17476 * ggc-none.c (ggc_alloc_typed_stat): Remove.
17477 (struct alloc_zone): Remove.
17478 (ggc_internal_alloc_zone_stat): Remove.
17479 (ggc_internal_cleared_alloc_zone_stat): Remove.
17480 * ggc-page.c (ggc_alloc_typed_stat): Remove.
17481 (ggc_pch_count_object): Remove last argument. Update all users.
17482 (ggc_pch_alloc_object): Remove last argument. Update all users.
17483 (struct alloc_zone): Remove.
17484 * ggc-zone.c: Remove.
17485 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
17486 (struct alloc_zone): Remove.
17487 (ggc_alloc_typed_stat): Remove.
17488 (ggc_alloc_typed): Remove.
17489 (ggc_splay_alloc): Remove first argument.
17490 (rtl_zone): Remove. Update all users.
17491 (tree_zone): Remove. Update all users.
17492 (tree_id_zone): Remove. Update all users.
17493 (ggc_internal_zone_alloc_stat): Remove. Update all users.
17494 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
17495 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
17496 * tree-ssanames.c: Remove references to zone allocator in comments.
17498 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
17500 * config/avr/avr.c (avr_out_fract): Make register numbers that
17501 might be outside of source operand signed.
17503 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
17505 * config/i386/constraints.md (Yf): New constraint.
17506 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
17507 of f constraint to conditionaly disable x87 register preferences.
17508 (*movdf_internal): Ditto.
17509 (*movsf_internal): Ditto.
17511 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
17513 PR inline-asm/55934
17514 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
17515 that have operands with impossible constraints.
17516 Add a FIXME for a speed-up opportunity.
17517 * lra-constraints.c (process_alt_operands): Verify that a class
17518 selected from constraints on asms is valid for the operand mode.
17519 (curr_insn_transform): Remove incorrect comment.
17521 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
17523 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
17524 TOC operand is a valid symbol ref in the constant pool.
17526 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
17528 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
17530 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
17533 * config/avr/stdfix.h: New file.
17534 * t-avr (stdfix-gcc.h): New rule to build it.
17535 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
17537 2013-01-23 Kostya Serebryany <kcc@google.com>
17539 * config/darwin.h: remove dependency on
17540 CoreFoundation (asan on Mac OS).
17542 2013-01-23 Jakub Jelinek <jakub@redhat.com>
17545 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
17546 instead of cmpdi_operand for first comparison operand.
17547 Don't assert that comparison operands aren't both constants.
17549 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
17551 * doc/install.texi (Downloading the Source): Update references to
17552 downloading separate components.
17554 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
17556 * doc/extend.texi (__int128): Improve grammar.
17558 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
17561 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
17562 alternative to (o,r).
17563 (*movdi_internal_rex64): Remove (!o,n) alternative.
17564 (DImode immediate->memory splitter): Remove.
17565 (DImode immediate->memory peephole2): Remove.
17566 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
17567 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
17568 alternative to (!o,*r).
17569 (*movtf_internal_sse): New pattern.
17570 (*movxf_internal_rex64): New pattern.
17571 (*movxf_internal): Disable for TARGET_64BIT.
17572 (*movdf_internal_rex64): Remove (!o,F) alternative.
17574 2013-01-22 Jakub Jelinek <jakub@redhat.com>
17576 PR middle-end/56074
17577 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
17578 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
17579 * tree-vect-loop-manip.c (find_loop_location): Also ignore
17580 stmt locations where LOCATION_LOCUS of the stmt location is
17581 UNKNOWN_LOCATION or BUILTINS_LOCATION.
17584 * config/i386/i386.md (UNSPEC_STOS): New.
17585 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
17586 *strsetqi_1): Add UNSPEC_STOS.
17588 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
17591 * doc/invoke.texi: Remove left over -Wsynth example.
17593 2013-01-21 Jakub Jelinek <jakub@redhat.com>
17595 PR tree-optimization/56051
17596 * fold-const.c (fold_binary_loc): Don't fold
17597 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
17598 a narrowing conversion, or widening conversion from signed
17601 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
17603 PR rtl-optimization/56023
17604 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
17605 dependent on debug instruction.
17607 2013-01-21 Martin Jambor <mjambor@suse.cz>
17609 PR middle-end/56022
17610 * function.c (allocate_struct_function): Call
17611 invoke_set_current_function_hook earlier.
17613 2013-01-21 Jakub Jelinek <jakub@redhat.com>
17615 * reload1.c (init_reload): Only initialize reload_obstack
17616 during the first call.
17618 2013-01-21 Marek Polacek <polacek@redhat.com>
17620 * cfgloop.c (verify_loop_structure): Fix up grammar.
17622 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
17624 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
17625 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
17627 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17630 * config/arm/marvell-pj4.md: Update copyright year.
17631 Fix up use of alu to alu_reg and simple_alu_imm.
17633 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
17635 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
17637 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
17640 * lra-constraints.c (curr_insn_transform): Don't reuse original
17641 insn for secondary memory move when memory mode should be different.
17643 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17645 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
17646 atomic_storedi_1): New patterns.
17648 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
17650 btver2 pipeline descriptions.
17651 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
17653 * config/i386/i386.md (btver2_decode): New type attributes.
17654 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
17656 * config/i386/btver2.md: New file describing btver2 pipelines.
17658 2013-01-19 Andrew Pinski <apinski@cavium.com>
17660 PR tree-optimization/52631
17661 * tree-ssa-sccvn (visit_use): Before looking up the original
17662 statement, try looking up the simplified expression.
17664 2013-01-19 Anthony Green <green@moxielogic.com>
17666 * config/moxie/moxie.c (moxie_expand_prologue): Set
17667 current_function_static_stack_size.
17669 2013-01-18 Jakub Jelinek <jakub@redhat.com>
17671 PR tree-optimization/56029
17672 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
17673 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
17675 2013-01-18 Sharad Singhai <singhai@google.com>
17677 PR tree-optimization/55995
17678 * dumpfile.c (dump_loc): Print location only if available.
17679 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
17681 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
17684 * lra-constraints.c (curr_insn_transform): Reuse original insn for
17685 secondary memory move.
17686 (inherit_reload_reg): Use rclass instead of cl for
17687 check_secondary_memory_needed_p.
17689 2013-01-18 Jakub Jelinek <jakub@redhat.com>
17691 PR middle-end/56015
17692 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
17693 the case where writing real complex part of target modifies op1.
17695 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
17697 * config/aarch64/aarch64-simd.md
17698 (aarch64_vcond_internal<mode>): Handle unordered cases.
17699 * config/aarch64/iterators.md (v_cmp_result): New.
17701 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
17702 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17704 * config/arm/marvell-pj4.md: New file.
17705 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
17706 * config/arm/arm.md (generic_sched): Add marvell_pj4.
17707 (generic_vfp): Likewise.
17708 * config/arm/arm-cores.def: Add marvell-pj4.
17709 * config/arm/arm-tune.md: Regenerate.
17710 * config/arm/arm-tables.opt: Regenerate.
17711 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
17712 * doc/invoke.texi: Document marvell-pj4.
17714 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
17716 * config/aarch64/arm_neon.h: Map scalar types to standard types.
17718 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
17723 * var-tracking.c (negative_power_of_two_p): New.
17724 (global_get_addr_cache, local_get_addr_cache): New.
17725 (get_addr_from_global_cache, get_addr_from_local_cache): New.
17726 (vt_canonicalize_addr): Rewrite using the above. Adjust the
17728 (vt_stack_offset_p): Remove.
17729 (vt_canon_true_dep): Always canonicalize loc's address.
17730 (clobber_overlapping_mems): Make sure we have a MEM.
17731 (local_get_addr_clear_given_value): New.
17732 (val_reset): Clear local cached entries.
17733 (compute_bb_dataflow): Create and release the local cache.
17734 Disable duplicate MEMs clobbering.
17735 (emit_notes_in_bb): Clobber MEMs likewise.
17736 (vt_emit_notes): Create and release the local cache.
17737 (vt_initialize, vt_finalize): Create and release the global
17738 cache, respectively.
17739 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
17741 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
17743 PR libmudflap/53359
17744 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
17745 not found in the symtab.
17747 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
17750 PR rtl-optimization/55547
17751 PR rtl-optimization/53827
17754 * alias.c (offset_overlap_p): New, factored out of...
17755 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
17756 the conservative special case for symbolic constants. Don't
17757 adjust zero sizes on alignment.
17759 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
17761 PR rtl-optimization/52573
17762 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
17763 REG_UNUSED for the same register.
17765 2013-01-17 Richard Biener <rguenther@suse.de>
17766 Marek Polacek <polacek@redhat.com>
17768 PR rtl-optimization/55833
17769 * loop-unswitch.c (unswitch_loops): Move loop verification...
17770 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
17771 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
17772 Set it to true when we're removing a loop from hierarchy tree in
17773 an irreducible region.
17774 (fix_bb_placements): Adjust caller.
17775 (fix_loop_placements): Likewise.
17777 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
17779 * config/avr/builtins.def (DEF_BUILTIN): Factor out
17780 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
17781 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
17782 Remove ID. Adjust comments.
17783 * config/avr/avr-c.c (avr_builtin_name): Remove.
17784 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
17785 * config/avr/avr.c (avr_tolower): New static function.
17786 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
17787 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
17788 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
17791 2013-01-17 Jan Hubicka <jh@suse.cz>
17793 PR tree-optimization/55273
17794 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
17796 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
17799 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
17800 store through atomic_store<mode>_1.
17801 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
17803 2013-01-17 Martin Jambor <mjambor@suse.cz>
17805 PR tree-optimizations/55264
17806 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
17807 for virtual methods.
17808 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
17809 virtual methods before inlining is over.
17810 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
17812 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
17815 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
17817 PR rtl-optimization/56005
17818 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
17819 pending reads for prefetch.
17821 2013-01-16 Ian Bolton <ian.bolton@arm.com>
17823 * config/aarch64/aarch64.md
17824 (*cstoresi_neg_uxtw): New pattern.
17825 (*cmovsi_insn_uxtw): New pattern.
17826 (*<optab>si3_uxtw): New pattern.
17827 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
17828 (*<optab>si3_insn_uxtw): New pattern.
17829 (*bswapsi2_uxtw): New pattern.
17831 2013-01-16 Richard Biener <rguenther@suse.de>
17833 * tree-inline.c (tree_function_versioning): Remove set but
17834 never used variable.
17836 2013-01-16 Richard Biener <rguenther@suse.de>
17838 PR tree-optimization/55964
17839 * tree-flow.h (rename_variables_in_loop): Remove.
17840 (rename_variables_in_bb): Likewise.
17841 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
17842 (copy_loop_before): Adjust and delete update-ssa status.
17843 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
17844 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
17845 (rename_variables_in_loop): Remove.
17846 (slpeel_update_phis_for_duplicate_loop): Likewise.
17847 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
17848 use available cfg machinery instead of duplicating it.
17849 Update PHI nodes and perform poor-mans SSA update here.
17850 (slpeel_tree_peel_loop_to_edge): Adjust.
17852 2013-01-16 Richard Biener <rguenther@suse.de>
17854 PR tree-optimization/54767
17855 PR tree-optimization/53465
17856 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
17857 (vrp_visit_phi_node): For PHI arguments coming via backedges
17858 drop all symbolical range information.
17859 (execute_vrp): Compute backedges.
17861 2013-01-16 Richard Biener <rguenther@suse.de>
17863 * doc/install.texi: Update CLooG and ISL requirements to
17866 2013-01-16 Christian Bruel <christian.bruel@st.com>
17869 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
17870 (broken_move): Handle UNSPECV_SP_SWITCH_B.
17871 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
17873 2013-01-16 DJ Delorie <dj@redhat.com>
17875 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
17876 (UNSPECV_SP_SWITCH_E): New.
17877 (sp_switch_1): Change to an unspec.
17878 (sp_switch_2): Change to an unspec. Don't use post-inc when we
17881 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
17883 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
17884 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
17885 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
17886 (expand_mem_thread_fence): Ditto.
17887 (expand_mem_signal_fence): Ditto.
17888 (expand_atomic_load): Ditto.
17889 (expand_atomic_store): Ditto.
17891 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
17893 PR rtl-optimization/55547
17894 PR rtl-optimization/53827
17897 * alias.c (memrefs_conflict_p): Set sizes to negative after
17900 2013-01-15 Jakub Jelinek <jakub@redhat.com>
17903 * function.c (thread_prologue_and_epilogue_insns): Always
17904 add crtl->drap_reg to set_up_by_prologue.set, even if
17905 stack_realign_drap is false.
17907 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
17909 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
17910 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
17911 *call): Fix indention.
17913 2013-01-15 Tom de Vries <tom@codesourcery.com>
17916 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
17919 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
17921 PR rtl-optimization/55153
17922 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
17924 2013-01-15 Martin Jambor <mjambor@suse.cz>
17926 PR tree-optimization/55920
17927 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
17928 accesses as grp_to_be_debug_replaced.
17930 2013-01-15 Jakub Jelinek <jakub@redhat.com>
17932 PR tree-optimization/55920
17933 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
17934 there is non-useless type conversion needed from debug rhs to lhs,
17935 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
17937 2013-01-15 Joseph Myers <joseph@codesourcery.com>
17938 Mikael Pettersson <mikpe@it.uu.se>
17941 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
17943 (ASM_OUTPUT_CASE_LABEL): Remove.
17944 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
17945 * final.c (shorten_branches): Update alignment of labels before
17946 jump tables if CASE_VECTOR_SHORTEN_MODE.
17948 2013-01-15 Richard Biener <rguenther@suse.de>
17951 * system.h: Do not include gmp.h for building host tools.
17953 2013-01-15 Richard Biener <rguenther@suse.de>
17955 PR middle-end/55882
17956 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
17957 account for bitpos when computing alignment.
17959 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
17961 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
17962 (ix86_target_macros_internal): Likewise.
17964 * config/i386/i386.c (m_CORE2I7): Removed.
17965 (m_CORE_HASWELL): New macro.
17966 (m_CORE_ALL): Likewise.
17967 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
17968 (initial_ix86_arch_features): Likewise.
17969 (processor_target_table): Initializations for Core avx2.
17970 (cpu_names): New names "core-avx2".
17971 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
17972 PROCESSOR_CORE_HASWELL.
17973 (ix86_issue_rate): New case.
17974 (ia32_multipass_dfa_lookahead): Likewise.
17975 (ix86_sched_init_global): Likewise.
17977 * config/i386/i386.h (TARGET_HASWELL): New macro.
17978 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
17979 (processor_type): New PROCESSOR_HASWELL.
17981 2013-01-15 Jakub Jelinek <jakub@redhat.com>
17983 PR tree-optimization/55955
17984 * tree-vect-loop.c (vectorizable_reduction): Give up early on
17985 *SHIFT_EXPR and *ROTATE_EXPR codes.
17987 PR tree-optimization/48766
17988 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
17989 -ftrapv disable -fwrapv.
17991 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
17994 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
17995 etc. to 1 and not to __flash.
17996 Use LL suffix for __INT24_MAX__ with -mint8.
17997 Use ULL suffix for __UINT24_MAX__ with -mint8.
17999 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
18001 * config/avr/avr-arch.h
18002 (struct base_arch_s): Use typedef avr_arch_t instead.
18003 (struct arch_info_s): Use typedef avr_arch_info_t instead.
18004 (struct mcu_type_s): Use typedef avr_mcu_t instead.
18005 * config/avr/avr.c: Same.
18006 * config/avr/avr-devices.c: Same.
18007 * config/avr/driver-avr.c: Same.
18008 * config/avr/gen-avr-mmcu-texi.c: Same.
18009 * config/avr/avr-mcus.def: Adjust comment.
18011 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
18013 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
18014 * config/aarch64/iterators.md (VALLDI): New.
18016 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
18017 Andi Kleen <ak@linux.intel.com>
18020 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
18021 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
18024 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
18026 * config/avr/avr-stdint.h: Remove trailing blanks.
18027 * config/avr/avr-log.h: Same.
18028 * config/avr/avr-arch.h: Same.
18029 * config/avr/avr-devices.c: Same.
18030 * config/avr/avr-dimode.md: Same.
18031 * config/avr/predicates.md: Same.
18032 * config/avr/avr-c.c: Same. And fix typo.
18034 * config/avr/avr-protos.h: Same. And:
18035 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
18036 (init_cumulative_args): Rename to avr_init_cumulative_args.
18037 (expand_prologue): Rename to avr_expand_prologue.
18038 (expand_epilogue): Rename to avr_expand_epilogue.
18039 (adjust_insn_length): Rename to avr_adjust_insn_length.
18040 (notice_update_cc): Rename to avr_notice_update_cc.
18041 (final_prescan_insn): Rename to avr_final_prescan_insn.
18042 * config/avr/avr.c: Same.
18043 * config/avr/avr.h: Same.
18044 * config/avr/avr.md: Remove trailing blanks.
18045 (prologue): Use avr_expand_prologue.
18046 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
18048 2013-01-14 Richard Biener <rguenther@suse.de>
18050 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
18051 verify_location, collect_subblocks): New functions.
18052 (verify_gimple_in_cfg): Verify that locations only reference
18053 BLOCKs in the functions BLOCK tree.
18055 2013-01-14 Richard Biener <rguenther@suse.de>
18057 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
18059 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
18061 (insert_out_of_ssa_copy_on_edge): Likewise.
18062 (rewrite_close_phi_out_of_ssa): Likewise.
18063 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
18065 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
18066 propagated constants.
18067 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
18070 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
18072 * config/avr/avr-modes.def: Add GPL copyright notice.
18074 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
18076 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
18077 MEMMODEL_MASK to determine memory model.
18078 (atomic_store<mode>): Ditto from operands[2].
18079 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
18081 2013-01-13 Jakub Jelinek <jakub@redhat.com>
18084 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
18085 (fold_gimple_assign): Don't call unshare_expr here.
18086 (fold_ctor_reference): Call unshare_expr.
18088 2013-01-13 Terry Guo <terry.guo@arm.com>
18090 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
18091 * doc/fragments.texi: Document MULTILIB_REUSE.
18092 * gcc.c (multilib_reuse): New internal spec.
18093 (set_multilib_dir): Also search multilib from multilib_reuse.
18094 * genmultilib (tmpmultilib3): Refactor code.
18095 (tmpmultilib4): Ditto.
18096 (multilib_reuse): New multilib argument.
18098 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
18100 * Makefile.in: Update copyright.
18102 2013-01-12 Tom de Vries <tom@codesourcery.com>
18104 PR middle-end/55890
18105 * calls.c (expand_call): Check if arg_nr is valid.
18107 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
18109 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
18110 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
18111 documentation. Add missing '__' in front of
18112 __builtin_ia32_packssdw256.
18114 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
18117 * config/s390/s390.c (s390_preferred_reload_class): Do not return
18118 NO_REGS for larl operands.
18119 (s390_reload_larl_operand): Use s390_load_address instead of
18122 2013-01-11 Richard Biener <rguenther@suse.de>
18124 * tree-cfg.c (verify_node_sharing_1): Split out from ...
18125 (verify_node_sharing): ... here.
18126 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
18128 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
18130 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
18131 Substitute TREECHECKING.
18132 * configure: Regenerate.
18133 * Makefile.in (TREECHECKING): New.
18135 2013-01-11 Richard Guenther <rguenther@suse.de>
18137 PR tree-optimization/44061
18138 * tree-vrp.c (extract_range_basic): Compute zero as
18139 value-range for __builtin_constant_p of function parameters.
18141 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
18143 Update copyright years.
18145 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
18147 PR rtl-optimization/55672
18148 * lra-eliminations.c (mark_not_eliminable): Permit addition with
18149 const to be eliminable.
18151 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
18153 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
18154 * configure: Regenerate.
18156 2013-01-10 Richard Biener <rguenther@suse.de>
18158 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
18160 2013-01-10 Richard Biener <rguenther@suse.de>
18163 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
18164 locations for virtual PHI arguments.
18165 (rewrite_update_phi_arguments): Likewise.
18167 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
18169 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
18172 2013-01-10 Jakub Jelinek <jakub@redhat.com>
18174 PR tree-optimization/55921
18175 * tree-complex.c (expand_complex_asm): New function.
18176 (expand_complex_operations_1): Call it for GIMPLE_ASM.
18178 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
18181 * config/s390/s390.c (s390_symref_operand_p)
18182 (s390_loadrelative_operand_p): Merge the two functions.
18183 (s390_check_qrst_address, print_operand_address): Add parameters
18184 to s390_loadrelative_operand_p invokation.
18185 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
18186 (s390_reload_larl_operand, s390_secondary_reload): Use
18187 s390_loadrelative_operand_p instead of s390_symref_operand_p.
18188 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
18190 2013-01-09 Mike Stump <mikestump@comcast.net>
18192 * dse.c (record_store): Remove unnecessary assert.
18194 2013-01-09 Jan Hubicka <jh@suse.cz>
18196 PR tree-optimization/55569
18197 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
18198 * cfgloop.h (scale_loop_profile): Likewise.
18200 2013-01-09 Jan Hubicka <jh@suse.cz>
18203 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
18205 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
18207 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
18209 PR middle-end/55114
18210 * expr.h (maybe_emit_group_store): Declare.
18211 * expr.c (maybe_emit_group_store): New function.
18212 * builtins.c (expand_builtin_int_roundingfn): Call it.
18213 (expand_builtin_int_roundingfn_2): Likewise.
18215 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
18217 PR rtl-optimization/55829
18218 * lra-constraints.c (match_reload): Add code for absent output.
18219 (curr_insn_transform): Add code for reloads of matched inputs
18222 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
18224 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
18225 attribute of movddup insn to DF.
18226 (*vec_interleave_lowv2df): Ditto.
18227 (vec_dupv2df): Ditto.
18229 2013-01-09 Jan Hubicka <jh@suse.cz>
18231 PR tree-optimiation/55875
18232 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
18233 EVERY_ITERATION parameter.
18234 (number_of_iterations_exit): Check if exit is executed every iteration.
18235 (idx_infer_loop_bounds): Similarly here.
18236 (n_of_executions_at_most): Simplify
18237 to only test for cases where statement is dominated by the
18238 particular bound; handle correctly the "postdominance" test.
18239 (scev_probably_wraps_p): Use max loop iterations info
18240 as a global bound first.
18242 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
18243 Nick Clifton <nickc@redhat.com>
18245 * config/v850/v850.md (cbranchsf4): New pattern.
18246 (cstoresf4): New pattern.
18247 (cbranchdf4): New pattern.
18248 (cstoredf4): New pattern.
18249 (movsicc): Disallow floating point comparisons.
18250 (cmpsf_le_insn): Fix order of operators.
18251 (cmpsf_lt_insn): Likewise.
18252 (cmpsf_eq_insn): Likewise.
18253 (cmpdf_le_insn): Likewise.
18254 (cmpdf_lt_insn): Likewise.
18255 (cmpdf_eq_insn): Likewise.
18256 (cmpsf_ge_insn): Use LE comparison.
18257 (cmpdf_ge_insn): Likewise.
18258 (cmpsf_gt_insn): Use LT comparison.
18259 (cmpdf_gt_insn): Likewise.
18260 (cmpsf_ne_insn): Delete pattern.
18261 (cmpdf_ne_insn): Delete pattern.
18262 * config/v850/v850.c (v850_gen_float_compare): Use
18263 gen_cmpdf_eq_insn for NE comparison.
18264 (v850_float_z_comparison_operator)
18265 (v850_float_nz_comparison_operator): Move from here ...
18266 * config/v850/predicates.md: ... to here. Move GT and GE
18267 comparisons into v850_float_z_comparison_operator.
18268 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
18270 (v850_float_nz_comparison_operator): Likewise.
18272 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
18274 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
18275 with calls to gen_insvsi/gen_insvdi.
18277 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
18279 * config/i386/i386.c (initial_ix86_tune_features): Set up
18280 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
18282 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
18283 Jakub Jelinek <jakub@redhat.com>
18285 PR tree-optimization/48189
18286 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
18287 If nitercst is 0, don't predict the exit edge.
18289 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
18291 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
18292 in asm_fprintf with reg_names.
18293 (aarch64_print_operand_address): Likewise.
18294 (aarch64_return_addr): Likewise.
18295 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
18297 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
18299 * config/pa/pa.h (VAL_U6_BITS_P): Define.
18300 (INT_U6_BITS): Likewise.
18301 * config/pa/predicates.md (uint6_operand): New predicate.
18302 (shift5_operand, shift6_operand): Likewise.
18303 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
18305 (lshrdi3): Use shift6_operand.
18306 (shrpsi4, shrpdi4): New insn patterns.
18307 (extzv): Delete expander.
18308 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
18309 predicates in unamed zero extract patterns. Tighten common constraint.
18310 (extv): Delete expander.
18311 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
18312 predicates in unamed sign extract patterns. Tighten common constraint.
18313 (insv): Delete expander.
18314 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
18315 predicates in unamed insert patterns. Tighten common constraint.
18316 Change uint32_operand predicate to uint6_operand predicate in unamed
18317 DImode pattern to insert constant values of type 1...1xxxx.
18319 2013-01-04 Jan Hubicka <jh@suse.cz>
18321 PR tree-optimization/55823
18322 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
18325 2013-01-08 Jakub Jelinek <jakub@redhat.com>
18326 Uros Bizjak <ubizjak@gmail.com>
18328 PR rtl-optimization/55845
18329 * df-problems.c (can_move_insns_across): Stop scanning at
18330 volatile_insn_p source instruction or give up if
18331 across_from .. across_to range contains any volatile_insn_p
18334 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
18336 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
18337 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
18339 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
18340 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
18342 2013-01-08 Jakub Jelinek <jakub@redhat.com>
18345 * asan.c (asan_clear_shadow): New function.
18346 (asan_emit_stack_protection): Use it.
18348 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
18350 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
18351 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
18352 with tab instead of space.
18354 2013-01-08 Nick Clifton <nickc@redhat.com>
18356 * config/rl78/rl78.c (rl78_expand_prologue): Always select
18357 register bank 0 at the start of an interrupt handler.
18358 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
18361 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
18363 * config/aarch64/aarch64-simd.md
18364 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
18365 (aarch64_simd_bsl): Likewise.
18366 (aarch64_vcond_internal<mode>): Likewise.
18367 (vcond<mode><mode>): Likewise.
18368 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
18369 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
18371 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
18373 * config/aarch64/aarch64-builtins.c
18374 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
18376 2013-01-08 Martin Jambor <mjambor@suse.cz>
18379 * tree-sra.c (analyze_access_subtree): Return true also after
18380 potentially creating a debug-only replacement.
18382 2013-01-08 Jakub Jelinek <jakub@redhat.com>
18384 PR middle-end/55890
18385 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
18387 PR tree-optimization/54120
18388 * tree-vrp.c (range_fits_type_p): Don't allow
18389 src_precision < precision from signed vr to unsigned_p
18390 if vr->min or vr->max is negative.
18391 (simplify_float_conversion_using_ranges): Test can_float_p
18392 against CODE_FOR_nothing.
18394 2013-01-08 Jakub Jelinek <jakub@redhat.com>
18395 Richard Biener <rguenther@suse.de>
18397 PR middle-end/55851
18398 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
18399 types instead of just INTEGER_TYPE types.
18401 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
18403 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
18406 2013-01-07 Steve Ellcey <sellcey@mips.com>
18409 * config/mips/mips.opt: Change mad to mmad to match documentation.
18411 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
18414 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
18415 .progmemx.data now.
18417 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
18420 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
18421 (avr_addrspace_t): Add .section_name field.
18422 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
18424 (avr_addrspace): Same. Initialize .section_name. Remove last
18425 NULL entry. Put __memx into .progmemx.data.
18426 (progmem_section_prefix): Remove.
18427 (avr_asm_init_sections): No need to initialize progmem_section.
18428 (avr_asm_named_section): Use avr_addrspace[].section_name to get
18429 section name prefix.
18430 (avr_asm_select_section): Ditto. And use get_unnamed_section to
18431 retrieve the progmem section.
18432 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
18433 boundary to run over avr_addrspace[].
18434 (avr_register_target_pragmas): Ditto.
18436 2013-01-06 Jakub Jelinek <jakub@redhat.com>
18438 * varasm.c (output_constant_def_contents): For asan_protect_global
18439 protected strings, adjust DECL_ALIGN if needed, before testing for
18441 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
18442 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
18444 (output_object_block): For asan protected decls, emit asan padding
18445 after their contents.
18446 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
18447 (asan_finish_file): Test it here instead.
18449 2013-01-07 Nick Clifton <nickc@redhat.com>
18450 Matthias Klose <doko@debian.org>
18451 Doug Kwan <dougkwan@google.com>
18452 H.J. Lu <hongjiu.lu@intel.com>
18455 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
18457 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
18459 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
18461 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
18463 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
18465 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
18468 * doc/install.texi (Cross-Compiler-Specific Options): Document
18471 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
18473 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
18474 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
18475 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
18476 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
18477 vqmovun_high_s64): Fix source operand number and update copyright.
18479 2013-01-07 Richard Biener <rguenther@suse.de>
18481 PR middle-end/55890
18482 * gimple.h (gimple_call_builtin_p): New overload.
18483 * gimple.c (validate_call): New function.
18484 (gimple_call_builtin_p): Likewise.
18485 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
18486 Use gimple_call_builtin_p.
18487 (find_func_clobbers): Likewise.
18488 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
18489 (strlen_optimize_stmt): Likewise.
18491 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
18493 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
18494 (vld1q_dup_*): Likewise.
18495 (vld1_*): Likewise.
18496 (vld1q_*): Likewise.
18497 (vld1_lane_*): Likewise.
18498 (vld1q_lane_*): Likewise.
18500 2013-01-07 Richard Biener <rguenther@suse.de>
18502 * lto-streamer.h (LTO_minor_version): Bump to 2.
18504 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
18506 * config/aarch64/aarch64-protos.h
18507 (aarch64_const_double_zero_rtx_p): Rename to...
18508 (aarch64_float_const_zero_rtx_p): ...this.
18509 (aarch64_float_const_representable_p): New.
18510 (aarch64_output_simd_mov_immediate): Likewise.
18511 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
18512 move immediate case.
18513 * config/aarch64/aarch64.c
18514 (aarch64_const_double_zero_rtx_p): Rename to...
18515 (aarch64_float_const_zero_rtx_p): ...this.
18516 (aarch64_print_operand): Allow printing of new constants.
18517 (aarch64_valid_floating_const): New.
18518 (aarch64_legitimate_constant_p): Check for valid floating-point
18520 (aarch64_simd_valid_immediate): Likewise.
18521 (aarch64_vect_float_const_representable_p): New.
18522 (aarch64_float_const_representable_p): Likewise.
18523 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
18524 (aarch64_output_simd_mov_immediate): New.
18525 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
18526 (*movdf_aarch64): Likewise.
18527 * config/aarch64/constraints.md (Ufc): New.
18528 (Y): call aarch64_float_const_zero_rtx.
18529 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
18531 2013-01-07 Richard Biener <rguenther@suse.de>
18533 PR tree-optimization/55888
18534 PR tree-optimization/55862
18535 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
18536 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
18537 not if it is contained therein.
18539 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
18541 * config/avr/t-avr: Typo.
18543 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
18546 * config/avr/t-avr: Don't automatically rebuild
18547 $(srcdir)/config/avr/t-multilib
18548 $(srcdir)/config/avr/avr-tables.opt
18549 $(srcdir)/doc/avr-mmcu.texi
18550 (avr-mcus): New phony target to build them on request.
18551 (s-avr-mlib, s-avr-mmcu-texi): Remove.
18552 * avr/avr-mcus.def: Adjust comments.
18554 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
18556 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
18558 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
18560 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
18562 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
18564 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
18566 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
18568 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
18569 to generate profiling.
18570 * config/rs6000/aix64.h (LIB_SPEC): Same.
18572 2013-01-04 Andrew Pinski <apinski@cavium.com>
18574 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
18576 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
18578 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
18580 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
18582 (ix86_expand_move): Ditto.
18583 (ix86_zero_extend_to_Pmode): Ditto.
18584 (ix86_expand_call): Ditto.
18585 (ix86_expand_special_args_builtin): Ditto.
18586 (ix86_expand_builtin): Ditto.
18588 2013-01-04 Richard Biener <rguenther@suse.de>
18590 PR tree-optimization/55862
18591 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
18592 translating them through PHI nodes.
18594 2013-01-04 Martin Jambor <mjambor@suse.cz>
18596 PR tree-optimization/55755
18597 * tree-sra.c (sra_modify_assign): Do not check that an access has no
18598 children when trying to avoid producing a VIEW_CONVERT_EXPR.
18600 2013-01-04 Marek Polacek <polacek@redhat.com>
18602 PR middle-end/55859
18603 * opts.c (default_options_optimization): Clarify error message.
18605 2013-01-04 Richard Biener <rguenther@suse.de>
18607 PR middle-end/55863
18608 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
18611 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
18614 * config/pa/pa.md (movsi): Revert previous change.
18615 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
18618 2013-01-03 Richard Henderson <rth@redhat.com>
18620 * config/i386/i386.c (ix86_expand_move): Always assign to op1
18621 after eliminating TLS symbols.
18623 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
18626 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
18627 * graphite-poly.c (debug_gmp_value): Likewise.
18629 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
18632 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
18633 selected code model, define __code_mode_small__, __code_model_medium__,
18634 __code_model_large__, __code_model_32__ or __code_model_kernel__.
18635 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
18636 xchg temporary register with %k. Declare temporary register as
18638 [__x86_64__]: For medium and large code models, preserve %rbx register.
18640 2013-01-03 Richard Biener <rguenther@suse.de>
18642 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
18643 (dump_subscript): Adjust.
18644 (finalize_ddr_dependent): Do not dump redundant info.
18645 (analyze_siv_subscript): Adjust.
18646 (subscript_dependence_tester): Likewise.
18647 (compute_affine_dependence): Likewise.
18649 2013-01-03 Richard Biener <rguenther@suse.de>
18652 2013-01-03 Richard Biener <rguenther@suse.de>
18654 PR tree-optimization/55857
18655 * tree-vect-stmts.c (vectorizable_load): Do not setup
18656 re-alignment for invariant loads.
18658 2013-01-02 Richard Biener <rguenther@suse.de>
18660 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
18661 invariant load do not generate a vector load from the scalar location.
18663 2013-01-03 Richard Biener <rguenther@suse.de>
18665 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
18666 for not vectorizing.
18667 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
18668 not build INDIRECT_REFs, call get_name once only.
18669 (vect_create_data_ref_ptr): Likewise. Dump base object kind
18670 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
18672 2013-01-03 Richard Biener <rguenther@suse.de>
18674 PR tree-optimization/55857
18675 * tree-vect-stmts.c (vectorizable_load): Do not setup
18676 re-alignment for invariant loads.
18678 2013-01-03 Richard Biener <rguenther@suse.de>
18681 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
18682 prefer a built-in decl.
18684 2013-01-03 Jakub Jelinek <jakub@redhat.com>
18686 * gcc.c (process_command): Update copyright notice dates.
18687 * gcov.c (print_version): Likewise.
18688 * gcov-dump.c (print_version): Likewise.
18690 PR rtl-optimization/55838
18691 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
18692 iv0.step, iv1.step and step.
18694 2013-01-03 Jakub Jelinek <jakub@redhat.com>
18695 Marc Glisse <marc.glisse@inria.fr>
18697 PR tree-optimization/55832
18698 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
18699 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
18700 integer_{one,zero}_node.
18702 2013-01-03 Jakub Jelinek <jakub@redhat.com>
18705 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
18706 * var-tracking.c (reverse_op): Don't add reverse ops to
18707 VALUEs that have already
18708 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
18710 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
18712 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
18714 2013-01-02 Teresa Johnson <tejohnson@google.com>
18716 * dumpfile.c (dump_loc): Print filename with location.
18717 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
18718 new location_t parameter to emit complete unroll message with
18719 new dump framework.
18720 (canonicalize_loop_induction_variables): Compute loops location
18721 and pass to try_unroll_loop_completely.
18722 * loop-unroll.c (report_unroll_peel): New function.
18723 (peel_loops_completely): Use new dump format with location
18724 for main dumpfile message, and invoke report_unroll_peel on success.
18725 (decide_unrolling_and_peeling): Ditto.
18726 (decide_peel_once_rolling): Remove old dumpfile message subsumed
18727 by report_unroll_peel.
18728 (decide_peel_completely): Ditto.
18729 (decide_unroll_constant_iterations): Ditto.
18730 (decide_unroll_runtime_iterations): Ditto.
18731 (decide_peel_simple): Ditto.
18732 (decide_unroll_stupid): Ditto.
18733 * cfgloop.c (get_loop_location): New function.
18734 * cfgloop.h (get_loop_location): Declare.
18736 2013-01-02 Sriraman Tallam <tmsriram@google.com>
18738 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
18741 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
18743 PR middle-end/55198
18744 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
18745 BLKmode objects when EXPAND_MEMORY is specified.
18747 2013-01-02 Sriraman Tallam <tmsriram@google.com>
18749 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
18751 (fold_builtin_cpu): Do not share cpu model decls across statements.
18753 2013-01-02 Jason Merrill <jason@redhat.com>
18756 * tree.c (build_array_type_1): Revert earlier change.
18758 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
18760 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
18762 * config/aarch64/aarch64-tune.md: Re-generate.
18764 2013-01-02 Richard Biener <rguenther@suse.de>
18766 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
18767 invariant load do not generate a vector load from the scalar location.
18769 2013-01-02 Richard Biener <rguenther@suse.de>
18772 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
18773 * configure: Regenerate.
18775 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
18777 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
18778 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
18779 (expand_builtin_int_roundingfn_2): Keep the original target around
18780 for the fallback case.
18782 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
18784 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
18785 to be clear for sign changes.
18787 2013-01-01 Jan Hubicka <jh@suse.cz>
18789 * ipa-inline-analysis.c: Fix formatting.
18791 2013-01-01 Jakub Jelinek <jakub@redhat.com>
18793 PR tree-optimization/55831
18794 * tree-vect-loop.c (get_initial_def_for_induction): Use
18795 gsi_after_labels instead of gsi_start_bb.
18797 Copyright (C) 2013 Free Software Foundation, Inc.
18799 Copying and distribution of this file, with or without modification,
18800 are permitted in any medium without royalty provided the copyright
18801 notice and this notice are preserved.