1 2013-09-27 Xinliang David Li <davidxl@google.com>
3 * opts.c (finish_options): Adjust parameters
4 according to vect cost model.
5 (common_handle_option): Set dynamic vect cost
7 targhooks.c (default_add_stmt_cost): Compute stmt cost
9 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
11 * tree-vectorizer.h (unlimited_cost_model): New function.
12 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
13 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
15 (vect_enhance_data_refs_alignment): Ditto.
16 * flag-types.h: New enum.
17 * common/config/i386/i386-common.c (ix86_option_init_struct):
18 No need to initialize vect_cost_model flag.
19 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
22 2013-09-27 Diego Novillo <dnovillo@google.com>
24 * gimple.h (enum ssa_mode): Remove.
26 2013-09-27 Paulo Matos <pmatos@broadcom.com>
29 * gcc.dg/pr58463.c: New test.
31 2013-09-27 Paulo Matos <pmatos@broadcom.com>
33 * cfgloop.h (number_of_loops): Fix typo in check for null.
35 2013-09-27 Jakub Jelinek <jakub@redhat.com>
38 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
39 are children of outermost saved_cfun's loop, and set it up to
40 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
41 if loop != loop0 and SESE region contains bbs that belong to loop0.
43 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
45 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
46 (binary_scale_code_p, get_base_term, get_index_term): New functions.
47 (set_address_segment, set_address_base, set_address_index)
48 (set_address_disp): Accept the argument unconditionally.
49 (baseness): Remove must_be_base_p and must_be_index_p checks.
50 (decompose_normal_address): Classify as much as possible in the
53 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
55 * cse.c (count_reg_usage): Handle INT_LIST.
56 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
57 * reginfo.c (reg_scan_mark_refs): Likewise.
58 * reload1.c (eliminate_regs_1): Likewise.
60 2013-09-27 Iain Sandoe <iain@codesourcery.com>
63 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
66 2013-09-27 Richard Biener <rguenther@suse.de>
68 PR tree-optimization/58459
69 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
70 restriction not propagating into loops.
72 2013-09-26 Florian Weimer <fw@deneb.enyo.de>
74 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
75 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
76 * doc/tree-ssa.texi (Walking use-def chains): Delete.
78 2013-09-26 Richard Biener <rguenther@suse.de>
80 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names
83 2013-09-26 Richard Biener <rguenther@suse.de>
85 * alias.h (component_uses_parent_alias_set): Rename to ...
86 (component_uses_parent_alias_set_from): ... this.
87 * alias.c (component_uses_parent_alias_set): Rename to ...
88 (component_uses_parent_alias_set_from): ... this and return
90 (reference_alias_ptr_type_1): Use the result from
91 component_uses_parent_alias_set_from instead of stripping
92 components one at a time.
93 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
95 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
97 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs): Move
99 * tree-ssa-ter.h: New File. Move prototypes here.
100 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
101 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
102 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
103 from is_replaceable_p.
104 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
105 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
106 (process_replaceable): Use ter_is_replaceable_p.
107 (find_replaceable_in_bb): Use ter_is_replaceable_p.
108 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
109 newly refactored ssa_is_replaceable_p.
110 * cfgexpand.c: Include tree-outof-ssa.h.
111 * ssaexpand.h: Delete.
113 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
115 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
116 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
118 (create_gimple_tmp): Delete.
119 (get_expr_type, build_assign, build_type_cast): Move to...
120 * gimple-builder.c: New File.
121 (get_expr_type): Relocate from gimple.c.
122 (build_assign, build_type_cast): Change to only create ssanames.
123 * gimple.h: Move prototypes to...
124 * gimple-builder.h: New File. Here.
125 * tree-ssa.h: And here.
126 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
127 count_uses_and_derefs): Relocate from gimple.c.
128 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
129 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
130 * tree-ssa-math-opts (execute_cse_reciprocals): Use
131 gimple_replace_ssa_lhs.
132 * asan.c: Include gimple-builder.h.
133 * Makefile.in: Add gimple-builder.o.
135 2013-09-26 Richard Biener <rguenther@suse.de>
137 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
138 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
139 (loe_visit_block): Use gcc_checking_assert.
140 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
141 gimple_assign_ssa_name_copy_p.
142 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
145 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
147 * config/rs6000/t-rs6000 (rs6000.o): Remove.
148 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
150 2013-09-26 Richard Biener <rguenther@suse.de>
152 PR tree-optimization/58539
153 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
154 the fact that debug statements are not taking part in loop-closed
157 2013-09-26 Nick Clifton <nickc@redhat.com>
159 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
160 time warning message.
161 (msp430_print_operand_raw): Delete unused letter parameter.
162 (TARGET_PRINT_OPERAND_ADDRESS): Define.
163 (msp430_print_operand_address): New function.
164 (msp430_print_operand): Move address printing code from here to
166 * config/msp430/msp430.md (movsipsi2): Add comment in generated
168 (zero_extendpsisi2): Likewise.
169 (extendpsisi2): New pattern.
170 (andneghi3): New pattern.
172 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
174 * config/aarch64/aarch64.opt (mlra): New option.
175 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
176 (TARGET_LRA_P): Define.
178 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
180 * expr.c (expand_assignment): Remove obsolete comment.
182 2013-09-25 Jeff Law <law@redhat.com>
184 * tree-flow.h (thread_through_all_blocks): Prototype moved into
185 tree-ssa-threadupdate.h.
186 (register_jump_thread): Similarly.
187 * tree-ssa-threadupdate.h: New header file.
188 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
189 * tree-vrp.c: Likewise.
190 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
191 (thread_around_empty_blocks): Change type of path vector argument to
192 an edge,type pair from just an edge. Initialize both elements when
193 appending to a jump threading path. Tweak references to elements
195 (thread_across_edge): Similarly. Release memory for the elements
197 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
198 (dump_jump_thread_path): New function broken out from
199 register_jump_thread.
200 (register_jump_thread): Use dump_jump_thread_path. Change type of
201 path vector entries. Search the path for NULL edges and dump
202 the path if one is found. Tweak the conversion of path to 3-edge
203 form to use the block copy type information embedded in the path.
205 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
207 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
209 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
210 Vladimir Makarov <vmakarov@redhat.com>
212 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
213 from the least significant bit.
214 (strip_address_mutations): Add bitfield operations handling.
215 (must_be_index_p): Add shifting and rotate operations handling.
216 (set_address_base): Use must_be_base_p predicate.
217 (set_address_index): Use must_be_index_p predicate.
219 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
220 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
221 Sergey Lega <sergey.s.lega@intel.com>
222 Anna Tikhonova <anna.tikhonova@intel.com>
223 Ilya Tocar <ilya.tocar@intel.com>
224 Andrey Turetskiy <andrey.turetskiy@intel.com>
225 Ilya Verbin <ilya.verbin@intel.com>
226 Kirill Yukhin <kirill.yukhin@intel.com>
227 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
229 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
231 (ix86_expand_vector_move_misalign): Support new unaligned load and
232 stores and use new names.
233 (CODE_FOR_sse2_storedqu): Rename to ...
234 (CODE_FOR_sse2_storedquv16qi): ... this.
235 (CODE_FOR_sse2_loaddqu): Rename to ...
236 (CODE_FOR_sse2_loaddquv16qi): ... this.
237 (CODE_FOR_avx_loaddqu256): Rename to ...
238 (CODE_FOR_avx_loaddquv32qi): ... this.
239 (CODE_FOR_avx_storedqu256): Rename to ...
240 (CODE_FOR_avx_storedquv32qi): ... this.
241 * config/i386/i386.md (fpint_logic): New.
242 * config/i386/sse.md (VMOVE): Extend for AVX512.
246 (VI_UNALIGNED_LOADSTORE): Ditto.
247 (sse2_avx_avx512f): Ditto.
248 (sse2_avx2): Extend for AVX512.
249 (sse4_1_avx2): Ditto.
251 (sse): Extend for AVX512.
254 (avxsizesuffix): Ditto.
255 (sseintvecmode): Ditto.
257 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
258 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
259 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
260 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
261 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
262 (<sse2_avx_avx512f>_storedqu<mode): ... this.
263 (<sse>_movnt<mode>): Replace constraint "x" with "v".
264 (STORENT_MODE): Extend for AVX512.
265 (*absneg<mode>2): Replace constraint "x" with "v".
266 (*mul<mode>3): Ditto.
267 (*ieee_smin<mode>3): Ditto.
268 (*ieee_smax<mode>3): Ditto.
269 (avx_cmp<mode>3): Replace VF with VF_128_256.
270 (*<sse>_maskcmp<mode>3_comm): Ditto.
271 (<sse>_maskcmp<mode>3): Ditto.
272 (<sse>_andnot<mode>3): Extend for AVX512.
273 (<code><mode>3, anylogic): Replace VF with VF_128_256.
274 (<code><mode>3, fpint_logic): New.
275 (*<code><mode>3): Extend for AVX512.
276 (avx512flogicsuff): New.
277 (avx512f_<logic><mode>): Ditto.
278 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
280 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
281 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
282 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
283 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
284 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
285 (xop_vpermil2<mode>3): Ditto.
286 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
287 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
288 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
289 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
291 2013-09-25 Tom Tromey <tromey@redhat.com>
293 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
294 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
295 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
296 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
297 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
298 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
299 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
300 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
301 (GRAPHITE_HTAB_H): Remove.
303 2013-09-25 Tom Tromey <tromey@redhat.com>
305 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
307 2013-09-25 Tom Tromey <tromey@redhat.com>
309 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
311 2013-09-25 Tom Tromey <tromey@redhat.com>
313 * config/i386/t-i386 (i386.o): Remove.
314 (i386-c.o): Use COMPILE and POSTCOMPILE.
316 2013-09-25 Tom Tromey <tromey@redhat.com>
318 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
320 2013-09-25 Tom Tromey <tromey@redhat.com>
322 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
323 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
324 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
325 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
326 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
327 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
328 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
329 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
330 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
331 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
332 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
333 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
334 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
335 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
336 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
337 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
338 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
339 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
340 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
341 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
342 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
343 (tree-ssa-pre.o, tree-ssa-sccvn.o)
344 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
345 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
346 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
347 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
348 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
349 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
350 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
351 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
352 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
353 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
354 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
355 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
356 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
357 (graphite-clast-to-gimple.o, graphite-dependences.o)
358 (graphite-interchange.o, graphite-poly.o)
359 (graphite-scop-detection.o, graphite-sese-to-poly.o)
360 (graphite-optimize-isl.o, tree-vect-loop.o)
361 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
362 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
363 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
364 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
365 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
366 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
367 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
368 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
369 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
370 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
371 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
372 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
373 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
374 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
375 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
376 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
377 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
378 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
379 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
380 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
381 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
382 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
383 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
384 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
385 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
386 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
387 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
388 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
389 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
390 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
391 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
392 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
393 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
394 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
395 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
396 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
397 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
398 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
399 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
400 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
401 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
402 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
403 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
404 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
405 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
406 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
407 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
408 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
409 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
410 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
411 (gcov-dump.o): Remove.
412 (default-c.o): Use COMPILE and POSTCOMPILE.
413 (CFLAGS-gcc.o): New variable.
414 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
416 2013-09-25 Tom Tromey <tromey@redhat.com>
418 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
419 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
420 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
421 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
422 (c-family/c-pragma.o, c-family/c-pretty-print.o)
423 (c-family/c-semantics.o, c-family/c-ada-spec.o)
424 (c-family/array-notation-common.o, c-family/stub-objc.o)
425 (c-family/c-ubsan.o): Remove.
427 2013-09-25 Tom Tromey <tromey@redhat.com>
429 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
431 2013-09-25 Tom Tromey <tromey@redhat.com>
433 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
434 to add -DENABLE_SHARED_LIBGCC.
435 (gcc.o): Don't use subshell.
437 2013-09-25 Tom Tromey <tromey@redhat.com>
439 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
440 * configure.ac: Don't invoke AM_PROG_CC_C_O.
441 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
442 * configure, config.in: Rebuild.
444 2013-09-25 Tom Tromey <tromey@redhat.com>
446 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
447 (COMPILE, POSTCOMPILE): New variables.
448 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
449 (DEPFILES): New variable.
451 * configure.ac: Add checks for dependency checking.
452 * configure, aclocal.m4: Regenerate.
454 2013-09-25 Tom Tromey <tromey@redhat.com>
456 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
457 ($(ALL_HOST_OBJS)): Move order-only dependency to end
460 2013-09-25 Tom Tromey <tromey@redhat.com>
462 * Makefile.in (generated_files): Add options.h,
463 target-hooks-def.h, insn-opinit.h,
464 common/common-target-hooks-def.h, pass-instances.def,
465 c-family/c-target-hooks-def.h.
467 2013-09-25 Jeff Law <law@redhat.com>
469 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
470 than foo[foo.length () - 1] to access last member in a vec.
471 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
473 2013-09-25 Richard Biener <rguenther@suse.de>
476 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
478 2013-09-25 Jan Hubicka <jh@suse.cz>
480 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
483 2013-09-25 Marek Polacek <polacek@redhat.com>
486 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
487 when determining the type name.
489 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
491 * config/sh/sh.md: Fix formatting.
493 2013-09-24 Xinliang David Li <davidxl@google.com>
495 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
496 Check max peel iterations parameter.
497 * param.def: New parameter.
498 * doc/invoke.texi: Document New parameter.
500 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
502 * gimple-pretty-print.c: Various whitespace tweaks.
503 * tree-core.h: Likewise.
504 * tree-pretty-print.c: Likewise.
505 * tree-ssa-alias.c: Likewise.
506 * tree-ssa-copy.c: Likewise.
507 * tree-ssanames.c: Likewise.
508 * tree-ssanames.h: Likewise.
509 * tree-vrp.c: Likewise.
511 2013-09-24 Alan Modra <amodra@gmail.com>
515 * stmt.c (expand_asm_operands): Call expand_expr with
516 EXPAND_MEMORY for output operands that disallow regs. Don't
517 use EXPAND_WRITE on inout operands.
519 2013-09-24 Richard Biener <rguenther@suse.de>
522 * tree.c (reference_alias_ptr_type): Move ...
523 * alias.c (reference_alias_ptr_type): ... here and implement
524 in terms of the new reference_alias_ptr_type_1.
525 (ref_all_alias_ptr_type_p): New helper.
526 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
527 use ref_all_alias_ptr_type_p.
528 (get_deref_alias_set): Add flag_strict_aliasing check here.
529 (reference_alias_ptr_type_1): New function, split out from ...
530 (get_alias_set): ... here.
531 (alias_ptr_types_compatible_p): New function.
532 * alias.h (reference_alias_ptr_type): Declare.
533 (alias_ptr_types_compatible_p): Likewise.
534 * tree.h (reference_alias_ptr_type): Remove.
535 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
536 to compare MEM_REF alias types.
538 2013-09-24 Richard Biener <rguenther@suse.de>
540 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
542 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
544 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
545 reload helper function arrays into a single array reg_addr.
546 (reload_fpr_gpr): Likewise.
547 (reload_gpr_vsx): Likewise.
548 (reload_vsx_gpr): Likewise.
549 (struct rs6000_reg_addr): Likewise.
550 (reg_addr): Likewise.
551 (rs6000_debug_reg_global): Change rs6000_vector_reload,
552 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
553 (rs6000_init_hard_regno_mode_ok): Likewise.
554 (rs6000_secondary_reload_direct_move): Likewise.
555 (rs6000_secondary_reload): Likewise.
557 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
558 constraints: wu, ww, and wy. Repurpose wv constraint added during
559 power8 changes. Put wg constraint in alphabetical order.
561 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
562 for future work to add ISA 2.07 VSX single precision support.
563 (-mvsx-scalar-double): Change default from -1 to 1, update
564 documentation comment.
565 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
566 (-mupper-regs-df): New debug switch to control whether DF values
567 can go in the traditional Altivec registers.
568 (-mupper-regs-sf): New debug switch to control whether SF values
569 can go in the traditional Altivec registers.
571 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
573 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
574 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
575 Add new constraints, wu/ww/wy. Repurpose wv constraint.
576 (rs6000_debug_legitimate_address_p): Print if we are running
577 before, during, or after reload.
578 (rs6000_secondary_reload): Add a comment.
579 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
581 * config/rs6000/constraints.md (wa constraint): Sort w<x>
582 constraints. Update documentation string.
583 (wd constraint): Likewise.
584 (wf constraint): Likewise.
585 (wg constraint): Likewise.
586 (wn constraint): Likewise.
587 (ws constraint): Likewise.
588 (wt constraint): Likewise.
589 (wx constraint): Likewise.
590 (wz constraint): Likewise.
591 (wu constraint): New constraint for ISA 2.07 SFmode scalar
593 (ww constraint): Likewise.
594 (wy constraint): Likewise.
595 (wv constraint): Repurpose ISA 2.07 constraint that we not used in
596 the previous submissions.
597 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
599 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
601 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
602 (REG_BR_PROB): Say that the probability is stored in an int_list.
603 * reg-notes.def: Update commentary to mention INT_LIST.
604 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
606 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
607 * rtlanal.c (int_reg_note_p): New function.
608 (alloc_reg_note): Assert that the note does not have an int argument.
609 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
610 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
611 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
612 rather than an INSN_LIST. Handle INT_LIST.
613 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
614 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
615 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
616 Manipulate them as ints rather than rtxes.
617 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
618 * regmove.c (copy_src_to_dest): Likewise.
619 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
621 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
622 into the cases that need it.
623 * config/arm/arm.c (arm_unwind_emit): Likewise.
625 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
626 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
627 * loop-doloop.c (add_test, doloop_modify): Likewise.
628 * loop-unswitch.c (compare_and_jump_seq): Likewise.
629 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
630 * predict.c (combine_predictions_for_insn): Likewise.
631 * print-rtl.c (print_rtx): Handle INT_LIST.
632 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
633 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
634 * config/arm/arm.c (emit_unlikely_jump): Likewise.
635 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
636 (ix86_split_fp_branch, predict_jump): Likewise.
637 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
638 * config/sh/sh.c (expand_cbranchsi4): Likewise.
639 * config/spu/spu.c (ea_load_store_inline): Likewise.
641 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
642 value of a REG_BR_PROB note.
643 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
644 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
645 * emit-rtl.c (try_split): Likewise.
646 * predict.c (br_prob_note_reliable_p): Likewise.
647 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
648 * reorg.c (mostly_true_jump): Likewise.
649 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
650 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
651 * config/i386/i386.c (ix86_print_operand): Likewise.
652 * config/ia64/ia64.c (ia64_print_operand): Likewise.
653 * config/mmix/mmix.c (mmix_print_operand): Likewise.
654 * config/rs6000/rs6000.c (output_cbranch): Likewise.
655 * config/s390/s390.c (s390_expand_tbegin): Likewise.
656 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
657 * config/sparc/sparc.c (output_cbranch): Likewise.
658 * config/spu/spu.c (get_branch_target): Likewise.
659 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
660 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
662 2013-09-23 Jan Hubicka <jh@suse.cz>
664 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
666 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
667 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
669 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
670 (gimple_fold_call): Dump inconsistent devirtualizations; add
671 sanity check for type based devirtualizations.
672 * ipa-prop.c: Include ipa-utils.h
673 (ipa_intraprocedural_devirtualization): Add sanity check.
674 (try_make_edge_direct_virtual_call): Likewise.
676 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
678 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
679 assignment statements.
681 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
683 * gimple-pretty-print.c (dump_ssaname_info): New function.
684 (dump_gimple_phi): Call it.
685 (pp_gimple_stmt_1): Likewise.
686 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
687 (range_info_def): Declare.
688 * tree-pretty-print.c (pp_double_int): New function.
689 (dump_generic_node): Call it.
690 * tree-pretty-print.h (pp_double_int): Declare.
691 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
692 * tree-ssanames.h (range_info_def): New structure.
693 (value_range_type): Move definition here.
694 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
696 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
698 (set_range_info): New function.
699 (get_range_info): Likewise.
700 (duplicate_ssa_name_range_info): Likewise.
701 (duplicate_ssa_name_fn): Check pointer type and call
702 duplicate_ssa_name_range_info.
703 * tree-ssa-copy.c (fini_copy_prop): Likewise.
704 * tree-vrp.c (value_range_type): Remove definition, now in
706 (vrp_finalize): Call set_range_info to update value range of
708 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
709 (SSA_NAME_RANGE_INFO): New macro.
711 2013-09-23 Richard Biener <rguenther@suse.de>
713 PR tree-optimization/58464
714 * tree-ssa-pre.c (phi_trans_lookup): Remove.
715 (phi_trans_add): Change to add conditionally on being not
717 (phi_translate_1): Remove recursion detection here.
718 (phi_translate): Pre-seed the cache with NULL to catch
719 recursion here in a more generic way.
720 (bitmap_find_leader): Adjust comment.
721 (get_representative_for): Dump value-numbers.
722 (create_expression_by_pieces): Likewise.
723 (insert_into_preds_of_block): Likewise.
725 2013-09-23 Christian Bruel <christian.bruel@st.com>
728 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
729 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
731 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
734 2013-09-20 Renlin Li <renlin.li@arm.com>
736 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
737 (aarch64_expand_epilogue): Likewise.
738 (aarch64_legitimize_reload_address): Likewise.
740 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
742 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
743 NULL_TREE before pushing them onto the vector. Likewise for labels.
745 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
747 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
748 comparison operators when -fno-trapping-math is in effect.
749 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
750 comparison operators in TFmode and assert that unsupported operators
752 (ia64_print_operand): Likewise.
754 2013-09-21 Jan Hubicka <jh@suse.cz>
756 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
757 (sse_typeless_stores): Enable for core
758 (sse_load0_by_pxor): Likewise.
759 (four_jump_limit): Disable for core.
760 (pad_returns): Likewise.
761 (avoid_vector_decode): Likewise.
762 (fuse_cmp_and_branch): Enable for cores.
763 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
765 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
768 * config/pa/pa.c (pa_option_override): Disable auto increment and
769 decrement instructions until reload is completed.
771 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
772 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
773 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
775 2013-09-20 DJ Delorie <dj@redhat.com>
776 Nick Clifton <nickc@redhat.com>
778 * config/rl78/rl78.c: Various whitespace and comment tweaks.
779 (need_to_save): Save bank 0 on interrupts.
780 (characterize_address): Strip far address wrappers.
781 (rl78_as_legitimate_address): Likewise.
782 (transcode_memory_rtx): Likewise.
783 (rl78_peep_movhi_p): Disable this peephole after devirt.
784 (rl78_propogate_register_origins): Forget all origins when a
786 * config/rl78/rl78-virt.md: Various whitespace tweaks.
787 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
789 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
790 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
791 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
793 * config/rl78/constraints.md: Various whitespace and paren tweaks.
795 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
797 * config/pa/pa.md: In "scc" insn patterns, change output template to
798 handle const0_rtx in reg_or_0_operand operands.
800 2013-09-20 Martin Husemann <martin@NetBSD.org>
803 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
804 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
806 2013-09-20 Richard Biener <rguenther@suse.de>
809 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
810 remembering SSA name version and block index.
811 (new_scev_info_str): Adjust.
812 (hash_scev_info): Likewise. Also hash the block index.
813 (eq_scev_info): Adjust.
814 (find_var_scev_info): Likewise.
815 (struct instantiate_cache_entry): Remove.
816 (struct instantiate_cache_type): Use a htab to map name, block
818 (instantiate_cache_type::~instantiate_cache_type): Adjust.
819 (get_instantiated_value_entry): Likewise.
820 (hash_idx_scev_info, eq_idx_scev_info): New functions.
821 (instantiate_scev_name): Adjust.
823 2013-09-20 Jeff Law <law@redhat.com>
825 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
827 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
829 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
830 Call aarch64_simd_expand_args to update op[argc].
832 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
834 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
837 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
839 * gengtype.c (file_rules): Added rule for *.cc files.
840 (get_output_file_with_visibility): Give fatal message when no
843 2013-09-20 Renlin Li <renlin.li@arm.com>
845 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
846 (aarch64_expand_epilogue): Likewise.
847 (aarch64_legitimize_reload_address): Likewise.
849 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
852 * expr.c (expand_assignment): Remove misalignp code path.
854 2013-09-20 Marek Polacek <polacek@redhat.com>
857 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
858 TYPE_PRECISION. Add asserts.
860 2013-09-20 Richard Biener <rguenther@suse.de>
862 PR tree-optimization/58453
863 * tree-loop-distribution.c (distribute_loop): Apply the cost
864 model for -ftree-loop-distribute-patterns, too.
866 2013-09-20 Richard Biener <rguenther@suse.de>
869 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
870 make type comparison less strict.
872 2013-09-20 Alan Modra <amodra@gmail.com>
874 * configure: Regenerate.
875 * aclocal.m4: Regenerate.
877 2013-09-20 Marek Polacek <polacek@redhat.com>
880 * doc/extend.texi: Document that attribute used is meant to be used
881 on variables with static storage duration.
883 2013-09-19 Jakub Jelinek <jakub@redhat.com>
885 PR tree-optimization/58472
886 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
887 simd_lane_access set inv_p = false.
888 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
889 the simduid magic VAR_DECL.
891 2013-09-19 Jan Hubicka <jh@suse.cz>
893 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
895 2013-09-17 Jeff Law <law@redhat.com>
897 * tree-ssa-dom.c (record_temporary_equivalences): New function
898 split out of dom_opt_dom_walker::after_dom_children.
899 (dom_opt_dom_walker::thread_across_edge): Move common code
900 in here from dom_opt_dom_walker::after_dom_children.
901 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
903 2013-09-19 Jan Hubicka <jh@suse.cz>
905 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
906 (TARGET_GENERIC): Use PROCESOR_GENERIC
907 (enum processor_type): Unify generic32 and 64.
908 * i386.md (cpu): Likewise.
909 * x86-tune.def (use_leave): Enable for generic32.
910 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
911 * athlon.md: Change generic64 to generic in all occurences.
912 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
913 (ix86_target_macros_internal): Likewise.
914 * driver-i386.c (host_detect_local_cpu): Likewise.
915 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename to ..
916 (generic_memcpy, generic_memset, generic_cost): This one.
917 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
918 (m_GENERIC32, m_GENERIC64): Remove.
919 (m_GENERIC): Turn into one flag.
920 (processor_target): Unify generic tunnings.
921 (ix86_option_override_internal): Replace generic32/64 by generic.
922 (ix86_issue_rate): Likewise.
923 (ix86_adjust_cost): Likewise.
925 2013-09-19 Jan Hubicka <jh@suse.cz>
927 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
930 2013-09-19 Jakub Jelinek <jakub@redhat.com>
932 * omp-low.c (expand_omp_sections): Always pass len - 1 to
933 GOMP_sections_start, even if !exit_reachable.
935 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
937 * lra-constraints.c (need_for_all_save_p): Use macro
938 HARD_REGNO_CALL_PART_CLOBBERED.
939 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
940 set up pseudo conflict hard regs.
942 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
945 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
948 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
950 PR rtl-optimization/58438
951 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
952 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
953 from previous subpasses.
955 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
957 * arm.c (arm_get_frame_offsets): Validate architecture supports
958 LDRD/STRD before accepting the tuning preference.
959 (arm_expand_prologue): Likewise.
960 (arm_expand_epilogue): Likewise.
962 2013-09-18 Richard Biener <rguenther@suse.de>
964 PR tree-optimization/58417
965 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
966 have chrecs with symbols defined in the loop as operands.
967 (chrec_fold_multiply): Likewise.
968 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
969 parameters before folding binary operations.
970 (struct instantiate_cache_entry_hasher): Remove.
971 (struct instantiate_cache_type): Use a pointer-map.
972 (instantiate_cache_type::instantiate_cache_type): New function.
973 (instantiate_cache_type::get): Likewise.
974 (instantiate_cache_type::set): Likewise.
975 (instantiate_cache_type::~instantiate_cache_type): Adjust.
976 (get_instantiated_value_entry): Likewise.
977 (global_cache): New global.
978 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
979 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
980 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
981 (instantiate_scev_name): Adjust.
982 (instantiate_scev): Construct global instead of local cache.
983 (resolve_mixers): Likewise.
985 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
986 Paolo Carlini <paolo.carlini@oracle.com>
989 * doc/implement-cxx.texi: Fix references to the C++ standards.
991 2013-09-18 Jakub Jelinek <jakub@redhat.com>
993 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
994 * tree-vect-data-refs.c (vect_analyze_data_refs): For
995 simd_lane_access drs, update also DR_ALIGNED_TO.
997 2013-09-18 Marek Polacek <polacek@redhat.com>
1000 * doc/extend.texi: Document no_sanitize_undefined attribute.
1001 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
1002 no_sanitize_undefined attribute.
1004 2013-09-18 Nick Clifton <nickc@redhat.com>
1006 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
1007 (ASM_DECLARE_FUNCTION_NAME): Define.
1009 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
1011 * compare-elim.c (find_comparison_dom_walker): New class
1012 (find_comparisons_in_bb): Rename to
1013 find_comparison_dom_walker::before_dom_children
1014 (find_comparisons): Adjust
1015 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
1017 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
1018 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
1019 (init_walk_dominator_tree): Remove declaration.
1020 (fini_walk_dominator_tree): Remove declaration.
1021 * fwprop.c (single_def_use_dom_walker): New class
1022 (single_def_use_enter_block): Convert to
1023 single_def_use_dom_walker::before_dom_children.
1024 (single_def_use_leave_block): Convert to
1025 single_def_use_dom_walker::after_dom_children.
1026 (build_single_def_use_links): Adjust.
1027 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
1029 (find_candidates_in_block): Convert to
1030 find_candidates_dom_walker::before_dom_children.
1031 (execute_strength_reduction): Adjust.
1032 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
1033 (sese_dom_walker): New class.
1034 (sese_dom_walker::sese_dom_walker): New constructor.
1035 (sese_dom_walker::~sese_dom_walker): New destructor.
1036 (build_sese_conditions_before): Convert to
1037 sese_dom_walker::before_dom_children.
1038 (build_sese_conditions_after): Convert to
1039 sese_dom_walker::after_dom_children.
1040 (build_poly_scop): Adjust
1041 * tree-into-ssa.c (rewrite_dom_walker): New class
1042 (rewrite_enter_block): Convert to
1043 rewrite_dom_walker::before_dom_children.
1044 (rewrite_leave_block): Convert to
1045 rewrite_dom_walker::after_dom_children.
1046 (rewrite_update_dom_walker): New class.
1047 (rewrite_update_enter_block): Convert to
1048 rewrite_update_dom_walker::before_dom_children.
1049 (rewrite_update_leave_block): Convert to
1050 rewrite_update_dom_walker::after_dom_children.
1051 (rewrite_blocks, rewrite_into_ssa): Adjust.
1052 (mark_def_dom_walker): New class.
1053 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
1054 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
1055 (mark_def_sites_blocks): Convert to
1056 mark_def_dom_walker::before_dom_children.
1057 (mark_def_site_blocks): Remove.
1058 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
1059 (tree_ssa_dominator_optimize): Adjust.
1060 (dom_thread_across_edge): Convert to method
1061 dom_opt_dom_walker::thread_across_edge.
1062 (dom_opt_enter_block): Convert to member function
1063 dom_opt_dom_walker::before_dom_children.
1064 (dom_opt_leave_block): Convert to member function
1065 dom_opt_dom_walker::after_dom_children.
1066 * tree-ssa-dse.c (dse_dom_walker): New class.
1067 (dse_enter_block): Convert to member function
1068 dse_dom_walker::before_dom_children.
1069 (tree_ssa_dse): Adjust.
1070 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
1071 (determine_invariantness_stmt): Convert to method
1072 invariantness_dom_walker::before_dom_children.
1073 (determine_invariantness): Remove
1074 (move_computations_dom_walker): New class.
1075 (move_computations_stmt): Convert to method
1076 move_computations_dom_walker::before_dom_children.
1077 (move_computations, tree_ssa_lim): Adjust.
1078 * tree-ssa-phiopt.c (nontrapping_dom_walker): new class
1079 (nt_init_block): Make method
1080 notrappping_dom_walker::before_dom_children.
1081 (nt_fini_block): Make
1082 method nontrapping_dom_walker::after_dom_children.
1083 (get_non_trapping): Adjust.
1084 * tree-ssa-pre.c (eliminate_dom_walker): New class.
1085 (eliminate_bb): Make method eliminate_dom_walker::before_dom_children.
1086 (eliminate_leave_block): Make method.
1087 eliminate_dom_walker::after_dom_children.
1089 * tree-ssa-strlen.c (strlen_dom_walker): New class.
1090 (strlen_enter_block): Make method
1091 strlen_dom_walker::before_dom_children.
1092 (strlen_leave_block): Make
1093 method strlen_dom_walker::after_dom_children.
1094 (tree_ssa_strlen): Adjust.
1095 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
1096 (tree_ssa_uncprop): Adjust.
1097 (uncprop_leave_block): Make method
1098 uncprop_dom_walker::after_dom_children.
1099 (uncprop_leave_block): Make method
1100 uncprop_dom_walker::before_dom_children.
1102 2013-09-18 Bin Cheng <bin.cheng@arm.com>
1104 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn
1105 before branch in same basic block. Check both src and dest of
1108 2013-09-17 Nick Clifton <nickc@redhat.com>
1110 * config/rl78/rl78-real.md (bf): New pattern.
1112 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
1113 (rl78_print_operand): Do not put a # before a %B.
1114 * config/rl78/rl78.opt: Tweak doc strings.
1116 2013-09-17 DJ Delorie <dj@redhat.com>
1118 * config/rl78/constraints.md (Wcv): Allow up to $r31.
1119 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
1120 (rl78_option_override): Likewise, if -mallregs.
1121 (is_virtual_register): Likewise.
1122 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
1123 (REGNO_OK_FOR_BASE_P): Likewise.
1124 * config/rl78/rl78.opt (-mallregs): New.
1126 2013-09-17 Nick Clifton <nickc@redhat.com>
1128 * config/rl78/rl78.c (need_to_save): Change return type to bool.
1129 For interrupt functions: save all call clobbered registers if the
1130 interrupt handler is not a leaf function.
1131 (rl78_expand_prologue): Always recompute the frame information.
1132 For interrupt functions: only select bank 0 if one of the bank 0
1133 registers is going to be psuhed.
1135 2013-09-17 DJ Delorie <dj@redhat.com>
1137 * config/rl78/constraints.md: For each W* constraint, rename to C*
1138 and create a W* constraint that checks for an optional ES: prefix
1140 * config/rl78/rl78.md (UNS_ES_ADDR): New.
1141 (es_addr): New. Used to wrap far addresses.
1142 * config/rl78/rl78-protos.h (rl78_es_addr): New.
1143 (rl78_es_base): New.
1144 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
1145 wrapped far addresses.
1146 (rl78_print_operand_1): Unwrap far addresses before processing.
1147 (rl78_lo16): Wrap far addresses in unspecs.
1148 (rl78_es_addr): New.
1149 (rl78_es_base): New.
1150 (insn_ok_now): Check for not-yet-wrapped far addresses.
1151 (transcode_memory_rtx): Properly re-wrap far addresses.
1153 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
1155 * config/sparc/t-rtems: Add leon3 multilibs.
1157 2013-09-17 Cong Hou <congh@google.com>
1159 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
1160 when checking the dot production pattern. The type of rhs operand
1161 of multiply is now checked correctly.
1163 2013-09-17 Jeff Law <law@redhat.com>
1165 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
1166 edge implied equivalences into successor phis.
1167 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
1168 here from tree-ssa-threadedge.c.
1169 (mark_threaded_blocks): When threading through a joiner, if both
1170 successors of the joiner's clone reach the same block, verify the
1171 PHI arguments are equal. If not, cancel the jump threading request.
1172 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
1173 tree-ssa-threadupdate.c
1174 (thread_across_edge): Don't check PHI argument equality when
1175 threading through joiner block here.
1177 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
1179 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
1180 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
1181 (warn_uninit, warn_uninitialized_vars, execute_early_warn_uninitialized,
1182 make_pass_early_warn_uninitialized): Move to tree-ssa-uninit.c.
1183 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c
1184 (has_undefined_value_p): New. Pass dependant parts of
1185 ssa_undefined_value_p.
1186 (uninit_undefined_value_p): Use has_undefined_value_p.
1187 (warn_uninit, warn_uninitialized_vars, execute_early_warn_uninitialized,
1188 make_pass_early_warn_uninitialized): Move from tree-ssa.c
1189 * tree-ssa.h: Adjust prototypes
1191 2013-09-17 Jan Hubicka <jh@suse.cz>
1194 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
1195 * ipa-inline.c (can_inline_edge_p): Do not downgrade
1196 FUNCTION_NOT_OPTIMIZED.
1197 * ipa-inline-analysis.c (compute_inline_parameters): Function
1198 not optimized is not inlinable unless it is alwaysinline.
1199 (inline_analyze_function): Force calls in not optimized
1200 function not inlinable.
1202 2013-09-17 Jan Hubicka <jh@suse.cz>
1205 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
1207 * ipa.c (function_and_variable_visibility): Likewise.
1208 * ipa-profile.c (ipa_profile): Likewise.
1210 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
1213 * cgraph.c (cgraph_function_body_availability): Check for ifunc
1214 attribute, and don't inline the resolver in this case.
1216 2013-09-17 Teresa Johnson <tejohnson@google.com>
1218 * coverage.c (get_coverage_counts): Add missing newline.
1220 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1222 PR tree-optimization/58088
1223 * fold-const.c (mask_with_trailing_zeros): New function.
1224 (fold_binary_loc): Make sure we don't recurse infinitely
1225 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
1226 Use mask_with_trailing_zeros where appropriate.
1228 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
1230 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
1231 of 'prev' var to get better distance estimation.
1233 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
1235 * tree-inline.h (struct copy_body_data): Add transform_parameter.
1236 * tree-inline.c (is_parameter_of): New predicate.
1237 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
1238 a parameter has been remapped.
1239 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
1240 (optimize_inline_calls): Initialize transform_parameter.
1241 (copy_gimple_seq_and_replace_locals): Likewise.
1242 (tree_function_versioning): Likewise.
1243 (maybe_inline_call_in_expr): Likewise.
1245 2013-09-17 Nick Clifton <nickc@redhat.com>
1247 * config/msp430/msp430-protos.h: Add prototypes for new functions.
1248 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
1250 (is_attr_func): New function.
1251 (msp430_is_interrupt_func): New function.
1252 (is_naked_func): New function.
1253 (is_reentrant_func): New function.
1254 (is_critical_func): New function.
1255 (msp430_start_function): Add annotations for function attributes.
1256 (msp430_attr): New function.
1257 (msp430_attribute_table): New.
1258 (msp430_function_section): New function.
1259 (TARGET_ASM_FUNCTION_SECTION): Define.
1260 (msp430_builtin): New enum.
1261 (msp430_init_builtins): New function.
1262 (msp430_builtin_devl): New function.
1263 (msp430_expand_builtin): New function.
1264 (TARGET_INIT_BUILTINS): Define.
1265 (TARGET_EXPAND_BUILTINS): Define.
1266 (TARGET_BUILTIN_DECL): Define.
1267 (msp430_expand_prologue): Add support for naked, interrupt,
1268 critical and reentrant functions.
1269 (msp430_expand_epilogue): Likewise.
1270 (msp430_print_operand): Handle 'O' character.
1271 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
1273 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
1274 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
1275 (pushm): Use a 'n' rather than an 'i' constraint.
1276 (msp_return): Add generation of the interrupt return instruction.
1277 (disable_interrupts): New pattern.
1278 (enable_interrupts): New pattern.
1279 (push_intr_state): New pattern.
1280 (pop_intr_state): New pattern.
1281 (bic_SR): New pattern.
1282 (bis_SR): New pattern.
1283 * doc/extend.texi: Document MSP430 function attributes and builtin
1286 2013-09-17 Richard Biener <rguenther@suse.de>
1288 PR tree-optimization/58432
1289 * tree-loop-distribution.c (tree_loop_distribution): Also
1290 scan PHIs for outside loop uses and seed a partition from them.
1292 2013-09-17 Bin Cheng <bin.cheng@arm.com>
1294 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
1295 (restructure_reference): Call backtrace_base_for_ref.
1297 2013-09-17 Alan Modra <amodra@gmail.com>
1300 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11 patch.
1302 2013-09-16 DJ Delorie <dj@redhat.com>
1304 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
1305 vregs location for RL78/G10.
1306 (rl78_expand_prologue): Avoid SEL on G10.
1307 (rl78_expand_epilogue): Likewise.
1308 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
1309 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
1310 __RL78_G10__ when appropriate.
1311 (ASM_SPEC): Pass -mg10 along to the assembler.
1312 * config/rl78/rl78.md (sel_rb): Disable for G10.
1313 * config/rl78/rl78.opt: Add -mg10 option.
1314 * config/rl78/t-rl78: Add -mg10 multilib.
1316 2013-09-16 Xinliang David Li <davidxl@google.com>
1318 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
1319 * omp-low.c (omp_max_vf): Ditto.
1320 (expand_omp_simd): Ditto.
1321 * tree-vectorizer.c (vectorize_loops): Ditto.
1322 (gate_vect_slp): Ditto.
1323 (gate_increase_alignment): Ditto.
1324 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
1325 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
1326 (gate_tree_vectorize): Name change.
1327 (tree_vectorize): Ditto.
1328 (pass_vectorize::gate): Call new function.
1329 (pass_vectorize::execute): Ditto.
1330 opts.c: O3 default setting change.
1331 (finish_options): Check new flag.
1332 * doc/invoke.texi: Document new flags.
1333 * common.opt: New flags.
1335 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
1337 * doc/tm.texi.in (Cond Exec Macros): Remove node.
1338 (Condition Code): Don't reference it.
1339 * doc/tm.texi: Regenerate.
1341 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
1344 * lra-constraints.c (undo_optional_reloads): Consider all optional
1345 reload even if it did not get a hard reg.
1347 2013-09-16 Teresa Johnson <tejohnson@google.com>
1349 * dumpfile.c (dump_loc): Remove newline emission.
1350 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
1351 emission to dump_printf_loc calls where missing.
1352 (vect_mark_for_runtime_alias_test): Ditto.
1353 (vect_analyze_data_ref_dependence): Ditto.
1354 (vect_analyze_data_ref_dependences): Ditto.
1355 (vect_slp_analyze_data_ref_dependence): Ditto.
1356 (vect_slp_analyze_data_ref_dependences): Ditto.
1357 (vect_compute_data_ref_alignment): Ditto.
1358 (vect_update_misalignment_for_peel): Ditto.
1359 (vect_verify_datarefs_alignment): Ditto.
1360 (vector_alignment_reachable_p): Ditto.
1361 (vect_get_data_access_cost): Ditto.
1362 (vect_enhance_data_refs_alignment): Ditto.
1363 (vect_find_same_alignment_drs): Ditto.
1364 (vect_analyze_data_refs_alignment): Ditto.
1365 (vect_analyze_group_access): Ditto.
1366 (vect_analyze_data_ref_access): Ditto.
1367 (vect_analyze_data_ref_accesses): Ditto.
1368 (vect_prune_runtime_alias_test_list): Ditto.
1369 (vect_analyze_data_refs): Ditto.
1370 (vect_create_addr_base_for_vector_ref): Ditto.
1371 (vect_create_data_ref_ptr): Ditto.
1372 (vect_grouped_store_supported): Ditto.
1373 (vect_grouped_load_supported): Ditto.
1374 * value-prof.c (check_counter): Ditto.
1375 (check_ic_target): Ditto.
1376 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
1377 (vect_recog_widen_mult_pattern): Ditto.
1378 (vect_recog_widen_sum_pattern): Ditto.
1379 (vect_recog_over_widening_pattern): Ditto.
1380 (vect_recog_widen_shift_pattern): Ditto.
1381 (vect_recog_rotate_pattern): Ditto.
1382 (vect_recog_vector_vector_shift_pattern): Ditto.
1383 (vect_recog_divmod_pattern): Ditto.
1384 (vect_recog_mixed_size_cond_pattern): Ditto.
1385 (vect_recog_bool_pattern): Ditto.
1386 (vect_pattern_recog_1): Ditto.
1387 (vect_pattern_recog): Ditto.
1388 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
1389 (vect_is_simple_iv_evolution): Ditto.
1390 (vect_analyze_scalar_cycles_1): Ditto.
1391 (vect_get_loop_niters): Ditto.
1392 (vect_analyze_loop_1): Ditto.
1393 (vect_analyze_loop_form): Ditto.
1394 (vect_analyze_loop_operations): Ditto.
1395 (vect_analyze_loop_2): Ditto.
1396 (vect_analyze_loop): Ditto.
1397 (report_vect_op): Ditto.
1398 (vect_is_slp_reduction): Ditto.
1399 (vect_is_simple_reduction_1): Ditto.
1400 (vect_get_known_peeling_cost): Ditto.
1401 (vect_estimate_min_profitable_iters): Ditto.
1402 (vect_model_reduction_cost): Ditto.
1403 (vect_model_induction_cost): Ditto.
1404 (get_initial_def_for_induction): Ditto.
1405 (vect_create_epilog_for_reduction): Ditto.
1406 (vectorizable_reduction): Ditto.
1407 (vectorizable_induction): Ditto.
1408 (vectorizable_live_operation): Ditto.
1409 (vect_loop_kill_debug_uses): Ditto.
1410 (vect_transform_loop): Ditto.
1411 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
1412 (vect_stmt_relevant_p): Ditto.
1413 (process_use): Ditto.
1414 (vect_mark_stmts_to_be_vectorized): Ditto.
1415 (vect_model_simple_cost): Ditto.
1416 (vect_model_promotion_demotion_cost): Ditto.
1417 (vect_model_store_cost): Ditto.
1418 (vect_get_store_cost): Ditto.
1419 (vect_model_load_cost): Ditto.
1420 (vect_get_load_cost): Ditto.
1421 (vect_init_vector_1): Ditto.
1422 (vect_get_vec_def_for_operand): Ditto.
1423 (vect_finish_stmt_generation): Ditto.
1424 (vectorizable_call): Ditto.
1425 (vectorizable_conversion): Ditto.
1426 (vectorizable_assignment): Ditto.
1427 (vectorizable_shift): Ditto.
1428 (vectorizable_operation): Ditto.
1429 (vectorizable_store): Ditto.
1430 (vectorizable_load): Ditto.
1431 (vectorizable_condition): Ditto.
1432 (vect_analyze_stmt): Ditto.
1433 (vect_transform_stmt): Ditto.
1434 (vect_is_simple_use): Ditto.
1435 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
1436 (vect_can_advance_ivs_p): Ditto.
1437 (vect_update_ivs_after_vectorizer): Ditto.
1438 (vect_do_peeling_for_loop_bound): Ditto.
1439 (vect_gen_niters_for_prolog_loop): Ditto.
1440 (vect_update_inits_of_drs): Ditto.
1441 (vect_create_cond_for_alias_checks): Ditto.
1442 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
1443 (vect_build_slp_tree_1): Ditto.
1444 (vect_supported_load_permutation_p): Ditto.
1445 (vect_analyze_slp_instance): Ditto.
1446 (vect_analyze_slp): Ditto.
1447 (vect_make_slp_decision): Ditto.
1448 (vect_detect_hybrid_slp): Ditto.
1449 (vect_bb_vectorization_profitable_p): Ditto.
1450 (vect_slp_analyze_bb_1): Ditto.
1451 (vect_update_slp_costs_according_to_vf): Ditto.
1452 (vect_get_mask_element): Ditto.
1453 (vect_transform_slp_perm_load): Ditto.
1454 (vect_schedule_slp_instance): Ditto.
1455 (vect_schedule_slp): Ditto.
1456 (vect_slp_transform_bb): Ditto.
1457 * profile.c (read_profile_edge_counts): Ditto.
1458 (compute_branch_probabilities): Ditto.
1459 * coverage.c (get_coverage_counts): Ditto.
1461 2013-09-16 Diego Novillo <dnovillo@google.com>
1463 * tree-core.h: Add missing comment lines from refactoring
1466 2013-09-16 Jan Hubicka <jh@suse.cz>
1468 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
1469 abstract functions; for static functions check the presence
1472 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
1474 * config/aarch64/aarch64-simd-builtins.def (fma): New.
1475 * config/aarch64/aarch64-simd.md
1476 (aarch64_mla_elt<mode>): New.
1477 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
1478 (aarch64_mls_elt<mode>): Likewise.
1479 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
1480 (aarch64_fma4_elt<mode>): Likewise.
1481 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
1482 (aarch64_fma4_elt_to_128v2df): Likewise.
1483 (aarch64_fma4_elt_to_64df): Likewise.
1484 (fnma<mode>4): Likewise.
1485 (aarch64_fnma4_elt<mode>): Likewise.
1486 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
1487 (aarch64_fnma4_elt_to_128v2df): Likewise.
1488 (aarch64_fnma4_elt_to_64df): Likewise.
1489 * config/aarch64/iterators.md (VDQSF): New.
1490 * config/aarch64/arm_neon.h
1491 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
1492 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
1494 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
1496 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
1497 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
1498 (aarch64_mul3_elt_to_128df): Likewise.
1499 (aarch64_mul3_elt_to_64v2df): Likewise.
1500 * config/aarch64/iterators.md (VEL): Also handle DFmode.
1502 (VMUL_CHANGE_NLANES) Likewise.
1505 * config/aarch64/arm_neon.h
1506 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
1508 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
1510 * config/aarch64/arm_neon.h
1511 (vcvtx_high_f32_f64): Fix parameters.
1513 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1514 Uros Bizjak <ubizjak@gmail.com>
1516 * config/alpha.c: Include tree-ssanames.h.
1518 2013-09-16 Richard Biener <rguenther@suse.de>
1520 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
1521 (dot_rdg_1): Handle control_dd.
1522 (create_edge_for_control_dependence): New function.
1523 (create_rdg_edges): Add control dependences if asked for.
1524 (build_rdg): Likewise.
1525 (generate_loops_for_partition): If there are not necessary
1526 control stmts remove all their dependencies.
1527 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
1528 (distribute_loop): Pass on control dependences.
1529 (tree_loop_distribution): Compute control dependences and remove
1530 restriction on number of loop nodes.
1532 2013-09-16 Jakub Jelinek <jakub@redhat.com>
1534 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
1537 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
1539 * cse.c (try_const_anchors): Punt on CC modes.
1541 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1543 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
1545 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
1548 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
1551 2013-09-14 Iain Sandoe <iain@codesourcery.com>
1554 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
1556 (darwin_objc1_section): Likewise.
1557 (darwin_file_end): Emit Image Info section when required.
1559 2013-09-14 Jan Hubicka <jh@suse.cz>
1561 * tree-into-ssa.c (gate_into_ssa): New.
1562 (pass_data_build_ssa): Use it.
1563 * cgraph.h (expand_thunk): Update prototype.
1564 * cgraphunit.c (analyze_function): Expand thunks early.
1565 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
1566 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
1567 set lowered flag; do not add new function.
1568 (assemble_thunks_and_aliases): Update.
1569 * tree-ssa.c (gate_init_datastructures): New gate.
1570 (pass_data_init_datastructures): Use it.
1572 2013-09-14 Iain Sandoe <iain@codesourcery.com>
1575 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
1576 xmm register set described in the psABI.
1578 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
1580 * dwarf2out.c (should_emit_struct_debug): Add check
1581 for type_decl variable is not NULL.
1583 2013-09-13 Jacek Caban <jacek@codeweavers.com>
1585 * config.gcc: Use new winnt-c.c target hooks
1586 * config/t-winnt: New file
1587 * config/winnt-c.c: New file
1588 * doc/tm.texi.in: Document new hook
1589 * doc/tm.texi: Regenerated
1591 2013-09-13 Jan Hubicka <jh@suse.cz>
1594 * ipa-inline.c (check_callers): New function.
1595 (check_caller_edge): Remove.
1596 (want_inline_function_to_all_callers_p): Also permit alises that are
1598 (inline_to_all_callers): Terminate the walk when devirtualization
1599 introduce new calls.
1601 2013-09-13 Jan Hubicka <jh@suse.cz>
1603 * ipa-inline-analysis.c (struct growth_data): Add node.
1604 (do_estimate_growth_1): Fix detection of recursion.
1606 2013-09-13 Jakub Jelinek <jakub@redhat.com>
1608 PR tree-optimization/58392
1609 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
1610 to avoid shadowing of outer loop variable. If
1611 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
1612 replace_by_duplicate_decl simduid of loops that have it set and
1613 set dest_cfun->has_simduid_loops and/or
1614 dest_cfun->has_force_vect_loops.
1615 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
1616 instead of maybe_lookup_decl.
1617 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
1618 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
1619 call. Copy over force_vect and copy and remap simduid. Set
1620 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
1621 (copy_cfg_body): Remove blocks_to_copy argument. Use
1622 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
1623 caller. Don't set cfun->has_simduid_loops and/or
1624 cfun->has_force_vect_loops here.
1625 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
1627 (expand_call_inline, tree_function_versioning): Adjust copy_body
1630 2013-09-13 Martin Jambor <mjambor@suse.cz>
1633 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
1634 the assert if the edge was a speculative one.
1636 2013-09-13 Richard Biener <rguenther@suse.de>
1638 * tree-data-ref.h (known_dependences_p): Move here ...
1639 * tree-loop-distribution.c (known_dependences_p): ... from here.
1640 (dump_rdg_component, debug_rdg_component): Remove.
1642 (generate_loops_for_partition): Use gimple_uid instead of
1643 relying on matching stmt visit order.
1644 (rdg_build_partitions): Take starting stmt vector.
1645 (ldist_gen): Merge into ...
1646 (distribute_loop): ... this function. Do not compute starting
1648 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
1650 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1652 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
1653 Set type attribute correctly. Set predicable_short_it attribute.
1654 (cmpsi_shiftsi): Remove %? from output template.
1656 2013-09-13 Richard Biener <rguenther@suse.de>
1658 * tree-loop-distribution.c (struct rdg_component,
1659 rdg_defs_used_in_other_loops_p, free_rdg_components,
1660 rdg_build_components): Remove.
1661 (stmts_from_loop): Do not record virtual PHIs.
1662 (generate_loops_for_partition): Skip virtual PHIs.
1663 (build_rdg_partition_for_component): Rename to ...
1664 (build_rdg_partition_for_vertex): ... this and adjust.
1665 (rdg_build_partitions): Take a vector of starting vertices
1666 instead of components. Remove unnecessary leftover handling.
1667 (ldist_gen): Do not build components or record other stores.
1668 (distribute_loop): Do not distribute loops containing stmts
1671 2013-09-13 Christian Bruel <christian.bruel@st.com>
1674 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
1676 2013-09-13 Kai Tietz <ktietz@redhat.com>
1678 * config.gcc: Separate cases for mingw and cygwin targets,
1679 and add 64-bit cygwin target case.
1681 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
1682 dll-export inline-functions.
1683 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
1686 2013-09-13 Jeff Law <law@redhat.com>
1690 2013-09-06 Jeff Law <law@redhat.com>
1692 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
1693 edge implied equivalences into successor phis.
1695 2013-09-12 DJ Delorie <dj@redhat.com>
1697 * config/rl78/rl78-virt.md: Change from | to \; for asm line
1700 2013-09-12 Brooks Moses <bmoses@google.com>
1703 * Makefile.in: Do not install driver binaries in $(target)/bin.
1705 2013-09-12 DJ Delorie <dj@redhat.com>
1707 * config/rl78/rl78.opt (mrelax): New.
1708 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
1709 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
1711 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
1713 (rl78_expand_epilogue): Likewise.
1715 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
1718 * lra-eliminations.c (remove_reg_equal_offset_note): New.
1719 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
1720 pointer elimination with using remove_reg_equal_offset_note.
1722 2013-09-12 DJ Delorie <dj@redhat.com>
1724 * config/msp430/: New port.
1725 * config.gcc (msp430): Added.
1726 * doc/invoke.texi: Document MSP430 options.
1727 * doc/install.texi: Document msp430-elf
1728 * doc/md.texi: Document msp430-elf
1729 * doc/contrib.texi: Document msp430-elf
1731 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
1734 2013-09-12 Martin Jambor <mjambor@suse.cz>
1737 * ipa-prop.c (remove_described_reference): Give up if the edge in the
1738 reference descriptor is NULL.
1739 (ipa_edge_removal_hook): If owning a reference descriptor, set its
1742 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
1744 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
1745 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
1746 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
1747 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
1748 make_temp_ssa_name): move to tree-ssanames.h
1749 * tree-ssa-alias.h: Move prototype.
1750 * tree-ssa.h: Include tree-ssanames.h.
1751 * tree-ssanames.c (FREE_SSANAMES): Move to here.
1752 * tree-ssanames.h: New. Move items from tree-flow*.h
1753 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
1755 2013-09-12 Richard Biener <rguenther@suse.de>
1757 PR tree-optimization/58404
1758 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
1759 propagate non-invariant addresses into dereferences wrapped
1760 in component references.
1762 2013-09-12 Richard Biener <rguenther@suse.de>
1764 PR tree-optimization/58402
1765 * passes.def: Move pass_late_warn_uninitialized later.
1767 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
1769 * tree-ssa.h: New. Move content from tree-flow.h and
1771 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
1772 Move prototypes belonging to tree-ssa.c.
1773 * tree-flow-inline.h (redirect_edge_var_map_def,
1774 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
1776 * gimple.h: Adjust prototypes.
1777 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
1779 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
1780 * tree.h: Move prototype to tree-ssa.h.
1781 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
1782 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
1783 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
1784 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
1785 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
1786 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
1787 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
1788 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
1789 graphite-blocking.c, graphite-clast-to-gimple.c,
1790 graphite-dependences.c, graphite-interchange.c,
1791 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
1792 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
1793 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
1794 ipa-reference.c, ipa-split.c, ipa-utils.c,
1795 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
1796 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
1797 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
1798 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
1799 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
1800 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
1801 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
1802 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
1803 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
1804 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
1805 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
1806 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
1807 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
1808 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
1809 value-prof.c, var-tracking.c,
1810 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
1812 2013-09-12 Richard Biener <rguenther@suse.de>
1814 PR tree-optimization/58396
1815 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
1816 (build_rdg): Take a loop-nest parameter, fix memleaks.
1817 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
1819 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
1821 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
1824 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
1826 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
1827 The FMA instruction names should have a 'v' prefix.
1829 2013-09-12 Richard Biener <rguenther@suse.de>
1831 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
1832 (dot_rdg): Use popen instead of system in optional code.
1833 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
1834 (already_processed_vertex_p): Adjust.
1835 (has_anti_or_output_dependence, predecessor_has_mem_write,
1836 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
1837 rdg_flag_uses): Remove.
1838 (rdg_flag_vertex): Simplify.
1839 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
1841 (build_rdg_partition_for_component): Process the first vertex
1842 of a component only.
1843 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
1845 2013-09-12 Alan Modra <amodra@gmail.com>
1847 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
1849 2013-09-11 DJ Delorie <dj@redhat.com>
1850 Nick Clifton <nickc@redhat.com>
1852 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
1853 (rl78_stack_based_mem): New.
1854 * config/rl78/constraints.md (Iv08): New.
1863 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
1865 (movsi): Change from expand to insn-and-split.
1866 (ashrsi3): Clobber AX.
1870 * config/rl78/rl78.md (CC_REG): Fix.
1871 (addsi3): Allow memory and immediate operands.
1872 (addsi3_internal): Split into...
1873 (addsi3_internal_virt): ...new, and ...
1874 (addsi3_internal_real): ...new.
1876 (subsi3_internal_virt): New.
1877 (subsi3_internal_real): New.
1878 (mulsi3): Add memory operand.
1879 (mulsi3_rl78): Likewise.
1880 (mulsi3_g13): Likewise.
1881 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
1882 (cbranchqi4_real): Add more constraint options.
1883 (cbranchhi4_real): Expand pattern.
1884 (cbranchhi4_real_signed): New.
1885 (cbranchhi4_real_inverted): New.
1886 (cbranchsi4_real_lt): New.
1887 (cbranchsi4_real_ge): New.
1888 (cbranchsi4_real_signed): New.
1889 (cbranchsi4_real): New.
1891 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
1893 (lshrsi3_virt): Likewise.
1894 (ashlsi3_virt): Likewise.
1895 (cbranchqi4_virt_signed): New.
1896 (cbranchhi4_virt_signed): New.
1897 (cbranchsi4_virt): New.
1898 * config/rl78/rl78.c: Whitespace fixes throughout.
1899 (move_elim_pass): New.
1900 (pass_data_rl78_move_elim): New.
1901 (pass_rl78_move_elim): New.
1902 (make_pass_rl78_move_elim): New.
1903 (rl78_devirt_info): Run devirt earlier.
1904 (rl78_move_elim_info): New.
1905 (rl78_asm_file_start): Register it.
1906 (rl78_split_movsi): New.
1907 (rl78_as_legitimate_address): Allow virtual base registers when
1909 (rl78_addr_space_convert): Remove spurious debug stuff.
1910 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
1911 (rl78_print_operand): More cases for not printing '#'.
1912 (rl78_expand_compare): Remove most of the logic.
1913 (content_memory): New.
1914 (clear_content_memory): New.
1915 (get_content_index): New.
1916 (get_content_name): New.
1917 (display_content_memory): New.
1918 (update_content): New.
1919 (record_content): New.
1920 (already_contains): New.
1921 (insn_ok_now): Re-recog insns with virtual registers.
1922 (add_postponed_content_update): New.
1923 (process_postponed_content_update): New.
1924 (gen_and_emit_move): New.
1925 (transcode_memory_rtx): Record new location content.
1926 Use gen_and_emit_move.
1927 (force_into_acc): New.
1928 (move_to_acc): Use gen_and_emit_move.
1929 (move_from_acc): Likewise.
1930 (move_acc_to_reg): Likewise.
1931 (move_to_x): Likewise.
1932 (move_to_hl): Likewise.
1933 (move_to_de): Likewise.
1934 (rl78_alloc_physical_registers_op1): Record location content.
1935 (has_constraint): New.
1936 (rl78_alloc_physical_registers_op2): Record location content.
1938 (rl78_alloc_physical_registers_ro1): Likewise.
1939 (rl78_alloc_physical_registers_cmp): Likewise.
1940 (rl78_alloc_physical_registers_umul): Likewise.
1941 (rl78_alloc_address_registers_macax): New.
1942 (rl78_alloc_physical_registers): Initialize and set location
1943 content memory as needed.
1944 (rl78_reorg): Make sure split2 is called.
1945 (rl78_rtx_costs): New.
1947 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
1949 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
1950 for (not (neg ...)) and (neg (not ...)) cases.
1952 2013-09-11 Richard Biener <rguenther@suse.de>
1955 * passes.def: Split critical edges before late uninit warning passes.
1956 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
1958 2013-09-11 Jakub Jelinek <jakub@redhat.com>
1960 PR tree-optimization/58385
1961 * fold-const.c (build_range_check): If both low and high are NULL,
1962 use omit_one_operand_loc to preserve exp side-effects.
1964 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1966 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
1968 2013-09-11 Richard Biener <rguenther@suse.de>
1970 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
1971 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
1972 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
1973 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
1974 stmts_from_loop, known_dependences_p, build_empty_rdg,
1975 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
1976 * tree-loop-distribution.c: ... here.
1977 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
1978 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
1979 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
1980 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
1981 * tree-loop-distribution.c: ... here.
1982 * tree-loop-distribution.c: Include gimple-pretty-print.h.
1983 (struct partition_s): Add loops member.
1984 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
1985 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
1986 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
1988 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1989 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1990 Sergey Lega <sergey.s.lega@intel.com>
1991 Anna Tikhonova <anna.tikhonova@intel.com>
1992 Ilya Tocar <ilya.tocar@intel.com>
1993 Andrey Turetskiy <andrey.turetskiy@intel.com>
1994 Ilya Verbin <ilya.verbin@intel.com>
1995 Kirill Yukhin <kirill.yukhin@intel.com>
1996 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1998 * config/i386/constraints.md (k): New.
2000 * config/i386/i386.c (const regclass_map): Add new mask registers.
2001 (dbx_register_map): Ditto.
2002 (dbx64_register_map): Ditto.
2003 (svr4_dbx_register_map): Ditto.
2004 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
2006 (ix86_preferred_reload_class): Disable constants for mask registers.
2007 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
2008 (ix86_hard_regno_mode_ok): Support new mask registers.
2009 (x86_order_regs_for_local_alloc): Ditto.
2010 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
2011 (FIXED_REGISTERS): Add new mask registers.
2012 (CALL_USED_REGISTERS): Ditto.
2013 (REG_ALLOC_ORDER): Ditto.
2014 (VALID_MASK_REG_MODE): New.
2015 (FIRST_MASK_REG): Ditto.
2016 (LAST_MASK_REG): Ditto.
2017 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
2018 (MAYBE_MASK_CLASS_P): New.
2019 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
2020 (REG_CLASS_CONTENTS): Ditto.
2021 (MASK_REGNO_P): New.
2022 (ANY_MASK_REG_P): Ditto.
2023 (HI_REGISTER_NAMES): Add new mask registers.
2024 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
2025 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
2027 (attribute "type"): Add mskmov, msklog.
2028 (attribute "length_immediate"): Support them.
2029 (attribute "memory"): Ditto.
2030 (attribute "prefix_0f"): Ditto.
2031 (*movhi_internal): Support new mask registers.
2032 (*movqi_internal): Ditto.
2033 (define_split): Split out clobber pattern is a logic
2034 insn on mask registers.
2035 (*k<logic><mode>): New.
2036 (*andhi_1): Extend to support mask regs.
2037 (*andqi_1): Extend to support mask regs.
2039 (define_split): Split and-not to and and not if operands
2041 (*<code><mode>_1): Separate HI mode to new pattern...
2042 (*<code>hi_1): This.
2043 (*<code>qi_1): Extend to support mask regs.
2045 (kortestzhi): Ditto.
2046 (kortestchi): Ditto.
2048 (*one_cmpl<mode>2_1): Remove HImode and handle it...
2049 (*one_cmplhi2_1): ...Here, now with mask registers support.
2050 (*one_cmplqi2_1): Support new mask registers.
2051 (HI/QImode arithmetics splitter): Don't split if mask registers
2053 (HI/QImode not splitter): Ditto.
2054 * config/i386/predicated.md (mask_reg_operand): New.
2055 (general_reg_operand): Ditto.
2057 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2059 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
2060 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
2062 2013-09-10 Jeff Law <law@redhat.com>
2064 PR tree-optimization/58380
2065 * tree-ssa-threadupdate.c (thread_block): Recognize another case
2066 of threading through a buried loop header.
2068 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
2069 return value for single successor case.
2071 2013-09-10 Jan Hubicka <jh@suse.cz>
2073 * ipa-devirt.c (ipa_devirt): Enable with LTO.
2075 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
2078 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
2079 support conditional execution.
2080 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
2082 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
2084 * lra.c (lra): Clear lra_optional_reload_pseudos before every
2086 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
2087 Check destination too to check move insn.
2088 (undo_optional_reloads): Add check that the original peudo did not
2089 changed its allocation and the optional reload was inherited on last
2090 inheritance pass. Break loop after deciding to keep optional reload.
2091 (lra_undo_inheritance): Add check that inherited pseudo still in
2094 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
2096 * config/aarch64/aarch64.md (generic_sched): New.
2097 * config/aarch64/aarch64-generic.md (load): Make conditional
2098 on generic_sched attribute.
2099 (nonload): Likewise.
2101 2013-09-10 Jan Hubicka <jh@suse.cz>
2103 * lto-cgraph.c: Include ipa-utils.h.
2104 (compute_ltrans_boundary): Also add possible targets into the boundary.
2106 2013-09-10 Jan Hubicka <jh@suse.cz>
2108 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
2109 VAR_DECL of vtable rather than full expression.
2111 2013-09-10 Jan Hubicka <jh@suse.cz>
2112 Paolo Carlini <paolo.carlini@oracle.com>
2114 * cgraphunit.c (analyze_functions): Save input_location, set it
2115 to UNKNOWN_LOCATION and restore it at the end.
2117 2013-09-10 Martin Jambor <mjambor@suse.cz>
2119 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
2120 represented by a thunk.
2122 2013-09-10 Jeff Law <law@redhat.com>
2124 PR tree-optimization/58343
2125 * tree-ssa-threadupdate.c (thread_block): Identify and disable
2126 jump threading requests through loop headers buried in the middle
2127 of a jump threading path.
2129 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
2130 in return value/type.
2132 2013-09-10 Jakub Jelinek <jakub@redhat.com>
2134 PR rtl-optimization/58365
2135 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
2136 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
2139 2013-09-10 Richard Biener <rguenther@suse.de>
2141 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
2142 * tree-data-ref.c (create_rdg_vertices): Collect all data
2143 references, signal failure to the caller, use data-ref API.
2144 (build_rdg): Compute data references only once. Maintain lifetime
2145 of data references and data dependences from within RDG.
2146 (free_rdg): Free dependence relations.
2147 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
2148 inventing extra dependences.
2149 (distribute_loop): Update for RDG API changes.
2151 2013-09-10 Kai Tietz <ktietz@redhat.com>
2153 * doc/invoke.texi (fms-extensions): Document changed
2154 behavior for ms-abi targets.
2155 * config/i386/i386.c (ix86_option_override_internal):
2156 Set default value of option -fms-extension for ms-abi targets.
2158 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
2160 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
2162 2013-09-10 Alan Modra <amodra@gmail.com>
2165 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
2167 2013-09-10 Alan Modra <amodra@gmail.com>
2169 * config/rs6000/predicates.md (add_cint_operand): New.
2170 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
2171 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
2172 using add_cint_operand.
2173 (largetoc_high_plus_aix): Likewise.
2175 2013-09-09 Jakub Jelinek <jakub@redhat.com>
2177 PR tree-optimization/58364
2178 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
2179 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
2180 the current range can't be an unconditional true or false.
2182 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
2184 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
2186 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
2188 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
2190 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
2193 * doc/invoke.texi (-Wdelete-incomplete): Document it.
2195 2013-09-09 Ian Bolton <ian.bolton@arm.com>
2197 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
2198 NO_REGS for immediate that can't be moved directly into FP_REGS.
2200 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2202 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
2203 comparison with negated operand.
2204 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
2207 2013-09-09 Richard Biener <rguenther@suse.de>
2210 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
2211 of a subloop record all its block as affecting loop-closed SSA form.
2213 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
2215 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
2216 of an rtx/bitpos pair.
2217 (store_fixed_bit_field): Update accordingly.
2219 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
2221 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
2223 * builtins.c (expand_errno_check): Likewise.
2224 * dwarf2cfi.c (init_return_column_size): Likewise.
2225 * except.c (sjlj_mark_call_sites): Likewise.
2226 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
2227 * lra-constraints.c (emit_inc): Likewise.
2228 * ree.c (combine_set_extension): Likewise.
2229 * regmove.c (fixup_match_2): Likewise.
2230 * reload1.c (inc_for_reload): Likewise.
2232 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
2234 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
2235 (force_to_mode, simplify_shift_const_1, simplify_comparison):
2236 Use gen_int_mode with the mode of the associated simplify_* call.
2237 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
2238 * expmed.c (expand_shift_1): Likewise.
2239 * function.c (instantiate_virtual_regs_in_insn): Likewise.
2240 * loop-iv.c (iv_number_of_iterations): Likewise.
2241 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
2242 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
2244 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
2246 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
2247 of the associated expand_* call.
2248 (asan_emit_stack_protection): Likewise.
2249 * builtins.c (round_trampoline_addr): Likewise.
2250 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
2251 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
2252 (emit_store_flag): Likewise.
2253 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
2254 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
2256 * function.c (instantiate_virtual_regs_in_insn): Likewise.
2257 * ifcvt.c (noce_try_store_flag_constants): Likewise.
2258 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
2259 * modulo-sched.c (generate_prolog_epilog): Likewise.
2260 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
2261 (expand_ctz, expand_ffs, expand_unop): Likewise.
2263 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
2265 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
2266 of the associated gen_rtx_* call.
2267 * caller-save.c (init_caller_save): Likewise.
2268 * combine.c (find_split_point, make_extraction): Likewise.
2269 (make_compound_operation): Likewise.
2270 * dwarf2out.c (mem_loc_descriptor): Likewise.
2271 * explow.c (plus_constant, probe_stack_range): Likewise.
2272 * expmed.c (expand_mult_const): Likewise.
2273 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
2274 * reload1.c (init_reload): Likewise.
2275 * valtrack.c (cleanup_auto_inc_dec): Likewise.
2276 * var-tracking.c (adjust_mems): Likewise.
2277 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
2278 rather than gen_rtx_fmt_ee.
2280 2013-09-09 Jan Hubicka <jh@suse.cz>
2283 * value-prof.c (gimple_ic): Copy also abnormal edges.
2285 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
2287 * asan.c (asan_shadow_cst): Use gen_int_mode.
2289 2013-09-08 Jan Hubicka <jh@suse.cz>
2291 * ipa-profile.c: Add toplevel comment.
2292 (ipa_propagate_frequency_1): Be more conservative when profile is read.
2293 (contains_hot_call_p): New function.
2294 (ipa_propagate_frequency): Set frequencies based on counts when
2296 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
2297 profile; do not tamper with profile after inlining if it is read.
2299 2013-09-08 Jan Hubicka <jh@suse.cz>
2301 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
2304 2013-09-08 Jan Hubicka <jh@suse.cz>
2306 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
2309 2013-09-08 Jeff Law <law@redhat.com>
2312 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
2315 2013-09-08 Andi Kleen <ak@linux.intel.com>
2317 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
2319 2013-09-08 Jan Hubicka <jh@suse.cz>
2321 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
2323 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
2325 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
2326 for non-debug insns.
2327 * lra.c (new_insn_reg): Take the containing insn as a parameter.
2328 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
2329 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
2332 2013-09-08 Jan Hubicka <jh@suse.cz>
2334 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
2335 targets and devirtualize to BUILT_IN_UNREACHABLE.
2336 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
2337 * ipa.c (walk_polymorphic_call_targets): New function.
2338 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
2339 functions; use the new timevar.
2340 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
2341 was removed from the program.
2342 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
2343 not consider it in the walk when its vtable is dead.
2344 (possible_polymorphic_call_targets_1): Pass anonymous flag to
2346 (devirt_variable_node_removal_hook): New function.
2347 (possible_polymorphic_call_targets): Also register
2348 devirt_variable_node_removal_hook.
2349 (ipa_devirt): Do not do non-speculative devirtualization.
2350 (gate_ipa_devirt): One execute if devirtualizing speculatively.
2352 2013-09-08 Jan Hubicka <jh@suse.cz>
2354 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
2355 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
2356 varpool_remove_variable_insertion_hook): Declare.
2357 * varpool.c (varpool_node_hook_list): New structure.
2358 (first_varpool_node_removal_hook,
2359 first_varpool_variable_insertion_hook): New variables.
2360 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
2361 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
2362 varpool_remove_variable_insertion_hook,
2363 varpool_call_variable_insertion_hooks): New functions.
2364 (varpool_remove_node): Use it.
2366 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
2369 * diagnostic.c (diagnostic_build_prefix): When s.file is
2370 "<built-in>" don't output line and column numbers.
2372 2013-09-06 Jan Hubicka <jh@suse.cz>
2374 * cgraphunit.c (expand_thunk): Get body before touching arguments.
2375 * lto-streamer-out.c: Stream thunks, too.
2376 * lto-streamer-in.c (input_function): Pop cfun here
2377 (lto_read_body): Instead of here.
2379 2013-09-06 Caroline Tice <cmtice@google.com>
2381 * doc/install.texi: Add documentation for the --enable-vtable-verify
2382 and the --disable-libvtv configure options.
2384 2013-09-06 Jeff Law <law@redhat.com>
2386 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
2387 edge implied equivalences into successor phis.
2389 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
2391 * resource.c (mark_referenced_resources): Handle COND_EXEC.
2393 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
2395 * resource.c (mark_target_live_regs): Compute resources taking
2396 into account if a call is predicated or not.
2398 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
2400 * toplev.c (output_stack_usage): Be prepared for suffixes created by
2401 the compiler in the function names.
2403 2013-09-06 Jan Hubicka <jh@suse.cz>
2406 * ipa-inline.c (has_caller_p): New function.
2407 (want_inline_function_to_all_callers_p): Use it.
2408 (sum_callers, inline_to_all_callers): Break out from ...
2409 (ipa_inline): ... here.
2411 2013-09-06 Jan Hubicka <jh@suse.cz>
2413 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
2414 only when AVX is enabled.
2416 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2418 * config/aarch64/aarch64.md
2419 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
2420 is fpsimd_<load/store>2.
2421 (load_pair<mode>): Likewise.
2422 (store_pair<mode>): Likewise.
2424 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2426 * config/arm/types.md (type): Add "mrs" type.
2427 * config/aarch64/aarch64.md
2428 (aarch64_load_tp_hard): Make type "mrs".
2430 (load_tp_hard): Make type "mrs".
2431 * config/arm/cortex-a15.md: Update with new attributes.
2432 * config/arm/cortex-a5.md: Update with new attributes.
2433 * config/arm/cortex-a53.md: Update with new attributes.
2434 * config/arm/cortex-a7.md: Update with new attributes.
2435 * config/arm/cortex-a8.md: Update with new attributes.
2436 * config/arm/cortex-a9.md: Update with new attributes.
2437 * config/arm/cortex-m4.md: Update with new attributes.
2438 * config/arm/cortex-r4.md: Update with new attributes.
2439 * config/arm/fa526.md: Update with new attributes.
2440 * config/arm/fa606te.md: Update with new attributes.
2441 * config/arm/fa626te.md: Update with new attributes.
2442 * config/arm/fa726te.md: Update with new attributes.
2444 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2446 * config/aarch64/aarch64.md
2447 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
2448 (*movtf_aarch64): Likewise.
2450 (thumb1_movdi_insn): Use "multiple" for type where more than one
2451 instruction is used for a move.
2452 (*arm32_movhf): Likewise.
2453 (*thumb_movdf_insn): Likewise.
2455 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2457 * config/arm/types.md (type): Rename fcpys to fmov.
2459 (*arm_movsi_vfp): Rename type fcpys as fmov.
2460 (*thumb2_movsi_vfp): Likewise
2461 (*movhf_vfp_neon): Likewise
2462 (*movhf_vfp): Likewise
2463 (*movsf_vfp): Likewise
2464 (*thumb2_movsf_vfp): Likewise
2465 (*movsfcc_vfp): Likewise
2466 (*thumb2_movsfcc_vfp): Likewise
2467 * config/aarch64/aarch64-simd.md
2468 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
2469 * config/aarch64/aarch64.md
2470 (*movsi_aarch64): Replace type mov_reg with fmovs.
2471 (*movdi_aarch64): Likewise
2472 (*movsf_aarch64): Likewise
2473 (*movdf_aarch64): Likewise
2475 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
2476 * config/arm/iwmmxt.md
2477 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
2478 * config/arm/arm1020e.md: Update with new attributes.
2479 * config/arm/cortex-a15-neon.md: Update with new attributes.
2480 * config/arm/cortex-a5.md: Update with new attributes.
2481 * config/arm/cortex-a53.md: Update with new attributes.
2482 * config/arm/cortex-a7.md: Update with new attributes.
2483 * config/arm/cortex-a8-neon.md: Update with new attributes.
2484 * config/arm/cortex-a9.md: Update with new attributes.
2485 * config/arm/cortex-m4-fpu.md: Update with new attributes.
2486 * config/arm/cortex-r4f.md: Update with new attributes.
2487 * config/arm/marvell-pj4.md: Update with new attributes.
2488 * config/arm/vfp11.md: Update with new attributes.
2490 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2492 * config/aarch64/aarch64.md
2493 (*madd<mode>): Fix type attribute.
2494 (*maddsi_uxtw): Likewise.
2495 (*msub<mode>): Likewise.
2496 (*msubsi_uxtw): Likewise.
2497 (<su_optab>maddsidi4): Likewise.
2498 (<su_optab>msubsidi4): Likewise.
2500 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2502 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
2503 * config/arm/arm.md (core_cycles): Remove fdiv.
2504 * config/arm/vfp.md:
2505 (*sqrtsf2_vfp): Update for attribute changes.
2506 (*sqrtdf2_vfp): Likewise.
2507 * config/aarch64/aarch64.md:
2508 (sqrt<mode>2): Update for attribute changes.
2509 * config/arm/arm1020e.md: Update with new attributes.
2510 * config/arm/cortex-a15-neon.md: Update with new attributes.
2511 * config/arm/cortex-a5.md: Update with new attributes.
2512 * config/arm/cortex-a53.md: Update with new attributes.
2513 * config/arm/cortex-a7.md: Update with new attributes.
2514 * config/arm/cortex-a8-neon.md: Update with new attributes.
2515 * config/arm/cortex-a9.md: Update with new attributes.
2516 * config/arm/cortex-m4-fpu.md: Update with new attributes.
2517 * config/arm/cortex-r4f.md: Update with new attributes.
2518 * config/arm/marvell-pj4.md: Update with new attributes.
2519 * config/arm/vfp11.md: Update with new attributes.
2521 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2523 * config/arm/types.md
2524 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
2525 * config/aarch64/aarch64.md
2526 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
2528 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
2529 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
2530 (float<GPI:mode><GPF:mode>2): Likewise.
2532 (*truncsisf2_vfp): Update with new attributes.
2533 (*truncsidf2_vfp): Likewise.
2534 (fixuns_truncsfsi2): Likewise.
2535 (fixuns_truncdfsi2): Likewise.
2536 (*floatsisf2_vfp): Likewise.
2537 (*floatsidf2_vfp): Likewise.
2538 (floatunssisf2): Likewise.
2539 (floatunssidf2): Likewise.
2540 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
2541 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
2542 * config/arm/arm1020e.md: Update with new attributes.
2543 * config/arm/cortex-a15-neon.md: Update with new attributes.
2544 * config/arm/cortex-a5.md: Update with new attributes.
2545 * config/arm/cortex-a53.md: Update with new attributes.
2546 * config/arm/cortex-a7.md: Update with new attributes.
2547 * config/arm/cortex-a8-neon.md: Update with new attributes.
2548 * config/arm/cortex-a9.md: Update with new attributes.
2549 * config/arm/cortex-m4-fpu.md: Update with new attributes.
2550 * config/arm/cortex-r4f.md: Update with new attributes.
2551 * config/arm/marvell-pj4.md: Update with new attributes.
2552 * config/arm/vfp11.md: Update with new attributes.
2554 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2556 * config/aarch64/arm_neon.h
2557 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
2558 (vqtbx<1,2,3,4><q>_s8): Likewise.
2560 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2562 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
2563 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
2564 (add<mode>3): Add type attribute.
2565 (add<mode>3): Likewise.
2566 (usadd<mode>3): Likewise.
2567 (ssadd<mode>3): Likewise.
2568 (sub<mode>3): Likewise.
2569 (sub<mode>3): Likewise.
2570 (ussub<mode>3): Likewise.
2571 (sssub<mode>3): Likewise.
2572 (ssmulsa3): Likewise.
2573 (usmulusa3): Likewise.
2574 (arm_usatsihi): Likewise.
2576 (*movdi_vfp): Add types for all instructions.
2577 (*movdi_vfp_cortexa8): Likewise.
2578 (*movhf_vfp_neon): Likewise.
2579 (*movhf_vfp): Likewise.
2580 (*movdf_vfp): Likewise.
2581 (*thumb2_movdf_vfp): Likewise.
2582 (*thumb2_movdfcc_vfp): Likewise.
2583 * config/arm/arm.md: Add type attribute to all insn patterns.
2584 (*thumb1_adddi3): Add type attribute.
2585 (*arm_adddi3): Likewise.
2586 (*adddi_sesidi_di): Likewise.
2587 (*adddi_zesidi_di): Likewise.
2588 (*thumb1_addsi3): Likewise.
2589 (addsi3_compare0): Likewise.
2590 (*addsi3_compare0_scratch): Likewise.
2591 (*compare_negsi_si): Likewise.
2592 (cmpsi2_addneg): Likewise.
2593 (*addsi3_carryin_<optab>): Likewise.
2594 (*addsi3_carryin_alt2_<optab>): Likewise.
2595 (*addsi3_carryin_clobercc_<optab>): Likewise.
2596 (*subsi3_carryin): Likewise.
2597 (*subsi3_carryin_const): Likewise.
2598 (*subsi3_carryin_compare): Likewise.
2599 (*subsi3_carryin_compare_const): Likewise.
2600 (*arm_subdi3): Likewise.
2601 (*thumb_subdi3): Likewise.
2602 (*subdi_di_zesidi): Likewise.
2603 (*subdi_di_sesidi): Likewise.
2604 (*subdi_zesidi_di): Likewise.
2605 (*subdi_sesidi_di): Likewise.
2606 (*subdi_zesidi_ze): Likewise.
2607 (thumb1_subsi3_insn): Likewise.
2608 (*arm_subsi3_insn): Likewise.
2609 (*anddi3_insn): Likewise.
2610 (*anddi_zesidi_di): Likewise.
2611 (*anddi_sesdi_di): Likewise.
2612 (*ne_zeroextracts): Likewise.
2613 (*ne_zeroextracts): Likewise.
2614 (*ite_ne_zeroextr): Likewise.
2615 (*ite_ne_zeroextr): Likewise.
2616 (*anddi_notdi_di): Likewise.
2617 (*anddi_notzesidi): Likewise.
2618 (*anddi_notsesidi): Likewise.
2619 (andsi_notsi_si): Likewise.
2620 (thumb1_bicsi3): Likewise.
2621 (*iordi3_insn): Likewise.
2622 (*iordi_zesidi_di): Likewise.
2623 (*iordi_sesidi_di): Likewise.
2624 (*thumb1_iorsi3_insn): Likewise.
2625 (*xordi3_insn): Likewise.
2626 (*xordi_zesidi_di): Likewise.
2627 (*xordi_sesidi_di): Likewise.
2628 (*arm_xorsi3): Likewise.
2629 (*andsi_iorsi3_no): Likewise.
2630 (*smax_0): Likewise.
2631 (*smax_m1): Likewise.
2632 (*arm_smax_insn): Likewise.
2633 (*smin_0): Likewise.
2634 (*arm_smin_insn): Likewise.
2635 (*arm_umaxsi3): Likewise.
2636 (*arm_uminsi3): Likewise.
2637 (*minmax_arithsi): Likewise.
2638 (*minmax_arithsi_): Likewise.
2639 (*satsi_<SAT:code>): Likewise.
2640 (arm_ashldi3_1bit): Likewise.
2641 (arm_ashrdi3_1bit): Likewise.
2642 (arm_lshrdi3_1bit): Likewise.
2643 (*arm_negdi2): Likewise.
2644 (*thumb1_negdi2): Likewise.
2645 (*arm_negsi2): Likewise.
2646 (*thumb1_negsi2): Likewise.
2647 (*negdi_extendsid): Likewise.
2648 (*negdi_zero_extend): Likewise.
2649 (*arm_abssi2): Likewise.
2650 (*thumb1_abssi2): Likewise.
2651 (*arm_neg_abssi2): Likewise.
2652 (*thumb1_neg_abss): Likewise.
2653 (one_cmpldi2): Likewise.
2654 (extend<mode>di2): Likewise.
2655 (*compareqi_eq0): Likewise.
2656 (*arm_extendhisi2addsi): Likewise.
2657 (*arm_movdi): Likewise.
2658 (*thumb1_movdi_insn): Likewise.
2659 (*arm_movt): Likewise.
2660 (*thumb1_movsi_insn): Likewise.
2661 (pic_add_dot_plus_four): Likewise.
2662 (pic_add_dot_plus_eight): Likewise.
2663 (tls_load_dot_plus_eight): Likewise.
2664 (*thumb1_movhi_insn): Likewise.
2665 (*thumb1_movsf_insn): Likewise.
2666 (*movdf_soft_insn): Likewise.
2667 (*thumb_movdf_insn): Likewise.
2668 (cbranchsi4_insn): Likewise.
2669 (cbranchsi4_scratch): Likewise.
2670 (*negated_cbranchsi4): Likewise.
2671 (*tbit_cbranch): Likewise.
2672 (*tlobits_cbranch): Likewise.
2673 (*tstsi3_cbranch): Likewise.
2674 (*cbranchne_decr1): Likewise.
2675 (*addsi3_cbranch): Likewise.
2676 (*addsi3_cbranch_scratch): Likewise.
2677 (*arm_cmpdi_insn): Likewise.
2678 (*arm_cmpdi_unsig): Likewise.
2679 (*arm_cmpdi_zero): Likewise.
2680 (*thumb_cmpdi_zero): Likewise.
2681 (*deleted_compare): Likewise.
2682 (*mov_scc): Likewise.
2683 (*mov_negscc): Likewise.
2684 (*mov_notscc): Likewise.
2685 (*cstoresi_eq0_thumb1_insn): Likewise.
2686 (cstoresi_nltu_thumb1): Likewise.
2687 (cstoresi_ltu_thu): Likewise.
2688 (thumb1_addsi3_addgeu): Likewise.
2689 (*arm_jump): Likewise.
2690 (*thumb_jump): Likewise.
2691 (*check_arch2): Likewise.
2692 (arm_casesi_internal): Likewise.
2693 (thumb1_casesi_dispatch): Likewise.
2694 (*arm_indirect_jump): Likewise.
2695 (*thumb1_indirect_jump): Likewise.
2697 (*and_scc): Likewise.
2698 (*ior_scc): Likewise.
2699 (*compare_scc): Likewise.
2700 (*cond_move): Likewise.
2701 (*cond_arith): Likewise.
2702 (*cond_sub): Likewise.
2703 (*cmp_ite0): Likewise.
2704 (*cmp_ite1): Likewise.
2705 (*cmp_and): Likewise.
2706 (*cmp_ior): Likewise.
2707 (*ior_scc_scc): Likewise.
2708 (*ior_scc_scc_cmp): Likewise.
2709 (*and_scc_scc): Likewise.
2710 (*and_scc_scc_cmp): Likewise.
2711 (*and_scc_scc_nod): Likewise.
2712 (*negscc): Likewise.
2713 (movcond_addsi): Likewise.
2714 (movcond): Likewise.
2715 (*ifcompare_plus_move): Likewise.
2716 (*if_plus_move): Likewise.
2717 (*ifcompare_move_plus): Likewise.
2718 (*if_move_plus): Likewise.
2719 (*ifcompare_arith_arith): Likewise.
2720 (*if_arith_arith): Likewise.
2721 (*ifcompare_arith_move): Likewise.
2722 (*if_arith_move): Likewise.
2723 (*ifcompare_move_arith): Likewise.
2724 (*if_move_arith): Likewise.
2725 (*ifcompare_move_not): Likewise.
2726 (*if_move_not): Likewise.
2727 (*ifcompare_not_move): Likewise.
2728 (*if_not_move): Likewise.
2729 (*ifcompare_shift_move): Likewise.
2730 (*if_shift_move): Likewise.
2731 (*ifcompare_move_shift): Likewise.
2732 (*if_move_shift): Likewise.
2733 (*ifcompare_shift_shift): Likewise.
2734 (*ifcompare_not_arith): Likewise.
2735 (*ifcompare_arith_not): Likewise.
2736 (*if_arith_not): Likewise.
2737 (*ifcompare_neg_move): Likewise.
2738 (*if_neg_move): Likewise.
2739 (*ifcompare_move_neg): Likewise.
2740 (*if_move_neg): Likewise.
2741 (prologue_thumb1_interwork): Likewise.
2742 (*cond_move_not): Likewise.
2743 (*sign_extract_onebit): Likewise.
2744 (*not_signextract_onebit): Likewise.
2745 (stack_tie): Likewise.
2746 (align_4): Likewise.
2747 (align_8): Likewise.
2748 (consttable_end): Likewise.
2749 (consttable_1): Likewise.
2750 (consttable_2): Likewise.
2751 (consttable_4): Likewise.
2752 (consttable_8): Likewise.
2753 (consttable_16): Likewise.
2754 (*thumb1_tablejump): Likewise.
2755 (prefetch): Likewise.
2756 (force_register_use): Likewise.
2757 (thumb_eh_return): Likewise.
2758 (load_tp_hard): Likewise.
2759 (load_tp_soft): Likewise.
2760 (tlscall): Likewise.
2761 (*arm_movtas_ze): Likewise.
2762 (*arm_rev): Likewise.
2763 (*arm_revsh): Likewise.
2764 (*arm_rev16): Likewise.
2765 * config/arm/thumb2.md
2766 (*thumb2_smaxsi3): Likewise.
2767 (*thumb2_sminsi3): Likewise.
2768 (*thumb32_umaxsi3): Likewise.
2769 (*thumb2_uminsi3): Likewise.
2770 (*thumb2_negdi2): Likewise.
2771 (*thumb2_abssi2): Likewise.
2772 (*thumb2_neg_abss): Likewise.
2773 (*thumb2_movsi_insn): Likewise.
2774 (tls_load_dot_plus_four): Likewise.
2775 (*thumb2_movhi_insn): Likewise.
2776 (*thumb2_mov_scc): Likewise.
2777 (*thumb2_mov_negs): Likewise.
2778 (*thumb2_mov_negs): Likewise.
2779 (*thumb2_mov_nots): Likewise.
2780 (*thumb2_mov_nots): Likewise.
2781 (*thumb2_movsicc_): Likewise.
2782 (*thumb2_movsfcc_soft_insn): Likewise.
2783 (*thumb2_indirect_jump): Likewise.
2784 (*thumb2_and_scc): Likewise.
2785 (*thumb2_ior_scc): Likewise.
2786 (*thumb2_ior_scc_strict_it): Likewise.
2787 (*thumb2_cond_move): Likewise.
2788 (*thumb2_cond_arith): Likewise.
2789 (*thumb2_cond_ari): Likewise.
2790 (*thumb2_cond_sub): Likewise.
2791 (*thumb2_negscc): Likewise.
2792 (*thumb2_movcond): Likewise.
2793 (thumb2_casesi_internal): Likewise.
2794 (thumb2_casesi_internal_pic): Likewise.
2795 (*thumb2_alusi3_short): Likewise.
2796 (*thumb2_mov<mode>_shortim): Likewise.
2797 (*thumb2_addsi_short): Likewise.
2798 (*thumb2_subsi_short): Likewise.
2799 (thumb2_addsi3_compare0): Likewise.
2800 (*thumb2_cbz): Likewise.
2801 (*thumb2_cbnz): Likewise.
2802 (*thumb2_one_cmplsi2_short): Likewise.
2803 (*thumb2_negsi2_short): Likewise.
2804 (*orsi_notsi_si): Likewise.
2805 * config/arm/arm1020e.md: Update with new attributes.
2806 * config/arm/arm1026ejs.md: Update with new attributes.
2807 * config/arm/arm1136jfs.md: Update with new attributes.
2808 * config/arm/arm926ejs.md: Update with new attributes.
2809 * config/arm/cortex-a15.md: Update with new attributes.
2810 * config/arm/cortex-a5.md: Update with new attributes.
2811 * config/arm/cortex-a53.md: Update with new attributes.
2812 * config/arm/cortex-a7.md: Update with new attributes.
2813 * config/arm/cortex-a8.md: Update with new attributes.
2814 * config/arm/cortex-a9.md: Update with new attributes.
2815 * config/arm/cortex-m4.md: Update with new attributes.
2816 * config/arm/cortex-r4.md: Update with new attributes.
2817 * config/arm/fa526.md: Update with new attributes.
2818 * config/arm/fa606te.md: Update with new attributes.
2819 * config/arm/fa626te.md: Update with new attributes.
2820 * config/arm/fa726te.md: Update with new attributes.
2822 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2824 * config/aarch64/aarch64-simd.md
2825 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
2826 <vwx> iterator to ensure correct register choice.
2827 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
2828 (aarch64_sqdmull_n<mode>): Likewise.
2829 (aarch64_sqdmull2_n<mode>_internal): Likewise.
2830 * config/aarch64/arm_neon.h
2831 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
2832 (vml<as><q>_n_<su>16): Likewise.
2833 (vml<as>l_high_lane<q>_<su>16): Likewise.
2834 (vml<as>l_high_n_<su>16): Likewise.
2835 (vml<as>l_lane<q>_<su>16): Likewise.
2836 (vml<as>l_n_<su>16): Likewise.
2837 (vmul<q>_lane<q>_<su>16): Likewise.
2838 (vmul<q>_n_<su>16): Likewise.
2839 (vmull_lane<q>_<su>16): Likewise.
2840 (vmull_n_<su>16): Likewise.
2841 (vmull_high_lane<q>_<su>16): Likewise.
2842 (vmull_high_n_<su>16): Likewise.
2843 (vqrdmulh<q>_n_s16): Likewise.
2845 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
2847 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
2848 have the correct lane parameter.
2850 2013-09-06 Richard Biener <rguenther@suse.de>
2852 * cfganal.c (control_dependences::~control_dependences):
2853 Properly free all of the vector.
2855 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
2858 * config/i386/i386.c (ix86_conditional_register_usage):
2859 Proper initialize extended SSE registers.
2861 2013-09-06 Jan Hubicka <jh@suse.cz>
2863 PR tree-optimization/58311
2864 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
2866 2013-09-06 Jan Hubicka <jh@suse.cz>
2868 * Makefile.in (tree-sra.o): Update dependencies.
2869 * tree-sra.c: Include ipa-utils.h
2870 (scan_function): Use recursive_call_p.
2871 (has_caller_p): New function.
2872 (cgraph_for_node_and_aliases): Count also callers of aliases.
2874 2013-09-06 Jan Hubicka <jh@suse.cz>
2877 * cgraph.h (symtab_semantically_equivalent_p): Declare.
2878 * tree-tailcall.c: Include ipa-utils.h.
2879 (find_tail_calls): Use it.
2880 * ipa-pure-const.c (check_call): Likewise.
2881 * ipa-utils.c (recursive_call_p): New function.
2882 * ipa-utils.h (recursive_call_p): Dclare.
2883 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
2884 (symtab_semantically_equivalent_p): New function.
2885 * Makefile.in (tree-tailcall.o): Update dependencies.
2887 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
2889 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
2892 2013-09-06 Richard Biener <rguenther@suse.de>
2894 * lto-streamer.h (lto_global_var_decls): Remove.
2895 * Makefile.in (OBJS): Remove lto-symtab.o.
2896 (lto-symtab.o): Remove.
2897 (GTFILES): Remove lto-symtab.c
2898 * lto-symtab.c: Move to lto/
2900 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2902 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
2903 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
2904 (UNSPEC_FPINT_RINT): New constant definitions.
2905 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
2906 definition with 2 attributes.
2907 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
2908 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
2911 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2913 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
2914 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
2915 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
2916 attribute to "z196_alone".
2917 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
2920 2013-09-06 Richard Biener <rguenther@suse.de>
2922 * basic-block.h (class control_dependences): New.
2923 * tree-ssa-dce.c (control_dependence_map): Remove.
2925 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
2926 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
2927 find_pdom, find_control_dependence, find_all_control_dependences):
2929 (mark_control_dependent_edges_necessary,
2930 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
2931 tree_dce_done, perform_tree_ssa_dce): Adjust.
2932 * cfganal.c (set_control_dependence_map_bit,
2933 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
2934 find_all_control_dependences): Move from tree-ssa-dce.c and
2935 implement as methods of control_dependences class.
2936 (control_dependences::control_dependences): New.
2937 (control_dependences::~control_dependences): Likewise.
2938 (control_dependences::get_edges_dependent_on): Likewise.
2939 (control_dependences::get_edge): Likewise.
2941 2013-09-04 Jan Hubicka <jh@suse.cz>
2943 * tree.c (types_same_for_odr): Drop overactive check.
2944 * ipa-devirt.c (hash_type_name): Likewise.
2946 2013-09-04 Jan Hubicka <jh@suse.cz>
2948 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
2949 (analyze_functions): ... here.
2951 2013-09-04 Jan Hubicka <jh@suse.cz>
2954 * cgraphunit.c (analyze_functions): Clear AUX fields
2955 after processing; initialize assembler name has.
2957 2013-09-05 Jeff Law <law@redhat.com>
2959 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
2960 from thread_around_empty_block. Record threading path into PATH.
2961 Recurse if threading through the initial block is successful.
2962 (thread_across_edge): Corresponding changes to slightly simplify.
2964 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
2966 * config/aarch64/aarch64.md
2967 (type): Remove frecpe, frecps, frecpx.
2968 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
2969 fix to be a TARGET_SIMD instruction.
2970 (aarch64_frecps): Remove.
2971 * config/aarch64/aarch64-simd.md
2972 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
2973 (aarch64_frecps<mode>): Handle all float/vector of float modes.
2975 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
2976 Sofiane Naci <sofiane.naci@arm.com>
2978 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
2979 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
2980 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
2981 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
2982 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
2983 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
2984 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
2985 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
2986 into "clz, "rbit". Rename "shift" to "shift_imm".
2987 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
2988 changes. Update for attribute changes all occurrences of arlo_* and
2990 * config/arm/arm-fixed.md: Update for attribute changes
2991 all occurrences of arlo_* types.
2992 * config/arm/thumb2.md: Update for attribute changes all occurrences
2994 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
2995 (cortexa7_older_only): Likewise.
2996 (cortexa7_younger): Likewise.
2997 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
2998 (1020alu_shift_op): Likewise.
2999 (1020alu_shift_reg_op): Likewise.
3000 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
3001 (alu_shift_op): Likewise.
3002 (alu_shift_reg_op): Likewise.
3003 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
3004 (11_alu_shift_op): Likewise.
3005 (11_alu_shift_reg_op): Likewise.
3006 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
3007 (9_alu_shift_reg_op): Likewise.
3008 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
3010 (cortex_a15_alu_shift): Likewise.
3011 (cortex_a15_alu_shift_reg): Likewise.
3012 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
3014 (cortex_a5_alu_shift): Likewise.
3015 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
3017 (cortex_a53_alu_shift): Likewise.
3018 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
3020 (cortex_a7_alu_reg): Likewise.
3021 (cortex_a7_alu_shift): Likewise.
3022 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
3024 (cortex_a8_alu_shift): Likewise.
3025 (cortex_a8_alu_shift_reg): Likewise.
3026 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
3027 (cortex_a9_dp_shift): Likewise.
3028 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
3030 * config/arm/cortex-r4.md
3031 (cortex_r4_alu): Update for attribute changes.
3032 (cortex_r4_mov): Likewise.
3033 (cortex_r4_alu_shift_reg): Likewise.
3034 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
3035 (526_alu_shift_op): Likewise.
3036 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
3037 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
3038 (626te_alu_shift_op): Likewise.
3039 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
3040 (726te_alu_shift_op): Likewise.
3041 (726te_alu_shift_reg_op): Likewise.
3042 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
3043 (mp626_alu_shift_op): Likewise.
3044 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
3045 (pj4_alu_conds): Likewise.
3046 (pj4_shift): Likewise.
3047 (pj4_shift_conds): Likewise.
3048 (pj4_alu_shift): Likewise.
3049 (pj4_alu_shift_conds): Likewise.
3050 * config/aarch64/aarch64.md: Update for attribute change
3051 all occurrences of arlo_* and shift* types.
3053 2013-09-05 Mike Stump <mikestump@comcast.net>
3055 * tree.h: Move documentation for tree_function_decl to tree-core.h
3056 with the declaration.
3058 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
3061 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
3062 looking for widest mode.
3064 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
3066 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
3068 2013-09-05 Richard Biener <rguenther@suse.de>
3070 PR tree-optimization/58137
3071 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
3072 Do not create vectors of pointers.
3073 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
3074 types for the components of the vector initializer.
3075 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
3076 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
3078 2013-09-05 Martin Jambor <mjambor@suse.cz>
3080 * ipa-prop.c (remove_described_reference): Accept missing references,
3081 return false if that hppens, otherwise return true.
3082 (cgraph_node_for_jfunc): New function.
3083 (try_decrement_rdesc_refcount): Likewise.
3084 (try_make_edge_direct_simple_call): Use them.
3085 (ipa_edge_removal_hook): Remove references from rdescs.
3086 (ipa_edge_duplication_hook): Clone rdescs and their references
3087 when the new edge has the same caller as the old one.
3088 * cgraph.c (cgraph_resolve_speculation): Remove speculative
3089 reference before removing any edges.
3091 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
3093 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
3095 * thumb2.md (thumb2_storewb_parisi): New pattern.
3097 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
3099 * config/aarch64/aarch64-option-extensions.def: Add
3100 AARCH64_OPT_EXTENSION of 'crc'.
3101 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
3102 (AARCH64_ISA_CRC): Ditto.
3103 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
3104 description of the CRC extension.
3106 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
3108 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
3109 * config/rs6000/linux.h: Ditto.
3110 * alpha/linux.h: Ditto.
3111 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
3112 no_c99_libc_has_function.
3113 * config/c6x/uclinux-elf.h: Ditto.
3114 * config/lm32/uclinux-elf.h: Ditto.
3115 * config/m68k/uclinux.h: Ditto.
3116 * config/moxie/uclinux.h: Ditto.
3117 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
3118 (crisv32-*-linux*, cris-*-linux*): Ditto.
3119 * config/bfin/bfin.c: Include "tm_p.h".
3121 2013-09-05 Richard Biener <rguenther@suse.de>
3123 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
3124 check for a definition without a basic-block.
3126 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
3127 Sofiane Naci <sofiane.naci@arm.com>
3129 * config/aarch64/aarch64.md
3130 (*movti_aarch64): Rename r_2_f and f_2_r.
3131 (*movsf_aarch64): Likewise.
3132 (*movdf_aarch64): Likewise.
3133 (*movtf_aarch64): Likewise.
3134 (aarch64_movdi_<mode>low): Likewise.
3135 (aarch64_movdi_<mode>high): Likewise.
3136 (aarch64_mov<mode>high_di): Likewise.
3137 (aarch64_mov<mode>low_di): Likewise.
3138 (aarch64_movtilow_tilow): Likewise.
3139 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
3140 values to config/arm/types.md
3141 (attribute "conds"): Update for attribute change.
3142 (anddi3_insn): Likewise.
3143 (iordi3_insn): Likewise.
3144 (xordi3_insn): Likewise.
3145 (one_cmpldi2): Likewise.
3146 * config/arm/types.md (type): Add Neon types.
3147 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
3148 use "type" attribute.
3149 (movmisalign<mode>_neon_store): Likewise.
3150 (movmisalign<mode>_neon_load): Likewise.
3151 (vec_set<mode>_internal): Likewise.
3152 (vec_setv2di_internal): Likewise.
3153 (vec_extract<mode>): Likewise.
3154 (vec_extractv2di): Likewise.
3155 (add<mode>3_neon): Likewise.
3156 (adddi3_neon): Likewise.
3157 (sub<mode>3_neon): Likewise.
3158 (subdi3_neon): Likewise.
3159 (mul<mode>3_neon): Likewise.
3160 (mul<mode>3add<mode>_neon): Likewise.
3161 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
3162 (fma<VCVTF:mode>4)): Likewise.
3163 (fma<VCVTF:mode>4_intrinsic): Likewise.
3164 (fmsub<VCVTF:mode>4)): Likewise.
3165 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
3166 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
3167 (ior<mode>3): Likewise.
3168 (and<mode>3): Likewise.
3169 (anddi3_neon): Likewise.
3170 (orn<mode>3_neon): Likewise.
3171 (orndi3_neon): Likewise.
3172 (bic<mode>3_neon): Likewise.
3173 (bicdi3_neon): Likewise.
3174 (xor<mode>3): Likewise.
3175 (one_cmpl<mode>2): Likewise.
3176 (abs<mode>2): Likewise.
3177 (neg<mode>2): Likewise.
3178 (umin<mode>3_neon): Likewise.
3179 (umax<mode>3_neon): Likewise.
3180 (smin<mode>3_neon): Likewise.
3181 (smax<mode>3_neon): Likewise.
3182 (vashl<mode>3): Likewise.
3183 (vashr<mode>3_imm): Likewise.
3184 (vlshr<mode>3_imm): Likewise.
3185 (ashl<mode>3_signed): Likewise.
3186 (ashl<mode>3_unsigned): Likewise.
3187 (neon_load_count): Likewise.
3188 (ashldi3_neon_noclobber): Likewise.
3189 (signed_shift_di3_neon): Likewise.
3190 (unsigned_shift_di3_neon): Likewise.
3191 (ashrdi3_neon_imm_noclobber): Likewise.
3192 (lshrdi3_neon_imm_noclobber): Likewise.
3193 (widen_ssum<mode>3): Likewise.
3194 (widen_usum<mode>3): Likewise.
3195 (quad_halves_<code>v4si): Likewise.
3196 (quad_halves_<code>v4sf): Likewise.
3197 (quad_halves_<code>v8hi): Likewise.
3198 (quad_halves_<code>v16qi): Likewise.
3199 (reduc_splus_v2di): Likewise.
3200 (neon_vpadd_internal<mode>): Likewise.
3201 (neon_vpsmin<mode>): Likewise.
3202 (neon_vpsmax<mode>): Likewise.
3203 (neon_vpumin<mode>): Likewise.
3204 (neon_vpumax<mode>): Likewise.
3205 (ss_add<mode>_neon): Likewise.
3206 (us_add<mode>_neon): Likewise.
3207 (ss_sub<mode>_neon): Likewise.
3208 (us_sub<mode>_neon): Likewise.
3209 (neon_vadd<mode>_unspec): Likewise.
3210 (neon_vaddl<mode>): Likewise.
3211 (neon_vaddw<mode>): Likewise.
3212 (neon_vhadd<mode>): Likewise.
3213 (neon_vqadd<mode>): Likewise.
3214 (neon_vaddhn<mode>): Likewise.
3215 (neon_vmul<mode>): Likewise.
3216 (neon_vmla<mode>): Likewise.
3217 (neon_vmlal<mode>): Likewise.
3218 (neon_vmls<mode>): Likewise.
3219 (neon_vmlsl<mode>): Likewise.
3220 (neon_vqdmulh<mode>): Likewise.
3221 (neon_vqdmlal<mode>): Likewise.
3222 (neon_vqdmlsl<mode>): Likewise.
3223 (neon_vmull<mode>): Likewise.
3224 (neon_vqdmull<mode>): Likewise.
3225 (neon_vsub<mode>_unspec): Likewise.
3226 (neon_vsubl<mode>): Likewise.
3227 (neon_vsubw<mode>): Likewise.
3228 (neon_vqsub<mode>): Likewise.
3229 (neon_vhsub<mode>): Likewise.
3230 (neon_vsubhn<mode>): Likewise.
3231 (neon_vceq<mode>): Likewise.
3232 (neon_vcge<mode>): Likewise.
3233 (neon_vcgeu<mode>): Likewise.
3234 (neon_vcgt<mode>): Likewise.
3235 (neon_vcgtu<mode>): Likewise.
3236 (neon_vcle<mode>): Likewise.
3237 (neon_vclt<mode>): Likewise.
3238 (neon_vcage<mode>): Likewise.
3239 (neon_vcagt<mode>): Likewise.
3240 (neon_vtst<mode>): Likewise.
3241 (neon_vabd<mode>): Likewise.
3242 (neon_vabdl<mode>): Likewise.
3243 (neon_vaba<mode>): Likewise.
3244 (neon_vabal<mode>): Likewise.
3245 (neon_vmax<mode>): Likewise.
3246 (neon_vmin<mode>): Likewise.
3247 (neon_vpaddl<mode>): Likewise.
3248 (neon_vpadal<mode>): Likewise.
3249 (neon_vpmax<mode>): Likewise.
3250 (neon_vpmin<mode>): Likewise.
3251 (neon_vrecps<mode>): Likewise.
3252 (neon_vrsqrts<mode>): Likewise.
3253 (neon_vqabs<mode>): Likewise.
3254 (neon_vqneg<mode>): Likewise.
3255 (neon_vcls<mode>): Likewise.
3256 (clz<mode>2): Likewise.
3257 (popcount<mode>2): Likewise.
3258 (neon_vrecpe): Likewise.
3259 (neon_vrsqrte): Likewise.
3260 (neon_vget_lane<mode>_sext_internal): Likewise.
3261 (neon_vget_lane<mode>_zext_internal): Likewise.
3262 (neon_vdup_n<mode>): Likewise.
3263 (neon_vdup_nv2di): Likewise.
3264 (neon_vdpu_lane<mode>_internal): Likewise.
3265 (neon_vswp<mode>): Likewise.
3266 (float<mode><V_cvtto>2): Likewise.
3267 (floatuns<mode><V_cvtto>2): Likewise.
3268 (fix_trunc<mode><V_cvtto>)2): Likewise
3269 (fixuns_trunc<mode><V_cvtto)2): Likewise.
3270 (neon_vcvt<mode>): Likewise.
3271 (neon_vcvtv4sfv4hf): Likewise.
3272 (neon_vcvtv4hfv4sf): Likewise.
3273 (neon_vcvt_n<mode>): Likewise.
3274 (neon_vmovn<mode>): Likewise.
3275 (neon_vqmovn<mode>): Likewise.
3276 (neon_vqmovun<mode>): Likewise.
3277 (neon_vmovl<mode>): Likewise.
3278 (neon_vmul_lane<mode>): Likewise.
3279 (neon_vmull_lane<mode>): Likewise.
3280 (neon_vqdmull_lane<mode>): Likewise.
3281 (neon_vqdmulh_lane<mode>): Likewise.
3282 (neon_vmla_lane<mode>): Likewise.
3283 (neon_vmlal_lane<mode>): Likewise.
3284 (neon_vqdmlal_lane<mode>): Likewise.
3285 (neon_vmls_lane<mode>): Likewise.
3286 (neon_vmlsl_lane<mode>): Likewise.
3287 (neon_vqdmlsl_lane<mode>): Likewise.
3288 (neon_vext<mode>): Likewise.
3289 (neon_vrev64<mode>): Likewise.
3290 (neon_vrev32<mode>): Likewise.
3291 (neon_vrev16<mode>): Likewise.
3292 (neon_vbsl<mode>_internal): Likewise.
3293 (neon_vshl<mode>): Likewise.
3294 (neon_vqshl<mode>): Likewise.
3295 (neon_vshr_n<mode>): Likewise.
3296 (neon_vshrn_n<mode>): Likewise.
3297 (neon_vqshrn_n<mode>): Likewise.
3298 (neon_vqshrun_n<mode>): Likewise.
3299 (neon_vshl_n<mode>): Likewise.
3300 (neon_vqshl_n<mode>): Likewise.
3301 (neon_vqshlu_n<mode>): Likewise.
3302 (neon_vshll_n<mode>): Likewise.
3303 (neon_vsra_n<mode>): Likewise.
3304 (neon_vsri_n<mode>): Likewise.
3305 (neon_vsli_n<mode>): Likewise.
3306 (neon_vtbl1v8qi): Likewise.
3307 (neon_vtbl2v8qi): Likewise.
3308 (neon_vtbl3v8qi): Likewise.
3309 (neon_vtbl4v8qi): Likewise.
3310 (neon_vtbx1v8qi): Likewise.
3311 (neon_vtbx2v8qi): Likewise.
3312 (neon_vtbx3v8qi): Likewise.
3313 (neon_vtbx4v8qi): Likewise.
3314 (neon_vtrn<mode>_internal): Likewise.
3315 (neon_vzip<mode>_internal): Likewise.
3316 (neon_vuzp<mode>_internal): Likewise.
3317 (neon_vld1<mode>): Likewise.
3318 (neon_vld1_lane<mode>): Likewise.
3319 (neon_vld1_dup<mode>): Likewise.
3320 (neon_vld1_dupv2di): Likewise.
3321 (neon_vst1<mode>): Likewise.
3322 (neon_vst1_lane<mode>): Likewise.
3323 (neon_vld2<mode>): Likewise.
3324 (neon_vld2_lane<mode>): Likewise.
3325 (neon_vld2_dup<mode>): Likewise.
3326 (neon_vst2<mode>): Likewise.
3327 (neon_vst2_lane<mode>): Likewise.
3328 (neon_vld3<mode>): Likewise.
3329 (neon_vld3qa<mode>): Likewise.
3330 (neon_vld3qb<mode>): Likewise.
3331 (neon_vld3_lane<mode>): Likewise.
3332 (neon_vld3_dup<mode>): Likewise.
3333 (neon_vst3<mode>): Likewise.
3334 (neon_vst3qa<mode>): Likewise.
3335 (neon_vst3qb<mode>): Likewise.
3336 (neon_vst3_lane<mode>): Likewise.
3337 (neon_vld4<mode>): Likewise.
3338 (neon_vld4qa<mode>): Likewise.
3339 (neon_vld4qb<mode>): Likewise.
3340 (neon_vld4_lane<mode>): Likewise.
3341 (neon_vld4_dup<mode>): Likewise.
3342 (neon_vst4<mode>): Likewise.
3343 (neon_vst4qa<mode>): Likewise.
3344 (neon_vst4qb<mode>): Likewise.
3345 (neon_vst4_lane<mode>): Likewise.
3346 (neon_vec_unpack<US>_lo_<mode>): Likewise.
3347 (neon_vec_unpack<US>_hi_<mode>): Likewise.
3348 (neon_vec_<US>mult_lo_<mode>): Likewise.
3349 (neon_vec_<US>mult_hi_<mode>): Likewise.
3350 (neon_vec_<US>shiftl_<mode>): Likewise.
3351 (neon_unpack<US>_<mode>): Likewise.
3352 (neon_vec_<US>mult_<mode>): Likewise.
3353 (vec_pack_trunc_<mode>): Likewise.
3354 (neon_vec_pack_trunk_<mode>): Likewise.
3355 (neon_vabd<mode>_2): Likewise.
3356 (neon_vabd<mode>_3): Likewise.
3357 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
3358 (thumb2_movsi_vfp): Likewise.
3359 (movdi_vfp): Likewise.
3360 (movdi_vfp_cortexa8): Likewise.
3361 (movhf_vfp_neon): Likewise.
3362 (movhf_vfp): Likewiwse.
3363 (movsf_vfp): Likewiwse.
3364 (thumb2_movsf_vfp): Likewiwse.
3365 (movdf_vfp): Likewise.
3366 (thumb2_movdf_vfp): Likewise.
3367 (movsfcc_vfp): Likewise.
3368 (thumb2_movsfcc_vfp): Likewise.
3369 (movdfcc_vfp): Likewise.
3370 (thumb2_movdfcc_vfp): Likewise.
3371 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
3372 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
3373 (v10_v2c): Likewise.
3374 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
3376 (cortex_a15_neon_int_2): Likewise.
3377 (cortex_a15_neon_int_3): Likewise.
3378 (cortex_a15_neon_int_4): Likewise.
3379 (cortex_a15_neon_int_5): Likewise.
3380 (cortex_a15_neon_vqneg_vqabs): Likewise.
3381 (cortex_a15_neon_vmov): Likewise.
3382 (cortex_a15_neon_vaba): Likewise.
3383 (cortex_a15_neon_vaba_qqq): Likewise.
3384 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3385 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
3386 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
3387 scalar_64_32_long_scalar): Likewise.
3388 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3389 (cortex_a15_neon_mla_qqq_8_16): Likewise.
3390 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
3391 lotype_qdd_64_32_long): Likewise.
3392 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
3393 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
3394 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
3395 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
3396 (cortex_a15_neon_shift_1): Likewise.
3397 (cortex_a15_neon_shift_2): Likewise.
3398 (cortex_a15_neon_shift_3): Likewise.
3399 (cortex_a15_neon_vshl_ddd): Likewise.
3400 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
3401 (cortex_a15_neon_vsra_vrsra): Likewise.
3402 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
3403 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
3404 (cortex_a15_neon_fp_vmul_ddd): Likewise.
3405 (cortex_a15_neon_fp_vmul_qqd): Likewise.
3406 (cortex_a15_neon_fp_vmla_ddd): Likewise.
3407 (cortex_a15_neon_fp_vmla_qqq): Likewise.
3408 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
3409 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
3410 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
3411 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
3412 (cortex_a15_neon_bp_simple): Likewise.
3413 (cortex_a15_neon_bp_2cycle): Likewise.
3414 (cortex_a15_neon_bp_3cycle): Likewise.
3415 (cortex_a15_neon_vld1_1_2_regs): Likewise.
3416 (cortex_a15_neon_vld1_3_4_regs): Likewise.
3417 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
3418 (cortex_a15_neon_vld2_4_regs): Likewise.
3419 (cortex_a15_neon_vld3_vld4): Likewise.
3420 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
3421 (cortex_a15_neon_vst1_3_4_regs): Likewise.
3422 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
3423 (cortex_a15_neon_vst3_vst4): Likewise.
3424 (cortex_a15_neon_vld1_vld2_lane): Likewise.
3425 (cortex_a15_neon_vld3_vld4_lane" 10
3426 (cortex_a15_neon_vst1_vst2_lane): Likewise.
3427 (cortex_a15_neon_vst3_vst4_lane): Likewise.
3428 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
3429 (cortex_a15_neon_ldm_2): Likewise.0
3430 (cortex_a15_neon_stm_2): Likewise.
3431 (cortex_a15_neon_mcr): Likewise.
3432 (cortex_a15_neon_mcr_2_mcrr): Likewise.
3433 (cortex_a15_neon_mrc): Likewise.
3434 (cortex_a15_neon_mrrc): Likewise.
3435 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
3437 (cortex_a15_alu_shift): Likewise.
3438 (cortex_a15_alu_shift_reg): Likewise.
3439 (cortex_a15_mult32): Likewise.
3440 (cortex_a15_mult64): Likewise.
3441 (cortex_a15_block): Likewise.
3442 (cortex_a15_branch): Likewise.
3443 (cortex_a15_load1): Likewise.
3444 (cortex_a15_load3): Likewise.
3445 (cortex_a15_store1): Likewise.
3446 (cortex_a15_store3): Likewise.
3447 (cortex_a15_call): Likewise.
3448 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
3449 (cortex_a5_f2r): Likewise.
3450 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
3452 (cortex_a53_f2r): Likewise.
3453 * config/arm/cortex-a7.md
3454 (cortex_a7_branch): Update for attribute change.
3455 (cortex_a7_call): Likewise.
3456 (cortex_a7_alu_imm): Likewise.
3457 (cortex_a7_alu_reg): Likewise.
3458 (cortex_a7_alu_shift): Likewise.
3459 (cortex_a7_mul): Likewise.
3460 (cortex_a7_load1): Likewise.
3461 (cortex_a7_store1): Likewise.
3462 (cortex_a7_load2): Likewise.
3463 (cortex_a7_store2): Likewise.
3464 (cortex_a7_load3): Likewise.
3465 (cortex_a7_store3): Likewise.
3466 (cortex_a7_load4): Likewise.
3467 (cortex_a7_store4): Likewise.
3468 (cortex_a7_fpalu): Likewise.
3469 (cortex_a7_fconst): Likewise.
3470 (cortex_a7_fpmuls): Likewise.
3471 (cortex_a7_neon_mul): Likewise.
3472 (cortex_a7_fpmacs): Likewise.
3473 (cortex_a7_neon_mla: Likewise.
3474 (cortex_a7_fpmuld: Likewise.
3475 (cortex_a7_fpmacd: Likewise.
3476 (cortex_a7_fpfmad: Likewise.
3477 (cortex_a7_fdivs: Likewise.
3478 (cortex_a7_fdivd: Likewise.
3479 (cortex_a7_r2f: Likewise.
3480 (cortex_a7_f2r: Likewise.
3481 (cortex_a7_f_flags: Likewise.
3482 (cortex_a7_f_loads: Likewise.
3483 (cortex_a7_f_loadd: Likewise.
3484 (cortex_a7_f_stores: Likewise.
3485 (cortex_a7_f_stored: Likewise.
3486 (cortex_a7_neon): Likewise.
3487 * config/arm/cortex-a8-neon.md
3488 (cortex_a8_neon_mrc): Update for attribute change.
3489 (cortex_a8_neon_mrrc): Likewise.
3490 (cortex_a8_neon_int_1): Likewise.
3491 (cortex_a8_neon_int_2): Likewise.
3492 (cortex_a8_neon_int_3): Likewise.
3493 (cortex_a8_neon_int_4): Likewise.
3494 (cortex_a8_neon_int_5): Likewise.
3495 (cortex_a8_neon_vqneg_vqabs): Likewise.
3496 (cortex_a8_neon_vmov): Likewise.
3497 (cortex_a8_neon_vaba): Likewise.
3498 (cortex_a8_neon_vaba_qqq): Likewise.
3499 (cortex_a8_neon_vsma): Likewise.
3500 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3501 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
3502 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
3503 long_scalar): Likewise.
3504 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3505 (cortex_a8_neon_mla_qqq_8_16): Likewise.
3506 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
3507 long_scalar_qdd_64_32_long): Likewise.
3508 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
3509 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
3510 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
3511 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
3512 (cortex_a8_neon_shift_1): Likewise.
3513 (cortex_a8_neon_shift_2): Likewise.
3514 (cortex_a8_neon_shift_3): Likewise.
3515 (cortex_a8_neon_vshl_ddd): Likewise.
3516 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
3517 (cortex_a8_neon_vsra_vrsra): Likewise.
3518 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
3519 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
3520 (cortex_a8_neon_fp_vsum): Likewise.
3521 (cortex_a8_neon_fp_vmul_ddd): Likewise.
3522 (cortex_a8_neon_fp_vmul_qqd): Likewise.
3523 (cortex_a8_neon_fp_vmla_ddd): Likewise.
3524 (cortex_a8_neon_fp_vmla_qqq): Likewise.
3525 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
3526 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
3527 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
3528 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
3529 (cortex_a8_neon_bp_simple): Likewise.
3530 (cortex_a8_neon_bp_2cycle): Likewise.
3531 (cortex_a8_neon_bp_3cycle): Likewise.
3532 (cortex_a8_neon_ldr): Likewise.
3533 (cortex_a8_neon_str): Likewise.
3534 (cortex_a8_neon_vld1_1_2_regs): Likewise.
3535 (cortex_a8_neon_vld1_3_4_regs): Likewise.
3536 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
3537 (cortex_a8_neon_vld2_4_regs): Likewise.
3538 (cortex_a8_neon_vld3_vld4): Likewise.
3539 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
3540 (cortex_a8_neon_vst1_3_4_regs): Likewise.
3541 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
3542 (cortex_a8_neon_vst3_vst4): Likewise.
3543 (cortex_a8_neon_vld1_vld2_lane): Likewise.
3544 (cortex_a8_neon_vld3_vld4_lane): Likewise.
3545 (cortex_a8_neon_vst1_vst2_lane): Likewise.
3546 (cortex_a8_neon_vst3_vst4_lane): Likewise.
3547 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
3548 (cortex_a8_neon_mcr): Likewise.
3549 (cortex_a8_neon_mcr_2_mcrr): Likewise.
3550 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
3551 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
3553 (ca9_neon_mrrc): Likewise.
3554 (cortex_a9_neon_int_1): Likewise.
3555 (cortex_a9_neon_int_2): Likewise.
3556 (cortex_a9_neon_int_3): Likewise.
3557 (cortex_a9_neon_int_4): Likewise.
3558 (cortex_a9_neon_int_5): Likewise.
3559 (cortex_a9_neon_vqneg_vqabs): Likewise.
3560 (cortex_a9_neon_vmov): Likewise.
3561 (cortex_a9_neon_vaba): Likewise.
3562 (cortex_a9_neon_vaba_qqq): Likewise.
3563 (cortex_a9_neon_vsma): Likewise.
3564 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3565 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
3566 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
3567 long_scalar): Likewise.
3568 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3569 (cortex_a9_neon_mla_qqq_8_16): Likewise.
3570 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
3571 long_scalar_qdd_64_32_long): Likewise.
3572 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
3573 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
3574 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
3575 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
3576 (cortex_a9_neon_shift_1): Likewise.
3577 (cortex_a9_neon_shift_2): Likewise.
3578 (cortex_a9_neon_shift_3): Likewise.
3579 (cortex_a9_neon_vshl_ddd): Likewise.
3580 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
3581 (cortex_a9_neon_vsra_vrsra): Likewise.
3582 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
3583 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
3584 (cortex_a9_neon_fp_vsum): Likewise.
3585 (cortex_a9_neon_fp_vmul_ddd): Likewise.
3586 (cortex_a9_neon_fp_vmul_qqd): Likewise.
3587 (cortex_a9_neon_fp_vmla_ddd): Likewise.
3588 (cortex_a9_neon_fp_vmla_qqq): Likewise.
3589 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
3590 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
3591 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
3592 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
3593 (cortex_a9_neon_bp_simple): Likewise.
3594 (cortex_a9_neon_bp_2cycle): Likewise.
3595 (cortex_a9_neon_bp_3cycle): Likewise.
3596 (cortex_a9_neon_ldr): Likewise.
3597 (cortex_a9_neon_str): Likewise.
3598 (cortex_a9_neon_vld1_1_2_regs): Likewise.
3599 (cortex_a9_neon_vld1_3_4_regs): Likewise.
3600 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
3601 (cortex_a9_neon_vld2_4_regs): Likewise.
3602 (cortex_a9_neon_vld3_vld4): Likewise.
3603 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
3604 (cortex_a9_neon_vst1_3_4_regs): Likewise.
3605 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
3606 (cortex_a9_neon_vst3_vst4): Likewise.
3607 (cortex_a9_neon_vld1_vld2_lane): Likewise.
3608 (cortex_a9_neon_vld3_vld4_lane): Likewise.
3609 (cortex_a9_neon_vst1_vst2_lane): Likewise.
3610 (cortex_a9_neon_vst3_vst4_lane): Likewise.
3611 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
3612 (cortex_a9_neon_mcr): Likewise.
3613 (cortex_a9_neon_mcr_2_mcrr): Likewise.
3614 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
3615 (cortex_a9_fps): Likewise.
3616 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
3618 (cortex_m4_fmuls): Likewise.
3619 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
3621 (cortex_r4_mrc): Likewise.
3622 * config/arm/iterators.md: Update comment referring to neon_type.
3623 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
3624 (iwmmxt_movsi_insn): Likewise.
3625 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
3627 (pj4_core_to_vfp): Likewise.
3628 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
3630 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
3631 (vfp_fstore): Likewise.
3632 * doc/md.texi: Change references to neon_type to refer to type.
3634 2013-09-04 Dodji Seketeli <dodji@redhat.com>
3636 * tree.h (DECL_BUILT_IN): Fix typo in comment.
3638 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
3640 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
3643 2013-09-04 Easwaran Raman <eraman@google.com>
3646 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
3647 (build_and_add_sum): Use it.
3648 (appears_later_in_bb): Simplify code.
3650 2013-09-04 Teresa Johnson <tejohnson@google.com>
3652 * dumpfile.c (dump_finish): Don't close stderr/stdout.
3654 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
3656 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
3658 2013-09-04 Jan Hubicka <jh@suse.cz>
3660 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
3661 * ipa-devirt.c: Include diganostic.h
3662 (odr_type_d): Add types and types_set.
3663 (hash_type_name): Work for types with vtables during LTO.
3664 (odr_hasher::remove): Fix comment; destroy types_set.
3665 (add_type_duplicate): New function,
3666 (get_odr_type): Use it.
3667 (dump_type_inheritance_graph): Dump type duplicates.
3668 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
3670 * tree.c (types_same_for_odr): Give exact answers on types with
3673 2013-09-04 Dodji Seketeli <dodji@redhat.com>
3675 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
3676 explaining their differences.
3678 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
3680 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
3682 2013-09-03 Jeff Law <law@redhat.com>
3684 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
3685 when not threading through a joiner block. Pass joiner/no joiner
3686 state to register_jump_thread.
3687 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
3688 state from argument rather than implying on path length.
3689 Dump the entire jump thread path into debugging dump.
3690 * tree-flow.h (register_jump_thread): Update prototype.
3692 2013-08-29 Xinliang David Li <davidxl@google.com>
3694 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
3695 Remove a trivial gcc_assert.
3697 2013-08-29 Xinliang David Li <davidxl@google.com>
3699 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
3700 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
3701 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
3702 Delay base decl alignment adjustment.
3703 * tree-vectorizer.c (vect_destroy_datarefs): New function.
3704 * tree-vectorizer.h: New data structure.
3705 (set_dr_misalignment): New function.
3706 (dr_misalignment): Ditto.
3707 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
3708 (vectorizable_load): Ditto.
3709 (ensure_base_align): New function.
3710 (vectorize_loops): Add dbg_cnt support.
3711 (execute_vect_slp): Ditto.
3712 * dbgcnt.def: New debug counter.
3713 * Makefile: New dependency.
3715 2013-09-03 Meador Inge <meadori@codesourcery.com>
3719 2013-08-30 Meador Inge <meadori@codesourcery.com>
3721 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
3723 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
3725 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
3726 function descriptor.
3728 2013-09-03 Richard Biener <rguenther@suse.de>
3730 * tree-affine.c (add_elt_to_tree): Fix association issue,
3731 avoid useless converts and make sure to always return a
3732 properly typed result.
3734 2013-09-03 Richard Biener <rguenther@suse.de>
3737 * fold-const.c (negate_expr_p): Fix division case.
3738 (negate_expr): Likewise.
3740 2013-09-03 Richard Biener <rguenther@suse.de>
3743 * tree-streamer-out.c: Include tm.h.
3744 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
3746 2013-09-03 Jan Hubicka <jh@suse.cz>
3748 * tree-profile.c (tree_profiling): Cleanup CFG when done.
3750 2013-09-03 Alan Modra <amodra@gmail.com>
3752 * config.gcc (powerpc*-*-linux*): Add support for little-endian
3753 multilibs to big-endian target and vice versa.
3754 * config/rs6000/t-linux64: Use := assignment on all vars.
3755 (MULTILIB_EXTRA_OPTS): Remove fPIC.
3756 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
3757 * config/rs6000/t-linux64le: New file.
3758 * config/rs6000/t-linux64bele: New file.
3759 * config/rs6000/t-linux64lebe: New file.
3761 2013-09-02 Jan Hubicka <jh@suse.cz>
3763 * ipa-inline-transform.c (inline_transform): Do not
3764 optimize_inline_calls when not optimizing.
3766 2013-09-02 Jan Hubicka <jh@suse.cz>
3768 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
3769 duplicated nodes for assembler names.
3770 * symtab.c (symtab_unregister_node): Do not attempt to unlink
3771 hard registers from assembler name hash.
3773 2013-09-02 Jan Hubicka <jh@suse.cz>
3775 * ipa-split.c (execute_split_functions): Split externally visible
3776 functions called once.
3778 2013-09-02 Martin Jambor <mjambor@suse.cz>
3781 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
3782 linked list. When finding the correct duplicate, also consider also
3783 the caller in additon to its inlined_to node.
3785 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
3787 * config/aarch64/aarch64-simd-builtins.def
3788 (dup_lane_scalar): Remove.
3789 * config/aarch64/aarch64-simd.md
3790 (aarch64_simd_dup): Add 'w->w' alternative.
3791 (aarch64_dup_lane<mode>): Allow for VALL.
3792 (aarch64_dup_lane_scalar<mode>): Remove.
3793 (aarch64_dup_lane_<vswap_width_name><mode>): New.
3794 (aarch64_get_lane_signed<mode>): Add w->w altenative.
3795 (aarch64_get_lane_unsigned<mode>): Likewise.
3796 (aarch64_get_lane<mode>): Likewise.
3797 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
3798 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
3799 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
3800 (VCON): Change container of V2SF.
3801 (vswap_width_name): Likewise.
3802 * config/aarch64/arm_neon.h
3803 (__aarch64_vdup_lane_any): New.
3804 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
3805 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
3806 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
3808 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
3811 * expr.c (emit_move_complex): Do not move complex FP values as parts if
3812 the source or the destination is a single hard register.
3814 2013-09-02 Richard Biener <rguenther@suse.de>
3817 * tree-scalar-evolution.c (instantiate_scev_name): Allow
3820 2013-09-02 Richard Biener <rguenther@suse.de>
3822 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
3823 arithmetic to sizetype.
3825 2013-09-02 Bin Cheng <bin.cheng@arm.com>
3827 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
3828 Find auto-increment use both before and after candidate.
3830 2013-09-02 Marek Polacek <polacek@redhat.com>
3832 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
3834 2013-09-01 Jan Hubicka <jh@suse.cz>
3836 * Makefile.in: Add ipa-profile.o
3837 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
3838 * cgraph.c (struct cgraph_propagate_frequency_data,
3839 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
3840 ipa-profile.c; replace cgraph_ by ipa_ prefix.
3841 * cgraph.h (cgraph_propagate_frequency): Remove.
3842 * ipa-inline-analysis.c: Include ipa-utils.h;
3843 drop duplicated cfgloop.h.
3844 (inline_update_callee_summaries): Update.
3845 * ipa-profile.c: New file.
3846 * ipa-utils.h (ipa_propagate_frequency): Declare.
3847 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
3848 data-streamer.h, value-prof.h.
3849 (symtab_remove_unreachable_nodes): Update profile.
3850 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
3851 account_time_size, cmp_counts, dump_histogram,
3852 ipa_profile_generate_summary, ipa_profile_write_summary,
3853 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
3854 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
3855 Move to ipa-profile.c.
3857 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
3859 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
3861 2013-09-01 Jan Hubicka <jh@suse.cz>
3863 * common.opt (fdevirtualize-speculatively): New function.
3864 * invoke.texi (fdevirtualize-speculatively): Document.
3865 * ipa-devirt.c: Include ipa-inline.h
3866 (likely_target_p): New function.
3867 (ipa_devirt): New function.
3868 (gate_ipa_devirt): New function.
3869 (pass_data_ipa_devirt): New static var.
3870 (pass_ipa_devirt): Likewise.
3871 (make_pass_ipa_devirt): New function.
3872 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
3873 (common_handle_option): Disable devirtualization when
3874 value range profiling is available.
3875 * passes.def (pass_ipa_devirt): Add.
3876 * timever.def (TV_IPA_DEVIRT): New timevar.
3877 * tree-pass.h (make_pass_ipa_devirt):
3879 2013-09-01 Iain Sandoe <iain@codesourcery.com>
3881 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
3882 include sanitize(undefined).
3884 2013-08-31 Diego Novillo <dnovillo@google.com>
3886 * Makefile.in (TREE_CORE_H): Define.
3888 (GTFILES): Add tree-core.h.
3889 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
3891 * tree-core.h: New file.
3892 Move all data structures, enum, typedefs, global
3893 declarations and constants from ...
3896 2013-08-31 Jan Hubicka <jh@suse.cz>
3898 * bulitins.c (expand_builtin): Do not early exit for gcov
3899 instrumented functions.
3901 2013-08-31 Marek Polacek <polacek@redhat.com>
3903 * ubsan.c: Include tm_p.h.
3905 2013-08-31 Jan Hubicka <jh@suse.cz>
3907 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
3910 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
3911 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
3912 * tree-cfg.c (verify_gimple_label): ... here.
3913 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
3914 (ipa_merge_profiles): New function.
3915 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
3916 (lto_input_function_body): Likewise.
3917 * ipa-utils.h (ipa_merge_profiles): Declare.
3918 * lto-streamer.h (lto_input_function_body): Update prototype.
3919 (emit_label_in_global_context_p): Remove.
3920 * lto-symtab.c: Include ipa-utils.h
3921 (lto_cgraph_replace_node): Use ipa_merge_profiles.
3923 2013-08-31 Jan Hubicka <jh@suse.cz>
3925 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
3927 2013-08-31 Jan Hubicka <jh@suse.cz>
3929 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
3931 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
3933 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
3934 "cmp" RTX before signed_comparison_operator check to account
3937 2013-08-30 Jan Hubicka <jh@suse.cz>
3939 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
3940 (detect_type_change_1): Rename to ...
3941 (detect_type_change): ... this one; early return on non-polymorphic
3943 (detect_type_change_ssa): Add comp_type parameter; update
3944 use of detect_type_change.
3945 (compute_complex_assign_jump_func): Add param_type parameter;
3946 update use of detect_type_change_ssa.
3947 (compute_complex_ancestor_jump_func): Likewise.
3948 (ipa_get_callee_param_type): New function.
3949 (ipa_compute_jump_functions_for_edge): Compute parameter type;
3950 update calls to the jump function computation functions.
3952 2013-08-30 Teresa Johnson <tejohnson@google.com>
3953 Steven Bosscher <steven@gcc.gnu.org>
3955 * cfgrtl.c (fixup_new_cold_bb): New routine.
3956 (commit_edge_insertions): Invoke fixup_partitions.
3957 (find_partition_fixes): New routine.
3958 (fixup_partitions): Ditto.
3959 (verify_hot_cold_block_grouping): Update comments.
3960 (rtl_verify_edges): Invoke find_partition_fixes.
3961 (rtl_verify_bb_pointers): Update comments.
3962 (rtl_verify_bb_layout): Ditto.
3963 * basic-block.h (probably_never_executed_edge_p): Declare.
3964 (fixup_partitions): Ditto.
3965 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
3966 * bb-reorder.c (sanitize_hot_paths): New function.
3967 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
3969 * predict.c (probably_never_executed_edge_p): New routine.
3970 * cfg.c (check_bb_profile): Add partition insanity warnings.
3972 2013-08-30 Meador Inge <meadori@codesourcery.com>
3974 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
3976 2013-08-30 Marek Polacek <polacek@redhat.com>
3978 * Makefile.in (ubsan.o): Add.
3979 (c-family/c-ubsan.o): Add.
3980 (builtins.o): Add ubsan.h dependency.
3981 * ubsan.h: New file.
3982 * ubsan.c: New file.
3983 * common.opt: Add -fsanitize=undefined option.
3984 (flag_sanitize): Add variable.
3985 (fsanitize=): Add option. Add Driver.
3986 (fsanitize=thread): Remove option.
3987 (fsanitize=address): Likewise.
3988 (static-libubsan): New option.
3989 * doc/invoke.texi: Document the new flag and -static-libubsan.
3990 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
3991 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
3992 * builtin-attrs.def (ATTR_COLD): Define.
3993 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
3994 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
3995 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
3996 * flag-types.h (sanitize_code): New enum.
3997 * opts.c (common_handle_option): Parse command line arguments
3998 of -fsanitize=. Add -fsanitize=unreachable option.
3999 * varasm.c (get_variable_section): Adjust.
4000 (assemble_noswitch_variable): Likewise.
4001 (assemble_variable): Likewise.
4002 (output_constant_def_contents): Likewise.
4003 (categorize_decl_for_section): Likewise.
4004 (place_block_symbol): Likewise.
4005 (output_object_block): Likewise.
4006 * builtins.def: Likewise.
4007 * toplev.c (compile_file): Likewise.
4008 (process_options): Likewise.
4009 * cppbuiltin.c: Likewise.
4010 * tsan.c (tsan_pass): Likewise.
4011 (tsan_gate): Likewise.
4012 (tsan_gate_O0): Likewise.
4013 * cfgexpand.c (partition_stack_vars): Likewise.
4014 (expand_stack_vars): Likewise.
4015 (defer_stack_allocation): Likewise.
4016 (expand_used_vars): Likewise.
4017 * cfgcleanup.c (old_insns_match_p): Likewise.
4018 * asan.c (asan_finish_file): Likewise.
4019 (asan_instrument): Likewise.
4020 (gate_asan): Likewise.
4021 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
4022 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
4023 (asan_global_struct): Use pointer_sized_int_node instead
4024 calling build_nonstandard_integer_type.
4025 (initialize_sanitizer_builtins): Likewise.
4026 (asan_finish_file): Likewise.
4027 * gcc.c: Document %{%:function(args):X}.
4028 (static_spec_functions): Add sanitize.
4029 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
4030 store funcval != NULL there.
4031 (do_spec_1): Adjust handle_spec_function caller.
4032 (handle_braces): Allow %:function(args) as condition.
4033 (sanitize_spec_function): New function.
4034 (ADD_STATIC_LIBUBSAN_LIBS): Define.
4035 (LIBUBSAN_SPEC): Likewise.
4036 (LIBUBSAN_EARLY_SPEC): Likewise.
4037 (SANITIZER_SPEC): Handle libubsan.
4038 (SANITIZER_EARLY_SPEC): Likewise.
4039 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
4040 instead of fsanitize=address.
4041 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
4042 instead of fsanitize=address*.
4043 * builtins.c: Include ubsan.h.
4044 (fold_builtin_0): Instrument __builtin_unreachable.
4045 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
4046 instead of flag_asan.
4047 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
4048 (pointer_sized_int_node): Define.
4049 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
4051 2013-08-30 Mike Stump <mikestump@comcast.net>
4053 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
4056 2013-08-29 Jan Hubicka <jh@suse.cz>
4058 * cgraph.c (cgraph_function_body_availability): Handle weakref
4060 * passes.def: Remove pass_fixup_cfg.
4061 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
4062 track when we need to remove functions.
4063 (gate_ipa_inline): Execute inlining always; add comment why.
4064 (pass_data_ipa_inline): Remove TODO_remove_functions.
4065 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
4066 do not produce summaries.
4067 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
4068 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
4069 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
4070 constant pool and vtable.
4072 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
4074 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
4075 New arm_neon.h's internal macros to specify 64-bit constants.
4076 Avoid using stdint.h's macros.
4078 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
4080 * recog.c (verify_changes): Verify that changes[i].old is non-zero
4081 before applying REG_P.
4083 2013-08-30 Jakub Jelinek <jakub@redhat.com>
4085 PR tree-optimization/58277
4086 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
4087 after seeing too many stmts with vdef in between dombb and current
4088 bb, invalidate everything.
4090 2013-08-30 Richard Biener <rguenther@suse.de>
4092 * fold-const.c (fold_single_bit_test): Fix overflow test.
4094 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
4096 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
4097 and which can live in a register, always retrieve the value on entry.
4098 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
4099 passed by invisible reference specially.
4100 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
4101 (vt_add_function_parameter): Correctly deal with a parameter passed by
4102 invisible reference.
4104 2013-08-30 Jan Hubicka <jh@suse.cz>
4106 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
4108 2013-08-30 Richard Biener <rguenther@suse.de>
4110 PR tree-optimization/58228
4111 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
4112 allow invariant loads in nested loop vectorization.
4114 2013-08-30 Richard Biener <rguenther@suse.de>
4116 PR tree-optimization/58223
4117 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
4118 (has_anti_or_output_dependence): ... this and adjust to also
4119 look for output dependences.
4120 (mark_nodes_having_upstream_mem_writes): Adjust.
4121 (rdg_flag_uses): Likewise.
4123 2013-08-30 Richard Biener <rguenther@suse.de>
4125 PR tree-optimization/58010
4126 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
4127 assert that we have a loop-closed PHI.
4129 2013-08-29 Jan Hubicka <jh@suse.cz>
4131 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
4132 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
4133 * lto-section-in.c (lto_free_function_in_decl_state): New function.
4134 (lto_free_function_in_decl_state_for_node): New function.
4136 2013-08-29 Xinliang David Li <davidxl@google.com>
4138 * loop-unroll.c (report_unroll_peel): Minor message change.
4139 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
4140 Emit alignment peeling message with default -fopt-info.
4141 (vect_loop_versioning): Emit loop version info message.
4142 * tree-vectorizer.c (vectorize_loops): Minor message change.
4143 (execute_vect_slp): Ditto.
4145 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
4147 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
4148 of the clone from the DECL_NAME of the original function.
4150 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
4152 * passes.c (register_pass): Add overload.
4153 * tree-pass.h (register_pass): Forward declare it. Add comment.
4155 2013-08-29 Jan Hubicka <jh@suse.cz>
4157 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
4158 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
4159 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
4160 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
4161 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
4162 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
4163 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
4164 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
4166 2013-08-29 Teresa Johnson <tejohnson@google.com>
4168 * dumpfile.c (dump_loc): Output column number.
4169 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
4170 * doc/invoke.texi: Document optall -fopt-info flag.
4171 * profile.c (read_profile_edge_counts): Use new dump framework.
4172 (compute_branch_probabilities): Ditto.
4173 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
4174 when pass not in any opt group.
4175 * pass_manager.h (pass_manager::get_pass_profile): New method.
4176 * value-prof.c (check_counter): Use new dump framework.
4177 (check_ic_target): Ditto.
4178 * coverage.c (get_coverage_counts): Ditto.
4179 (coverage_init): Setup new dump framework.
4181 2013-08-29 Richard Biener <rguenther@suse.de>
4183 PR tree-optimization/58246
4184 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
4185 handle the dominance check inside a basic-block.
4187 2013-08-29 Richard Biener <rguenther@suse.de>
4190 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
4191 of default defs that appear in abnormal PHI nodes.
4193 2013-08-29 Richard Biener <rguenther@suse.de>
4195 PR tree-optimization/57685
4196 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
4197 single-use operands to avoid exponential complexity.
4199 2013-08-28 Dehao Chen <dehao@google.com>
4201 * ipa-inline.c (edge_badness): Fix integer underflow.
4203 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
4205 * gtm-builtins.def (_ITM_free): Declare leaf.
4207 2013-08-28 Jakub Jelinek <jakub@redhat.com>
4210 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
4211 (*tls_local_dynamic_base_64_largepic): Likewise.
4212 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
4213 Remove predicate from call operand.
4214 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
4215 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
4217 2013-08-28 Jeff Law <law@redhat.com>
4219 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
4220 checks for the number of predecessors and successors allowed.
4221 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
4222 which require copying a joiner block if there is a request which
4223 is a subpath that requires no joiner block copying.
4225 2013-08-28 Jan Hubicka <jh@suse.cz>
4227 * lto-streamer-out.c (DFS_write_tree_body): Drop
4228 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
4229 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
4230 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
4231 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
4232 Do not read DECL_ERROR_ISSUED.
4233 (unpack_ts_decl_with_vis_value_fields): Do not read
4235 (lto_input_ts_binfo_tree_pointers): Do not read
4236 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
4237 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
4238 write DECL_ERROR_ISSUED..
4239 (pack_ts_decl_with_vis_value_fields): Do not write
4241 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
4242 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
4243 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
4244 * tree.h (tree_decl_common): Update comment.
4245 (DECL_ERROR_ISSUED): Remove.
4247 2013-08-28 Jakub Jelinek <jakub@redhat.com>
4250 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
4252 2013-08-28 Jan Hubicka <jh@suse.cz>
4254 * builtins.def (free): Declare leaf.
4256 2013-08-27 David Malcolm <dmalcolm@redhat.com>
4259 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
4260 * configure: Regenerate.
4262 2013-08-27 Martin Jambor <mjambor@suse.cz>
4264 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
4265 (ipa_ancestor_jf_data): Likewise.
4266 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
4267 (ipa_get_jf_pass_through_type_preserved): New function.
4268 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
4269 (ipa_get_jf_ancestor_type_preserved): New function.
4270 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
4271 (ipa_get_jf_ancestor_result): Likewise.
4272 (propagate_vals_accross_pass_through): Use
4273 ipa_get_jf_pass_through_result to do all the value mappings.
4274 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
4275 type_preserved flag.
4276 (ipa_set_jf_cst_copy): New function.
4277 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
4278 (ipa_set_jf_arith_pass_through): Likewise.
4279 (ipa_set_ancestor_jf): Likewise.
4280 (compute_complex_assign_jump_func): Set type_preserved instead of
4282 (ipa_compute_jump_functions_for_edge): Likewise.
4283 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
4284 (update_jump_functions_after_inlining): Update type_preserved.
4285 Explicitely create jump functions when combining one with pass_through.
4286 (ipa_write_jump_function): Stream the type_preserved flags.
4287 (ipa_read_jump_function): Likewise.
4289 2013-08-27 Jakub Jelinek <jakub@redhat.com>
4290 Aldy Hernandez <aldyh@redhat.com>
4292 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
4293 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
4294 * function.h (struct function): Add has_force_vect_loops and
4296 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
4297 * gimple.c (gimple_build_omp_critical): Add KIND argument and
4299 * gimple.def: Update CLAUSES comments.
4300 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
4301 (gimple_build_omp_for): Add argument to prototype.
4302 (gimple_omp_for_kind): New.
4303 (gimple_omp_for_set_kind): New.
4304 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
4305 GOVD_DATA_SHARE_CLASS.
4306 (enum omp_region_type): Add ORT_SIMD.
4307 (gimple_add_tmp_var): Handle ORT_SIMD.
4308 (gimplify_var_or_parm_decl): Same.
4309 (is_gimple_stmt): Same.
4310 (omp_firstprivatize_variable): Same.
4311 (omp_add_variable): Only use splay_tree_insert if lookup failed.
4312 (omp_notice_variable): Handle ORT_SIMD.
4313 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
4314 (omp_check_private): Handle ORT_SIMD.
4315 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
4317 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
4318 Handle OMP_CLAUSE_LASTPRIVATE.
4319 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
4321 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
4322 (gimplify_expr): Handle OMP_SIMD.
4323 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
4324 (expand_GOMP_SIMD_VF): New.
4325 (expand_GOMP_SIMD_LAST_LANE): New.
4326 * internal-fn.def (GOMP_SIMD_LANE): New.
4327 (GOMP_SIMD_VF): New.
4328 (GOMP_SIMD_LAST_LANE): New.
4329 * omp-low.c: Include target.h.
4330 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
4332 (check_omp_nesting_restrictions): Same.
4334 (lower_rec_simd_input_clauses): New.
4335 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
4337 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
4338 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
4339 (expand_omp_build_assign): New.
4340 (expand_omp_for_init_counts): New.
4341 (expand_omp_for_init_vars): New.
4342 (extract_omp_for_update_vars): New.
4343 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
4344 and rewrite accordingly.
4345 (expand_omp_simd): New.
4346 (expand_omp_for): Use expand_omp_simd.
4347 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
4348 (lower_omp_for): Do not lower the body.
4349 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
4351 * tree-flow.h (find_omp_clause): Remove prototype.
4352 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
4353 forcing vectorization of the loop, or if flag_tree_vectorize.
4354 (gate_tree_if_conversion): Similarly.
4355 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
4356 gimple_build_omp_for.
4357 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
4358 * tree-parloops (create_parallel_loop): Pass kind argument to
4359 gimple_build_omp_for.
4360 * tree-pretty-print.c (dump_omp_clause): Add cases for
4361 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
4362 OMP_CLAUSE__SIMDUID_.
4363 (dump_generic_node): Handle OMP_SIMD.
4364 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
4365 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
4366 unroll OMP_SIMD loops here.
4367 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
4368 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
4370 (vect_analyze_data_refs): Handle simd loops.
4371 * tree-vect-loop.c (vectorizable_live_operation): Handle
4373 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
4374 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
4375 (vectorizable_load): Same.
4376 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
4377 (struct simduid_to_vf): New.
4378 (simduid_to_vf::hash): New.
4379 (simduid_to-vf::equal): New.
4380 (struct simd_array_to_simduid): New.
4381 (simd_array_to_simduid::hash): New.
4382 (simd_array_to_simduid::equal): New.
4383 (adjust_simduid_builtins): New.
4384 (struct note_simd_array_uses_struct): New.
4385 (note_simd_array_uses_cb): New.
4386 (note_simd_array_uses): New.
4387 (vectorize_loops): Handle simd hints and adjust simd builtins
4389 * tree-vectorizer.h (struct _stmt_vec_info): Add
4390 simd_lane_access_p field.
4391 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
4392 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
4393 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
4394 (omp_clause_code_name): Same.
4395 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
4396 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
4397 * tree.def (OMP_SIMD): New entry.
4398 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
4399 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
4400 (OMP_CLAUSE_DECL): Adjust range for new clauses.
4401 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
4402 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
4403 (OMP_CLAUSE_LINEAR_STEP): New.
4404 (OMP_CLAUSE_SAFELEN_EXPR): New.
4405 (OMP_CLAUSE__SIMDUID__DECL): New.
4406 (find_omp_clause): New prototype.
4408 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
4410 * config/i386/driver-i386.c (host_detect_local_cpu): Update
4411 Haswell processor detection.
4413 2013-08-27 Christian Widmer <shadow@umbrox.de>
4416 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
4417 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
4418 AVX2 capable processors.
4420 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
4422 * config/aarch64/arm_neon.h: Replace all inline asm implementations
4423 of vget_low_* with implementations in terms of other intrinsics.
4425 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
4428 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
4429 values to -1, 0 and 1.
4431 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
4433 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
4434 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
4435 (<optab><mode>3_insn): Remove.
4436 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
4437 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
4438 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
4439 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
4440 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
4441 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
4442 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
4443 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
4444 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
4445 (ror<mode>3_insn): Likewise.
4446 * config/aarch64/predicates.md (aarch64_simd_register): New.
4448 2013-08-27 Richard Biener <rguenther@suse.de>
4450 PR tree-optimization/57521
4451 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
4452 one edge is non-critical.
4453 (find_phi_replacement_condition): Make sure to use a non-critical
4454 edge. Cleanup and remove old bug workarounds.
4455 (bb_postdominates_preds): Remove.
4456 (if_convertible_loop_p_1): Do not compute post-dominators.
4457 (combine_blocks): Do not free post-dominators.
4458 (main_tree_if_conversion): Likewise.
4459 (pass_data_if_conversion): Add TODO_verify_ssa.
4461 2013-08-27 DJ Delorie <dj@redhat.com>
4463 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
4465 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
4467 * function.c (assign_parm_find_data_types): Set passed_mode and
4468 nominal_mode to the TYPE_MODE of nominal_type for the built
4469 pointer type in case of the struct-pass-by-reference.
4471 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
4473 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
4474 (UINT16_TYPE): Change default to "unsigned int".
4476 * config/avr/avr.opt (mfract-convert-truncate): New option.
4477 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
4478 is set, round negative fractional integers according to n1169
4479 when converting to integer types.
4481 2013-08-26 Jan Hubicka <jh@suse.cz>
4483 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
4484 methods can not be called indirectly when their address is not taken.
4486 2013-08-26 Jan Hubicka <jh@suse.cz>
4488 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
4491 2013-08-26 Jan Hubicka <jh@suse.cz>
4493 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
4496 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
4498 * reload.c (find_valid_class): Allow classes that do not include
4499 FIRST_PSEUDO_REGISTER - 1.
4501 2013-08-26 Jan Hubicka <jh@suse.cz>
4503 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
4504 fix edge count/frequency when speculation failed; fix type check
4505 for the direct call.
4507 2013-08-26 Jan Hubicka <jh@suse.cz>
4509 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
4511 2013-08-26 Jan Hubicka <jh@suse.cz>
4513 * ipa-inline-transform.c (inline_transform): Be ready for basic block
4514 to be changed by edge redirection.
4516 2013-08-26 Jan Hubicka <jh@suse.cz>
4518 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
4519 formating; add sanity check.
4520 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
4521 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
4522 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
4523 (resolve_noninline_speculation): Update callee keys, too.
4525 2013-08-26 Jan Hubicka <jh@suse.cz>
4527 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
4528 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
4530 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
4532 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
4535 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
4537 * config/i386/i386.c (ix86_debug_options): Remove prototype.
4538 (x86_64_elf_select_section): Ditto.
4539 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
4541 (ix86_pass_by_reference): Ditto.
4542 (output_set_got): Ditto.
4543 (ix86_unary_operator_ok): Ditto.
4544 (ix86_expand_builtin): Ditto.
4546 2013-08-23 Jan Hubicka <jh@suse.cz>
4548 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
4550 2013-08-23 Jan Hubicka <jh@suse.cz>
4552 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
4553 (tree_decl_with_vis): Add FINAL field.
4555 2013-08-23 Jeff Law <law@redhat.com>
4557 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
4558 the debugging dump when the expression is fully redundant.
4560 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
4562 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
4563 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
4564 * pretty-print.c (pp_formatted_text_data): Likewise.
4565 (pp_write_text_to_stream): Likewise.
4566 (pp_write_text_as_dot_label_to_stream): Likewise.
4567 (pp_append_r): Likewise.
4568 (pp_format): Likewise.
4569 (pp_flush): Likewise.
4570 (pp_clear_output_area): Likewise.
4571 (pp_append_text): Likewise.
4572 (pp_formatted_text): Likewise.
4573 (pp_remaining_character_count_for_line): Likewise.
4574 (pp_newline): Likewise.
4575 (pp_character): Likewise.
4576 (output_buffer::~output_buffer): Define.
4577 (pretty_printer::~pretty_printer): Destruct output buffer.
4578 * pretty-print.h (output_buffer::~output_buffer): Declare.
4579 (pretty_printer::~pretty_printer): Declare virtual.
4581 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
4584 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
4585 HOST_WIDE_INT_M1U): New macros.
4586 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
4587 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
4588 unsigned -1 for lshift.
4589 * cse.c (cse_insn): Likewise.
4590 * double-int.c (rshift_double, lshift_double): Likewise.
4591 * builtins.c (fold_builtin_bitop): Likewise.
4592 * combine.c (force_to_mode): Likewise.
4593 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
4594 * simplify-rtx.c (simplify_const_unary_operation,
4595 simplify_const_binary_operation): Likewise.
4596 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
4597 check_va_list_escapes): Likewise.
4598 * rtlanal.c (nonzero_bits1): Likewise.
4599 * expmed.c (expand_smod_pow2): Likewise.
4600 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
4602 2013-08-23 Jan Hubicka <jh@suse.cz>
4604 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
4605 as having address taken.
4607 2013-08-23 Jan Hubicka <jh@suse.cz>
4609 * ipa-utils.h (method_class_type): Declare.
4610 * ipa-devirt.c (method_class_type): Export.
4612 * cgraphunit.c (analyze_functions): Do basic devirtualization;
4613 do not walk base classes of anonymous types.
4615 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
4617 PR rtl-optimization/58220
4619 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
4620 handle SEQUENCE insns properly.
4622 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
4624 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
4626 (pp_newline_and_indent): Likewise.
4627 (pp_separate_with): Likewise.
4628 * pretty-print.c (pp_newline_and_flush): Define.
4629 (pp_newline_and_indent): Likewise.
4630 (pp_separate_with): Likewise.
4632 2013-08-23 Jakub Jelinek <jakub@redhat.com>
4635 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
4636 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
4638 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
4640 * config/i386/predicates.md (ext_sse_reg_operand): New.
4641 * config/i386/i386.md (*movti_internal): Use
4642 predicate to determine if EVEX is needed.
4643 (*movsi_internal): Ditto.
4644 (*movdf_internal): Ditto.
4645 (*movsf_internal): Ditto.
4646 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
4648 2013-08-23 Jakub Jelinek <jakub@redhat.com>
4650 PR tree-optimization/58209
4651 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
4652 (find_tail_calls): Give up for pointer result types if m is non-NULL.
4653 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
4654 emit POINTER_PLUS_EXPR.
4655 (create_tailcall_accumulator): For pointer result type accumulate in
4658 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
4660 * configure.ac: Add backslashes missing from the last change.
4661 * configure: Regenerate.
4663 2013-08-22 Jan Hubicka <jh@susue.cz>
4665 * ipa.c (function_and_variable_visibility): First remember function
4666 was global and then make it local.
4668 2013-08-22 Julian Brown <julian@codesourcery.com>
4670 * configure.ac: Add aarch64 to list of arches which use "nop" in
4672 * configure: Regenerate.
4674 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4676 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
4677 gnu_libc_has_function.
4678 * config/s390/tpf.h: Likewise.
4680 2013-08-22 Jan Hubicka <jh@susue.cz>
4682 * timevar.c (validate_phases): Add cast.
4684 2013-08-22 Jan Hubicka <jh@susue.cz>
4686 * timevar.c (validate_phases): Use size_t for memory.
4687 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
4689 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
4691 * pretty-print.h (output_buffer::output_buffer): Declare.
4692 (pretty_printer::pretty_printer): Likewise.
4693 (pp_construct): Remove.
4694 * pretty-print.c (output_buffer::output_buffer): Define.
4695 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
4696 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
4697 (print_gimple_expr): Likewise.
4698 (print_gimple_seq): Likewise.
4699 (gimple_dump_bb): Likewise.
4700 * sched-vis.c (dump_value_slim): Likewise.
4701 (dump_insn_slim): Likewise.
4702 (dump_rtl_slim): Likewise.
4703 (str_pattern_slim): Likewise.
4704 * tree-mudflap.c (mf_varname_tree): Likewise.
4705 * graph.c (print_graph_cfg): Likewise.
4706 (start_graph_dump): Likewise.
4707 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
4709 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
4710 pretty printer initialization.
4711 * coretypes.h (diagnostic_context): Remove superflous type alias
4713 (pretty_printer): Likewise. Declare directly as a class.
4714 (pretty_print_info): Remove declaration as class.
4715 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
4716 and pp_clear_output_area.
4717 (asan_add_global): Likewise.
4719 2013-08-22 Jan Hubicka <jh@suse.cz>
4721 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
4722 * ipa-utils.h (update_type_inheritance_graph): Declare.
4723 (possible_polymorphic_call_target_p): Declare.
4724 (possible_polymorphic_call_target_p): New.
4725 * ipa-devirt.c: Update toplevel comments.
4726 (cached_polymorphic_call_targets): Move up.
4727 (odr_type_d): Move ID down.
4728 (polymorphic_type_binfo_p): Update comment.
4729 (odr_hasher::remove): Likewise;
4730 (get_odr_type): Set anonymous_namespace.
4731 (dump_odr_type): Dump it.
4732 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
4733 (maybe_record_node): Record node in cached_polymorphic_call_targets.
4734 (record_binfo): Add comment.
4735 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
4737 (devirt_node_removal_hook): Do not iCE when cache is freed.
4738 (possible_polymorphic_call_target_p): New predicate.
4739 (update_type_inheritance_graph): New function.
4741 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
4742 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4743 Sergey Lega <sergey.s.lega@intel.com>
4744 Anna Tikhonova <anna.tikhonova@intel.com>
4745 Ilya Tocar <ilya.tocar@intel.com>
4746 Andrey Turetskiy <andrey.turetskiy@intel.com>
4747 Ilya Verbin <ilya.verbin@intel.com>
4748 Kirill Yukhin <kirill.yukhin@intel.com>
4749 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4751 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
4752 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
4753 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
4754 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
4755 (OPTION_MASK_ISA_AVX2_UNSET): Update.
4756 (OPTION_MASK_ISA_AVX512F_UNSET): New.
4757 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
4758 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
4759 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
4760 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
4761 OPT_mavx512pf, OPT_mavx512er cases.
4762 * config/i386/constraints.md (v): New constraint.
4763 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
4764 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
4765 (bit_AVX512CD): New.
4766 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
4767 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
4768 * config/i386/i386-c.c (ix86_target_macros_internal):
4769 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
4771 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
4772 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
4773 * config/i386/i386.c (regclass_map, dbx_register_map)
4774 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
4775 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
4776 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
4778 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
4779 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
4780 -mavx512pf options. Fix formatting.
4781 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
4782 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
4783 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
4784 -mavx512cd, -mavx512pf options.
4785 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
4786 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
4787 (ix86_preferred_output_reload_class): Replace SSE_REGS with
4789 (ix86_hard_regno_mode_ok): Support 512-bit registers.
4790 (ix86_set_reg_reg_cost): Ditto.
4791 (x86_order_regs_for_local_alloc): Ditto.
4792 (MAX_VECT_LEN): Extend to 64-byte.
4793 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
4794 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
4795 (TARGET_AVX512ER, TARGET_AVX512CD): New.
4796 (BIGGEST_ALIGNMENT): Extend to 512-bits.
4797 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
4798 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
4799 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
4800 (SSE_REG_MODE_P): Support new modes.
4801 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
4802 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
4803 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
4804 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
4805 (REG_CLASS_CONTENTS): Add new registers.
4806 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
4807 (EXT_REX_SSE_REGNO_P): New.
4808 (HI_REGISTER_NAMES): Add new registers.
4809 * config/i386/i386.md: Define constants for new registers.
4810 (mode): Add new 512-bit modes.
4811 (prefix): Support evex prefix.
4812 (isa): Support avx512f, noavx512f, fma_avx512f.
4813 (ssemodesuffix): Add new 512-bit modes.
4815 (*movxi_internal_avx512f): Ditto.
4816 (*movdi_internal): Replace constraint "x" with the new constraint "v".
4818 (*movsi_internal): Likewise.
4819 (*movdf_internal): Likewise.
4820 (*movsf_internal): Likewise.
4821 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
4822 (<code><mode>3): Likewise.
4823 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
4825 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
4826 with the new constraint "v".
4827 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
4829 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
4830 with the new constraint "v".
4831 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
4832 (<sse2>_storedqu<avxsizesuffix>): Likewise.
4833 (*<plusminus_insn><mode>3): Likewise.
4834 (<sse>_vm<plusminus_insn><mode>3): Likewise.
4835 (*mul<mode>3): Likewise.
4836 (<sse>_vmmul<mode>3): Likewise.
4837 (<sse>_div<mode>3): Likewise.
4838 (<sse>_vmdiv<mode>3): Likewise.
4839 (<sse>_sqrt<mode>2): Likewise.
4840 (<sse>_vmsqrt<mode>2): Likewise.
4841 (*<code><mode>3_finite): Likewise.
4842 (*<code><mode>3) <smaxmin>: Likewise.
4843 (<sse>_vm<code><mode>3): Likewise.
4844 (*<code><mode>3) <any_logic>: Likewise.
4845 (*fma_fmadd_<mode>): Likewise.
4846 (*fma_fmsub_<mode>): Likewise.
4847 (*fma_fnmadd_<mode>): Likewise.
4848 (*fma_fnmsub_<mode>): Likewise.
4849 (*fma_fmaddsub_<mode>): Likewise.
4850 (*fma_fmsubadd_<mode>): Likewise.
4851 (*fmai_fmadd_<mode>): Likewise.
4852 (*fmai_fmsub_<mode>): Likewise.
4853 (*fmai_fnmadd_<mode>): Likewise.
4854 (*fmai_fnmsub_<mode>): Likewise.
4855 (sse_cvtsi2ss): Likewise.
4856 (sse_cvtsi2ssq): Likewise.
4857 (sse_cvtss2si): Likewise.
4858 (sse_cvtss2si_2): Likewise.
4859 (sse_cvtss2siq): Likewise.
4860 (sse_cvtss2siq_2): Likewise.
4861 (sse_cvttss2si): Likewise.
4862 (sse_cvtss2siq_2): Likewise.
4863 (float<sseintvecmodelower><mode>2): Likewise.
4864 (sse2_cvtsd2si_2): Likewise.
4865 (sse2_cvtsd2siq_2): Likewise.
4866 (*<plusminus_insn><mode>3): Likewise.
4867 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
4868 (*<sse4_1_avx2>_mul<mode>3): Likewise.
4869 (ashr<mode>3): Likewise.
4870 (<shift_insn><mode>3): Likewise.
4871 (avx2_<code><mode>3): Likewise.
4872 (*avx2_<code><mode>3): Likewise.
4873 (*andnot<mode>3): Likewise.
4874 (*<code><mode>3) <any_logic>: Likewise.
4875 (abs<mode>2): Likewise.
4876 (avx2_permvar<mode>): Likewise.
4877 (avx2_perm<mode>_1): Likewise.
4878 (*avx_vpermilp<mode>): Likewise.
4879 (avx_vpermilvar<mode>3): Likewise.
4880 (avx2_ashrv<mode>): Likewise.
4881 (avx2_<shift_insn>v<mode>): Likewise.
4882 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
4884 * doc/rtl.texi: Document XImode.
4886 2013-08-21 Jeff Law <law@redhat.com>
4888 * tree-flow.h (register_jump_thread): Pass vector of edges
4889 instead of each important edge.
4890 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
4891 thread path into a vector and pass that to register_jump_thread.
4892 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
4893 passed in edge vector to the current 3-edge form.
4896 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
4898 * dce.c (fini_dce): Call df_analyze again just in case
4899 delete_unmarked_insns removed anything.
4901 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
4903 * reload.h (struct reg_equivs): Rename to ..
4904 (struct reg_equivs_s): .. this.
4906 2013-08-20 Martin Liska <marxin.liska@gmail.com>
4908 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
4910 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4912 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
4914 2013-08-21 Jeff Law <law@redhat.com>
4916 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
4917 simplify assignments too. If the RHS collapses to a singleton
4918 range, then return the value for the range.
4920 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
4922 * config/i386/sse.md (V16): Rename to...
4924 (mov<mode>): Update iterator name.
4925 (*mov<mode>_internal): Ditto.
4926 (push<mode>1): Ditto.
4927 (movmisalign<mode>): Ditto.
4929 2013-08-20 Jan Hubicka <jh@suse.cz>
4932 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
4933 entry for direct edges.
4934 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
4936 2013-08-20 David Malcolm <dmalcolm@redhat.com>
4938 Revert my last two changes, r201865 and r201864:
4941 2013-08-20 David Malcolm <dmalcolm@redhat.com>
4943 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
4944 instances can own GC refs.
4946 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
4947 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
4948 (gcc::context::gt_pch_nx): Likewise.
4949 (gcc::context::gt_pch_nx): Likewise.
4950 * ggc.h (gt_ggc_mx <T>): New.
4951 (gt_pch_nx_with_op <T>): New.
4952 (gt_pch_nx <T>): New.
4953 * passes.c (opt_pass::gt_ggc_mx): New.
4954 (opt_pass::gt_pch_nx): New.
4955 (opt_pass::gt_pch_nx_with_op): New.
4956 (pass_manager::gt_ggc_mx): New.
4957 (pass_manager::gt_pch_nx): New.
4958 (pass_manager::gt_pch_nx_with_op): New.
4959 (pass_manager::operator new): Use
4960 ggc_internal_cleared_alloc_stat rather than xcalloc.
4961 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
4962 (pass_manager::gt_ggc_mx): New.
4963 (pass_manager::gt_pch_nx): New.
4964 (pass_manager::gt_pch_nx_with_op): New.
4965 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
4966 (opt_pass::operator new): New.
4967 (opt_pass::gt_ggc_mx): New.
4968 (opt_pass::gt_pch_nx): New.
4969 (opt_pass::gt_pch_nx_with_op): New.
4972 2013-08-20 David Malcolm <dmalcolm@redhat.com>
4974 * Makefile.in (GTFILES): Add context.h.
4975 * context.c (gcc::context::operator new): New.
4976 (gcc::context::gt_ggc_mx): New.
4977 (gcc::context::gt_pch_nx): New.
4978 (gcc::context::gt_pch_nx): New.
4979 * context.h (gcc::context): Add GTY((user)) marking.
4980 (gcc::context::operator new): New.
4981 (gcc::context::gt_ggc_mx): New.
4982 (gcc::context::gt_pch_nx): New.
4983 (gcc::context::gt_pch_nx): New.
4984 (g): Add GTY marking.
4985 (gt_ggc_mx (gcc::context *)): New.
4986 (gt_pch_nx (gcc::context *)): New.
4987 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
4988 void *cookie)): New.
4989 * gengtype.c (open_base_files) <ifiles>: Add context.h.
4991 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
4993 * dce.c (fini_dce): Call df_analyze again just in case
4994 delete_unmarked_insns removed anything.
4996 2013-08-20 Teresa Johnson <tejohnson@google.com>
4998 PR rtl-optimizations/57451
4999 * final.c (reemit_insn_block_notes): Prevent lexical blocks
5000 from crossing split section boundaries.
5002 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
5004 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
5005 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
5006 with MULTLIB_DEFAULTS.
5008 2013-08-20 Nick Clifton <nickc@redhat.com>
5010 * target.def (narrow_volatile_bitfield): Note that the default
5011 value is false, not !TARGET_STRICT_ALIGN.
5012 * doc/tm.texi: Regenerate.
5014 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
5016 Fix LIB_SPEC for systems without libpthread.
5018 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
5019 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
5021 * config/i386/linux-common.h: Likewise.
5022 * config/mips/linux-common.h: Likewise.
5024 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
5026 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
5029 2013-08-20 David Malcolm <dmalcolm@redhat.com>
5031 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
5032 instances can own GC refs.
5034 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
5035 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
5036 (gcc::context::gt_pch_nx): Likewise.
5037 (gcc::context::gt_pch_nx): Likewise.
5038 * ggc.h (gt_ggc_mx <T>): New.
5039 (gt_pch_nx_with_op <T>): New.
5040 (gt_pch_nx <T>): New.
5041 * passes.c (opt_pass::gt_ggc_mx): New.
5042 (opt_pass::gt_pch_nx): New.
5043 (opt_pass::gt_pch_nx_with_op): New.
5044 (pass_manager::gt_ggc_mx): New.
5045 (pass_manager::gt_pch_nx): New.
5046 (pass_manager::gt_pch_nx_with_op): New.
5047 (pass_manager::operator new): Use
5048 ggc_internal_cleared_alloc_stat rather than xcalloc.
5049 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
5050 (pass_manager::gt_ggc_mx): New.
5051 (pass_manager::gt_pch_nx): New.
5052 (pass_manager::gt_pch_nx_with_op): New.
5053 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
5054 (opt_pass::operator new): New.
5055 (opt_pass::gt_ggc_mx): New.
5056 (opt_pass::gt_pch_nx): New.
5057 (opt_pass::gt_pch_nx_with_op): New.
5059 2013-08-20 David Malcolm <dmalcolm@redhat.com>
5061 * Makefile.in (GTFILES): Add context.h.
5062 * context.c (gcc::context::operator new): New.
5063 (gcc::context::gt_ggc_mx): New.
5064 (gcc::context::gt_pch_nx): New.
5065 (gcc::context::gt_pch_nx): New.
5066 * context.h (gcc::context): Add GTY((user)) marking.
5067 (gcc::context::operator new): New.
5068 (gcc::context::gt_ggc_mx): New.
5069 (gcc::context::gt_pch_nx): New.
5070 (gcc::context::gt_pch_nx): New.
5071 (g): Add GTY marking.
5072 (gt_ggc_mx (gcc::context *)): New.
5073 (gt_pch_nx (gcc::context *)): New.
5074 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
5075 void *cookie)): New.
5076 * gengtype.c (open_base_files) <ifiles>: Add context.h.
5078 2013-08-20 Alan Modra <amodra@gmail.com>
5081 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
5082 (rs6000_emit_epilogue): Likewise.
5084 2013-08-19 Dehao Chen <dehao@google.com>
5086 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
5088 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
5089 Jakub Jelinek <jakub@redhat.com>
5091 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
5092 (BUILT_IN_FABSD64): Likewise.
5093 (BUILT_IN_FABSD128): Likewise.
5094 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
5095 (fold_builtin_1): Likewise.
5096 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
5097 destination and source operands.
5098 (*abstd2_fpr): Likewise.
5099 (*nabstd2_fpr): Likewise.
5101 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
5103 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
5106 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
5108 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
5111 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
5113 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
5114 * builtins.c (default_libc_has_function): New.
5115 (gnu_libc_has_function): Ditto.
5116 (no_c99_libc_has_function): Ditto.
5117 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
5118 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
5119 (fold_builtin_sincos): Likewise.
5120 (fold_builtin_cexp): Likewise.
5121 * builtins.def (DEF_C94_BUILTIN): Likewise.
5122 (DEF_C99_BUILTIN): Likewise.
5123 (DEF_C99_C90RES_BUILTIN): Likewise.
5124 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
5125 definitions to using this define.
5126 * config/darwin-protos.h (darwin_libc_has_function): New.
5127 * config/darwin.c (darwin_libc_has_function): Ditto.
5128 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
5129 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
5130 * config/darwin.h: Ditto.
5131 * config/elfos.h: Ditto.
5132 * config/freebsd.h: Ditto.
5133 * config/i386/cygming.h: Ditto.
5134 * config/i386/djgpp.h: Ditto.
5135 * config/i386/i386-interix.h: Ditto.
5136 * config/microblaze/microblaze.h: Ditto.
5137 * config/mmix/mmix.h: Ditto.
5138 * config/gnu-user.h: Ditto.
5139 * config/ia64/hpux.h: Ditto.
5140 * config/pa/pa-hpux.h: Ditto.
5141 * config/pdp11/pdp11.h: Ditto.
5142 * config/picochip/picochip.h: Ditto.
5143 * config/linux.h: Ditto.
5144 * config/netbsd.h: Ditto.
5145 * config/openbsd.h: Ditto.
5146 * config/rs6000/aix43.h: Ditto.
5147 * config/rs6000/aix51.h: Ditto.
5148 * config/rs6000/aix52.h: Ditto.
5149 * config/rs6000/aix53.h: Ditto.
5150 * config/rs6000/aix61.h: Ditto.
5151 * config/rs6000/darwin.h: Ditto.
5152 * config/rs6000/linux.h: Ditto.
5153 * config/rs6000/linux64.h: Ditto.
5154 * config/s390/tpf.h: Ditto.
5155 * config/sol2-10.h: Ditto.
5156 * config/sol2.h: Ditto.
5157 * config/vms/vms.h: Ditto.
5158 * config/vxworks.h: Ditto.
5159 * config/linux-android.c (linux_android_libc_has_function):
5160 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
5161 * config/linux-protos.h (linux_android_libc_has_function):
5163 * config/i386/i386.c (ix86_libc_has_function): New.
5164 * config/i386/i386-protos.h
5165 (ix86_libc_has_function): New declaration.
5166 * config/i386/i386.md
5167 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
5168 ("isinf<mode>2): Likewise.
5169 * convert.c (convert_to_integer): Using new target hook
5170 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
5171 TARGET_C99_FUNCTIONS.
5172 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
5173 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
5174 * coretypes.h (function_class): New enum for different
5175 classes of functions.
5176 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
5177 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
5178 (TARGET_HAS_SINCOS): Likewise.
5179 (TARGET_LIBC_HAS_FUNCTION): New.
5180 * doc/tm.texi: Regenerated.
5181 * targhooks.h (default_libc_has_function): New declaration.
5182 (no_c99_libc_has_function): Ditto.
5183 (gnu_libc_has_function): Ditto.
5184 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
5185 and TARGET_HAS_SINCOS.
5187 2013-08-18 Jan Hubicka <jh@suse.cz>
5189 * Makeifle-in (ipa-devirt.o): New.
5190 (GTFILES): Add ipa-utils.h and ipa-devirt.c
5191 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
5192 (analyze_functions): Look into possible targets of polymorphic call.
5193 * dumpfile.c (dump_files): Add type-inheritance dump.
5194 * dumpfile.h (TDI_inheritance): New.
5195 * ipa-devirt.c: New file.
5196 * ipa-utils.h (odr_type_d): Forward declare.
5197 (odr_type): New type.
5198 (build_type_inheritance_graph): Declare.
5199 (possible_polymorphic_call_targets): Declare and introduce inline
5200 variant when only edge is pased.
5201 (dump_possible_polymorphic_call_targets): Likewise.
5202 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
5203 * tree.c (type_in_anonymous_namespace_p): Break out from ...
5204 (types_same_for_odr): ... here.
5205 * tree.h (type_in_anonymous_namespace_p): Declare.
5207 2013-08-18 Jakub Jelinek <jakub@redhat.com>
5209 PR tree-optimization/58006
5210 * tree-parloops.c (take_address_of): Don't ICE if get_name
5212 (eliminate_local_variables_stmt): Remove clobber stmts.
5214 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
5216 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
5217 error message is issued for an alias to undefined symbol.
5219 2013-08-18 Jan Hubicka <jh@suse.cz>
5221 * cgraph.c (cgraph_create_indirect_edge): Discover
5222 polymorphic calls and record basic info into indirect_info.
5223 * gimple-fold.c (gimple_fold_call): When doing BINFO based
5224 devirtualization, ignore objc function calls.
5225 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
5226 call with no parm index info.
5227 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
5228 * tree.c (virtual_method_call_p): New function.
5229 * tree.h (virtual_method_call_p): Declare.
5231 2013-08-16 Jan Hubicka <jh@suse.cz>
5234 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
5236 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
5238 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
5241 2013-08-16 David Malcolm <dmalcolm@redhat.com>
5243 * gengtype.c (type_for_name): Add special-case support for
5244 locating types within the "gcc::" namespace.
5245 (open_base_files): Emit a "using namespace gcc" directive.
5247 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
5250 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
5251 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
5253 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
5254 array instead of each individual operand as a separate argument.
5255 (emit_fusion_gpr_load): Likewise.
5256 (expand_fusion_gpr_load): Add new function declaration.
5258 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
5259 signature to have the operands passed as an array, instead of as
5260 separate arguments. Allow ZERO_EXTEND to be in the memory
5261 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
5262 depend on the register live/dead flags when peepholes are run.
5263 (expand_fusion_gpr_load): New function to be called from the
5264 peephole2 pass, to change the register that addis sets to be the
5266 (emit_fusion_gpr_load): Change the calling signature to have the
5267 operands passed as an array, instead of as separate arguments.
5268 Allow ZERO_EXTEND to be in the memory address, and also
5269 SIGN_EXTEND if -mpower8-fusion-sign.
5271 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
5273 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
5274 adjust the register addis loads up in the peephole2 pass. Do not
5275 depend on the register live/dead state when the peephole pass is done.
5277 2013-08-16 David Malcolm <dmalcolm@redhat.com>
5279 * gengtype.c (create_user_defined_type): Ensure that the kind
5280 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
5281 declaration is seen before the GTY((user)) marking.
5283 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
5286 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
5288 2013-08-16 Jan Hubicka <jh@suse.cz>
5290 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
5291 arugment expected_type.
5292 (gimple_fold_call): Use it.
5293 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
5294 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
5295 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
5296 (try_make_edge_direct_virtual_call): Likewise.
5297 * tree.c (obj_type_ref_class): New.
5298 * tree.h (obj_type_ref_class): Use it.
5300 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
5302 * sched-vis.c (rtl_slim_pp_initialized): Remove.
5303 (rtl_slim_pp): Likewise.
5304 (init_rtl_slim_pretty_print): Likewise.
5305 (dump_value_slim): Don't call it. Use local pretty printer.
5306 (dump_insn_slim): Likewise.
5307 (dump_rtl_slim): Likewise.
5308 (str_pattern_slim): Likewise.
5309 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
5312 2013-08-16 Jakub Jelinek <jakub@redhat.com>
5314 PR tree-optimization/58164
5315 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
5316 walk gimple_goto_dest of GIMPLE_GOTO.
5318 PR tree-optimization/58165
5319 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
5320 bi_call must be the last stmt in a bb, don't split_block, instead
5321 use fallthru edge from it and give up if there is none.
5322 Release conds vector when returning early.
5324 2013-08-14 Xinliang David Li <davidxl@google.com>
5326 * config/i386/i386.c (ix86_option_override_internal):
5327 Remove unused variable and field.
5329 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5332 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
5333 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
5334 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
5335 and Linux, correct BLKmode alignment when 128-bit alignment is
5336 required and compatibility flag is not set.
5337 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
5338 for zero-size arguments when compatibility flag is not set.
5340 2013-08-14 Jakub Jelinek <jakub@redhat.com>
5342 PR tree-optimization/58145
5343 * tree-sra.c (build_ref_for_offset): If prev_base has
5344 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
5346 2013-08-14 Xinliang David Li <davidxl@google.com>
5348 * config/i386/i386.c (ix86_option_override_internal):
5349 Fix uninitialized variable error.
5351 2013-08-14 Xinliang David Li <davidxl@google.com>
5353 * config/i386/i386.opt: Define two new options.
5354 * config/i386/x86-tune.def: Add arch selector field in macros.
5355 * config/i386/i386.h: Adjust macro definition.
5356 * config/i386/i386.c (ix86_option_override_internal):
5358 (parse_mtune_ctrl_str): New function.
5359 (set_ix86_tune_features): New function.
5360 (ix86_function_specific_restore): Call the new helper function.
5362 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
5364 PR rtl-optimization/57662
5365 * sel-sched.c (code_motion_process_successors): When the current insn
5366 is removed after the recursive traversal, break from the loop.
5367 Add comments and debug printouts.
5369 2013-08-14 Jakub Jelinek <jakub@redhat.com>
5370 Alexandre Oliva <aoliva@redhat.com>
5373 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
5374 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
5375 there also UNSPEC_PLTOFF.
5377 2013-08-14 Marek Polacek <polacek@redhat.com>
5379 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
5380 NUM_CONDITIONS bit positions.
5382 2013-08-13 Cary Coutant <ccoutant@google.com>
5384 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
5385 (attr_checksum): Hash vector contents instead of pointer.
5386 (attr_checksum_ordered): Likewise.
5388 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
5390 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
5391 when Pmode != word_mode. Add length_address attribute.
5392 (sse3_monitor_<mode>): Merge from sse3_monitor and
5393 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
5394 Pmode != word_mode. Update insn length attribute.
5395 * config/i386/i386.c (ix86_option_override_internal): Update
5396 ix86_gen_monitor selection for merged sse3_monitor insn.
5398 2013-08-13 Julian Brown <julian@codesourcery.com>
5400 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
5401 perform invalid legitimization on greater-than-word-size modes for
5404 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
5406 * ira.c (setup_class_translate_array): Use aclass instead of cl
5407 for classes not fully covered by allocno classes.
5409 2013-08-13 Jakub Jelinek <jakub@redhat.com>
5411 PR tree-optimization/57661
5412 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
5413 * tree-inline.c (tree_function_versioning): Initialize it.
5414 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
5415 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
5416 that is not being copied.
5419 * asan.c (instrument_strlen_call): Fix typo in comment.
5420 Use char * type even for the lhs of POINTER_PLUS_EXPR.
5422 2013-08-13 Steve Ellcey <sellcey@mips.com>
5424 * config/mips/mips.md (prefetch): Use lw instead of ld on
5425 loongson in 32bit mode.
5427 2013-08-13 Nick Clifton <nickc@redhat.com>
5429 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
5431 2013-08-13 Jan Hubicka <jh@suse.cz>
5433 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
5434 introduced edge; fix typo in sanity check.
5435 (cgraph_resolve_speculation): Export; improve diagnostic.
5436 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
5437 speculation at type mismatch.
5438 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
5439 (cgraph_resolve_speculation): Declare.
5440 (symtab_can_be_discarded): New function.
5441 * value-prof.c (gimple_ic_transform): Remove actual transform code.
5442 * ipa-inline-transform.c (speculation_removed): New global var.
5443 (clone_inlined_nodes): See if speculation can be removed.
5444 (inline_call): If speculations was removed, we growths may not match.
5445 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
5446 (speculation_useful_p): New function.
5447 (resolve_noninline_speculation): New function.
5448 (inline_small_functions): Resolve useless speculations.
5449 * ipa-inline.h (speculation_useful_p): Declare
5450 * ipa.c (can_replace_by_local_alias): Simplify.
5451 (ipa_profile): Produce speculative calls in non-lto, too;
5452 add simple cost model; produce local aliases.
5454 2013-08-13 David Malcolm <dmalcolm@redhat.com>
5456 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
5459 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
5461 * config/i386/i386.c (ix86_function_versions): Use error + inform.
5463 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
5465 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
5466 iterator instead of X87MODEF.
5468 2013-08-12 Perez Read <netfirewall@gmail.com>
5471 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
5472 operand 0 for intel asm alternative.
5473 (*movabs<mode>_2): Ditto for operand 1.
5475 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
5477 * config/aarch64/arm_none.h
5478 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
5480 2013-08-12 Nick Clifton <nickc@redhat.com>
5482 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
5484 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
5486 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
5487 expand for QI/HImode operand to produce more effictive code for
5488 unsigned char(short) --> float(double) conversion.
5490 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
5492 * doc/invoke.texi: Mention that -ftls-model does not force the final
5495 2013-08-12 Marek Polacek <polacek@redhat.com>
5496 Marc Glisse <marc.glisse@inria.fr>
5498 PR tree-optimization/57980
5499 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
5500 when creating -1 constant.
5502 2013-08-10 Jan Hubicka <jh@suse.cz>
5504 Workaround binutils PR14342.
5505 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
5506 (gimple_init_edge_profiler): Likewise.
5507 (gimple_gen_ic_func_profiler): Likewise.
5509 2013-08-09 Jan Hubicka <jh@suse.cz>
5511 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
5513 2013-08-09 Xinliang David Li <davidxl@google.com>
5515 * config/i386/stringop.def: New file.
5516 * config/i386/stringop.opt: New file.
5517 * config/i386/i386-opts.h: Include stringopt.def.
5518 * config/i386/i386.opt: Include stringopt.opt.
5519 * config/i386/i386.c (ix86_option_override_internal):
5520 Override default size based stringop inline strategies with options.
5521 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
5524 2013-08-09 Jan Hubicka <jh@suse.cz>
5526 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
5528 2013-08-09 Jan Hubicka <jh@suse.cz>
5530 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
5532 (dump_cgraph_node): Dump profile-id.
5533 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
5534 and common_target_probability.
5535 * lto-cgraph.c (lto_output_edge): Stream common targets.
5536 (lto_output_node): Stream profile ids.
5537 (input_node): Stream profile ids.
5538 (input_edge): Stream common targets.
5539 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
5540 * ipa.c: Include value-prof.h
5541 (ipa_profile_generate_summary): Turn indirect call statement histograms
5542 into common targets.
5543 (ipa_profile): Turn common targets into speculative edges.
5545 2013-08-09 Jan Hubicka <jh@suse.cz>
5547 * cgraph.h (cgraph_node): Add profile_id.
5548 * value-prof.c (cgraph_node_map): Turn into pointer_map.
5549 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
5550 (del_node_map): Update.
5551 (find_func_by_funcdef_no): Replace by ...
5552 (find_func_by_profile_id): ... this one.
5553 (gimple_ic_transform): Do not remove useful histograms when
5554 speculation is not done; dump info when indirect call removal
5556 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
5557 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
5558 (__gcov_indirect_call_profiler_v2): .. this one.
5559 * profile.h (init_node_map): Update.
5560 * coverage.c (coverage_compute_profile_id): New function.
5561 * coverage.h (coverage_compute_profile_id): Declare.
5562 * tree-profile.c (init_ic_make_global_vars): Make
5563 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
5564 (gimple_init_edge_profiler): Update prototype of
5565 __gcov_indirect_call_profiler.
5566 (gimple_gen_ic_func_profiler): Simplify.
5567 (tree_profiling): Use init_node_map
5569 2013-08-09 Jan Hubicka <jh@suse.cz>
5571 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
5572 non-speculative refs.
5573 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
5574 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
5575 (cgraph_set_call_stmt): Likewise.
5576 (cgraph_create_edge_1): Fix release checking compilatoin;
5578 (cgraph_free_edge): Free indirect info.
5579 (cgraph_turn_edge_to_speculative): New function.
5580 (cgraph_speculative_call_info): New function.
5581 (cgraph_make_edge_direct): Return direct edge; handle speculation.
5582 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
5583 (dump_cgraph_node): Dump speculation.
5584 (verify_edge_count_and_frequency): Accept speculative edges.
5585 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
5586 (verify_cgraph_node): Handle speculation.
5587 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
5588 (cgraph_set_call_stmt): Update prototype.
5589 (cgraph_make_edge_direct): Update prototype.
5590 (cgraph_speculative_call_info): Declare.
5591 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
5592 to change; update call of ipa_find_references.
5593 * ipa-ref.c (ipa_record_reference): Fix return value; clear
5594 lto_stmt_uid and speculative flags.
5595 (ipa_dump_references): Dump speculation.
5596 (ipa_clone_references): Clone speculative flag.
5597 (ipa_clone_referring): Likewise.
5598 (ipa_clone_ref): New function.
5599 (ipa_find_reference): Look into lto_stmt_uids
5600 (ipa_clear_stmts_in_references): Do not clear speculative calls.
5601 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
5602 (ipa_find_reference): Update declaration.
5603 (ipa_clone_ref): Declare.
5604 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
5605 stream speculative flag.
5606 (lto_output_ref): Stream statements uids and speculation.
5607 (input_ref): Likewise.
5608 (input_edge): Stream speuclation.
5609 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
5610 (cgraph_set_call_stmt_including_clones): Handle speculation.
5611 * ipa-inline.c (heap_edge_removal_hook): New function.
5612 (inline_small_functions): Register it.
5613 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
5614 also initialize refs.
5615 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
5617 (try_make_edge_direct_simple_call): Likewise.
5618 (try_make_edge_direct_simple_call): Likewise.
5619 (update_indirect_edges_after_inlining): Likewise.
5620 (remove_described_reference): Look proper lto_stmt_uid.
5621 (propagate_controlled_uses): Likewise.
5622 (propagate_controlled_uses): Liekwise.
5623 * tree-inline.c (copy_bb): Copy speculative edges.
5624 (redirect_all_calls): New function.
5625 (copy_cfg_body): Do redirection after loop info is updated.
5626 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
5628 2013-08-09 Jan Hubicka <jh@suse.cz>
5630 * lto-streamer-out.c (output_function): Renumber PHIs.
5631 * lto-streamer-in.c (input_function): Likewise.
5633 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
5635 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
5636 (get_lane_unsigned): Likewise.
5637 (dup_lane_scalar): Likewise.
5638 (get_lane): enable for VALL.
5639 * config/aarch64/aarch64-simd.md
5640 (aarch64_dup_lane_scalar<mode>): Remove.
5641 (aarch64_get_lane_signed<mode>): Likewise.
5642 (aarch64_get_lane_unsigned<mode>): Likewise.
5643 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
5644 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
5645 (aarch64_get_lane<mode>): Enable for all vector modes.
5646 (aarch64_get_lanedi): Remove misleading constraints.
5647 * config/aarch64/arm_neon.h
5648 (__aarch64_vget_lane_any): Define.
5649 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
5650 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
5651 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
5652 * config/aarch64/iterators.md (VDQQH): New.
5656 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
5658 * configure.ac: Add GAS check for LEON instructions on SPARC.
5659 * configure: Regenerate.
5660 * config.in: Likewise.
5661 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
5663 * config/sparc/sparc.opt (LEON, LEON3): New masks.
5664 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
5666 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
5667 (AS_LEON_FLAG): New macro.
5668 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
5669 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
5670 Deal with LEON and LEON3 for the memory model.
5671 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
5672 (atomic_compare_and_swap<mode>_1): Likewise.
5673 (*atomic_compare_and_swap<mode>_1): Likewise.
5675 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
5677 * config/arm/neon.md (vcond): Fix floating-point vector
5678 comparisons against 0.
5680 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
5682 * lra-constraints.c (emit_spill_move): Remove assert.
5683 (process_alt_operands): Add more debugging
5684 output. Increase reject for spilling into memory. Decrease
5685 reject for reloading scratch.
5686 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
5688 2013-08-08 Steve Ellcey <sellcey@mips.com>
5690 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
5691 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
5692 micromips incompatible. Add nan2008.
5693 (MULTILIB_DIRNAMES): Add nan2008.
5694 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
5695 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
5696 and micromips incompatible. Add nan2008.
5697 (MULTILIB_DIRNAMES): Add nan2008.
5698 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
5700 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
5702 PR rtl-optimization/58079
5703 * combine.c (combine_simplify_rtx): Avoid using SUBST if
5704 simplify_comparison has widened a comparison with an integer.
5706 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5708 * config/arm/neon.md (movmisalign<mode>): Disable when we
5709 don't allow unaligned accesses.
5710 (*movmisalign<mode>_neon_store): Likewise.
5711 (*movmisalign<mode>_neon_load): Likewise.
5712 (*movmisalign<mode>_neon_store): Likewise.
5713 (*movmisalign<mode>_neon_load): Likewise.
5715 2013-08-08 Jan Hubicka <jh@suse.cz>
5717 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
5718 (make_pass_rebuild_cgraph_edges): Also clear references.
5719 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
5720 * ipa-inline-transform.c (inline_transform): Remove all references
5722 * cgraphunit.c (expand_function): Remove all references after
5724 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
5725 (ipa_find_reference): Rewrite to iterator.
5726 (remove_stmt_references): Likewise.
5727 (ipa_clear_stmts_in_references): New function.
5728 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
5729 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
5731 * ipa-split.c (split_function): Remove references in split function.
5733 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
5736 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
5737 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
5739 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
5742 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
5743 suggested mode for the assignment isn't compatible with the
5746 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
5749 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
5751 2013-08-07 Xinliang David Li <davidxl@google.com>
5753 * config/i386/i386.opt: New option -mtune-ctrl=.
5754 * config/i386/x86-tune.def: New file.
5755 * config/i386/i386.h: include x86-tune.def.
5756 * config/i386/i386.c (ix86_option_override_internal):
5757 Parsing -mtune-ctrl= option and set tune features.
5759 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
5762 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
5763 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
5766 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
5768 * rtl.h (update_alignments): Declare.
5769 * final.c (grow_label_align): New function extracted from...
5770 (shorten_branches): ...here. Call it.
5771 (update_alignments): New function.
5772 * reorg.c (sibling_labels): New variable.
5773 (get_label_before): Add SIBLING parameter. If it is non-zero, push
5774 the new label along with it onto the sibling_labels vector.
5775 (fill_simple_delay_slots): Adjust call to get_label_before.
5776 (fill_slots_from_thread): Likewise.
5777 (relax_delay_slots): Likewise.
5778 (make_return_insns): Likewise.
5779 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
5781 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
5783 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
5784 document its semantics.
5785 (diagnostic_report_diagnostic): Adjust accordingly.
5787 2013-08-07 David Malcolm <dmalcolm@redhat.com>
5789 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
5790 (sparc_option_override): ...and port to new C++ pass API.
5791 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
5793 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
5795 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
5797 2013-08-06 Caroline Tice <cmtice@google.com>
5799 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
5800 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
5801 * tree-pass.h: Add pass_vtable_verify.
5802 * varasm.c (assemble_variable): Add code to properly set the comdat
5803 section and name for the .vtable_map_vars section.
5804 (assemble_vtyv_preinit_initializer): New function.
5805 (default_sectin_type_flags): Make sure .vtable_map_vars section has
5807 * output.h: Add function decl for assemble_vtv_preinit_initializer.
5808 * vtable-verify.c: New file.
5809 * vtable-verify.h: New file.
5810 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
5811 initialiation levels.
5812 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
5813 * passes.def: Insert pass_vtable_verify.
5814 * aclocal.m4: Reorder includes.
5815 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
5816 -fvtv-counts options.
5817 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
5818 as appropriate, if -fvtable-verify=... is used.
5819 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
5820 -fvtable-verify=... is used.
5821 * Makefile.in (OBJS): Add vtable-verify.o to list.
5822 (vtable-verify.o): Add new build rule.
5823 (GTFILES): Add vtable-verify.c to list.
5824 * common.opt (fvtable-verify=): New flag.
5825 (vtv_priority): Values for fvtable-verify= flag.
5826 (fvtv-counts): New flag.
5827 (fvtv-debug): New flag.
5828 * tree.h (save_vtable_map_decl): New extern function decl.
5830 2013-08-07 David Malcolm <dmalcolm@redhat.com>
5832 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
5833 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
5834 (pass_data_rl78_devirt): ...new pass_data instance and...
5835 (make_pass_rl78_devirt): ...new function.
5836 (rl78_asm_file_start): Port pass registration to new C++ API.
5838 2013-08-07 David Malcolm <dmalcolm@redhat.com>
5840 * coretypes.h (rtl_opt_pass): Add.
5841 (gcc::context): Add.
5842 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
5843 (epiphany_init): Port to new C++ pass API.
5844 (epiphany_optimize_mode_switching): Likewise.
5845 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
5846 (pass_manager::get_pass_mode_switching): New.
5847 (pass_manager::get_pass_peephole2): New.
5848 * mode-switching.c (pass_mode_switching): Add clone method.
5849 * recog.c (pass_peephole2): Add clone method.
5850 (pass_split_all_insns): Add clone method.
5852 2013-08-06 David Malcolm <dmalcolm@redhat.com>
5854 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
5855 (mips_option_override): ...here, porting to new C++ API for passes.
5857 2013-08-06 Jan Hubicka <jh@suse.cz>
5859 * cgraph.c (cgraph_get_body): New function based on lto.c
5861 * cgraph.h (cgraph_get_body): Declare.
5862 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
5864 * cgraphunit.c (expand_function): Get body prior expanding.
5865 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
5866 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
5868 * passes.c (do_per_function_toporder): Get body.
5869 * tree-inline.c (expand_call_inline): Get body prior inlining it.
5870 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
5872 2013-08-06 Martin Jambor <mjambor@suse.cz>
5875 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
5876 is not re-finalized. Rename second parameter to no_collect.
5878 2013-08-06 Martin Jambor <mjambor@suse.cz>
5881 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
5882 MEM_REF has proper alignment information.
5884 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
5887 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
5889 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
5890 rtx (*) (rtx, ...) with insn_gen_fn.
5891 * genoutput.c (output_insn_data): Cast gen_? function pointers to
5892 insn_gen_fn::stored_funcptr. Add initializer braces.
5894 2013-08-05 David Malcolm <dmalcolm@redhat.com>
5896 Rewrite how instances of passes are cloned to remove assumptions
5897 about their sizes (thus allowing pass subclasses to have
5898 additional data fields, albeit non-GC-managed ones at this point).
5900 * passes.c (make_pass_instance): Now that passes have clone
5901 methods, rewrite this function to eliminate XNEW and memcpy
5902 calls that used hardcoded sizes. Since this function no longer
5903 creates pass instances, rename it to...
5904 (add_pass_instance): ...this. Document the old way that passes were
5905 numbered and flagged, and rework this function to continue using it.
5906 (next_pass_1): Add an initial_pass argument for use by
5908 (position_pass): When adding multiple instances of a pass, use
5909 the pass's clone method, rather than relying on the XNEW/memcpy
5910 within the former make_pass_instance (now add_pass_instance).
5911 (pass_manager::pass_manager): When invoking next_pass_1, also supply
5912 the initial instance of the current pass within the pass manager.
5914 2013-08-05 David Malcolm <dmalcolm@redhat.com>
5916 This is the automated part of the conversion of passes from C
5917 structs to C++ classes.
5919 Patch autogenerated by refactor_passes.py from
5920 https://github.com/davidmalcolm/gcc-refactoring-scripts
5921 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
5923 * asan.c (pass_asan): Convert from a global struct to a subclass of
5924 gimple_opt_pass along with...
5925 (pass_data_asan): ...new pass_data instance and...
5926 (make_pass_asan): ...new function.
5927 (pass_asan_O0): Convert from a global struct to a subclass of
5928 gimple_opt_pass along with...
5929 (pass_data_asan_O0): ...new pass_data instance and...
5930 (make_pass_asan_O0): ...new function.
5931 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
5932 subclass of rtl_opt_pass along with...
5933 (pass_data_inc_dec): ...new pass_data instance and...
5934 (make_pass_inc_dec): ...new function.
5935 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
5936 a subclass of rtl_opt_pass along with...
5937 (pass_data_reorder_blocks): ...new pass_data instance and...
5938 (make_pass_reorder_blocks): ...new function.
5939 (pass_duplicate_computed_gotos): Convert from a global struct to a
5940 subclass of rtl_opt_pass along with...
5941 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
5942 (make_pass_duplicate_computed_gotos): ...new function.
5943 (pass_partition_blocks): Convert from a global struct to a subclass of
5944 rtl_opt_pass along with...
5945 (pass_data_partition_blocks): ...new pass_data instance and...
5946 (make_pass_partition_blocks): ...new function.
5947 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
5948 struct to a subclass of rtl_opt_pass along with...
5949 (pass_data_branch_target_load_optimize1): ...new pass_data instance
5951 (make_pass_branch_target_load_optimize1): ...new function.
5952 (pass_branch_target_load_optimize2): Convert from a global struct to a
5953 subclass of rtl_opt_pass along with...
5954 (pass_data_branch_target_load_optimize2): ...new pass_data instance
5956 (make_pass_branch_target_load_optimize2): ...new function.
5957 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
5958 of rtl_opt_pass along with...
5959 (pass_data_jump): ...new pass_data instance and...
5960 (make_pass_jump): ...new function.
5961 (pass_jump2): Convert from a global struct to a subclass of
5962 rtl_opt_pass along with...
5963 (pass_data_jump2): ...new pass_data instance and...
5964 (make_pass_jump2): ...new function.
5965 * cfgexpand.c (pass_expand): Convert from a global struct to a
5966 subclass of rtl_opt_pass along with...
5967 (pass_data_expand): ...new pass_data instance and...
5968 (make_pass_expand): ...new function.
5969 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
5970 of rtl_opt_pass along with...
5971 (pass_data_free_cfg): ...new pass_data instance and...
5972 (make_pass_free_cfg): ...new function.
5973 (pass_into_cfg_layout_mode): Convert from a global struct to a
5974 subclass of rtl_opt_pass along with...
5975 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
5976 (make_pass_into_cfg_layout_mode): ...new function.
5977 (pass_outof_cfg_layout_mode): Convert from a global struct to a
5978 subclass of rtl_opt_pass along with...
5979 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
5980 (make_pass_outof_cfg_layout_mode): ...new function.
5981 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
5982 struct to a subclass of gimple_opt_pass along with...
5983 (pass_data_build_cgraph_edges): ...new pass_data instance and...
5984 (make_pass_build_cgraph_edges): ...new function.
5985 (pass_rebuild_cgraph_edges): Convert from a global struct to a
5986 subclass of gimple_opt_pass along with...
5987 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
5988 (make_pass_rebuild_cgraph_edges): ...new function.
5989 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
5990 subclass of gimple_opt_pass along with...
5991 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
5993 (make_pass_remove_cgraph_callee_edges): ...new function.
5994 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
5995 struct to a subclass of rtl_opt_pass along with...
5996 (pass_data_stack_adjustments): ...new pass_data instance and...
5997 (make_pass_stack_adjustments): ...new function.
5998 * combine.c (pass_combine): Convert from a global struct to a subclass
5999 of rtl_opt_pass along with...
6000 (pass_data_combine): ...new pass_data instance and...
6001 (make_pass_combine): ...new function.
6002 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
6003 global struct to a subclass of rtl_opt_pass along with...
6004 (pass_data_compare_elim_after_reload): ...new pass_data instance
6006 (make_pass_compare_elim_after_reload): ...new function.
6007 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
6008 of rtl_opt_pass along with...
6009 (pass_data_rtl_cprop): ...new pass_data instance and...
6010 (make_pass_rtl_cprop): ...new function.
6011 * cse.c (pass_cse): Convert from a global struct to a subclass of
6012 rtl_opt_pass along with...
6013 (pass_data_cse): ...new pass_data instance and...
6014 (make_pass_cse): ...new function.
6015 (pass_cse2): Convert from a global struct to a subclass of
6016 rtl_opt_pass along with...
6017 (pass_data_cse2): ...new pass_data instance and...
6018 (make_pass_cse2): ...new function.
6019 (pass_cse_after_global_opts): Convert from a global struct to a
6020 subclass of rtl_opt_pass along with...
6021 (pass_data_cse_after_global_opts): ...new pass_data instance and...
6022 (make_pass_cse_after_global_opts): ...new function.
6023 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
6024 of rtl_opt_pass along with...
6025 (pass_data_ud_rtl_dce): ...new pass_data instance and...
6026 (make_pass_ud_rtl_dce): ...new function.
6027 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
6028 rtl_opt_pass along with...
6029 (pass_data_fast_rtl_dce): ...new pass_data instance and...
6030 (make_pass_fast_rtl_dce): ...new function.
6031 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
6032 a subclass of rtl_opt_pass along with...
6033 (pass_data_df_initialize_opt): ...new pass_data instance and...
6034 (make_pass_df_initialize_opt): ...new function.
6035 (pass_df_initialize_no_opt): Convert from a global struct to a
6036 subclass of rtl_opt_pass along with...
6037 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
6038 (make_pass_df_initialize_no_opt): ...new function.
6039 (pass_df_finish): Convert from a global struct to a subclass of
6040 rtl_opt_pass along with...
6041 (pass_data_df_finish): ...new pass_data instance and...
6042 (make_pass_df_finish): ...new function.
6043 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
6044 rtl_opt_pass along with...
6045 (pass_data_rtl_dse1): ...new pass_data instance and...
6046 (make_pass_rtl_dse1): ...new function.
6047 (pass_rtl_dse2): Convert from a global struct to a subclass of
6048 rtl_opt_pass along with...
6049 (pass_data_rtl_dse2): ...new pass_data instance and...
6050 (make_pass_rtl_dse2): ...new function.
6051 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
6052 subclass of rtl_opt_pass along with...
6053 (pass_data_dwarf2_frame): ...new pass_data instance and...
6054 (make_pass_dwarf2_frame): ...new function.
6055 * except.c (pass_set_nothrow_function_flags): Convert from a global
6056 struct to a subclass of rtl_opt_pass along with...
6057 (pass_data_set_nothrow_function_flags): ...new pass_data instance
6059 (make_pass_set_nothrow_function_flags): ...new function.
6060 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
6061 subclass of rtl_opt_pass along with...
6062 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
6064 (make_pass_convert_to_eh_region_ranges): ...new function.
6065 * final.c (pass_compute_alignments): Convert from a global struct to a
6066 subclass of rtl_opt_pass along with...
6067 (pass_data_compute_alignments): ...new pass_data instance and...
6068 (make_pass_compute_alignments): ...new function.
6069 (pass_final): Convert from a global struct to a subclass of
6070 rtl_opt_pass along with...
6071 (pass_data_final): ...new pass_data instance and...
6072 (make_pass_final): ...new function.
6073 (pass_shorten_branches): Convert from a global struct to a subclass of
6074 rtl_opt_pass along with...
6075 (pass_data_shorten_branches): ...new pass_data instance and...
6076 (make_pass_shorten_branches): ...new function.
6077 (pass_clean_state): Convert from a global struct to a subclass of
6078 rtl_opt_pass along with...
6079 (pass_data_clean_state): ...new pass_data instance and...
6080 (make_pass_clean_state): ...new function.
6081 * function.c (pass_instantiate_virtual_regs): Convert from a global
6082 struct to a subclass of rtl_opt_pass along with...
6083 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
6084 (make_pass_instantiate_virtual_regs): ...new function.
6085 (pass_leaf_regs): Convert from a global struct to a subclass of
6086 rtl_opt_pass along with...
6087 (pass_data_leaf_regs): ...new pass_data instance and...
6088 (make_pass_leaf_regs): ...new function.
6089 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
6090 subclass of rtl_opt_pass along with...
6091 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
6093 (make_pass_thread_prologue_and_epilogue): ...new function.
6094 (pass_match_asm_constraints): Convert from a global struct to a
6095 subclass of rtl_opt_pass along with...
6096 (pass_data_match_asm_constraints): ...new pass_data instance and...
6097 (make_pass_match_asm_constraints): ...new function.
6098 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
6099 subclass of rtl_opt_pass along with...
6100 (pass_data_rtl_fwprop): ...new pass_data instance and...
6101 (make_pass_rtl_fwprop): ...new function.
6102 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
6103 rtl_opt_pass along with...
6104 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
6105 (make_pass_rtl_fwprop_addr): ...new function.
6106 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
6107 rtl_opt_pass along with...
6108 (pass_data_rtl_pre): ...new pass_data instance and...
6109 (make_pass_rtl_pre): ...new function.
6110 (pass_rtl_hoist): Convert from a global struct to a subclass of
6111 rtl_opt_pass along with...
6112 (pass_data_rtl_hoist): ...new pass_data instance and...
6113 (make_pass_rtl_hoist): ...new function.
6114 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
6115 subclass of gimple_opt_pass along with...
6116 (pass_data_lower_cf): ...new pass_data instance and...
6117 (make_pass_lower_cf): ...new function.
6118 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
6119 from a global struct to a subclass of gimple_opt_pass along with...
6120 (pass_data_strength_reduction): ...new pass_data instance and...
6121 (make_pass_strength_reduction): ...new function.
6122 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
6123 of rtl_opt_pass along with...
6124 (pass_data_rtl_ifcvt): ...new pass_data instance and...
6125 (make_pass_rtl_ifcvt): ...new function.
6126 (pass_if_after_combine): Convert from a global struct to a subclass of
6127 rtl_opt_pass along with...
6128 (pass_data_if_after_combine): ...new pass_data instance and...
6129 (make_pass_if_after_combine): ...new function.
6130 (pass_if_after_reload): Convert from a global struct to a subclass of
6131 rtl_opt_pass along with...
6132 (pass_data_if_after_reload): ...new pass_data instance and...
6133 (make_pass_if_after_reload): ...new function.
6134 * init-regs.c (pass_initialize_regs): Convert from a global struct to
6135 a subclass of rtl_opt_pass along with...
6136 (pass_data_initialize_regs): ...new pass_data instance and...
6137 (make_pass_initialize_regs): ...new function.
6138 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
6139 of ipa_opt_pass_d along with...
6140 (pass_data_ipa_cp): ...new pass_data instance and...
6141 (make_pass_ipa_cp): ...new function.
6142 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
6143 global struct to a subclass of gimple_opt_pass along with...
6144 (pass_data_inline_parameters): ...new pass_data instance and...
6145 (make_pass_inline_parameters): ...new function.
6146 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
6147 subclass of gimple_opt_pass along with...
6148 (pass_data_early_inline): ...new pass_data instance and...
6149 (make_pass_early_inline): ...new function.
6150 (pass_ipa_inline): Convert from a global struct to a subclass of
6151 ipa_opt_pass_d along with...
6152 (pass_data_ipa_inline): ...new pass_data instance and...
6153 (make_pass_ipa_inline): ...new function.
6154 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
6155 struct to a subclass of gimple_opt_pass along with...
6156 (pass_data_local_pure_const): ...new pass_data instance and...
6157 (make_pass_local_pure_const): ...new function.
6158 (pass_ipa_pure_const): Convert from a global struct to a subclass of
6159 ipa_opt_pass_d along with...
6160 (pass_data_ipa_pure_const): ...new pass_data instance and...
6161 (make_pass_ipa_pure_const): ...new function.
6162 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
6163 to a subclass of ipa_opt_pass_d along with...
6164 (pass_data_ipa_reference): ...new pass_data instance and...
6165 (make_pass_ipa_reference): ...new function.
6166 * ipa-split.c (pass_split_functions): Convert from a global struct to
6167 a subclass of gimple_opt_pass along with...
6168 (pass_data_split_functions): ...new pass_data instance and...
6169 (make_pass_split_functions): ...new function.
6170 (pass_feedback_split_functions): Convert from a global struct to a
6171 subclass of gimple_opt_pass along with...
6172 (pass_data_feedback_split_functions): ...new pass_data instance and...
6173 (make_pass_feedback_split_functions): ...new function.
6174 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
6175 global struct to a subclass of simple_ipa_opt_pass along with...
6176 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
6178 (make_pass_ipa_function_and_variable_visibility): ...new function.
6179 (pass_ipa_free_inline_summary): Convert from a global struct to a
6180 subclass of simple_ipa_opt_pass along with...
6181 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
6182 (make_pass_ipa_free_inline_summary): ...new function.
6183 (pass_ipa_whole_program_visibility): Convert from a global struct to a
6184 subclass of ipa_opt_pass_d along with...
6185 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
6187 (make_pass_ipa_whole_program_visibility): ...new function.
6188 (pass_ipa_profile): Convert from a global struct to a subclass of
6189 ipa_opt_pass_d along with...
6190 (pass_data_ipa_profile): ...new pass_data instance and...
6191 (make_pass_ipa_profile): ...new function.
6192 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
6193 ipa_opt_pass_d along with...
6194 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
6195 (make_pass_ipa_cdtor_merge): ...new function.
6196 * ira.c (pass_ira): Convert from a global struct to a subclass of
6197 rtl_opt_pass along with...
6198 (pass_data_ira): ...new pass_data instance and...
6199 (make_pass_ira): ...new function.
6200 (pass_reload): Convert from a global struct to a subclass of
6201 rtl_opt_pass along with...
6202 (pass_data_reload): ...new pass_data instance and...
6203 (make_pass_reload): ...new function.
6204 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
6205 subclass of rtl_opt_pass along with...
6206 (pass_data_cleanup_barriers): ...new pass_data instance and...
6207 (make_pass_cleanup_barriers): ...new function.
6208 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
6209 of rtl_opt_pass along with...
6210 (pass_data_loop2): ...new pass_data instance and...
6211 (make_pass_loop2): ...new function.
6212 (pass_rtl_loop_init): Convert from a global struct to a subclass of
6213 rtl_opt_pass along with...
6214 (pass_data_rtl_loop_init): ...new pass_data instance and...
6215 (make_pass_rtl_loop_init): ...new function.
6216 (pass_rtl_loop_done): Convert from a global struct to a subclass of
6217 rtl_opt_pass along with...
6218 (pass_data_rtl_loop_done): ...new pass_data instance and...
6219 (make_pass_rtl_loop_done): ...new function.
6220 (pass_rtl_move_loop_invariants): Convert from a global struct to a
6221 subclass of rtl_opt_pass along with...
6222 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
6223 (make_pass_rtl_move_loop_invariants): ...new function.
6224 (pass_rtl_unswitch): Convert from a global struct to a subclass of
6225 rtl_opt_pass along with...
6226 (pass_data_rtl_unswitch): ...new pass_data instance and...
6227 (make_pass_rtl_unswitch): ...new function.
6228 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
6229 subclass of rtl_opt_pass along with...
6230 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
6232 (make_pass_rtl_unroll_and_peel_loops): ...new function.
6233 (pass_rtl_doloop): Convert from a global struct to a subclass of
6234 rtl_opt_pass along with...
6235 (pass_data_rtl_doloop): ...new pass_data instance and...
6236 (make_pass_rtl_doloop): ...new function.
6237 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
6238 a subclass of rtl_opt_pass along with...
6239 (pass_data_lower_subreg): ...new pass_data instance and...
6240 (make_pass_lower_subreg): ...new function.
6241 (pass_lower_subreg2): Convert from a global struct to a subclass of
6242 rtl_opt_pass along with...
6243 (pass_data_lower_subreg2): ...new pass_data instance and...
6244 (make_pass_lower_subreg2): ...new function.
6245 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
6246 struct to a subclass of ipa_opt_pass_d along with...
6247 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
6248 (make_pass_ipa_lto_gimple_out): ...new function.
6249 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
6250 of ipa_opt_pass_d along with...
6251 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
6252 (make_pass_ipa_lto_finish_out): ...new function.
6253 * mode-switching.c (pass_mode_switching): Convert from a global struct
6254 to a subclass of rtl_opt_pass along with...
6255 (pass_data_mode_switching): ...new pass_data instance and...
6256 (make_pass_mode_switching): ...new function.
6257 * modulo-sched.c (pass_sms): Convert from a global struct to a
6258 subclass of rtl_opt_pass along with...
6259 (pass_data_sms): ...new pass_data instance and...
6260 (make_pass_sms): ...new function.
6261 * omp-low.c (pass_expand_omp): Convert from a global struct to a
6262 subclass of gimple_opt_pass along with...
6263 (pass_data_expand_omp): ...new pass_data instance and...
6264 (make_pass_expand_omp): ...new function.
6265 (pass_lower_omp): Convert from a global struct to a subclass of
6266 gimple_opt_pass along with...
6267 (pass_data_lower_omp): ...new pass_data instance and...
6268 (make_pass_lower_omp): ...new function.
6269 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
6270 of gimple_opt_pass along with...
6271 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
6272 (make_pass_diagnose_omp_blocks): ...new function.
6273 * passes.c (pass_early_local_passes): Convert from a global struct to
6274 a subclass of simple_ipa_opt_pass along with...
6275 (pass_data_early_local_passes): ...new pass_data instance and...
6276 (make_pass_early_local_passes): ...new function.
6277 (pass_all_early_optimizations): Convert from a global struct to a
6278 subclass of gimple_opt_pass along with...
6279 (pass_data_all_early_optimizations): ...new pass_data instance and...
6280 (make_pass_all_early_optimizations): ...new function.
6281 (pass_all_optimizations): Convert from a global struct to a subclass
6282 of gimple_opt_pass along with...
6283 (pass_data_all_optimizations): ...new pass_data instance and...
6284 (make_pass_all_optimizations): ...new function.
6285 (pass_all_optimizations_g): Convert from a global struct to a subclass
6286 of gimple_opt_pass along with...
6287 (pass_data_all_optimizations_g): ...new pass_data instance and...
6288 (make_pass_all_optimizations_g): ...new function.
6289 (pass_rest_of_compilation): Convert from a global struct to a subclass
6290 of rtl_opt_pass along with...
6291 (pass_data_rest_of_compilation): ...new pass_data instance and...
6292 (make_pass_rest_of_compilation): ...new function.
6293 (pass_postreload): Convert from a global struct to a subclass of
6294 rtl_opt_pass along with...
6295 (pass_data_postreload): ...new pass_data instance and...
6296 (make_pass_postreload): ...new function.
6297 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
6298 subclass of rtl_opt_pass along with...
6299 (pass_data_gcse2): ...new pass_data instance and...
6300 (make_pass_gcse2): ...new function.
6301 * postreload.c (pass_postreload_cse): Convert from a global struct to
6302 a subclass of rtl_opt_pass along with...
6303 (pass_data_postreload_cse): ...new pass_data instance and...
6304 (make_pass_postreload_cse): ...new function.
6305 * predict.c (pass_profile): Convert from a global struct to a subclass
6306 of gimple_opt_pass along with...
6307 (pass_data_profile): ...new pass_data instance and...
6308 (make_pass_profile): ...new function.
6309 (pass_strip_predict_hints): Convert from a global struct to a subclass
6310 of gimple_opt_pass along with...
6311 (pass_data_strip_predict_hints): ...new pass_data instance and...
6312 (make_pass_strip_predict_hints): ...new function.
6313 * recog.c (pass_peephole2): Convert from a global struct to a subclass
6314 of rtl_opt_pass along with...
6315 (pass_data_peephole2): ...new pass_data instance and...
6316 (make_pass_peephole2): ...new function.
6317 (pass_split_all_insns): Convert from a global struct to a subclass of
6318 rtl_opt_pass along with...
6319 (pass_data_split_all_insns): ...new pass_data instance and...
6320 (make_pass_split_all_insns): ...new function.
6321 (pass_split_after_reload): Convert from a global struct to a subclass
6322 of rtl_opt_pass along with...
6323 (pass_data_split_after_reload): ...new pass_data instance and...
6324 (make_pass_split_after_reload): ...new function.
6325 (pass_split_before_regstack): Convert from a global struct to a
6326 subclass of rtl_opt_pass along with...
6327 (pass_data_split_before_regstack): ...new pass_data instance and...
6328 (make_pass_split_before_regstack): ...new function.
6329 (pass_split_before_sched2): Convert from a global struct to a subclass
6330 of rtl_opt_pass along with...
6331 (pass_data_split_before_sched2): ...new pass_data instance and...
6332 (make_pass_split_before_sched2): ...new function.
6333 (pass_split_for_shorten_branches): Convert from a global struct to a
6334 subclass of rtl_opt_pass along with...
6335 (pass_data_split_for_shorten_branches): ...new pass_data instance
6337 (make_pass_split_for_shorten_branches): ...new function.
6338 * ree.c (pass_ree): Convert from a global struct to a subclass of
6339 rtl_opt_pass along with...
6340 (pass_data_ree): ...new pass_data instance and...
6341 (make_pass_ree): ...new function.
6342 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
6343 subclass of rtl_opt_pass along with...
6344 (pass_data_stack_regs): ...new pass_data instance and...
6345 (make_pass_stack_regs): ...new function.
6346 (pass_stack_regs_run): Convert from a global struct to a subclass of
6347 rtl_opt_pass along with...
6348 (pass_data_stack_regs_run): ...new pass_data instance and...
6349 (make_pass_stack_regs_run): ...new function.
6350 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
6351 subclass of rtl_opt_pass along with...
6352 (pass_data_cprop_hardreg): ...new pass_data instance and...
6353 (make_pass_cprop_hardreg): ...new function.
6354 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
6355 subclass of rtl_opt_pass along with...
6356 (pass_data_reginfo_init): ...new pass_data instance and...
6357 (make_pass_reginfo_init): ...new function.
6358 * regmove.c (pass_regmove): Convert from a global struct to a subclass
6359 of rtl_opt_pass along with...
6360 (pass_data_regmove): ...new pass_data instance and...
6361 (make_pass_regmove): ...new function.
6362 * regrename.c (pass_regrename): Convert from a global struct to a
6363 subclass of rtl_opt_pass along with...
6364 (pass_data_regrename): ...new pass_data instance and...
6365 (make_pass_regrename): ...new function.
6366 * reorg.c (pass_delay_slots): Convert from a global struct to a
6367 subclass of rtl_opt_pass along with...
6368 (pass_data_delay_slots): ...new pass_data instance and...
6369 (make_pass_delay_slots): ...new function.
6370 (pass_machine_reorg): Convert from a global struct to a subclass of
6371 rtl_opt_pass along with...
6372 (pass_data_machine_reorg): ...new pass_data instance and...
6373 (make_pass_machine_reorg): ...new function.
6374 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
6375 of rtl_opt_pass along with...
6376 (pass_data_sched): ...new pass_data instance and...
6377 (make_pass_sched): ...new function.
6378 (pass_sched2): Convert from a global struct to a subclass of
6379 rtl_opt_pass along with...
6380 (pass_data_sched2): ...new pass_data instance and...
6381 (make_pass_sched2): ...new function.
6382 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
6383 to a subclass of rtl_opt_pass along with...
6384 (pass_data_stack_ptr_mod): ...new pass_data instance and...
6385 (make_pass_stack_ptr_mod): ...new function.
6386 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
6387 to a subclass of rtl_opt_pass along with...
6388 (pass_data_rtl_store_motion): ...new pass_data instance and...
6389 (make_pass_rtl_store_motion): ...new function.
6390 * tracer.c (pass_tracer): Convert from a global struct to a subclass
6391 of gimple_opt_pass along with...
6392 (pass_data_tracer): ...new pass_data instance and...
6393 (make_pass_tracer): ...new function.
6394 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
6395 to a subclass of gimple_opt_pass along with...
6396 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
6397 (make_pass_diagnose_tm_blocks): ...new function.
6398 (pass_lower_tm): Convert from a global struct to a subclass of
6399 gimple_opt_pass along with...
6400 (pass_data_lower_tm): ...new pass_data instance and...
6401 (make_pass_lower_tm): ...new function.
6402 (pass_tm_init): Convert from a global struct to a subclass of
6403 gimple_opt_pass along with...
6404 (pass_data_tm_init): ...new pass_data instance and...
6405 (make_pass_tm_init): ...new function.
6406 (pass_tm_mark): Convert from a global struct to a subclass of
6407 gimple_opt_pass along with...
6408 (pass_data_tm_mark): ...new pass_data instance and...
6409 (make_pass_tm_mark): ...new function.
6410 (pass_tm_edges): Convert from a global struct to a subclass of
6411 gimple_opt_pass along with...
6412 (pass_data_tm_edges): ...new pass_data instance and...
6413 (make_pass_tm_edges): ...new function.
6414 (pass_tm_memopt): Convert from a global struct to a subclass of
6415 gimple_opt_pass along with...
6416 (pass_data_tm_memopt): ...new pass_data instance and...
6417 (make_pass_tm_memopt): ...new function.
6418 (pass_ipa_tm): Convert from a global struct to a subclass of
6419 simple_ipa_opt_pass along with...
6420 (pass_data_ipa_tm): ...new pass_data instance and...
6421 (make_pass_ipa_tm): ...new function.
6422 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
6423 subclass of gimple_opt_pass along with...
6424 (pass_data_call_cdce): ...new pass_data instance and...
6425 (make_pass_call_cdce): ...new function.
6426 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
6427 subclass of gimple_opt_pass along with...
6428 (pass_data_build_cfg): ...new pass_data instance and...
6429 (make_pass_build_cfg): ...new function.
6430 (pass_split_crit_edges): Convert from a global struct to a subclass of
6431 gimple_opt_pass along with...
6432 (pass_data_split_crit_edges): ...new pass_data instance and...
6433 (make_pass_split_crit_edges): ...new function.
6434 (pass_warn_function_return): Convert from a global struct to a
6435 subclass of gimple_opt_pass along with...
6436 (pass_data_warn_function_return): ...new pass_data instance and...
6437 (make_pass_warn_function_return): ...new function.
6438 (pass_warn_function_noreturn): Convert from a global struct to a
6439 subclass of gimple_opt_pass along with...
6440 (pass_data_warn_function_noreturn): ...new pass_data instance and...
6441 (make_pass_warn_function_noreturn): ...new function.
6442 (pass_warn_unused_result): Convert from a global struct to a subclass
6443 of gimple_opt_pass along with...
6444 (pass_data_warn_unused_result): ...new pass_data instance and...
6445 (make_pass_warn_unused_result): ...new function.
6446 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
6447 a subclass of gimple_opt_pass along with...
6448 (pass_data_merge_phi): ...new pass_data instance and...
6449 (make_pass_merge_phi): ...new function.
6450 * tree-complex.c (pass_lower_complex): Convert from a global struct to
6451 a subclass of gimple_opt_pass along with...
6452 (pass_data_lower_complex): ...new pass_data instance and...
6453 (make_pass_lower_complex): ...new function.
6454 (pass_lower_complex_O0): Convert from a global struct to a subclass of
6455 gimple_opt_pass along with...
6456 (pass_data_lower_complex_O0): ...new pass_data instance and...
6457 (make_pass_lower_complex_O0): ...new function.
6458 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
6459 subclass of gimple_opt_pass along with...
6460 (pass_data_lower_eh): ...new pass_data instance and...
6461 (make_pass_lower_eh): ...new function.
6462 (pass_refactor_eh): Convert from a global struct to a subclass of
6463 gimple_opt_pass along with...
6464 (pass_data_refactor_eh): ...new pass_data instance and...
6465 (make_pass_refactor_eh): ...new function.
6466 (pass_lower_resx): Convert from a global struct to a subclass of
6467 gimple_opt_pass along with...
6468 (pass_data_lower_resx): ...new pass_data instance and...
6469 (make_pass_lower_resx): ...new function.
6470 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
6471 of gimple_opt_pass along with...
6472 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
6473 (make_pass_lower_eh_dispatch): ...new function.
6474 (pass_cleanup_eh): Convert from a global struct to a subclass of
6475 gimple_opt_pass along with...
6476 (pass_data_cleanup_eh): ...new pass_data instance and...
6477 (make_pass_cleanup_eh): ...new function.
6478 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
6479 to a subclass of simple_ipa_opt_pass along with...
6480 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
6481 (make_pass_ipa_lower_emutls): ...new function.
6482 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
6483 a subclass of gimple_opt_pass along with...
6484 (pass_data_if_conversion): ...new pass_data instance and...
6485 (make_pass_if_conversion): ...new function.
6486 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
6487 subclass of gimple_opt_pass along with...
6488 (pass_data_build_ssa): ...new pass_data instance and...
6489 (make_pass_build_ssa): ...new function.
6490 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
6491 global struct to a subclass of gimple_opt_pass along with...
6492 (pass_data_loop_distribution): ...new pass_data instance and...
6493 (make_pass_loop_distribution): ...new function.
6494 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
6495 subclass of gimple_opt_pass along with...
6496 (pass_data_mudflap_1): ...new pass_data instance and...
6497 (make_pass_mudflap_1): ...new function.
6498 (pass_mudflap_2): Convert from a global struct to a subclass of
6499 gimple_opt_pass along with...
6500 (pass_data_mudflap_2): ...new pass_data instance and...
6501 (make_pass_mudflap_2): ...new function.
6502 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
6503 subclass of gimple_opt_pass along with...
6504 (pass_data_mudflap_1): ...new pass_data instance and...
6505 (make_pass_mudflap_1): ...new function.
6506 (pass_mudflap_2): Convert from a global struct to a subclass of
6507 gimple_opt_pass along with...
6508 (pass_data_mudflap_2): ...new pass_data instance and...
6509 (make_pass_mudflap_2): ...new function.
6510 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
6511 gimple_opt_pass along with...
6512 (pass_data_nrv): ...new pass_data instance and...
6513 (make_pass_nrv): ...new function.
6514 (pass_return_slot): Convert from a global struct to a subclass of
6515 gimple_opt_pass along with...
6516 (pass_data_return_slot): ...new pass_data instance and...
6517 (make_pass_return_slot): ...new function.
6518 * tree-object-size.c (pass_object_sizes): Convert from a global struct
6519 to a subclass of gimple_opt_pass along with...
6520 (pass_data_object_sizes): ...new pass_data instance and...
6521 (make_pass_object_sizes): ...new function.
6522 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
6523 global struct to a subclass of gimple_opt_pass along with...
6524 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
6526 (make_pass_cleanup_cfg_post_optimizing): ...new function.
6527 (pass_fixup_cfg): Convert from a global struct to a subclass of
6528 gimple_opt_pass along with...
6529 (pass_data_fixup_cfg): ...new pass_data instance and...
6530 (make_pass_fixup_cfg): ...new function.
6531 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
6532 (make_pass_mudflap_1): ...new function.
6533 (pass_mudflap_2): Replace declaration with that of...
6534 (make_pass_mudflap_2): ...new function.
6535 (pass_asan): Replace declaration with that of...
6536 (make_pass_asan): ...new function.
6537 (pass_asan_O0): Replace declaration with that of...
6538 (make_pass_asan_O0): ...new function.
6539 (pass_tsan): Replace declaration with that of...
6540 (make_pass_tsan): ...new function.
6541 (pass_tsan_O0): Replace declaration with that of...
6542 (make_pass_tsan_O0): ...new function.
6543 (pass_lower_cf): Replace declaration with that of...
6544 (make_pass_lower_cf): ...new function.
6545 (pass_refactor_eh): Replace declaration with that of...
6546 (make_pass_refactor_eh): ...new function.
6547 (pass_lower_eh): Replace declaration with that of...
6548 (make_pass_lower_eh): ...new function.
6549 (pass_lower_eh_dispatch): Replace declaration with that of...
6550 (make_pass_lower_eh_dispatch): ...new function.
6551 (pass_lower_resx): Replace declaration with that of...
6552 (make_pass_lower_resx): ...new function.
6553 (pass_build_cfg): Replace declaration with that of...
6554 (make_pass_build_cfg): ...new function.
6555 (pass_early_tree_profile): Replace declaration with that of...
6556 (make_pass_early_tree_profile): ...new function.
6557 (pass_cleanup_eh): Replace declaration with that of...
6558 (make_pass_cleanup_eh): ...new function.
6559 (pass_sra): Replace declaration with that of...
6560 (make_pass_sra): ...new function.
6561 (pass_sra_early): Replace declaration with that of...
6562 (make_pass_sra_early): ...new function.
6563 (pass_early_ipa_sra): Replace declaration with that of...
6564 (make_pass_early_ipa_sra): ...new function.
6565 (pass_tail_recursion): Replace declaration with that of...
6566 (make_pass_tail_recursion): ...new function.
6567 (pass_tail_calls): Replace declaration with that of...
6568 (make_pass_tail_calls): ...new function.
6569 (pass_tree_loop): Replace declaration with that of...
6570 (make_pass_tree_loop): ...new function.
6571 (pass_tree_loop_init): Replace declaration with that of...
6572 (make_pass_tree_loop_init): ...new function.
6573 (pass_lim): Replace declaration with that of...
6574 (make_pass_lim): ...new function.
6575 (pass_tree_unswitch): Replace declaration with that of...
6576 (make_pass_tree_unswitch): ...new function.
6577 (pass_predcom): Replace declaration with that of...
6578 (make_pass_predcom): ...new function.
6579 (pass_iv_canon): Replace declaration with that of...
6580 (make_pass_iv_canon): ...new function.
6581 (pass_scev_cprop): Replace declaration with that of...
6582 (make_pass_scev_cprop): ...new function.
6583 (pass_empty_loop): Replace declaration with that of...
6584 (make_pass_empty_loop): ...new function.
6585 (pass_record_bounds): Replace declaration with that of...
6586 (make_pass_record_bounds): ...new function.
6587 (pass_graphite): Replace declaration with that of...
6588 (make_pass_graphite): ...new function.
6589 (pass_graphite_transforms): Replace declaration with that of...
6590 (make_pass_graphite_transforms): ...new function.
6591 (pass_if_conversion): Replace declaration with that of...
6592 (make_pass_if_conversion): ...new function.
6593 (pass_loop_distribution): Replace declaration with that of...
6594 (make_pass_loop_distribution): ...new function.
6595 (pass_vectorize): Replace declaration with that of...
6596 (make_pass_vectorize): ...new function.
6597 (pass_slp_vectorize): Replace declaration with that of...
6598 (make_pass_slp_vectorize): ...new function.
6599 (pass_complete_unroll): Replace declaration with that of...
6600 (make_pass_complete_unroll): ...new function.
6601 (pass_complete_unrolli): Replace declaration with that of...
6602 (make_pass_complete_unrolli): ...new function.
6603 (pass_parallelize_loops): Replace declaration with that of...
6604 (make_pass_parallelize_loops): ...new function.
6605 (pass_loop_prefetch): Replace declaration with that of...
6606 (make_pass_loop_prefetch): ...new function.
6607 (pass_iv_optimize): Replace declaration with that of...
6608 (make_pass_iv_optimize): ...new function.
6609 (pass_tree_loop_done): Replace declaration with that of...
6610 (make_pass_tree_loop_done): ...new function.
6611 (pass_ch): Replace declaration with that of...
6612 (make_pass_ch): ...new function.
6613 (pass_ccp): Replace declaration with that of...
6614 (make_pass_ccp): ...new function.
6615 (pass_phi_only_cprop): Replace declaration with that of...
6616 (make_pass_phi_only_cprop): ...new function.
6617 (pass_build_ssa): Replace declaration with that of...
6618 (make_pass_build_ssa): ...new function.
6619 (pass_build_alias): Replace declaration with that of...
6620 (make_pass_build_alias): ...new function.
6621 (pass_build_ealias): Replace declaration with that of...
6622 (make_pass_build_ealias): ...new function.
6623 (pass_dominator): Replace declaration with that of...
6624 (make_pass_dominator): ...new function.
6625 (pass_dce): Replace declaration with that of...
6626 (make_pass_dce): ...new function.
6627 (pass_dce_loop): Replace declaration with that of...
6628 (make_pass_dce_loop): ...new function.
6629 (pass_cd_dce): Replace declaration with that of...
6630 (make_pass_cd_dce): ...new function.
6631 (pass_call_cdce): Replace declaration with that of...
6632 (make_pass_call_cdce): ...new function.
6633 (pass_merge_phi): Replace declaration with that of...
6634 (make_pass_merge_phi): ...new function.
6635 (pass_split_crit_edges): Replace declaration with that of...
6636 (make_pass_split_crit_edges): ...new function.
6637 (pass_pre): Replace declaration with that of...
6638 (make_pass_pre): ...new function.
6639 (pass_profile): Replace declaration with that of...
6640 (make_pass_profile): ...new function.
6641 (pass_strip_predict_hints): Replace declaration with that of...
6642 (make_pass_strip_predict_hints): ...new function.
6643 (pass_lower_complex_O0): Replace declaration with that of...
6644 (make_pass_lower_complex_O0): ...new function.
6645 (pass_lower_complex): Replace declaration with that of...
6646 (make_pass_lower_complex): ...new function.
6647 (pass_lower_vector): Replace declaration with that of...
6648 (make_pass_lower_vector): ...new function.
6649 (pass_lower_vector_ssa): Replace declaration with that of...
6650 (make_pass_lower_vector_ssa): ...new function.
6651 (pass_lower_omp): Replace declaration with that of...
6652 (make_pass_lower_omp): ...new function.
6653 (pass_diagnose_omp_blocks): Replace declaration with that of...
6654 (make_pass_diagnose_omp_blocks): ...new function.
6655 (pass_expand_omp): Replace declaration with that of...
6656 (make_pass_expand_omp): ...new function.
6657 (pass_expand_omp_ssa): Replace declaration with that of...
6658 (make_pass_expand_omp_ssa): ...new function.
6659 (pass_object_sizes): Replace declaration with that of...
6660 (make_pass_object_sizes): ...new function.
6661 (pass_strlen): Replace declaration with that of...
6662 (make_pass_strlen): ...new function.
6663 (pass_fold_builtins): Replace declaration with that of...
6664 (make_pass_fold_builtins): ...new function.
6665 (pass_stdarg): Replace declaration with that of...
6666 (make_pass_stdarg): ...new function.
6667 (pass_early_warn_uninitialized): Replace declaration with that of...
6668 (make_pass_early_warn_uninitialized): ...new function.
6669 (pass_late_warn_uninitialized): Replace declaration with that of...
6670 (make_pass_late_warn_uninitialized): ...new function.
6671 (pass_cse_reciprocals): Replace declaration with that of...
6672 (make_pass_cse_reciprocals): ...new function.
6673 (pass_cse_sincos): Replace declaration with that of...
6674 (make_pass_cse_sincos): ...new function.
6675 (pass_optimize_bswap): Replace declaration with that of...
6676 (make_pass_optimize_bswap): ...new function.
6677 (pass_optimize_widening_mul): Replace declaration with that of...
6678 (make_pass_optimize_widening_mul): ...new function.
6679 (pass_warn_function_return): Replace declaration with that of...
6680 (make_pass_warn_function_return): ...new function.
6681 (pass_warn_function_noreturn): Replace declaration with that of...
6682 (make_pass_warn_function_noreturn): ...new function.
6683 (pass_cselim): Replace declaration with that of...
6684 (make_pass_cselim): ...new function.
6685 (pass_phiopt): Replace declaration with that of...
6686 (make_pass_phiopt): ...new function.
6687 (pass_forwprop): Replace declaration with that of...
6688 (make_pass_forwprop): ...new function.
6689 (pass_phiprop): Replace declaration with that of...
6690 (make_pass_phiprop): ...new function.
6691 (pass_tree_ifcombine): Replace declaration with that of...
6692 (make_pass_tree_ifcombine): ...new function.
6693 (pass_dse): Replace declaration with that of...
6694 (make_pass_dse): ...new function.
6695 (pass_nrv): Replace declaration with that of...
6696 (make_pass_nrv): ...new function.
6697 (pass_rename_ssa_copies): Replace declaration with that of...
6698 (make_pass_rename_ssa_copies): ...new function.
6699 (pass_sink_code): Replace declaration with that of...
6700 (make_pass_sink_code): ...new function.
6701 (pass_fre): Replace declaration with that of...
6702 (make_pass_fre): ...new function.
6703 (pass_check_data_deps): Replace declaration with that of...
6704 (make_pass_check_data_deps): ...new function.
6705 (pass_copy_prop): Replace declaration with that of...
6706 (make_pass_copy_prop): ...new function.
6707 (pass_vrp): Replace declaration with that of...
6708 (make_pass_vrp): ...new function.
6709 (pass_uncprop): Replace declaration with that of...
6710 (make_pass_uncprop): ...new function.
6711 (pass_return_slot): Replace declaration with that of...
6712 (make_pass_return_slot): ...new function.
6713 (pass_reassoc): Replace declaration with that of...
6714 (make_pass_reassoc): ...new function.
6715 (pass_rebuild_cgraph_edges): Replace declaration with that of...
6716 (make_pass_rebuild_cgraph_edges): ...new function.
6717 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
6718 (make_pass_remove_cgraph_callee_edges): ...new function.
6719 (pass_build_cgraph_edges): Replace declaration with that of...
6720 (make_pass_build_cgraph_edges): ...new function.
6721 (pass_local_pure_const): Replace declaration with that of...
6722 (make_pass_local_pure_const): ...new function.
6723 (pass_tracer): Replace declaration with that of...
6724 (make_pass_tracer): ...new function.
6725 (pass_warn_unused_result): Replace declaration with that of...
6726 (make_pass_warn_unused_result): ...new function.
6727 (pass_diagnose_tm_blocks): Replace declaration with that of...
6728 (make_pass_diagnose_tm_blocks): ...new function.
6729 (pass_lower_tm): Replace declaration with that of...
6730 (make_pass_lower_tm): ...new function.
6731 (pass_tm_init): Replace declaration with that of...
6732 (make_pass_tm_init): ...new function.
6733 (pass_tm_mark): Replace declaration with that of...
6734 (make_pass_tm_mark): ...new function.
6735 (pass_tm_memopt): Replace declaration with that of...
6736 (make_pass_tm_memopt): ...new function.
6737 (pass_tm_edges): Replace declaration with that of...
6738 (make_pass_tm_edges): ...new function.
6739 (pass_split_functions): Replace declaration with that of...
6740 (make_pass_split_functions): ...new function.
6741 (pass_feedback_split_functions): Replace declaration with that of...
6742 (make_pass_feedback_split_functions): ...new function.
6743 (pass_strength_reduction): Replace declaration with that of...
6744 (make_pass_strength_reduction): ...new function.
6745 (pass_ipa_lower_emutls): Replace declaration with that of...
6746 (make_pass_ipa_lower_emutls): ...new function.
6747 (pass_ipa_function_and_variable_visibility): Replace declaration with
6749 (make_pass_ipa_function_and_variable_visibility): ...new function.
6750 (pass_ipa_tree_profile): Replace declaration with that of...
6751 (make_pass_ipa_tree_profile): ...new function.
6752 (pass_early_local_passes): Replace declaration with that of...
6753 (make_pass_early_local_passes): ...new function.
6754 (pass_ipa_whole_program_visibility): Replace declaration with that
6756 (make_pass_ipa_whole_program_visibility): ...new function.
6757 (pass_ipa_lto_gimple_out): Replace declaration with that of...
6758 (make_pass_ipa_lto_gimple_out): ...new function.
6759 (pass_ipa_increase_alignment): Replace declaration with that of...
6760 (make_pass_ipa_increase_alignment): ...new function.
6761 (pass_ipa_inline): Replace declaration with that of...
6762 (make_pass_ipa_inline): ...new function.
6763 (pass_ipa_free_lang_data): Replace declaration with that of...
6764 (make_pass_ipa_free_lang_data): ...new function.
6765 (pass_ipa_free_inline_summary): Replace declaration with that of...
6766 (make_pass_ipa_free_inline_summary): ...new function.
6767 (pass_ipa_cp): Replace declaration with that of...
6768 (make_pass_ipa_cp): ...new function.
6769 (pass_ipa_reference): Replace declaration with that of...
6770 (make_pass_ipa_reference): ...new function.
6771 (pass_ipa_pure_const): Replace declaration with that of...
6772 (make_pass_ipa_pure_const): ...new function.
6773 (pass_ipa_pta): Replace declaration with that of...
6774 (make_pass_ipa_pta): ...new function.
6775 (pass_ipa_lto_finish_out): Replace declaration with that of...
6776 (make_pass_ipa_lto_finish_out): ...new function.
6777 (pass_ipa_tm): Replace declaration with that of...
6778 (make_pass_ipa_tm): ...new function.
6779 (pass_ipa_profile): Replace declaration with that of...
6780 (make_pass_ipa_profile): ...new function.
6781 (pass_ipa_cdtor_merge): Replace declaration with that of...
6782 (make_pass_ipa_cdtor_merge): ...new function.
6783 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
6785 (make_pass_cleanup_cfg_post_optimizing): ...new function.
6786 (pass_init_datastructures): Replace declaration with that of...
6787 (make_pass_init_datastructures): ...new function.
6788 (pass_fixup_cfg): Replace declaration with that of...
6789 (make_pass_fixup_cfg): ...new function.
6790 (pass_expand): Replace declaration with that of...
6791 (make_pass_expand): ...new function.
6792 (pass_instantiate_virtual_regs): Replace declaration with that of...
6793 (make_pass_instantiate_virtual_regs): ...new function.
6794 (pass_rtl_fwprop): Replace declaration with that of...
6795 (make_pass_rtl_fwprop): ...new function.
6796 (pass_rtl_fwprop_addr): Replace declaration with that of...
6797 (make_pass_rtl_fwprop_addr): ...new function.
6798 (pass_jump): Replace declaration with that of...
6799 (make_pass_jump): ...new function.
6800 (pass_jump2): Replace declaration with that of...
6801 (make_pass_jump2): ...new function.
6802 (pass_lower_subreg): Replace declaration with that of...
6803 (make_pass_lower_subreg): ...new function.
6804 (pass_cse): Replace declaration with that of...
6805 (make_pass_cse): ...new function.
6806 (pass_fast_rtl_dce): Replace declaration with that of...
6807 (make_pass_fast_rtl_dce): ...new function.
6808 (pass_ud_rtl_dce): Replace declaration with that of...
6809 (make_pass_ud_rtl_dce): ...new function.
6810 (pass_rtl_dce): Replace declaration with that of...
6811 (make_pass_rtl_dce): ...new function.
6812 (pass_rtl_dse1): Replace declaration with that of...
6813 (make_pass_rtl_dse1): ...new function.
6814 (pass_rtl_dse2): Replace declaration with that of...
6815 (make_pass_rtl_dse2): ...new function.
6816 (pass_rtl_dse3): Replace declaration with that of...
6817 (make_pass_rtl_dse3): ...new function.
6818 (pass_rtl_cprop): Replace declaration with that of...
6819 (make_pass_rtl_cprop): ...new function.
6820 (pass_rtl_pre): Replace declaration with that of...
6821 (make_pass_rtl_pre): ...new function.
6822 (pass_rtl_hoist): Replace declaration with that of...
6823 (make_pass_rtl_hoist): ...new function.
6824 (pass_rtl_store_motion): Replace declaration with that of...
6825 (make_pass_rtl_store_motion): ...new function.
6826 (pass_cse_after_global_opts): Replace declaration with that of...
6827 (make_pass_cse_after_global_opts): ...new function.
6828 (pass_rtl_ifcvt): Replace declaration with that of...
6829 (make_pass_rtl_ifcvt): ...new function.
6830 (pass_into_cfg_layout_mode): Replace declaration with that of...
6831 (make_pass_into_cfg_layout_mode): ...new function.
6832 (pass_outof_cfg_layout_mode): Replace declaration with that of...
6833 (make_pass_outof_cfg_layout_mode): ...new function.
6834 (pass_loop2): Replace declaration with that of...
6835 (make_pass_loop2): ...new function.
6836 (pass_rtl_loop_init): Replace declaration with that of...
6837 (make_pass_rtl_loop_init): ...new function.
6838 (pass_rtl_move_loop_invariants): Replace declaration with that of...
6839 (make_pass_rtl_move_loop_invariants): ...new function.
6840 (pass_rtl_unswitch): Replace declaration with that of...
6841 (make_pass_rtl_unswitch): ...new function.
6842 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
6843 (make_pass_rtl_unroll_and_peel_loops): ...new function.
6844 (pass_rtl_doloop): Replace declaration with that of...
6845 (make_pass_rtl_doloop): ...new function.
6846 (pass_rtl_loop_done): Replace declaration with that of...
6847 (make_pass_rtl_loop_done): ...new function.
6848 (pass_web): Replace declaration with that of...
6849 (make_pass_web): ...new function.
6850 (pass_cse2): Replace declaration with that of...
6851 (make_pass_cse2): ...new function.
6852 (pass_df_initialize_opt): Replace declaration with that of...
6853 (make_pass_df_initialize_opt): ...new function.
6854 (pass_df_initialize_no_opt): Replace declaration with that of...
6855 (make_pass_df_initialize_no_opt): ...new function.
6856 (pass_reginfo_init): Replace declaration with that of...
6857 (make_pass_reginfo_init): ...new function.
6858 (pass_inc_dec): Replace declaration with that of...
6859 (make_pass_inc_dec): ...new function.
6860 (pass_stack_ptr_mod): Replace declaration with that of...
6861 (make_pass_stack_ptr_mod): ...new function.
6862 (pass_initialize_regs): Replace declaration with that of...
6863 (make_pass_initialize_regs): ...new function.
6864 (pass_combine): Replace declaration with that of...
6865 (make_pass_combine): ...new function.
6866 (pass_if_after_combine): Replace declaration with that of...
6867 (make_pass_if_after_combine): ...new function.
6868 (pass_ree): Replace declaration with that of...
6869 (make_pass_ree): ...new function.
6870 (pass_partition_blocks): Replace declaration with that of...
6871 (make_pass_partition_blocks): ...new function.
6872 (pass_match_asm_constraints): Replace declaration with that of...
6873 (make_pass_match_asm_constraints): ...new function.
6874 (pass_regmove): Replace declaration with that of...
6875 (make_pass_regmove): ...new function.
6876 (pass_split_all_insns): Replace declaration with that of...
6877 (make_pass_split_all_insns): ...new function.
6878 (pass_fast_rtl_byte_dce): Replace declaration with that of...
6879 (make_pass_fast_rtl_byte_dce): ...new function.
6880 (pass_lower_subreg2): Replace declaration with that of...
6881 (make_pass_lower_subreg2): ...new function.
6882 (pass_mode_switching): Replace declaration with that of...
6883 (make_pass_mode_switching): ...new function.
6884 (pass_sms): Replace declaration with that of...
6885 (make_pass_sms): ...new function.
6886 (pass_sched): Replace declaration with that of...
6887 (make_pass_sched): ...new function.
6888 (pass_ira): Replace declaration with that of...
6889 (make_pass_ira): ...new function.
6890 (pass_reload): Replace declaration with that of...
6891 (make_pass_reload): ...new function.
6892 (pass_clean_state): Replace declaration with that of...
6893 (make_pass_clean_state): ...new function.
6894 (pass_branch_prob): Replace declaration with that of...
6895 (make_pass_branch_prob): ...new function.
6896 (pass_value_profile_transformations): Replace declaration with that
6898 (make_pass_value_profile_transformations): ...new function.
6899 (pass_postreload_cse): Replace declaration with that of...
6900 (make_pass_postreload_cse): ...new function.
6901 (pass_gcse2): Replace declaration with that of...
6902 (make_pass_gcse2): ...new function.
6903 (pass_split_after_reload): Replace declaration with that of...
6904 (make_pass_split_after_reload): ...new function.
6905 (pass_branch_target_load_optimize1): Replace declaration with that
6907 (make_pass_branch_target_load_optimize1): ...new function.
6908 (pass_thread_prologue_and_epilogue): Replace declaration with that
6910 (make_pass_thread_prologue_and_epilogue): ...new function.
6911 (pass_stack_adjustments): Replace declaration with that of...
6912 (make_pass_stack_adjustments): ...new function.
6913 (pass_peephole2): Replace declaration with that of...
6914 (make_pass_peephole2): ...new function.
6915 (pass_if_after_reload): Replace declaration with that of...
6916 (make_pass_if_after_reload): ...new function.
6917 (pass_regrename): Replace declaration with that of...
6918 (make_pass_regrename): ...new function.
6919 (pass_cprop_hardreg): Replace declaration with that of...
6920 (make_pass_cprop_hardreg): ...new function.
6921 (pass_reorder_blocks): Replace declaration with that of...
6922 (make_pass_reorder_blocks): ...new function.
6923 (pass_branch_target_load_optimize2): Replace declaration with that
6925 (make_pass_branch_target_load_optimize2): ...new function.
6926 (pass_leaf_regs): Replace declaration with that of...
6927 (make_pass_leaf_regs): ...new function.
6928 (pass_split_before_sched2): Replace declaration with that of...
6929 (make_pass_split_before_sched2): ...new function.
6930 (pass_compare_elim_after_reload): Replace declaration with that of...
6931 (make_pass_compare_elim_after_reload): ...new function.
6932 (pass_sched2): Replace declaration with that of...
6933 (make_pass_sched2): ...new function.
6934 (pass_stack_regs): Replace declaration with that of...
6935 (make_pass_stack_regs): ...new function.
6936 (pass_stack_regs_run): Replace declaration with that of...
6937 (make_pass_stack_regs_run): ...new function.
6938 (pass_df_finish): Replace declaration with that of...
6939 (make_pass_df_finish): ...new function.
6940 (pass_compute_alignments): Replace declaration with that of...
6941 (make_pass_compute_alignments): ...new function.
6942 (pass_duplicate_computed_gotos): Replace declaration with that of...
6943 (make_pass_duplicate_computed_gotos): ...new function.
6944 (pass_variable_tracking): Replace declaration with that of...
6945 (make_pass_variable_tracking): ...new function.
6946 (pass_free_cfg): Replace declaration with that of...
6947 (make_pass_free_cfg): ...new function.
6948 (pass_machine_reorg): Replace declaration with that of...
6949 (make_pass_machine_reorg): ...new function.
6950 (pass_cleanup_barriers): Replace declaration with that of...
6951 (make_pass_cleanup_barriers): ...new function.
6952 (pass_delay_slots): Replace declaration with that of...
6953 (make_pass_delay_slots): ...new function.
6954 (pass_split_for_shorten_branches): Replace declaration with that of...
6955 (make_pass_split_for_shorten_branches): ...new function.
6956 (pass_split_before_regstack): Replace declaration with that of...
6957 (make_pass_split_before_regstack): ...new function.
6958 (pass_convert_to_eh_region_ranges): Replace declaration with that
6960 (make_pass_convert_to_eh_region_ranges): ...new function.
6961 (pass_shorten_branches): Replace declaration with that of...
6962 (make_pass_shorten_branches): ...new function.
6963 (pass_set_nothrow_function_flags): Replace declaration with that of...
6964 (make_pass_set_nothrow_function_flags): ...new function.
6965 (pass_dwarf2_frame): Replace declaration with that of...
6966 (make_pass_dwarf2_frame): ...new function.
6967 (pass_final): Replace declaration with that of...
6968 (make_pass_final): ...new function.
6969 (pass_rtl_seqabstr): Replace declaration with that of...
6970 (make_pass_rtl_seqabstr): ...new function.
6971 (pass_release_ssa_names): Replace declaration with that of...
6972 (make_pass_release_ssa_names): ...new function.
6973 (pass_early_inline): Replace declaration with that of...
6974 (make_pass_early_inline): ...new function.
6975 (pass_inline_parameters): Replace declaration with that of...
6976 (make_pass_inline_parameters): ...new function.
6977 (pass_update_address_taken): Replace declaration with that of...
6978 (make_pass_update_address_taken): ...new function.
6979 (pass_convert_switch): Replace declaration with that of...
6980 (make_pass_convert_switch): ...new function.
6981 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
6982 to a subclass of simple_ipa_opt_pass along with...
6983 (pass_data_ipa_tree_profile): ...new pass_data instance and...
6984 (make_pass_ipa_tree_profile): ...new function.
6985 * tree-sra.c (pass_sra_early): Convert from a global struct to a
6986 subclass of gimple_opt_pass along with...
6987 (pass_data_sra_early): ...new pass_data instance and...
6988 (make_pass_sra_early): ...new function.
6989 (pass_sra): Convert from a global struct to a subclass of
6990 gimple_opt_pass along with...
6991 (pass_data_sra): ...new pass_data instance and...
6992 (make_pass_sra): ...new function.
6993 (pass_early_ipa_sra): Convert from a global struct to a subclass of
6994 gimple_opt_pass along with...
6995 (pass_data_early_ipa_sra): ...new pass_data instance and...
6996 (make_pass_early_ipa_sra): ...new function.
6997 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
6998 subclass of gimple_opt_pass along with...
6999 (pass_data_ccp): ...new pass_data instance and...
7000 (make_pass_ccp): ...new function.
7001 (pass_fold_builtins): Convert from a global struct to a subclass of
7002 gimple_opt_pass along with...
7003 (pass_data_fold_builtins): ...new pass_data instance and...
7004 (make_pass_fold_builtins): ...new function.
7005 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
7006 subclass of gimple_opt_pass along with...
7007 (pass_data_copy_prop): ...new pass_data instance and...
7008 (make_pass_copy_prop): ...new function.
7009 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
7010 global struct to a subclass of gimple_opt_pass along with...
7011 (pass_data_rename_ssa_copies): ...new pass_data instance and...
7012 (make_pass_rename_ssa_copies): ...new function.
7013 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
7014 subclass of gimple_opt_pass along with...
7015 (pass_data_dce): ...new pass_data instance and...
7016 (make_pass_dce): ...new function.
7017 (pass_dce_loop): Convert from a global struct to a subclass of
7018 gimple_opt_pass along with...
7019 (pass_data_dce_loop): ...new pass_data instance and...
7020 (make_pass_dce_loop): ...new function.
7021 (pass_cd_dce): Convert from a global struct to a subclass of
7022 gimple_opt_pass along with...
7023 (pass_data_cd_dce): ...new pass_data instance and...
7024 (make_pass_cd_dce): ...new function.
7025 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
7026 subclass of gimple_opt_pass along with...
7027 (pass_data_dominator): ...new pass_data instance and...
7028 (make_pass_dominator): ...new function.
7029 (pass_phi_only_cprop): Convert from a global struct to a subclass of
7030 gimple_opt_pass along with...
7031 (pass_data_phi_only_cprop): ...new pass_data instance and...
7032 (make_pass_phi_only_cprop): ...new function.
7033 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
7034 subclass of gimple_opt_pass along with...
7035 (pass_data_dse): ...new pass_data instance and...
7036 (make_pass_dse): ...new function.
7037 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
7038 a subclass of gimple_opt_pass along with...
7039 (pass_data_forwprop): ...new pass_data instance and...
7040 (make_pass_forwprop): ...new function.
7041 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
7042 struct to a subclass of gimple_opt_pass along with...
7043 (pass_data_tree_ifcombine): ...new pass_data instance and...
7044 (make_pass_tree_ifcombine): ...new function.
7045 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
7046 subclass of gimple_opt_pass along with...
7047 (pass_data_ch): ...new pass_data instance and...
7048 (make_pass_ch): ...new function.
7049 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
7050 subclass of gimple_opt_pass along with...
7051 (pass_data_tree_loop): ...new pass_data instance and...
7052 (make_pass_tree_loop): ...new function.
7053 (pass_tree_loop_init): Convert from a global struct to a subclass of
7054 gimple_opt_pass along with...
7055 (pass_data_tree_loop_init): ...new pass_data instance and...
7056 (make_pass_tree_loop_init): ...new function.
7057 (pass_lim): Convert from a global struct to a subclass of
7058 gimple_opt_pass along with...
7059 (pass_data_lim): ...new pass_data instance and...
7060 (make_pass_lim): ...new function.
7061 (pass_tree_unswitch): Convert from a global struct to a subclass of
7062 gimple_opt_pass along with...
7063 (pass_data_tree_unswitch): ...new pass_data instance and...
7064 (make_pass_tree_unswitch): ...new function.
7065 (pass_predcom): Convert from a global struct to a subclass of
7066 gimple_opt_pass along with...
7067 (pass_data_predcom): ...new pass_data instance and...
7068 (make_pass_predcom): ...new function.
7069 (pass_vectorize): Convert from a global struct to a subclass of
7070 gimple_opt_pass along with...
7071 (pass_data_vectorize): ...new pass_data instance and...
7072 (make_pass_vectorize): ...new function.
7073 (pass_graphite): Convert from a global struct to a subclass of
7074 gimple_opt_pass along with...
7075 (pass_data_graphite): ...new pass_data instance and...
7076 (make_pass_graphite): ...new function.
7077 (pass_graphite_transforms): Convert from a global struct to a subclass
7078 of gimple_opt_pass along with...
7079 (pass_data_graphite_transforms): ...new pass_data instance and...
7080 (make_pass_graphite_transforms): ...new function.
7081 (pass_check_data_deps): Convert from a global struct to a subclass of
7082 gimple_opt_pass along with...
7083 (pass_data_check_data_deps): ...new pass_data instance and...
7084 (make_pass_check_data_deps): ...new function.
7085 (pass_iv_canon): Convert from a global struct to a subclass of
7086 gimple_opt_pass along with...
7087 (pass_data_iv_canon): ...new pass_data instance and...
7088 (make_pass_iv_canon): ...new function.
7089 (pass_scev_cprop): Convert from a global struct to a subclass of
7090 gimple_opt_pass along with...
7091 (pass_data_scev_cprop): ...new pass_data instance and...
7092 (make_pass_scev_cprop): ...new function.
7093 (pass_record_bounds): Convert from a global struct to a subclass of
7094 gimple_opt_pass along with...
7095 (pass_data_record_bounds): ...new pass_data instance and...
7096 (make_pass_record_bounds): ...new function.
7097 (pass_complete_unroll): Convert from a global struct to a subclass of
7098 gimple_opt_pass along with...
7099 (pass_data_complete_unroll): ...new pass_data instance and...
7100 (make_pass_complete_unroll): ...new function.
7101 (pass_complete_unrolli): Convert from a global struct to a subclass of
7102 gimple_opt_pass along with...
7103 (pass_data_complete_unrolli): ...new pass_data instance and...
7104 (make_pass_complete_unrolli): ...new function.
7105 (pass_parallelize_loops): Convert from a global struct to a subclass
7106 of gimple_opt_pass along with...
7107 (pass_data_parallelize_loops): ...new pass_data instance and...
7108 (make_pass_parallelize_loops): ...new function.
7109 (pass_loop_prefetch): Convert from a global struct to a subclass of
7110 gimple_opt_pass along with...
7111 (pass_data_loop_prefetch): ...new pass_data instance and...
7112 (make_pass_loop_prefetch): ...new function.
7113 (pass_iv_optimize): Convert from a global struct to a subclass of
7114 gimple_opt_pass along with...
7115 (pass_data_iv_optimize): ...new pass_data instance and...
7116 (make_pass_iv_optimize): ...new function.
7117 (pass_tree_loop_done): Convert from a global struct to a subclass of
7118 gimple_opt_pass along with...
7119 (pass_data_tree_loop_done): ...new pass_data instance and...
7120 (make_pass_tree_loop_done): ...new function.
7121 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
7122 struct to a subclass of gimple_opt_pass along with...
7123 (pass_data_cse_reciprocals): ...new pass_data instance and...
7124 (make_pass_cse_reciprocals): ...new function.
7125 (pass_cse_sincos): Convert from a global struct to a subclass of
7126 gimple_opt_pass along with...
7127 (pass_data_cse_sincos): ...new pass_data instance and...
7128 (make_pass_cse_sincos): ...new function.
7129 (pass_optimize_bswap): Convert from a global struct to a subclass of
7130 gimple_opt_pass along with...
7131 (pass_data_optimize_bswap): ...new pass_data instance and...
7132 (make_pass_optimize_bswap): ...new function.
7133 (pass_optimize_widening_mul): Convert from a global struct to a
7134 subclass of gimple_opt_pass along with...
7135 (pass_data_optimize_widening_mul): ...new pass_data instance and...
7136 (make_pass_optimize_widening_mul): ...new function.
7137 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
7138 subclass of gimple_opt_pass along with...
7139 (pass_data_phiopt): ...new pass_data instance and...
7140 (make_pass_phiopt): ...new function.
7141 (pass_cselim): Convert from a global struct to a subclass of
7142 gimple_opt_pass along with...
7143 (pass_data_cselim): ...new pass_data instance and...
7144 (make_pass_cselim): ...new function.
7145 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
7146 subclass of gimple_opt_pass along with...
7147 (pass_data_phiprop): ...new pass_data instance and...
7148 (make_pass_phiprop): ...new function.
7149 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
7150 subclass of gimple_opt_pass along with...
7151 (pass_data_pre): ...new pass_data instance and...
7152 (make_pass_pre): ...new function.
7153 (pass_fre): Convert from a global struct to a subclass of
7154 gimple_opt_pass along with...
7155 (pass_data_fre): ...new pass_data instance and...
7156 (make_pass_fre): ...new function.
7157 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
7158 subclass of gimple_opt_pass along with...
7159 (pass_data_reassoc): ...new pass_data instance and...
7160 (make_pass_reassoc): ...new function.
7161 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
7162 subclass of gimple_opt_pass along with...
7163 (pass_data_sink_code): ...new pass_data instance and...
7164 (make_pass_sink_code): ...new function.
7165 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
7166 subclass of gimple_opt_pass along with...
7167 (pass_data_strlen): ...new pass_data instance and...
7168 (make_pass_strlen): ...new function.
7169 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
7170 struct to a subclass of gimple_opt_pass along with...
7171 (pass_data_build_alias): ...new pass_data instance and...
7172 (make_pass_build_alias): ...new function.
7173 (pass_build_ealias): Convert from a global struct to a subclass of
7174 gimple_opt_pass along with...
7175 (pass_data_build_ealias): ...new pass_data instance and...
7176 (make_pass_build_ealias): ...new function.
7177 (pass_ipa_pta): Convert from a global struct to a subclass of
7178 simple_ipa_opt_pass along with...
7179 (pass_data_ipa_pta): ...new pass_data instance and...
7180 (make_pass_ipa_pta): ...new function.
7181 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
7182 subclass of gimple_opt_pass along with...
7183 (pass_data_uncprop): ...new pass_data instance and...
7184 (make_pass_uncprop): ...new function.
7185 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
7186 global struct to a subclass of gimple_opt_pass along with...
7187 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
7188 (make_pass_late_warn_uninitialized): ...new function.
7189 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
7190 to a subclass of gimple_opt_pass along with...
7191 (pass_data_init_datastructures): ...new pass_data instance and...
7192 (make_pass_init_datastructures): ...new function.
7193 (pass_early_warn_uninitialized): Convert from a global struct to a
7194 subclass of gimple_opt_pass along with...
7195 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
7196 (make_pass_early_warn_uninitialized): ...new function.
7197 (pass_update_address_taken): Convert from a global struct to a
7198 subclass of gimple_opt_pass along with...
7199 (pass_data_update_address_taken): ...new pass_data instance and...
7200 (make_pass_update_address_taken): ...new function.
7201 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
7202 struct to a subclass of gimple_opt_pass along with...
7203 (pass_data_release_ssa_names): ...new pass_data instance and...
7204 (make_pass_release_ssa_names): ...new function.
7205 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
7206 subclass of gimple_opt_pass along with...
7207 (pass_data_stdarg): ...new pass_data instance and...
7208 (make_pass_stdarg): ...new function.
7209 * tree-switch-conversion.c (pass_convert_switch): Convert from a
7210 global struct to a subclass of gimple_opt_pass along with...
7211 (pass_data_convert_switch): ...new pass_data instance and...
7212 (make_pass_convert_switch): ...new function.
7213 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
7214 to a subclass of gimple_opt_pass along with...
7215 (pass_data_tail_recursion): ...new pass_data instance and...
7216 (make_pass_tail_recursion): ...new function.
7217 (pass_tail_calls): Convert from a global struct to a subclass of
7218 gimple_opt_pass along with...
7219 (pass_data_tail_calls): ...new pass_data instance and...
7220 (make_pass_tail_calls): ...new function.
7221 * tree-vect-generic.c (pass_lower_vector): Convert from a global
7222 struct to a subclass of gimple_opt_pass along with...
7223 (pass_data_lower_vector): ...new pass_data instance and...
7224 (make_pass_lower_vector): ...new function.
7225 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
7226 gimple_opt_pass along with...
7227 (pass_data_lower_vector_ssa): ...new pass_data instance and...
7228 (make_pass_lower_vector_ssa): ...new function.
7229 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
7230 to a subclass of gimple_opt_pass along with...
7231 (pass_data_slp_vectorize): ...new pass_data instance and...
7232 (make_pass_slp_vectorize): ...new function.
7233 (pass_ipa_increase_alignment): Convert from a global struct to a
7234 subclass of simple_ipa_opt_pass along with...
7235 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
7236 (make_pass_ipa_increase_alignment): ...new function.
7237 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
7238 gimple_opt_pass along with...
7239 (pass_data_vrp): ...new pass_data instance and...
7240 (make_pass_vrp): ...new function.
7241 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
7242 subclass of simple_ipa_opt_pass along with...
7243 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
7244 (make_pass_ipa_free_lang_data): ...new function.
7245 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
7246 gimple_opt_pass along with...
7247 (pass_data_tsan): ...new pass_data instance and...
7248 (make_pass_tsan): ...new function.
7249 (pass_tsan_O0): Convert from a global struct to a subclass of
7250 gimple_opt_pass along with...
7251 (pass_data_tsan_O0): ...new pass_data instance and...
7252 (make_pass_tsan_O0): ...new function.
7253 * var-tracking.c (pass_variable_tracking): Convert from a global
7254 struct to a subclass of rtl_opt_pass along with...
7255 (pass_data_variable_tracking): ...new pass_data instance and...
7256 (make_pass_variable_tracking): ...new function.
7257 * web.c (pass_web): Convert from a global struct to a subclass of
7258 rtl_opt_pass along with...
7259 (pass_data_web): ...new pass_data instance and...
7260 (make_pass_web): ...new function.
7261 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
7262 declaration with that of...
7263 (make_pass_mode_switch_use): ...new function.
7264 (pass_resolve_sw_modes): Replace declaration with that of...
7265 (make_pass_resolve_sw_modes): ...new function.
7266 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
7267 from a global struct to a subclass of rtl_opt_pass along with...
7268 (pass_data_mode_switch_use): ...new pass_data instance and...
7269 (make_pass_mode_switch_use): ...new function.
7270 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
7271 from a global struct to a subclass of rtl_opt_pass along with...
7272 (pass_data_resolve_sw_modes): ...new pass_data instance and...
7273 (make_pass_resolve_sw_modes): ...new function.
7274 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
7275 struct to a subclass of rtl_opt_pass along with...
7276 (pass_data_insert_vzeroupper): ...new pass_data instance and...
7277 (make_pass_insert_vzeroupper): ...new function.
7278 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
7279 global struct to a subclass of rtl_opt_pass along with...
7280 (pass_data_work_around_errata): ...new pass_data instance and...
7281 (make_pass_work_around_errata): ...new function.
7282 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
7283 struct to a subclass of rtl_opt_pass along with...
7284 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
7285 (make_pass_mips_machine_reorg2): ...new function.
7287 2013-08-05 David Malcolm <dmalcolm@redhat.com>
7289 * passes.c (pass_manager::operator new): New.
7291 2013-08-05 David Malcolm <dmalcolm@redhat.com>
7293 Handwritten part of conversion of passes to C++ classes.
7295 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
7296 (toplev.o): Add dep on PASS_MANAGER_H.
7297 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
7298 of early local pases to reflect this moving from a global to a
7299 member of gcc::pass_manager.
7300 (cgraph_add_new_function): Likewise.
7301 * lto-cgraph.c (lto_output_node): Update for conversion of
7302 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
7303 * passes.c (opt_pass::clone): New.
7304 (opt_pass::gate): New.
7305 (opt_pass::execute): New.
7306 (opt_pass::opt_pass): New.
7307 (pass_manager::execute_early_local_passes): New.
7308 (pass_manager::execute_pass_mode_switching): new.
7309 (finish_optimization_passes): Convert to...
7310 (pass_manager::finish_optimization_passes): ...this.
7311 (finish_optimization_passes): Update for conversion of passes to
7313 (register_dump_files_1): Use has_gate since we cannot portably
7314 check a vtable entry against NULL.
7315 (dump_one_pass): Likewise.
7316 (ipa_write_summaries_2): Likewise.
7317 (ipa_write_optimization_summaries_1): Likewise.
7318 (ipa_read_summaries_1): Likewise.
7319 (ipa_read_optimization_summaries_1): Likewise.
7320 (execute_ipa_stmt_fixups): Likewise.
7321 (pass_manager::pass_manager): Rewrite pass-creation, invoking
7322 pass-creation functions rather than wiring up globals, and
7323 storing the results in fields of pass_manager generated using
7325 (pass_manager::dump_profile_report): Update for conversion of
7326 passes to C++ classes.
7327 (pass_manager::execute_ipa_summary_passes): Likewise.
7328 (execute_one_ipa_transform_pass): Likewise.
7329 (execute_one_pass): Use has_gate and has_execute since we cannot
7330 portably check a vtable entry against NULL.
7331 * pass_manager.h (pass_manager::finish_optimization_passes): New.
7332 (pass_manager): Use pass-instances.def to add fields for the
7333 various pass instances.
7334 * toplev.c (finalize): Update for move of
7335 finish_optimization_passes to a method of gcc::pass_manager.
7336 * toplev.h (finish_optimization_passes): Move to method of class
7338 * tree-pass.h (struct pass_data): New.
7339 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
7340 (opt_pass::gate): Convert to virtual function.
7341 (opt_pass::~opt_pass): New.
7342 (opt_pass::clone): New.
7343 (opt_pass::execute): Convert to virtual function.
7344 (opt_pass::opt_pass): New.
7345 (opt_pass::ctxt_): new.
7346 (gimple_opt_pass): Convert to subclass of opt_pass.
7347 (gimple_opt_pass::gimple_opt_pass): New.
7348 (rtl_opt_pass): Convert to subclass of opt_pass.
7349 (rtl_opt_pass::rtl_opt_pass): New.
7350 (ipa_opt_pass_d): Convert to subclass of opt_pass.
7351 (ipa_opt_pass_d::ipa_opt_pass_d): New.
7352 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
7353 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
7354 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
7355 invocation of pass_mode_switching to reflect this moving from a
7356 global to a member of gcc::pass_manager.
7357 (ix86_option_override): Rework how pass_insert_vzeroupper is
7358 added to the pass_manager to reflect autogenerated changes.
7359 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
7361 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
7363 PR rtl-optimization/57708
7364 * recog.c (peep2_find_free_register): Validate all regs in a
7367 2013-08-05 Jan Hubicka <jh@suse.cz>
7370 * cgraph.c (verify_cgraph_node): Accept local flags from other
7372 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
7373 (function_and_variable_visibility): Likewise.
7374 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
7376 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
7378 * graph.c (init_graph_slim_pretty_print): Remove.
7379 (print_graph_cfg): Do not call it. Use local pretty printer.
7380 (start_graph_dump): Likewise.
7382 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
7384 * gimple-pretty-print.c (buffer): Remove.
7385 (initialized): Likewise.
7386 (maybe_init_pretty_print): Likewise.
7387 (print_gimple_stmt): Do not call it. Use non-static local
7388 pretty_printer variable.
7389 (print_gimple_expr): Likewise.
7390 (print_gimple_seq): Likewise.
7391 (gimple_dump_bb): Likewise.
7393 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
7395 * asan.c (asan_pp): Remove.
7396 (asan_pp_initialized): Likewise.
7397 (asan_pp_initialize): Likewise.
7398 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
7399 (asan_emit_stack_protection): Tidy. Use local pretty printer.
7400 (asan_add_global): Likewise.
7402 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
7404 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
7405 * diagnostic.h (diagnostic_flush_buffer): Adjust.
7406 * pretty-print.c (pp_formatted_text_data): Likewise.
7407 (pp_indent): Rename from pp_base_indent.
7408 (pp_format): Rename from pp_base_format.
7409 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
7410 (pp_format_verbatim): Rename from pp_base_format_verbatim.
7411 (pp_flush): Rename from pp_base_flush.
7412 (pp_set_line_maximum_length): Rename from
7413 pp_base_set_line_maximum_length.
7414 (pp_clear_output_area): Rename from pp_base_clear_output_area.
7415 (pp_set_prefix): Rename from pp_base_set_prefix.
7416 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
7417 (pp_emit_prefix): Rename from pp_base_emit_prefix.
7418 (pp_append_text): Rename from pp_base_append_text.
7419 (pp_formatted_text): Rename from pp_base_formatted_text.
7420 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
7421 (pp_remaining_character_count_for_line): Rename from
7422 pp_base_remaining_character_count_for_line.
7423 (pp_newline): Rename from pp_base_newline.
7424 (pp_character): Rename from pp_base_character.
7425 (pp_string): Rename from pp_base_string.
7426 (pp_maybe_space): Rename from pp_base_maybe_space.
7427 * asan.c (asan_pp_string): Adjust.
7428 (asan_emit_stack_protection): Likewise.
7429 (asan_add_global): Likewise.
7430 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
7431 * tree-mudflap.c (mf_varname_tree): Likewise.
7432 * tree-pretty-print.c (pp_tree_identifier): Rename from
7433 pp_base_tree_identifier.
7434 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
7435 Declare as function.
7437 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
7439 * pretty-print.h (pp_bar_bar): New.
7440 (pp_ampersand_ampersand): Likewise.
7441 (pp_less_equal): Likewise.
7442 (pp_greater_equal): Likewise.
7443 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
7444 printer functions instead of pp_string or operators and punctuators.
7445 (dump_gimple_call): Likewise.
7446 (dump_gimple_omp_for): Likewise.
7447 (dump_gimple_transaction): Likewise.
7448 (dump_gimple_phi): Likewise.
7449 (pp_gimple_stmt_1): Likewise.
7450 * sched-vis.c (print_insn): Likewise.
7451 * tree-mudflap.c (mf_varname_tree): Likewise.
7452 * tree-pretty-print.c (dump_block_node): Likewise.
7453 (dump_generic_node): Likewise.
7455 2013-08-02 Jan Hubicka <jh@suse.cz>
7457 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
7459 * lto-streamer-out.c (tree_is_indexable): Results decls and
7460 parm decls are not indexable.
7461 (DFS_write_tree_body): Do not follow args and results.
7462 (hash_tree): Likewise.
7463 (output_functions): Rearrange so struct function is needed
7464 only when real body is output; be able to also ouptut abstract
7465 functions; output DECL_ARGUMENTS and DECL_RESULT.
7466 (lto_output): When not in WPA, ale store abstract functions.
7467 (write_symbol): Do not care about RESULT_DECL.
7468 (output_symbol_p): Handle correctly sbtract decls.
7469 * lto-streamer-in.c (input_function): Rearrange so struct
7470 function can be NULL at entry; allow streaming of
7471 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
7472 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
7473 sanity check during LTO.
7474 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
7475 RESULT_DECl and DECL_ARGUMENTS.
7476 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
7479 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
7481 * pretty-print.h (pp_underscore): New.
7483 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
7484 printer functions instead of pp_character.
7485 (dump_binary_rhs): Likewise.
7486 (dump_ternary_rhs): Likewise.
7487 (dump_gimple_call_args): Likewise.
7488 (pp_points_to_solution): Likewise.
7489 (dump_gimple_call): Likewise.
7490 (dump_gimple_switch): Likewise.
7491 (dump_gimple_cond): Likewise.
7492 (dump_gimple_bind): Likewise.
7493 (dump_gimple_try): Likewise.
7494 (dump_gimple_omp_for): Likewise.
7495 (dump_gimple_omp_continue): Likewise.
7496 (dump_gimple_omp_single): Likewise.
7497 (dump_gimple_omp_sections): Likewise.
7498 (dump_gimple_omp_block): Likewise.
7499 (dump_gimple_omp_critical): Likewise.
7500 (dump_gimple_transaction): Likewise.
7501 (dump_gimple_asm): Likewise.
7502 (dump_gimple_phi): Likewise.
7503 (dump_gimple_omp_parallel): Likewise.
7504 (dump_gimple_omp_task): Likewise.
7505 (dump_gimple_omp_atomic_load): Likewise.
7506 (dump_gimple_omp_atomic_store): Likewise.
7507 (dump_gimple_mem_ops): Likewise.
7508 (pp_gimple_stmt_1): Likewise.
7509 (pp_cfg_jump): Likewise.
7510 (dump_implicit_edges): Likewise.
7511 (gimple_dump_bb_for_graph): Likewise.
7512 * graph.c (draw_cfg_node): Likewise.
7513 * langhooks.c (lhd_print_error_function): Likewise.
7514 * sched-vis.c (print_exp): Likewise.
7515 (print_value): Likewise.
7516 (print_pattern): Likewise.
7517 (print_insn): Likewise.
7518 (rtl_dump_bb_for_graph): Likewise.
7519 * tree-pretty-print.c (dump_function_declaration): Likewise.
7520 (dump_array_domain): Likewise.
7521 (dump_omp_clause): Likewise.
7522 (dump_location): Likewise.
7523 (dump_generic_node): Likewise.
7524 (print_struct_decl): Likewise.
7525 * diagnostic.c (diagnostic_show_locus): Use pp_space.
7527 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
7529 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
7530 candidate table when replacing a candidate statement.
7531 (replace_rhs_if_not_dup): Likewise.
7532 (replace_one_candidate): Likewise.
7534 2013-08-02 Jan Hubicka <jh@suse.cz>
7535 Martin Liska <marxin.liska@gmail.com>
7537 * cgraphunit.c (add_new_function): Fix logic when adding from
7539 (assemble_thunk): Rename to ...
7540 (expand_thunk); .. this one; export; get it working with
7541 general functions; make produced gimple valid.
7542 * cgraph.h (expand_thunk): Declare.
7544 2013-08-02 Jan Hubicka <jh@suse.cz>
7546 * ipa-cp.c (gather_context_independent_values): Use
7547 ipa_get_param_move_cost.
7548 (get_replacement_map): Remove PARAM; move parameter folding
7550 (create_specialized_node): Update.
7551 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
7552 assert that we have gimple body; update move_cost.
7553 (count_formal_params): Assert that we have gimple body.
7554 (ipa_dump_param): New function.
7555 (ipa_alloc_node_params): Break out from ...
7556 (ipa_initialize_node_params): ... here.
7557 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
7558 (ipa_write_node_info): Stream move costs.
7559 (ipa_read_node_info): Read move costs.
7560 (ipa_update_after_lto_read): Do not recompute node params.
7561 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
7562 (ipa_get_param): Check we are not in WPA.
7563 (ipa_get_param_move_cost): New.
7564 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
7565 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
7566 parm numbers to be present.
7568 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
7570 PR rtl-optimization/58048
7571 * lra-constraints.c (process_alt_operands): Don't check asm
7572 operand on register.
7574 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
7576 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
7577 the implied StoreLoad barrier for atomic operations if before.
7579 2013-08-02 Jan Hubicka <jh@suse.cz>
7580 Martin Liska <marxin.liska@gmail.com>
7582 * cgraph.c (cgraph_function_body_availability): Do not check
7584 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
7585 symtab_node_availability): Declare.
7586 * ipa.c (can_replace_by_local_alias): New.
7587 (function_and_variable_visibility): Use it.
7588 * symtab.c (symtab_for_node_and_aliases,
7589 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
7591 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
7593 PR rtl-optimization/57963
7594 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
7595 (lra_constraints): Use them.
7597 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
7599 * config/arm/types.md (define_attr "type"): Add "load_acq"
7601 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
7603 (cortex_a53_store1): Likewise.
7605 2013-08-01 Jan Hubicka <jh@suse.cz>
7607 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
7608 partitions are not needed.
7610 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
7612 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
7613 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
7614 MAYBE_NON_Q_CLASS_P where appropriate.
7616 2013-08-01 Jan Hubicka <jh@suse.cz>
7618 * cgraph.h (release_function_body): Declare.
7619 * tree.c (free_lang_data_in_decl): Free, parameters and return values
7620 of unused delcarations.
7622 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7624 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
7625 RTL form when subtracting a constant.
7627 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7629 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
7630 Generate canonical plus rtx with negated immediate instead of minus
7632 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
7634 2013-08-01 Jan Hubicka <jh@suse.cz>
7636 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
7637 (cgraph_release_function_body): Likewise.
7638 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
7639 * cgraph.h (cgrpah_node): Rename abstract_and_needed
7640 to used_as_abstract_origin.
7641 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
7642 symbols used as abstract origins.
7643 * cgraphunit.c (analyze_functions): Update.
7644 * ipa.c (symtab_remove_unreachable_nodes): Recompute
7645 used_as_abstract_origin.
7646 * tree-inline.c (tree_function_versioning): Update
7647 used_as_abstract_origin; be ready for DECL_RESULT and
7648 DECL_ARGUMENTS to be NULL.
7650 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
7651 for abstract functions.
7652 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
7655 2013-08-01 Jan Hubicka <jh@suse.cz>
7657 * profile.c (compute_value_histograms): Fix thinko.
7659 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
7661 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
7662 aarch-common-protos.h to extra_headers.
7663 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
7664 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
7665 * config/aarch64/t-aarch64 (aarch-common.o): Define.
7667 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
7669 * config/aarch64/aarch64.md (define_attr "type"): Delete.
7670 Include "../arm/types.md". Define "type" attribute for all patterns.
7671 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
7674 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
7676 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
7677 to support power8 load fusion.
7678 (fusion_gpr_mem_load): Likewise.
7680 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
7682 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
7683 declarations for power8 load fusion.
7684 (emit_fusion_gpr_load): Likewise.
7686 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
7687 tuning for power8, turn on fusion mode by default. Turn on sign
7688 extending fusion mode if normal fusion mode is on, and we are at
7690 (fusion_gpr_load_p): New function, return true if we can fuse an
7691 addis instruction with a dependent load to a GPR.
7692 (emit_fusion_gpr_load): Emit the instructions for power8 load
7695 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
7696 (VSX load fusion peepholes): New peepholes to fuse together an
7697 addi instruction with a VSX load instruction.
7699 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
7700 peepholes to fuse an addis instruction with a load to a GPR base
7701 register. If we are supporting sign extending fusions, convert
7702 sign extending loads to zero extending loads and add an explicit
7705 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
7707 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
7708 aarch-common-protos.h to extra_headers.
7709 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
7710 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
7711 (arm_early_store_addr_dep): Likewise.
7712 (arm_no_early_alu_shift_dep): Likewise.
7713 (arm_no_early_alu_shift_value_dep): Likewise.
7714 (arm_no_early_mul_dep): Likewise.
7715 (arm_no_early_store_addr_dep): Likewise.
7716 (arm_mac_accumulator_is_mul_result): Likewise.
7717 (arm_mac_accumulator_is_result): Likewise.
7718 * config/arm/aarch-common.c: ... here. New file.
7719 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
7721 (arm_early_store_addr_dep): Likewise.
7722 (arm_no_early_alu_shift_dep): Likewise.
7723 (arm_no_early_alu_shift_value_dep): Likewise.
7724 (arm_no_early_mul_dep): Likewise.
7725 (arm_no_early_store_addr_dep): Likewise.
7726 (arm_mac_accumulator_is_mul_result): Likewise.
7727 (arm_mac_accumulator_is_result): Likewise.
7728 * config/arm/aarch-common-protos.h: ... here. New file.
7729 * config/arm/t-arm (aarch-common.o): Define.
7731 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
7733 * config/arm/arm.md: Include new file "types.md".
7734 (define_attr "type"): Move from here to ...
7735 (define_attr "mul32"): Likewise.
7736 (define_attr "mul64"): Likewise.
7737 * config/arm/types.md: ... here. New file.
7739 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
7741 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
7742 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
7744 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7746 * gen-pass-instances.awk: Fix offset of substr().
7748 2013-07-31 David Malcolm <dmalcolm@redhat.com>
7750 * Makefile.in (pass-instances.def): New.
7751 (passes.o): Replace dependency on passes.def with one on
7754 * gen-pass-instances.awk: New.
7756 * passes.c (pass_manager::pass_manager): Use pass-instances.def
7757 rather than passes.def, updating local definition of NEXT_PASS
7758 macro to add an extra NUM parameter (currently unused).
7760 2013-07-30 David Malcolm <dmalcolm@redhat.com>
7762 * Makefile.in (PASS_MANAGER_H): New.
7763 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
7764 (passes.o): Likewise.
7765 (statistics.o): Likewise.
7766 (cgraphunit.o): Likewise.
7767 (context.o): Depend on PASS_MANAGER_H.
7769 * pass_manager.h: New.
7771 * cgraphunit.c (cgraph_add_new_function): Update for moves
7772 of globals to fields of pass_manager.
7773 (analyze_function): Likewise.
7774 (expand_function): Likewise.
7775 (ipa_passes): Likewise.
7776 (compile): Likewise.
7778 * context.c (context::context): New.
7779 * context.h (context::context): New.
7780 (context::get_passes): New.
7781 (context::passes_): New.
7783 * lto-cgraph.c (input_node): Update for moves of globals to
7784 fields of pass_manager.
7786 * passes.c (all_passes): Remove, in favor of a field of the
7787 same name within the new class pass_manager.
7788 (all_small_ipa_passes): Likewise.
7789 (all_lowering_passes): Likewise.
7790 (all_regular_ipa_passes): Likewise.
7791 (all_late_ipa_passes): Likewise.
7792 (all_lto_gen_passes): Likewise.
7793 (passes_by_id): Likewise.
7794 (passes_by_id_size): Likewise.
7795 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
7796 the new class pass_manager.
7797 (set_pass_for_id): Convert to...
7798 (pass_manager::set_pass_for_id): ...method.
7799 (get_pass_for_id): Convert to...
7800 (pass_manager::get_pass_for_id): ...method.
7801 (register_one_dump_file): Move body of implementation into...
7802 (pass_manager::register_one_dump_file): ...here.
7803 (register_dump_files_1): Convert to...
7804 (pass_manager::register_dump_files_1): ...method.
7805 (register_dump_files): Convert to...
7806 (pass_manager::register_dump_files): ...method.
7807 (create_pass_tab): Update for moves of globals to fields of
7809 (dump_passes): Move body of implementation into...
7810 (pass_manager::dump_passes): ...here.
7811 (register_pass): Move body of implementation into...
7812 (pass_manager::register_pass): ...here.
7813 (init_optimization_passes): Convert into...
7814 (pass_manager::pass_manager): ...constructor for new
7815 pass_manager class, and initialize the pass_lists array.
7816 (check_profile_consistency): Update for moves of globals to
7817 fields of pass_manager.
7818 (dump_profile_report): Move body of implementation into...
7819 (pass_manager::dump_profile_report): ...here.
7820 (ipa_write_summaries_1): Update for moves of pass lists from
7821 being globals to fields of pass_manager.
7822 (ipa_write_optimization_summaries): Likewise.
7823 (ipa_read_summaries): Likewise.
7824 (ipa_read_optimization_summaries): Likewise.
7825 (execute_all_ipa_stmt_fixups): Likewise.
7827 * statistics.c (statistics_fini): Update for moves of globals to
7828 fields of pass_manager.
7830 * toplev.c (general_init): Replace call to
7831 init_optimization_passes with construction of the pass_manager
7834 * tree-pass.h (all_passes): Remove, in favor of a field of the
7835 same name within the new class pass_manager.
7836 (all_small_ipa_passes): Likewise.
7837 (all_lowering_passes): Likewise.
7838 (all_regular_ipa_passes): Likewise.
7839 (all_lto_gen_passes): Likewise.
7840 (all_late_ipa_passes): Likewise.
7841 (passes_by_id): Likewise.
7842 (passes_by_id_size): Likewise.
7843 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
7844 the new class pass_manager.
7845 (get_pass_for_id): Remove.
7847 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
7849 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
7852 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
7854 * arm.md (mulhi3): New expand pattern.
7856 2013-07-30 Jan Hubicka <jh@suse.cz>
7857 Martin Liska <marxin.liska@gmail.com>
7859 * profile.c (compute_value_histograms): Do not ICE when
7860 there is mismatch only on some counters.
7862 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
7864 PR rtl-optimization/57637
7865 * function.c (move_insn_for_shrink_wrap): Also check the
7866 GEN set of the LIVE problem for the liveness analysis
7867 if it exists, otherwise give up.
7869 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
7871 PR tree-optimization/57993
7872 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
7873 replaced statement in the candidate table.
7874 (phi_add_costs): Return infinite cost when the hidden basis does
7875 not dominate all phis on which the candidate is dependent.
7876 (replace_one_candidate): Record replaced statement in the
7879 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
7881 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
7882 (ashlv2si3): New expander.
7883 (*ashlv2si3_i): New define_insn_and_split.
7884 * predicates.md (float_operation): Allow patterns with three
7887 PR rtl-optimization/58021
7888 * mode-switching.c (create_pre_exit): Always split off preceding
7889 insns if we are not at the basic block head.
7891 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
7893 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
7894 (UCLIBC_DYNAMIC_LINKER): New macro.
7895 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
7897 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
7898 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
7900 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
7901 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
7902 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
7903 for SF and DF modes. Use ieee_quad_format for TF mode.
7904 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
7905 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
7906 (mips_option_override): Handle `-mnan=legacy'.
7907 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
7908 `-mabs=2008' and `-mnan=2008'.
7909 (OPTION_DEFAULT_SPECS): Add "nan" default.
7910 (ASM_SPEC): Handle `-mnan='.
7911 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
7912 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
7913 comment accordingly.
7914 (neg<mode>2): Likewise.
7915 * config/mips/mips.opt (mabs, mnan): New options.
7916 * doc/install.texi (Configuration): Document `--with-nan=' option.
7917 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
7919 (MIPS Options): Document them.
7920 * config.gcc <mips*-*-*>: Handle `--with-nan='.
7921 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
7922 * configure: Regenerate.
7923 * config.in: Regenerate.
7925 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
7927 * config/i386/i386.md (float post-reload splitters): Do not check
7928 for subregs of SSE registers.
7930 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
7931 H.J. Lu <hongjiu.lu@intel.com>
7935 * config/i386/i386.md (post-reload splitter
7936 to avoid partial SSE reg dependency stalls): New pattern.
7938 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
7940 * config/s390/s390.md ("movcc"): Swap load and store instructions.
7942 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
7944 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
7945 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
7947 2013-07-26 Cary Coutant <ccoutant@google.com>
7949 * dwarf2out.c (die_checksum_ordered): Don't include template
7950 instantiations in signature.
7951 (is_template_parameter): New function.
7952 (is_template_instantiation): New function.
7953 (generate_skeleton_bottom_up): Don't include template instantiations
7955 (generate_skeleton): Likewise.
7956 (break_out_comdat_types): Move recursive call to break out nested
7958 (prune_unused_types_mark_generic_parms_dies): Call
7959 is_template_parameter.
7961 2013-07-26 Ian Bolton <ian.bolton@arm.com>
7963 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
7964 uses vector registers.
7965 * config/aarch64/iterators.md: Add attributes rtn and vas.
7967 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7968 Richard Earnshaw <richard.earnshaw@arm.com>
7970 * combine.c (simplify_comparison): Re-canonicalize operands
7972 * config/arm/arm.md (movcond_addsi): New splitter.
7974 2013-07-25 Sterling Augustine <saugustine@google.com>
7976 * dwarf2out.c (size_of_pubnames): Move code to...
7977 (include_pubname_in_output): ...here. New.
7978 (want_pubnames): Rearrange.
7979 (output_pubnames): Call include_pubname_in_output. Move assertion.
7981 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
7983 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
7985 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
7988 * doc/extend.texi: Remove obsolete builtins. Fix
7989 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
7991 2013-07-25 Jan Hubicka <jh@suse.cz>
7993 * cgraph.c (release_function_body): Break out from ...
7994 (cgraph_release_function_body): ... this one; also release DECL_RESULT
7996 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
7997 old_tree in the map.
7998 (create_specialized_node): Update.
7999 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
8001 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
8002 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
8003 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
8004 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
8007 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8009 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
8010 addsi3_carryin_alt2_<optab>): Correct output template.
8012 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8014 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
8015 Adjust for arm_restrict_it.
8016 Remove trailing whitespace.
8018 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
8020 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
8021 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
8023 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
8025 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
8027 PR rtl-optimization/57960
8028 * lra-constraints.c (process_alt_operands): Use the right mode
8029 when checking strict_low.
8031 2013-07-25 Jan Hubicka <jh@suse.cz>
8033 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
8034 * cgraph.c (cgraph_remove_node): Do not release function body
8035 when in cgraph streaming.
8036 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
8037 in other partitions are not considered reachable; fix handling of
8040 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
8042 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
8044 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
8049 * config/arm/arm.md ("*sibcall_insn): Replace use of
8050 Ss with US. Adjust output for v5 and v4t.
8051 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
8053 * config/arm/constraints.md ("Ss"): Rename to US.
8055 2013-07-25 Terry Guo <terry.guo@arm.com>
8057 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
8058 shift_add/shift_sub0/shift_sub1 RTXs.
8060 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
8061 Anton Blanchard <anton@au1.ibm.com>
8063 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
8064 (altivec_vpks<VI_char>ss): Likewise.
8065 (altivec_vpks<VI_char>us): Likewise.
8066 (altivec_vpku<VI_char>us): Likewise.
8067 (altivec_vpku<VI_char>um): Likewise.
8069 2013-07-24 David Malcolm <dmalcolm@redhat.com>
8071 Introduce context class.
8073 * Makefile.in (CONTEXT_H): New.
8074 (OBJS): Add context.o.
8075 (toplev.o): Add CONTEXT_H to dependencies.
8078 * toplev.c (general_init): Create the singleton gcc::context instance.
8084 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
8086 PR rtl-optimization/57968
8087 * mode-switching.c (create_pre_exit): Allow instructions that
8088 don't set a return register to need a non-exit mode.
8090 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
8091 Anton Blanchard <anton@au1.ibm.com>
8093 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
8094 operands to vperm for little endian.
8095 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
8096 of lvsl to create the control mask for a vperm for little endian.
8098 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8099 Anton Blanchard <anton@au1.ibm.com>
8101 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
8102 two operands for little-endian.
8104 2013-07-23 Steve Ellcey <sellcey@mips.com>
8106 * config/mips/mips.c (mips_case_values_threshold): New.
8107 (TARGET_CASE_VALUES_THRESHOLD): Define.
8109 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8110 Anton Blanchard <anton@au1.ibm.com>
8112 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
8113 selection of field for vector splat in little endian mode.
8115 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
8117 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
8118 expanders to rs6000.md.
8119 (ior<mode>3): Likewise.
8120 (and<mode>3): Likewise.
8121 (one_cmpl<mode>2): Likewise.
8122 (nor<mode>3): Likewise.
8123 (andc<mode>3): Likewise.
8124 (eqv<mode>3): Likewise.
8125 (nand<mode>3): Likewise.
8126 (orc<mode>3): Likewise.
8128 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
8131 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
8132 to split multi-word logical operations.
8133 (rs6000_split_logical_di): Likewise.
8134 (rs6000_split_logical): Likewise.
8136 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
8137 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
8138 and allow TImode operations in 32-bit.
8139 (vsx_and<mode>3_64bit): Likewise.
8140 (vsx_ior<mode>3_32bit): Likewise.
8141 (vsx_ior<mode>3_64bit): Likewise.
8142 (vsx_xor<mode>3_32bit): Likewise.
8143 (vsx_xor<mode>3_64bit): Likewise.
8144 (vsx_one_cmpl<mode>2_32bit): Likewise.
8145 (vsx_one_cmpl<mode>2_64bit): Likewise.
8146 (vsx_nor<mode>3_32bit): Likewise.
8147 (vsx_nor<mode>3_64bit): Likewise.
8148 (vsx_andc<mode>3_32bit): Likewise.
8149 (vsx_andc<mode>3_64bit): Likewise.
8150 (vsx_eqv<mode>3_32bit): Likewise.
8151 (vsx_eqv<mode>3_64bit): Likewise.
8152 (vsx_nand<mode>3_32bit): Likewise.
8153 (vsx_nand<mode>3_64bit): Likewise.
8154 (vsx_orc<mode>3_32bit): Likewise.
8155 (vsx_orc<mode>3_64bit): Likewise.
8157 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
8158 logical types in GPRs.
8160 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
8161 logical insns to rs6000.md, and allow TImode operations in
8163 (altivec_ior<mode>3): Likewise.
8164 (altivec_xor<mode>3): Likewise.
8165 (altivec_one_cmpl<mode>2): Likewise.
8166 (altivec_nor<mode>3): Likewise.
8167 (altivec_andc<mode>3): Likewise.
8169 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
8170 attributes for moving the 128-bit logical operations into
8172 (BOOL_REGS_OUTPUT): Likewise.
8173 (BOOL_REGS_OP1): Likewise.
8174 (BOOL_REGS_OP2): Likewise.
8175 (BOOL_REGS_UNARY): Likewise.
8176 (BOOL_REGS_AND_CR0): Likewise.
8177 (one_cmpl<mode>2): Add support for DI logical operations on
8178 32-bit, splitting the operations to 32-bit.
8182 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
8183 changes to combine the 32/64-bit code, allow logical operations on
8184 TI mode in 32-bit, and to use similar match_operator patterns like
8185 scalar mode uses. Combine the Altivec and VSX code for logical
8186 operations, and move it here.
8187 (ior<mode>3, 128-bit types): Likewise.
8188 (xor<mode>3, 128-bit types): Likewise.
8189 (one_cmpl<mode>3, 128-bit types): Likewise.
8190 (nor<mode>3, 128-bit types): Likewise.
8191 (andc<mode>3, 128-bit types): Likewise.
8192 (eqv<mode>3, 128-bit types): Likewise.
8193 (nand<mode>3, 128-bit types): Likewise.
8194 (orc<mode>3, 128-bit types): Likewise.
8195 (and<mode>3_internal): Likewise.
8196 (bool<mode>3_internal): Likewise.
8197 (boolc<mode>3_internal1): Likewise.
8198 (boolc<mode>3_internal2): Likewise.
8199 (boolcc<mode>3_internal1): Likewise.
8200 (boolcc<mode>3_internal2): Likewise.
8201 (eqv<mode>3_internal1): Likewise.
8202 (eqv<mode>3_internal2): Likewise.
8203 (one_cmpl1<mode>3_internal): Likewise.
8205 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
8207 * config/microblaze/microblaze.c (microblaze_expand_prologue):
8208 Rename flag_stack_usage to flag_stack_usage_info.
8210 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
8212 * config/microblaze/sync.md: New file.
8213 * config/microblaze/microblaze.md: Include sync.md
8214 * config/microblaze/microblaze.c: Add print_operand 'y'.
8215 * config/microblaze/constraints.md: Add memory_contraint
8216 'Q' which is a single register.
8218 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
8220 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
8222 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
8224 * reload.c (find_reloads): Exit loop once we find this operand
8225 cannot be reloaded somehow for this alternative.
8227 * reload.c (find_reloads): Exit loop once we find a hard register.
8229 * rtlanal.c (computed_jump_p): Exit loop once we find label
8232 * i386.c (ix86_pad_returns): Exit loop after setting replace.
8234 * cfgloopmanip.c (remove_path): Exit loop after setting
8237 * gensupport.c (subst_dup): Avoid loop if code is not
8238 MATCH_DUP nor MATCH_OP_DUP.
8240 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
8242 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
8244 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
8246 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
8247 true for SP_REGNUM if mode == ptr_mode.
8248 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
8249 with value R0_REGNUM + 31.
8251 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
8253 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
8254 pad pointer-typed argument downward.
8256 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
8258 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
8259 and __ILP32__ when the ILP32 model is in use.
8261 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
8263 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
8264 (aarch64_load_symref_appropriately): In the case of
8265 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
8266 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
8267 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
8268 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
8269 if 'mode' doesn't equal to 'ptr_mode'.
8270 (aarch64_output_mi_thunk): Add an assertion on the alignment of
8271 'vcall_offset'; change to call aarch64_emit_move differently depending
8272 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
8273 to calculate the upper bound of 'vcall_offset'.
8274 (aarch64_cannot_force_const_mem): Change to also return true if
8276 (aarch64_legitimize_reload_address): In the case of large
8277 displacements, add new local variable 'xmode' and an assertion
8278 based on it; change to use 'xmode' to generate the new rtx and
8280 (aarch64_asm_trampoline_template): Change to generate the template
8281 differently depending on TARGET_ILP32 or not; change to use
8282 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
8283 (aarch64_trampoline_size): Removed.
8284 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
8285 and replace immediate literals with it. Change to use 'ptr_mode'
8286 instead of 'DImode' and call convert_memory_address if the mode
8287 of 'fnaddr' doesn't equal to 'ptr_mode'.
8288 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
8290 (aarch64_elf_asm_destructor): Likewise.
8291 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
8292 on TARGET_ILP32 instead of aarch64_trampoline_size.
8293 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
8294 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
8295 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
8296 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
8297 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
8298 (storewb_pair<GPI:mode>_<P:mode>): ... this.
8299 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
8300 depending on the value of 'mode'.
8301 (add_losym_<mode>): New.
8302 (ldr_got_small_<mode>): New, based on ldr_got_small.
8303 (ldr_got_small): Remove.
8304 (ldr_got_small_sidi): New.
8305 * config/aarch64/iterators.md (P): New.
8306 (PTR): Change to 'ptr_mode' in the condition.
8308 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
8310 * config.gcc (aarch64*-*-*): Support --with-abi.
8311 (aarch64*-*-elf): Support --with-multilib-list.
8312 (aarch64*-*-linux*): Likewise.
8313 (supported_defaults): Add abi to aarch64*-*-*.
8314 * configure.ac: Mention AArch64 for --with-multilib-list.
8315 * configure: Re-generated.
8316 * config/aarch64/biarchilp32.h: New file.
8317 * config/aarch64/biarchlp64.h: New file.
8318 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
8320 (MULTILIB_DEFAULTS): Ditto.
8321 (DRIVER_SELF_SPECS): Ditto.
8322 (ASM_SPEC): Update to also substitute -mabi.
8323 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
8324 file whose name depends on -mabi= and -mbig-endian.
8325 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
8327 (POINTER_SIZE): New define.
8328 (POINTERS_EXTEND_UNSIGNED): Ditto.
8329 (enum aarch64_abi_type): New enumeration tag.
8330 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
8331 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
8332 (TARGET_ILP32): New define.
8333 * config/aarch64/aarch64.opt (mabi): New.
8335 (ilp32, lp64): New values for -mabi.
8336 * config/aarch64/t-aarch64 (comma): New define.
8337 (MULTILIB_OPTIONS): Ditto.
8338 (MULTILIB_DIRNAMES): Ditto.
8339 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
8340 * doc/invoke.texi: Document -mabi for AArch64.
8342 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
8344 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
8346 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
8347 Anton Blanchard <anton@au1.ibm.com>
8349 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
8350 endianness when selecting field to splat.
8352 2013-07-22 Eric Christopher <echristo@gmail.com>
8354 * dwarf2out.c (die_odr_checksum): New function to use
8355 CHECKSUM_ macros and ULEB128 for DIE tag.
8356 (generate_type_signature): Use.
8358 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
8360 * config.gcc (sparc*-*-*): Accept leon3 processor.
8361 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
8362 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
8363 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
8364 * config/sparc/sparc.opt (enum processor_type): Add leon3.
8365 (mfix-ut699): Adjust comment.
8366 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
8367 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
8368 (CPP_CPU_SPEC): Likewise.
8369 (ASM_CPU_SPEC): Likewise.
8370 * config/sparc/sparc.c (leon3_cost): New constant.
8371 (sparc_option_override): Add leon3 support.
8372 (mem_ref): New function.
8373 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
8374 (sparc_do_work_around_errata): Look into the instruction in the delay
8375 slot and adjust accordingly. Add fix for the data cache nullify issues
8376 of the UT699. Change insertion position for the NOP.
8377 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
8378 (leon3_load): New reservation.
8379 (leon_store): Bump latency to 2.
8380 (grfpu): New automaton.
8381 (grfpu_alu): New unit.
8382 (grfpu_ds): Likewise.
8383 (leon_fp_alu): Adjust.
8384 (leon_fp_mult): Delete.
8385 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
8386 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
8387 * config/sparc/sparc.md (cpu): Add leon3.
8388 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
8390 (atomic_test_and_set): Likewise.
8393 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
8395 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
8396 default for R5900 targets.
8397 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
8398 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
8399 * config/mips/mips.c (mips_option_override): Report an error for
8400 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
8401 for -march=r5900 -mhard-float.
8403 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
8405 * df-problems.c (can_move_insns_across): Exit loop once we
8406 find a non-fixed, non-global register.
8408 * ipa-pure-const.c (propagate_nothrow): Exit loop after
8411 * omega.c (omega_eliminate_red): Break after setting red_found.
8412 (omega_problem_has_red_equations): Similarly after setting found.
8413 (omega_query_variable): Similarly after setting coupled.
8415 2013-07-22 Marek Polacek <polacek@redhat.com>
8417 * gimplify.c: Don't include gimple.h twice.
8419 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8421 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
8422 instead of TARGET_THUMB1.
8423 (Pz): New constraint.
8424 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
8426 (compare_negsi_si): Likewise.
8427 (compare_addsi2_op0): Likewise.
8428 (compare_addsi2_op1): Likewise.
8429 (addsi3_carryin_<optab>): Likewise.
8430 (addsi3_carryin_alt2_<optab>): Likewise.
8431 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
8432 for arm_restrict_it.
8433 (subsi3_carryin): Likewise.
8434 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
8435 (minmax_arithsi): Disable for arm_restrict_it.
8436 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
8437 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
8438 (satsi_<SAT:code>_shift): Likewise.
8439 (arm_shiftsi3): Add alternative for 16-bit encoding.
8440 (arm32_movhf): Disable for arm_restrict_it.
8441 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
8442 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
8444 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
8446 * config/arm/arm.md (attribute "insn"): Delete.
8447 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
8448 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
8449 (not_shiftsi): Update for attribute change.
8450 (not_shiftsi_compare0): Likewise.
8451 (not_shiftsi_compare0_scratch): Likewise.
8452 (arm_one_cmplsi2): Likewise.
8453 (thumb1_one_cmplsi2): Likewise.
8454 (notsi_compare0): Likewise.
8455 (notsi_compare0_scratch): Likewise.
8456 (thumb1_movdi_insn): Likewise.
8457 (arm_movsi_insn): Likewise.
8458 (movhi_insn_arch4): Likewise.
8459 (movhi_bytes): Likewise.
8460 (arm_movqi_insn): Likewise.
8461 (thumb1_movqi_insn): Likewise.
8462 (arm32_movhf): Likewise.
8463 (thumb1_movhf): Likewise.
8464 (arm_movsf_soft_insn): Likewise.
8465 (thumb1_movsf_insn): Likewise.
8466 (thumb_movdf_insn): Likewise.
8467 (movsicc_insn): Likewise.
8468 (movsfcc_soft_insn): Likewise.
8469 (and_scc): Likewise.
8470 (cond_move): Likewise.
8471 (if_move_not): Likewise.
8472 (if_not_move): Likewise.
8473 (if_shift_move): Likewise.
8474 (if_move_shift): Likewise.
8475 (if_shift_shift): Likewise.
8476 (if_not_arith): Likewise.
8477 (if_arith_not): Likewise.
8478 (cond_move_not): Likewise.
8479 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
8480 (neon_mov<mode>): Likewise.
8481 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
8482 (thumb2_movsi_vfp): Likewise.
8483 (movsf_vfp): Likewise.
8484 (thumb2_movsf_vfp): Likewise.
8485 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
8487 (cortexa7_older_only): Likewise.
8488 (cortexa7_younger): Likewise.
8489 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
8490 (1020alu_shift_op): Likewise.
8491 (1020alu_shift_reg_op): Likewise.
8492 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
8493 (alu_shift_op): Likewise.
8494 (alu_shift_reg_op): Likewise.
8495 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
8496 (11_alu_shift_op): Likewise.
8497 (11_alu_shift_reg_op): Likewise.
8498 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
8499 (9_alu_shift_reg_op): Likewise.
8500 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
8502 (cortex_a15_alu_shift): Likewise.
8503 (cortex_a15_alu_shift_reg): Likewise.
8504 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
8505 (cortex_a5_alu_shift): Likewise.
8506 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
8508 (cortex_a53_alu_shift): Likewise.
8509 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
8511 (cortex_a7_alu_reg): Likewise.
8512 (cortex_a7_alu_shift): Likewise.
8513 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
8514 (cortex_a8_alu_shift): Likewise.
8515 (cortex_a8_alu_shift_reg): Likewise.
8516 (cortex_a8_mov): Likewise.
8517 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
8518 (cortex_a9_dp_shift): Likewise.
8519 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
8520 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
8521 (cortex_r4_mov): Likewise.
8522 (cortex_r4_alu_shift): Likewise.
8523 (cortex_r4_alu_shift_reg): Likewise.
8524 * config/arm/fa526.md (526_alu_op): Update for attribute change.
8525 (526_alu_shift_op): Likewise.
8526 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
8527 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
8528 (626te_alu_shift_op): Likewise.
8529 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
8530 (726te_alu_op): Likewise.
8531 (726te_alu_shift_op): Likewise.
8532 (726te_alu_shift_reg_op): Likewise.
8533 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
8534 (mp626_alu_shift_op): Likewise.
8535 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
8536 (pj4_alu_e1_conds): Likewise.
8537 (pj4_alu): Likewise.
8538 (pj4_alu_conds): Likewise.
8539 (pj4_shift): Likewise.
8540 (pj4_shift_conds): Likewise.
8541 (pj4_alu_shift): Likewise.
8542 (pj4_alu_shift_conds): Likewise.
8544 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8546 * config/arm/predicates.md (shiftable_operator_strict_it):
8548 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
8549 Disable cond_exec version for arm_restrict_it.
8550 (thumb2_smaxsi3): Convert to generate cond_exec.
8551 (thumb2_sminsi3): Likewise.
8552 (thumb32_umaxsi3): Likewise.
8553 (thumb2_uminsi3): Likewise.
8554 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
8555 (thumb2_neg_abssi2): Likewise.
8556 (thumb2_mov_scc): Add alternative for 16-bit encoding.
8557 (thumb2_movsicc_insn): Adjust alternatives.
8558 (thumb2_mov_negscc): Disable for arm_restrict_it.
8559 (thumb2_mov_negscc_strict_it): New pattern.
8560 (thumb2_mov_notscc_strict_it): New pattern.
8561 (thumb2_mov_notscc): Disable for arm_restrict_it.
8562 (thumb2_ior_scc): Likewise.
8563 (thumb2_ior_scc_strict_it): New pattern.
8564 (thumb2_cond_move): Adjust for arm_restrict_it.
8565 (thumb2_cond_arith): Disable for arm_restrict_it.
8566 (thumb2_cond_arith_strict_it): New pattern.
8567 (thumb2_cond_sub): Adjust for arm_restrict_it.
8568 (thumb2_movcond): Likewise.
8569 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
8570 (thumb2_zero_extendhisi2_v6): Likewise.
8571 (thumb2_zero_extendqisi2_v6): Likewise.
8572 (orsi_notsi_si): Likewise.
8573 (orsi_not_shiftsi_si): Likewise.
8575 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
8577 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
8578 instruction sequence is 1 byte shorter.
8580 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
8582 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
8583 it is not needed after split.
8585 2013-07-20 Iain Sandoe <iain@codesourcery.com>
8588 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
8589 second label for nonlocal goto receivers. Don't output pic base labels
8590 unless we're producing PIC; mark that action unreachable().
8591 (ix86_save_reg): If the function contains a nonlocal label, save the
8593 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
8594 * config/darwin.c (emitted_pic_label_num): New GTY.
8595 (update_pic_label_number_if_needed): New.
8596 (machopic_output_function_base_name): Adjust for nonlocal receiver
8598 (machopic_should_output_picbase_label): New.
8599 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
8600 (nonlocal_goto_receiver): New insn and split.
8602 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
8604 * config/aarch64/aarch64-builtins.c
8605 (aarch64_fold_builtin): Fold abs in all modes.
8606 * config/aarch64/aarch64-simd-builtins.def
8607 (abs): Enable for all modes.
8608 * config/aarch64/arm_neon.h
8609 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
8610 (vabs_f64): Add missing intrinsic.
8612 2013-07-19 Ian Bolton <ian.bolton@arm.com>
8614 * config/aarch64/arm_neon.h (vabs_s64): New function
8616 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
8619 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
8620 * config/avr/avr.md (adjust_len): Add `round'.
8621 * config/avr/avr-protos.h (avr_out_round): New prototype.
8622 (avr_out_plus): Add `out_label' argument.
8623 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
8624 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
8625 Handle the case where `insn' is just a pattern.
8626 (avr_out_bitop): Handle the case where `insn' is just a pattern.
8627 (avr_out_round): New function.
8628 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
8630 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
8632 * config/microblaze/microblaze.c (microblaze_expand_prologue):
8633 Add check for flag_stack_usage to handle -fstack-usage support
8635 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
8637 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
8638 interaction for new Power8 flags and VSX.
8640 2013-07-18 Sriraman Tallam <tmsriram@google.com>
8643 * tree-inline.c (expand_call_inline): Emit errors during
8644 early_inlining only if optimization is not turned on.
8646 2013-07-18 David Malcolm <dmalcolm@redhat.com>
8650 * passes.c (init_optimization_passes): Move the construction of
8651 the pass hierarchy into a new passes.def file.
8653 * Makefile.in (passes.o): Add dependency on passes.def.
8655 2013-07-18 David Malcolm <dmalcolm@redhat.com>
8657 * passes.c (init_optimization_passes): Introduce macros for
8658 constructing the tree of passes (INSERT_PASSES_AFTER,
8659 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
8660 TERMINATE_PASS_LIST).
8662 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
8663 Wei Mi <wmi@google.com>
8665 PR rtl-optimization/57878
8666 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
8668 (reload_pseudo_compare_func): Check nregs first for reload
8671 2013-07-18 David Malcolm <dmalcolm@redhat.com>
8673 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
8675 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
8677 * read-rtl.c (validate_const_int): Once an invalid character is
8678 seen, quit the loop.
8680 * gengtype.c (write_roots): Similarly once we find the "deletable"
8681 or "if_marked" option.
8683 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
8685 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
8686 "xtab" and "sat". Move value "clz" from here to ...
8687 (attriubte "type"): ... here.
8688 (satsi_<SAT:code>): Delete "insn" attribute.
8689 (satsi_<SAT:code>_shift): Likewise.
8690 (arm_zero_extendqisi2addsi): Likewise.
8691 (arm_extendqisi2addsi): Likewise.
8692 (clzsi2): Update for attribute changes.
8693 (rbitsi2): Likewise.
8694 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
8696 (arm_usatsihi): Likewise.
8697 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
8699 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
8701 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
8702 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
8703 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
8704 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
8705 in alphabetical order.
8706 (attribute "core_cycles"): Update for attribute changes.
8707 (arm_addsi3): Likewise.
8708 (addsi3_compare0): Likewise.
8709 (addsi3_compare0_scratch): Likewise.
8710 (addsi3_compare_op1): Likewise.
8711 (addsi3_compare_op2): Likewise.
8712 (compare_addsi2_op0): Likewise.
8713 (compare_addsi2_op1): Likewise.
8714 (addsi3_carryin_shift_<optab>): Likewise.
8715 (subsi3_carryin_shift): Likewise.
8716 (rsbsi3_carryin_shift): Likewise.
8717 (arm_subsi3_insn): Likewise.
8718 (subsi3_compare0): Likewise.
8719 (subsi3_compare): Likewise.
8720 (arm_andsi3_insn): Likewise.
8721 (thumb1_andsi3_insn): Likewise.
8722 (andsi3_compare0): Likewise.
8723 (andsi3_compare0_scratch): Likewise.
8724 (zeroextractsi_compare0_scratch
8725 (andsi_not_shiftsi_si): Likewise.
8726 (iorsi3_insn): Likewise.
8727 (iorsi3_compare0): Likewise.
8728 (iorsi3_compare0_scratch): Likewise.
8729 (arm_xorsi3): Likewise.
8730 (thumb1_xorsi3_insn): Likewise.
8731 (xorsi3_compare0): Likewise.
8732 (xorsi3_compare0_scratch): Likewise.
8733 (satsi_<SAT:code>_shift): Likewise.
8735 (arm_shiftsi3): Likewise.
8736 (shiftsi3_compare0): Likewise.
8737 (not_shiftsi): Likewise.
8738 (not_shiftsi_compare0): Likewise.
8739 (not_shiftsi_compare0_scratch): Likewise.
8740 (arm_one_cmplsi2): Likewise.
8741 (thumb_one_complsi2): Likewise.
8742 (notsi_compare0): Likewise.
8743 (notsi_compare0_scratch): Likewise.
8744 (thumb1_zero_extendhisi2): Likewise.
8745 (arm_zero_extendhisi2): Likewise.
8746 (arm_zero_extendhisi2_v6): Likewise.
8747 (arm_zero_extendhisi2addsi): Likewise.
8748 (thumb1_zero_extendqisi2): Likewise.
8749 (thumb1_zero_extendqisi2_v6): Likewise.
8750 (arm_zero_extendqisi2): Likewise.
8751 (arm_zero_extendqisi2_v6): Likewise.
8752 (arm_zero_extendqisi2addsi): Likewise.
8753 (thumb1_extendhisi2): Likewise.
8754 (arm_extendhisi2): Likewise.
8755 (arm_extendhisi2_v6): Likewise.
8756 (arm_extendqisi): Likewise.
8757 (arm_extendqisi_v6): Likewise.
8758 (arm_extendqisi2addsi): Likewise.
8759 (thumb1_extendqisi2): Likewise.
8760 (thumb1_movdi_insn): Likewise.
8761 (arm_movsi_insn): Likewise.
8762 (movsi_compare0): Likewise.
8763 (movhi_insn_arch4): Likewise.
8764 (movhi_bytes): Likewise.
8765 (arm_movqi_insn): Likewise.
8766 (thumb1_movqi_insn): Likewise.
8767 (arm32_movhf): Likewise.
8768 (thumb1_movhf): Likewise.
8769 (arm_movsf_soft_insn): Likewise.
8770 (thumb1_movsf_insn): Likewise.
8771 (movdf_soft_insn): Likewise.
8772 (thumb_movdf_insn): Likewise.
8773 (arm_cmpsi_insn): Likewise.
8774 (cmpsi_shiftsi): Likewise.
8775 (cmpsi_shiftsi_swp): Likewise.
8776 (arm_cmpsi_negshiftsi_si): Likewise.
8777 (movsicc_insn): Likewise.
8778 (movsfcc_soft_insn): Likewise.
8779 (arith_shiftsi): Likewise.
8780 (arith_shiftsi_compare0
8781 (arith_shiftsi_compare0_scratch
8782 (sub_shiftsi): Likewise.
8783 (sub_shiftsi_compare0
8784 (sub_shiftsi_compare0_scratch
8785 (and_scc): Likewise.
8786 (cond_move): Likewise.
8787 (if_plus_move): Likewise.
8788 (if_move_plus): Likewise.
8789 (if_move_not): Likewise.
8790 (if_not_move): Likewise.
8791 (if_shift_move): Likewise.
8792 (if_move_shift): Likewise.
8793 (if_shift_shift): Likewise.
8794 (if_not_arith): Likewise.
8795 (if_arith_not): Likewise.
8796 (cond_move_not): Likewise.
8797 (thumb1_ashlsi3): Set type attribute.
8798 (thumb1_ashrsi3): Likewise.
8799 (thumb1_lshrsi3): Likewise.
8800 (thumb1_rotrsi3): Likewise.
8801 (shiftsi3_compare0_scratch): Likewise.
8802 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
8803 (neon_mov<mode>): Likewise.
8804 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
8806 (thumb2_movsi_insn): Likewise.
8807 (thumb2_cmpsi_neg_shiftsi): Likewise.
8808 (thumb2_extendqisi_v6): Likewise.
8809 (thumb2_zero_extendhisi2_v6): Likewise.
8810 (thumb2_zero_extendqisi2_v6): Likewise.
8811 (thumb2_shiftsi3_short): Likewise.
8812 (thumb2_addsi3_compare0_scratch): Likewise.
8813 (orsi_not_shiftsi_si): Likewise.
8814 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
8815 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
8817 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
8818 (1020alu_shift_op): Likewise.
8819 (1020alu_shift_reg_op): Likewise.
8820 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
8821 (alu_shift_op): Likewise.
8822 (alu_shift_reg_op): Likewise.
8823 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
8824 (11_alu_shift_op): Likewise.
8825 (11_alu_shift_reg_op): Likewise.
8826 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
8827 (9_alu_shift_reg_op): Likewise.
8828 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
8830 (cortex_a15_alu_shift): Likewise.
8831 (cortex_a15_alu_shift_reg): Likewise.
8832 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
8834 (cortex_a5_alu_shift): Likewise.
8835 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
8837 (cortex_a53_alu_shift): Likewise.
8838 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
8840 (cortex_a7_alu_reg): Likewise.
8841 (cortex_a7_alu_shift): Likewise.
8842 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
8844 (cortex_a8_alu_shift): Likewise.
8845 (cortex_a8_alu_shift_reg): Likewise.
8846 (cortex_a8_mov): Likewise.
8847 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
8848 (cortex_a9_dp_shift): Likewise.
8849 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
8851 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
8853 (cortex_r4_mov): Likewise.
8854 (cortex_r4_alu_shift): Likewise.
8855 (cortex_r4_alu_shift_reg): Likewise.
8856 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
8857 (526_alu_shift_op): Likewise.
8858 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
8859 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
8860 (626te_alu_shift_op): Likewise.
8861 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
8862 (726te_alu_op): Likewise.
8863 (726te_alu_shift_op): Likewise.
8864 (726te_alu_shift_reg_op): Likewise.
8865 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
8866 (mp626_alu_shift_op): Likewise.
8867 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
8868 (pj4_alu_e1_conds): Likewise.
8869 (pj4_alu): Likewise.
8870 (pj4_alu_conds): Likewise.
8871 (pj4_shift): Likewise.
8872 (pj4_shift_conds): Likewise.
8873 (pj4_alu_shift): Likewise.
8874 (pj4_alu_shift_conds): Likewise.
8875 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
8877 (cortexa7_older_only): Likewise.
8878 (cortexa7_younger): Likewise.
8880 2013-07-18 David Malcolm <dmalcolm@redhat.com>
8882 * ipa-pure-const.c (generate_summary): Rename to...
8883 (pure_const_generate_summary): ... this.
8885 2013-07-17 Iain Sandoe <iain@codesourcery.com>
8887 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
8889 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
8892 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
8895 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8897 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
8898 enabled without -march=zEC12.
8899 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
8902 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
8904 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
8905 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
8906 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
8907 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
8908 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
8909 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
8910 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
8912 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
8913 and ISA_HAS_NMADD3_NMSUB3.
8914 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
8915 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
8916 (nmsub4<mode>, nmsub3<mode>): Likewise.
8917 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
8919 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
8921 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
8922 TARGET_MIPS5400 checking.
8924 2013-07-16 Jakub Jelinek <jakub@redhat.com>
8925 Peter Bergner <bergner@vnet.ibm.com>
8927 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
8928 registers in the comment.
8929 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
8930 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
8931 rather than FIRST_PSEUDO_REGISTERS.
8933 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
8935 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
8936 enable extra ISA flags with TARGET_HTM.
8938 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
8940 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
8943 2013-07-15 Cong Hou <congh@google.com>
8945 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
8946 in compare function for sorting.
8948 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
8950 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
8951 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
8952 * config/rs6000/rs6000.opt: Add -mhtm option.
8953 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
8954 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
8955 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
8956 __HTM__ if the HTM instructions are available.
8957 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
8958 htm_spr_reg_operand): New define_predicates.
8959 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
8960 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
8962 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
8963 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
8964 HTM builtin functions.
8965 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
8966 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
8967 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
8968 (rs6000_builtin_mask_calculate): Likewise.
8969 (rs6000_option_override_internal): Likewise.
8970 (bdesc_htm): Add new HTM builtin support.
8971 (htm_spr_num): New function.
8972 (htm_spr_regno): Likewise.
8973 (rs6000_htm_spr_icode): Likewise.
8974 (htm_expand_builtin): Likewise.
8975 (htm_init_builtins): Likewise.
8976 (rs6000_expand_builtin): Add support for HTM builtin functions.
8977 (rs6000_init_builtins): Likewise.
8978 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
8980 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
8981 (TARGET_HTM, MASK_HTM): Define macros.
8982 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
8983 (FIXED_REGISTERS): Likewise.
8984 (CALL_USED_REGISTERS): Likewise.
8985 (CALL_REALLY_USED_REGISTERS): Likewise.
8986 (REG_ALLOC_ORDER): Likewise.
8987 (enum reg_class): Likewise.
8988 (REG_CLASS_NAMES): Likewise.
8989 (REG_CLASS_CONTENTS): Likewise.
8990 (REGISTER_NAMES): Likewise.
8991 (ADDITIONAL_REGISTER_NAMES): Likewise.
8992 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
8993 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
8994 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
8995 * config/rs6000/htm.md: New file.
8996 * config/rs6000/htmintrin.h: New file.
8997 * config/rs6000/htmxlintrin.h: New file.
8999 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
9001 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
9002 Define SYMBOL_TINY_GOT, update comment.
9003 * config/aarch64/aarch64.c
9004 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
9005 (aarch64_expand_mov_immediate): Likewise.
9006 (aarch64_print_operand): Likewise.
9007 (aarch64_classify_symbol): Likewise.
9008 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
9009 (ldr_got_tiny): Define.
9011 2013-07-13 Tobias Grosser <tobias@grosser.es>
9013 PR tree-optimization/54094
9014 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
9015 scheduling dimension for the parallelism check from the polyhedral
9016 information in the AST.
9017 * graphite-dependences.c (carries_deps): Do not assume the schedule is
9020 2013-07-13 Jason Merrill <jason@redhat.com>
9022 * print-tree.c (debug_vec_tree): Use debug_raw.
9023 (debug_raw (vec<tree, va_gc> &)): New.
9024 (debug_raw (vec<tree, va_gc> *)): New.
9025 * tree.h: Declare them.
9027 2013-07-13 Bin Cheng <bin.cheng@arm.com>
9029 * ifcvt.c (ifcvt_after_combine): New static variable.
9030 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
9032 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
9033 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
9034 rest_of_handle_if_after_reload): Pass new argument for if_convert.
9036 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
9038 * config/mips/mips.c (mips_expand_call): Remove empty statement.
9040 2013-07-12 Michael Matz <matz@suse.de>
9043 * convert.c (convert_to_real): Reject non-float inner types.
9045 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
9047 * config/aarch64/aarch64-protos.h
9048 (aarch64_simd_immediate_valid_for_move): Remove.
9049 * config/aarch64/aarch64.c (simd_immediate_info): New member.
9050 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
9052 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
9054 2013-07-11 Steve Ellcey <sellcey@mips.com>
9056 * config/mips/mips.c (mips_conditional_register_usage): Do not
9057 use t[0-7] registers in MIPS16 mode when optimizing for size.
9059 2013-07-11 Sriraman Tallam <tmsriram@google.com>
9061 * config/i386/i386.c (dispatch_function_versions): Fix array
9062 indexing of function_version_info to match actual_versions.
9064 2013-07-11 Teresa Johnson <tejohnson@google.com>
9066 * vec.h (struct va_gc): Move release out-of-line.
9067 (va_gc::release): Call ggc_free on released vec.
9069 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
9071 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
9072 Require GOT register as additional operand in UNSPEC.
9073 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
9074 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
9075 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
9076 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
9077 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
9078 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
9079 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
9081 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
9084 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
9085 name seen by assembler/linker rather if available.
9087 2013-07-11 Andreas Schwab <schwab@suse.de>
9089 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
9091 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
9093 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
9095 2013-07-10 Joseph Myers <joseph@codesourcery.com>
9097 * doc/tm.texi.in: Move hook documentation to ....
9098 * target.def: ... here.
9100 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
9102 * doc/tm.texi: Regenerate.
9104 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
9107 * doc/invoke.texi: Document -Wconditionally-supported.
9109 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
9112 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
9115 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
9118 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
9119 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
9120 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
9121 Remove duplicate devices.
9122 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
9123 * config/avr/t-multilib: Regenerate.
9124 * config/avr/avr-tables.opt: Regenerate.
9125 * doc/avr-mmcu.texi: Regenerate.
9127 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
9130 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
9132 2013-07-10 Graham Stott <graham.stott@btinternet.com>
9134 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
9135 the cost of MULT when optimizing for size.
9137 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
9139 * config/cr16/cr16-protos.h: Don't include target.h.
9141 2013-07-09 Joseph Myers <joseph@codesourcery.com>
9143 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
9144 adjust register size for TDmode and TFmode for VSX registers.
9146 2013-07-08 Kai Tietz <ktietz@redhat.com>
9149 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
9150 hook_bool_const_tree_true.
9152 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9154 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
9155 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
9156 * config/s390/s390.md: Define FPR*_REGNUM constants.
9157 Fix FPR2_REGNUM constant (18 -> 17).
9158 ("*trunc<BFP:mode><DFP_ALL:mode>2")
9159 ("*trunc<DFP_ALL:mode><BFP:mode>2")
9160 ("trunc<BFP:mode><DFP_ALL:mode>2")
9161 ("trunc<DFP_ALL:mode><BFP:mode>2")
9162 ("*extend<BFP:mode><DFP_ALL:mode>2")
9163 ("*extend<DFP_ALL:mode><BFP:mode>2")
9164 ("extend<BFP:mode><DFP_ALL:mode>2")
9165 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
9168 2013-07-08 Graham Stott <graham.stott@btinternet.com>
9170 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
9172 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9174 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
9175 and cfun_fpr_bit_p to cfun_fpr_save_p.
9176 (s390_frame_area, s390_register_info, s390_frame_info)
9177 (s390_emit_prologue, s390_emit_epilogue)
9178 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
9180 * config/s390/s390.h: Define *_REGNUM macros for floating point
9183 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
9185 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
9187 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
9189 PR rtl-optimization/57786
9190 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
9191 and break out of the loop when it is set to false.
9193 2013-07-08 Jakub Jelinek <jakub@redhat.com>
9196 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
9197 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
9198 (const_int 63)) 0)).
9199 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
9200 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
9201 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
9203 PR rtl-optimization/57829
9204 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
9205 mask bits outside of mode are just sign-extension from mode to HWI.
9207 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
9209 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
9210 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
9211 adjust_address instead of change_address to keep info about alignment.
9212 (emit_strmov): Remove.
9213 (emit_memmov): New function.
9214 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
9215 (expand_movmem_epilogue): Likewise and return updated rtx for
9217 (expand_constant_movmem_prologue): Likewise and return updated rtx for
9218 destination and source.
9219 (decide_alignment): Refactor, handle vector_loop.
9220 (ix86_expand_movmem): Likewise.
9221 (ix86_expand_setmem): Likewise.
9222 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
9224 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
9226 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
9227 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
9229 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
9231 * config/i386/sse.md (sse_movlhps): Change alternative 3
9232 of operand 2 to "m".
9234 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
9237 * config/i386/sse.md (iptr): New mode attribute.
9238 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
9239 (<sse>_vm<plusminus_insn><mode>3): Ditto.
9240 (<sse>_vmmul<mode>3): Ditto.
9241 (<sse>_vmdiv<mode>3): Ditto.
9242 (sse_vmrcpv4sf2): Ditto.
9243 (<sse>_vmsqrt<mode>2): Ditto.
9244 (sse_vmrsqrtv4sf2): Ditto.
9245 (<sse>_vm<code><mode>3): Ditto.
9246 (avx_vmcmp<mode>3): Ditto.
9247 (<sse>_vmmaskcmp<mode>3): Ditto.
9248 (<sse>_comi): Ditto.
9249 (<sse>_ucomi): Ditto.
9250 (*xop_vmfrcz_<mode>): Ditto.
9251 (*fmai_fmadd_<mode>): Ditto.
9252 (*fmai_fmsub_<mode>): Ditto.
9253 (*fmai_fnmadd_<mode>): Ditto.
9254 (*fmai_fnmsub_<mode>): Ditto.
9255 (*fma4i_vmfmadd_<mode>): Ditto.
9256 (*fma4i_vmfmsub_<mode>): Ditto.
9257 (*fma4i_vmfnmadd_<mode>): Ditto.
9258 (*fma4i_vmfnmsub_<mode>): Ditto.
9259 (*xop_vmfrcz_<mode>): Ditto.
9260 (sse_cvtps2pi): Ditto.
9261 (sse_cvttps2pi): Ditto.
9262 (sse_cvtss2si): Ditto.
9263 (sse_cvtss2si_2): Ditto.
9264 (sse_cvtss2siq_2): Ditto.
9265 (sse_cvttss2si): Ditto.
9266 (sse_cvttss2siq): Ditto.
9267 (sse_cvtsd2si): Ditto.
9268 (sse_cvtsd2si_2): Ditto.
9269 (sse_cvtsd2siq_2): Ditto.
9270 (sse_cvttsd2si): Ditto.
9271 (sse_cvttsd2siq): Ditto.
9272 (sse_cvtsd2ss): Ditto.
9273 (sse_cvtss2sd): Ditto.
9274 (avx2_pbroadcast<mode>): Ditto.
9275 (avx2_pbroadcast<mode>_1): Ditto.
9276 (*avx_vperm_broadcast_v4sf): Ditto.
9278 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
9279 (sse_movlhps): Ditto.
9280 (sse_storehps): Ditto.
9281 (sse_loadhps): Ditto.
9282 (sse_storelps): Ditto.
9283 (sse_loadlps): Ditto.
9284 (*vec_concatv4sf): Ditto.
9285 (*vec_interleave_highv2df): Ditto.
9286 (*vec_interleave_lowv2df): Ditto.
9287 (*vec_extractv2df_1_sse): Ditto.
9288 (*vec_extractv2df_0_sse): Ditto.
9289 (sse2_storelpd): Ditto.
9290 (sse2_loadlpd): Ditto.
9291 (sse2_movsd): Ditto.
9292 (*vec_concatv4si): Ditto.
9293 (vec_concatv2di): Ditto.
9295 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
9296 for Intel asm dialect.
9297 (mmx_punpcklwd): Ditto.
9298 (mmx_punpckldq): Ditto.
9300 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
9301 for intel assembler dialect.
9303 2013-07-06 Jakub Jelinek <jakub@redhat.com>
9306 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
9307 for BUILT_IN_C{LZ,LRSB}*.
9308 * tree.h (CASE_INT_FN): Add FN##IMAX case.
9309 * tree-vrp.c (extract_range_basic): Handle
9310 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
9311 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
9312 fall thru to code calling set_value*.
9313 * builtins.c (expand_builtin): Remove *IMAX cases.
9314 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
9315 if width is bigger than 2*HWI.
9317 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
9319 PR rtl-optimization/55342
9320 * lra-int.h (lra_subreg_reload_pseudos): New.
9321 * lra.c: Add undoing optional reloads to the block diagram.
9322 (lra_subreg_reload_pseudos): New.
9323 (lra_optional_reload_pseudos): Change comments.
9324 (lra): Init and clear lra_subreg_reload_pseudos. Clear
9325 lra_optional_reload_pseudos after undo transformations.
9326 * lra-assigns.c (pseudo_prefix_title): New.
9327 (lra_setup_reg_renumber): Use it.
9328 (spill_for): Ditto. Check subreg reload pseudos too.
9329 (assign_by_spills): Consider subreg reload pseudos too.
9330 * lra-constraints.c (simplify_operand_subreg): Use
9331 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
9332 (curr_insn_transform): Recognize and do optional reloads.
9333 (undo_optional_reloads): New.
9334 (lra_undo_inheritance): Call undo_optional_reloads.
9336 2013-07-05 Thomas Quinot <quinot@adacore.com>
9338 * tree-complex.c (expand_complex_operations_1): Fix typo.
9340 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
9342 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
9343 (tune_params): New member 'const vec_costs'.
9344 * config/aarch64/aarch64.c (generic_vector_cost): New.
9345 (generic_tunings): New member 'generic_vector_cost'.
9346 (aarch64_builtin_vectorization_cost): New.
9347 (aarch64_add_stmt_cost): New.
9348 (TARGET_VECTORIZE_ADD_STMT_COST): New.
9349 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
9351 2013-07-03 Jakub Jelinek <jakub@redhat.com>
9354 * config/i386/predicates.md (vsib_address_operand): Disallow
9355 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
9357 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
9360 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
9361 expand_builtin_setjmp_receiver.
9362 (expand_label): Adjust, call expand_builtin_setjmp_receiver
9363 with NULL for the label parameter.
9364 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
9365 the frame-pointer. Adjust comments.
9366 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
9367 only if LABEL is non-NULL.
9369 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
9371 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
9372 (ARM_ABI_AAPCS64): Ditto.
9374 (ARM_DEFAULT_ABI): Ditto.
9376 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
9378 * config/aarch64/aarch64-builtins.c
9379 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
9380 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
9382 * config/aarch64/aarch64-simd.md
9383 (aarch64_ld1<VALL:mode>): New.
9384 (aarch64_st1<VALL:mode>): Likewise.
9385 * config/aarch64/arm_neon.h
9386 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
9388 2013-07-02 Sriraman Tallam <tmsriram@google.com>
9390 * config/i386/i386.c (gate_insert_vzeroupper): Check if
9392 (ix86_option_override_internal):Turn on all -mavx target flags by
9393 default as they are dependent on AVX anyway.
9395 2013-07-02 Cary Coutant <ccoutant@google.com>
9397 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
9399 (loc_checksum_ordered): Likewise.
9400 (hash_loc_operands): Remove inline keyword.
9402 2013-07-02 Jakub Jelinek <jakub@redhat.com>
9404 PR tree-optimization/57741
9405 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
9406 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
9407 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
9408 Allow REAL_CST step_exprs if flag_associative_math.
9409 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
9411 2013-07-02 Ian Bolton <ian.bolton@arm.com>
9413 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
9415 2013-07-02 Ian Bolton <ian.bolton@arm.com>
9417 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
9419 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9421 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
9423 (iorsi3_insn): Likewise.
9424 (arm_xorsi3): Likewise.
9426 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
9428 * arm.md (attribute "wtype"): Delete. Move attribute values from here
9430 (attribute "type"): ... here, and prefix with "wmmx_".
9431 (attribute "core_cycles"): Update for attribute changes.
9432 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
9433 (tbcstv4hi): Likewise.
9434 (tbcstv2si): Likewise.
9435 (iwmmxt_iordi3): Likewise.
9436 (iwmmxt_xordi3): Likewise.
9437 (iwmmxt_anddi3): Likewise.
9438 (iwmmxt_nanddi3): Likewise.
9439 (iwmmxt_arm_movdi): Likewise.
9440 (iwmmxt_movsi_insn): Likewise.
9441 (mov<mode>_internal): Likewise.
9442 (and<mode>3_iwmmxt): Likewise.
9443 (ior<mode>3_iwmmxt): Likewise.
9444 (xor<mode>3_iwmmxt): Likewise.
9445 (add<mode>3_iwmmxt): Likewise.
9446 (ssaddv8qi3): Likewise.
9447 (ssaddv4hi3): Likewise.
9448 (ssaddv2si3): Likewise.
9449 (usaddv8qi3): Likewise.
9450 (usaddv4hi3): Likewise.
9451 (usaddv2si3): Likewise.
9452 (sub<mode>3_iwmmxt): Likewise.
9453 (sssubv8qi3): Likewise.
9454 (sssubv4hi3): Likewise.
9455 (sssubv2si3): Likewise.
9456 (ussubv8qi3): Likewise.
9457 (ussubv4hi3): Likewise.
9458 (ussubv2si3): Likewise.
9459 (mulv4hi3_iwmmxt): Likewise.
9460 (smulv4hi3_highpart): Likewise.
9461 (umulv4hi3_highpart): Likewise.
9462 (iwmmxt_wmacs): Likewise.
9463 (iwmmxt_wmacsz): Likewise.
9464 (iwmmxt_wmacu): Likewise.
9465 (iwmmxt_wmacuz): Likewise.
9466 (iwmmxt_clrdi): Likewise.
9467 (iwmmxt_clrv8qi): Likewise.
9468 (iwmmxt_clr4hi): Likewise.
9469 (iwmmxt_clr2si): Likewise.
9470 (iwmmxt_uavgrndv8qi3): Likewise.
9471 (iwmmxt_uavgrndv4hi3): Likewise.
9472 (iwmmxt_uavgv8qi3): Likewise.
9473 (iwmmxt_uavgv4hi3): Likewise.
9474 (iwmmxt_tinsrb): Likewise.
9475 (iwmmxt_tinsrh): Likewise.
9476 (iwmmxt_tinsrw): Likewise.
9477 (iwmmxt_textrmub): Likewise.
9478 (iwmmxt_textrmsb): Likewise.
9479 (iwmmxt_textrmuh): Likewise.
9480 (iwmmxt_textrmsh): Likewise.
9481 (iwmmxt_textrmw): Likewise.
9482 (iwmxxt_wshufh): Likewise.
9483 (eqv8qi3): Likewise.
9484 (eqv4hi3): Likewise.
9485 (eqv2si3): Likewise.
9486 (gtuv8qi3): Likewise.
9487 (gtuv4hi3): Likewise.
9488 (gtuv2si3): Likewise.
9489 (gtv8qi3): Likewise.
9490 (gtv4hi3): Likewise.
9491 (gtv2si3): Likewise.
9492 (smax<mode>3_iwmmxt): Likewise.
9493 (umax<mode>3_iwmmxt): Likewise.
9494 (smin<mode>3_iwmmxt): Likewise.
9495 (umin<mode>3_iwmmxt): Likewise.
9496 (iwmmxt_wpackhss): Likewise.
9497 (iwmmxt_wpackwss): Likewise.
9498 (iwmmxt_wpackdss): Likewise.
9499 (iwmmxt_wpackhus): Likewise.
9500 (iwmmxt_wpackwus): Likewise.
9501 (iwmmxt_wpackdus): Likewise.
9502 (iwmmxt_wunpckihb): Likewise.
9503 (iwmmxt_wunpckihh): Likewise.
9504 (iwmmxt_wunpckihw): Likewise.
9505 (iwmmxt_wunpckilb): Likewise.
9506 (iwmmxt_wunpckilh): Likewise.
9507 (iwmmxt_wunpckilw): Likewise.
9508 (iwmmxt_wunpckehub): Likewise.
9509 (iwmmxt_wunpckehuh): Likewise.
9510 (iwmmxt_wunpckehuw): Likewise.
9511 (iwmmxt_wunpckehsb): Likewise.
9512 (iwmmxt_wunpckehsh): Likewise.
9513 (iwmmxt_wunpckehsw): Likewise.
9514 (iwmmxt_wunpckelub): Likewise.
9515 (iwmmxt_wunpckeluh): Likewise.
9516 (iwmmxt_wunpckeluw): Likewise.
9517 (iwmmxt_wunpckelsb): Likewise.
9518 (iwmmxt_wunpckelsh): Likewise.
9519 (iwmmxt_wunpckelsw): Likewise.
9520 (ror<mode>3): Likewise.
9521 (ashr<mode>3_iwmmxt): Likewise.
9522 (lshr<mode>3_iwmmxt): Likewise.
9523 (ashl<mode>3_iwmmxt): Likewise.
9524 (ror<mode>3_di): Likewise.
9525 (ashr<mode>3_di): Likewise.
9526 (lshr<mode>3_di): Likewise.
9527 (ashl<mode>3_di): Likewise.
9528 (iwmmxt_wmadds): Likewise.
9529 (iwmmxt_wmaddu): Likewise.
9530 (iwmmxt_tmia): Likewise.
9531 (iwmmxt_tmiaph): Likewise.
9532 (iwmmxt_tmiabb): Likewise.
9533 (iwmmxt_tmiatb): Likewise.
9534 (iwmmxt_tmiabt): Likewise.
9535 (iwmmxt_tmiatt): Likewise.
9536 (iwmmxt_tmovmskb): Likewise.
9537 (iwmmxt_tmovmskh): Likewise.
9538 (iwmmxt_tmovmskw): Likewise.
9539 (iwmmxt_waccb): Likewise.
9540 (iwmmxt_wacch): Likewise.
9541 (iwmmxt_waccw): Likewise.
9542 (iwmmxt_waligni): Likewise.
9543 (iwmmxt_walignr): Likewise.
9544 (iwmmxt_walignr0): Likewise.
9545 (iwmmxt_walignr1): Likewise.
9546 (iwmmxt_walignr2): Likewise.
9547 (iwmmxt_walignr3): Likewise.
9548 (iwmmxt_wsadb): Likewise.
9549 (iwmmxt_wsadh): Likewise.
9550 (iwmmxt_wsadbz): Likewise.
9551 (iwmmxt_wsadhz): Likewise.
9552 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
9553 (iwmmxt_wabsdiffb): Likewise.
9554 (iwmmxt_wabsdiffh): Likewise.
9555 (iwmmxt_wabsdiffw): Likewise.
9556 (iwmmxt_waddsubhx): Likewise
9557 (iwmmxt_wsubaddhx): Likewise.
9558 (addc<mode>3): Likewise.
9559 (iwmmxt_avg4): Likewise.
9560 (iwmmxt_avg4r): Likewise.
9561 (iwmmxt_wmaddsx): Likewise.
9562 (iwmmxt_wmaddux): Likewise.
9563 (iwmmxt_wmaddsn): Likewise.
9564 (iwmmxt_wmaddun): Likewise.
9565 (iwmmxt_wmulwsm): Likewise.
9566 (iwmmxt_wmulwum): Likewise.
9567 (iwmmxt_wmulsmr): Likewise.
9568 (iwmmxt_wmulumr): Likewise.
9569 (iwmmxt_wmulwsmr): Likewise.
9570 (iwmmxt_wmulwumr): Likewise.
9571 (iwmmxt_wmulwl): Likewise.
9572 (iwmmxt_wqmulm): Likewise.
9573 (iwmmxt_wqmulwm): Likewise.
9574 (iwmmxt_wqmulmr): Likewise.
9575 (iwmmxt_wqmulwmr): Likewise.
9576 (iwmmxt_waddbhusm): Likewise.
9577 (iwmmxt_waddbhusl): Likewise.
9578 (iwmmxt_wqmiabb): Likewise.
9579 (iwmmxt_wqmiabt): Likewise.
9580 (iwmmxt_wqmiatb): Likewise.
9581 (iwmmxt_wqmiatt): Likewise.
9582 (iwmmxt_wqmiabbn): Likewise.
9583 (iwmmxt_wqmiabtn): Likewise.
9584 (iwmmxt_wqmiatbn): Likewise.
9585 (iwmmxt_wqmiattn): Likewise.
9586 (iwmmxt_wmiabb): Likewise.
9587 (iwmmxt_wmiabt): Likewise.
9588 (iwmmxt_wmiatb): Likewise.
9589 (iwmmxt_wmiatt): Likewise.
9590 (iwmmxt_wmiabbn): Likewise.
9591 (iwmmxt_wmiabtn): Likewise.
9592 (iwmmxt_wmiatbn): Likewise.
9593 (iwmmxt_wmiattn): Likewise.
9594 (iwmmxt_wmiawbb): Likewise.
9595 (iwmmxt_wmiawbt): Likewise.
9596 (iwmmxt_wmiawtb): Likewise.
9597 (iwmmxt_wmiawtt): Likewise.
9598 (iwmmxt_wmiawbbn): Likewise.
9599 (iwmmxt_wmiawbtn): Likewise.
9600 (iwmmxt_wmiawtbn): Likewise.
9601 (iwmmxt_wmiawttn): Likewise.
9602 (iwmmxt_wmerge): Likewise.
9603 (iwmmxt_tandc<mode>3): Likewise.
9604 (iwmmxt_torc<mode>3): Likewise.
9605 (iwmmxt_torvsc<mode>3): Likewise.
9606 (iwmmxt_textrc<mode>3): Likewise.
9607 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
9608 (wmmxt_pack): Likewise.
9609 (wmmxt_mult_c1): Likewise.
9610 (wmmxt_mult_c2): Likewise.
9611 (wmmxt_alu_c1): Likewise.
9612 (wmmxt_alu_c2): Likewise.
9613 (wmmxt_alu_c3): Likewise.
9614 (wmmxt_transfer_c1): Likewise.
9615 (wmmxt_transfer_c2): Likewise.
9616 (wmmxt_transfer_c3): Likewise.
9617 (marvell_f_iwmmxt_wstr): Likewise.
9618 (marvell_f_iwmmxt_wldr): Likewise.
9620 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
9622 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
9624 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
9627 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
9628 * lra-constraints.c (need_for_split_p): Check call used hard regs
9629 living through calls.
9631 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
9632 call used regs for call insn.
9634 2013-06-28 Jakub Jelinek <jakub@redhat.com>
9637 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
9638 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
9640 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
9642 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
9644 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
9646 * lra-constraints.c (need_for_split_p): Check call used hard regs
9647 living through calls.
9649 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
9652 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
9653 to tie with any other modes. Eliminate Altivec vector mode tests,
9654 since these are a subset of ALTIVEC or VSX vector modes. Simplify
9655 code, to return 0 if testing MODE2 for a condition, if we've
9656 already tested MODE1 for the same condition.
9658 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
9660 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
9663 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
9665 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
9666 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
9668 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
9670 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
9672 (aarch64_symbolic_constant_p): Remove.
9673 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
9674 static. Fix line length and white space.
9675 (aarch64_symbolic_constant_p): Remove.
9676 * config/aarch64/predicates.md (aarch64_valid_symref):
9677 Use aarch64_classify_symbol_expression.
9679 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9681 * config/arm/constraints.md (Ts): New constraint.
9682 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
9684 (compare_scc): Use "Ts" constraint for operand 0.
9685 (ior_scc_scc): Likewise.
9686 (and_scc_scc): Likewise.
9687 (and_scc_scc_nodom): Likewise.
9688 (ior_scc_scc_cmp): Likewise for operand 7.
9689 (and_scc_scc_cmp): Likewise.
9690 * config/arm/thumb2.md (thumb2_movsi_insn):
9691 Add alternatives for 16-bit encodings.
9692 (thumb2_movhi_insn): Likewise.
9693 (thumb2_movsicc_insn): Likewise.
9694 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
9695 (thumb2_negscc): Use "Ts" constraint.
9696 Move mvn instruction outside cond_exec block.
9697 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
9698 for 16-bit encodings.
9700 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9702 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
9704 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
9705 (mulsi3subsi): Likewise.
9706 (mulsidi3adddi): Likewise.
9707 (mulsidi3_v6): Likewise.
9708 (umulsidi3_v6): Likewise.
9709 (umulsidi3adddi_v6): Likewise.
9710 (smulsi3_highpart_v6): Likewise.
9711 (umulsi3_highpart_v6): Likewise.
9712 (mulhisi3tb): Likewise.
9713 (mulhisi3bt): Likewise.
9714 (mulhisi3tt): Likewise.
9715 (maddhisi4): Likewise.
9716 (maddhisi4tb): Likewise.
9717 (maddhisi4tt): Likewise.
9718 (maddhidi4): Likewise.
9719 (maddhidi4tb): Likewise.
9720 (maddhidi4tt): Likewise.
9721 (zeroextractsi_compare0_scratch): Likewise.
9722 (insv_zero): Likewise.
9723 (insv_t2): Likewise.
9724 (anddi_notzesidi_di): Likewise.
9725 (anddi_notsesidi_di): Likewise.
9726 (andsi_notsi_si): Likewise.
9727 (iordi_zesidi_di): Likewise.
9728 (xordi_zesidi_di): Likewise.
9729 (andsi_iorsi3_notsi): Likewise.
9731 (smax_m1): Likewise.
9733 (not_shiftsi): Likewise.
9734 (unaligned_loadsi): Likewise.
9735 (unaligned_loadhis): Likewise.
9736 (unaligned_loadhiu): Likewise.
9737 (unaligned_storesi): Likewise.
9738 (unaligned_storehi): Likewise.
9739 (extv_reg): Likewise.
9740 (extzv_t2): Likewise.
9742 (udivsi3): Likewise.
9743 (arm_zero_extendhisi2addsi): Likewise.
9744 (arm_zero_extendqisi2addsi): Likewise.
9745 (compareqi_eq0): Likewise.
9746 (arm_extendhisi2_v6): Likewise.
9747 (arm_extendqisi2addsi): Likewise.
9748 (arm_movt): Likewise.
9749 (thumb2_ldrd): Likewise.
9750 (thumb2_ldrd_base): Likewise.
9751 (thumb2_ldrd_base_neg): Likewise.
9752 (thumb2_strd): Likewise.
9753 (thumb2_strd_base): Likewise.
9754 (thumb2_strd_base_neg): Likewise.
9755 (arm_negsi2): Add alternative for 16-bit encoding.
9756 (arm_one_cmplsi2): Likewise.
9758 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9760 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
9761 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
9762 (movdfcc): Likewise.
9763 * config/arm/vfp.md (*thumb2_movsf_vfp):
9764 Disable predication for arm_restrict_it.
9765 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
9766 (*thumb2_movdfcc_vfp): Likewise.
9767 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
9768 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
9769 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
9770 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
9771 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
9772 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
9773 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
9774 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
9775 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
9776 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
9777 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
9778 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
9779 Disable predication for arm_restrict_it.
9781 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
9783 * config/i386/bmiintrin.h (_bextr_u32): New.
9784 (_bextr_u64): Ditto.
9786 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
9788 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
9789 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
9790 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
9791 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
9792 * config/mips/n32-elf.h: ...this new file.
9794 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
9797 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
9798 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
9800 2013-06-27 Catherine Moore <clm@codesourcery.com>
9802 * config/mips/mips-tables.opt: Regenerate.
9803 * config/mips/mips-cpus.def: Add m14ke and m14kec.
9804 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
9805 * doc/invoke.texi: Add -m14kc.
9807 2013-06-27 Jakub Jelinek <jakub@redhat.com>
9810 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
9811 constraints of operand 1 and 2.
9814 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
9815 to match RTL canonicalization. Swap predicates and
9816 constraints of operand 1 and 2.
9818 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
9820 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
9821 Process OP_INOUT regs for splitting too.
9823 2013-06-27 Jakub Jelinek <jakub@redhat.com>
9825 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
9826 decl before the loop, initialize to NULL.
9827 (vectorizable_load): Initialize ptr_incr to NULL.
9829 2013-06-27 Martin Jambor <mjambor@suse.cz>
9832 * ipa-ref.h (ipa_maybe_record_reference): Declare.
9833 * ipa-ref.c (ipa_maybe_record_reference): New function.
9834 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
9835 * ipa-cp.c (create_specialized_node): Record potential references from
9837 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
9839 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
9841 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
9842 parameter 'mode' of type 'enum machine_mode mode'; change to pass
9843 'mode' to force_reg.
9844 (aarch64_add_offset): Update calls to aarch64_force_temporary.
9845 (aarch64_expand_mov_immediate): Likewise.
9847 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
9849 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
9850 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
9852 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9854 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
9855 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
9856 (struct machine_function): Add tbegin_p.
9857 (s390_canonicalize_comparison): Fold CC mode compares to
9858 conditional jump if possible.
9859 (s390_emit_jump): Return the emitted jump.
9860 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
9861 Handle CCRAWmode compares.
9862 (s390_option_override): Default to -mhtm if available.
9863 (s390_reg_clobbered_rtx): Handle floating point regs as well.
9864 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
9865 FPRs instead of df_regs_ever_live_p.
9866 (s390_optimize_nonescaping_tx): New function.
9867 (s390_init_frame_layout): Extend clobbered_regs array to cover
9869 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
9870 (s390_expand_tbegin): New function.
9871 (enum s390_builtin): New enum definition.
9872 (code_for_builtin): New array definition.
9873 (s390_init_builtins): New function.
9874 (s390_expand_builtin): New function.
9875 (TARGET_INIT_BUILTINS): Define.
9876 (TARGET_EXPAND_BUILTIN): Define.
9877 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
9878 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
9879 (s390_alc_comparison): Likewise.
9880 * config/s390/s390-modes.def: Add CCRAWmode.
9881 * config/s390/s390.h (processor_flags): Add PF_TX.
9882 (TARGET_CPU_HTM): Define macro.
9883 (TARGET_HTM): Define macro.
9884 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
9885 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
9886 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
9887 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
9888 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
9889 (TBEGIN_MASK, TBEGINC_MASK): New constants.
9890 ("*cc_to_int"): Move up.
9891 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
9892 constants other than 0.
9893 ("*ccraw_to_int"): New insn and splitter definition.
9894 ("tbegin", "tbegin_nofloat", "tbegin_retry")
9895 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
9896 ("tx_assist"): New expander.
9897 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
9898 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
9899 * config/s390/s390.opt: Add -mhtm option.
9900 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
9901 * config/s390/htmxlintrin.h: New file.
9902 * config/s390/htmintrin.h: New file.
9903 * config/s390/s390intrin.h: New file.
9904 * doc/extend.texi: Document htm builtins.
9905 * config.gcc: Add the new header files to extra_headers.
9907 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
9909 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
9910 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
9912 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
9913 Pat Haugen <pthaugen@us.ibm.com>
9914 Peter Bergner <bergner@vnet.ibm.com>
9916 * config/rs6000/power8.md: New.
9917 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
9918 setting for power8 entry.
9919 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
9920 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
9921 test for Power4/Power5 only.
9922 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
9924 (force_new_group): Adjust comment.
9925 * config/rs6000/rs6000.md: Include power8.md.
9927 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
9929 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
9930 * config/arm/arm-protos.h (arm_max_conditional_execute): New
9932 (tune_params): Update comment.
9933 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
9934 (arm_max_conditional_execute): New function.
9935 (thumb2_final_prescan_insn): Use max_insn_skipped and
9936 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
9938 2013-06-25 Jakub Jelinek <jakub@redhat.com>
9940 PR tree-optimization/57705
9941 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
9942 SSA_NAME step, provided that it is not defined inside the loop.
9943 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
9944 (get_initial_def_for_induction): Handle SSA_NAME IV step.
9946 2013-06-25 Martin Jambor <mjambor@suse.cz>
9949 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
9950 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
9952 (ipa_note_param_call): Initialize member_ptr flag.
9953 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
9954 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
9955 (ipa_write_indirect_edge_info): Stream member_ptr flag.
9956 (ipa_read_indirect_edge_info): Likewise.
9958 2013-06-25 Richard Biener <rguenther@suse.de>
9961 * passes.c (init_optimization_passes): Move pass_fold_builtins
9962 and pass_dce earlier with -Og.
9964 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
9966 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
9967 <BIT_FIELD_REF>: Remove trailing TAB.
9968 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
9971 2013-06-24 Martin Jambor <mjambor@suse.cz>
9973 PR tree-optimization/57358
9974 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
9975 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
9976 (ipa_analyze_params_uses): Generate pessimistic info when true.
9978 2013-06-24 Martin Jambor <mjambor@suse.cz>
9980 PR tree-optimization/57539
9981 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
9982 global.inlined_to of the new node to it. All callers changed.
9983 * ipa-inline-transform.c (clone_inlined_nodes): New variable
9984 inlining_into, pass it to cgraph_clone_node.
9985 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
9986 ipa_free_edge_args_substructures.
9987 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
9988 rdesc linked list. Do not assert rdesc edges have inlined caller.
9989 Assert we have found an rdesc in the rdesc list.
9991 2013-06-24 Richard Biener <rguenther@suse.de>
9993 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
9994 (pointer_set_lookup): Declare.
9995 (class pointer_map): New template class implementing a
9996 generic pointer to T map.
9997 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
9998 pointer_map<T>::contains, pointer_map<T>::insert,
9999 pointer_map<T>::traverse): New functions.
10000 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
10001 (pointer_set_lookup): New function.
10002 (pointer_set_contains): Use pointer_set_lookup.
10003 (pointer_set_insert): Likewise.
10004 (insert_aux): Remove.
10005 (struct pointer_map_t): Embed a pointer_set_t.
10006 (pointer_map_create): Adjust.
10007 (pointer_map_destroy): Likewise.
10008 (pointer_map_contains): Likewise.
10009 (pointer_map_insert): Likewise.
10010 (pointer_map_traverse): Likewise.
10011 * tree-streamer.h (struct streamer_tree_cache_d): Use a
10012 pointer_map<unsigned> instead of a pointer_map_t.
10013 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
10014 (streamer_tree_cache_lookup): Likewise.
10015 (streamer_tree_cache_create): Likewise.
10016 (streamer_tree_cache_delete): Likewise.
10017 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
10018 pointer_map<unsigned> instead of a pointer_map_t.
10019 (lto_init_tree_ref_encoder): Adjust.
10020 (lto_destroy_tree_ref_encoder): Likewise.
10021 * lto-section-out.c (lto_output_decl_index): Likewise.
10022 (lto_record_function_out_decl_state): Likewise.
10023 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
10025 2013-06-24 Richard Biener <rguenther@suse.de>
10027 PR tree-optimization/57488
10028 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
10030 2013-06-24 Alan Modra <amodra@gmail.com>
10032 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
10033 (gen_easy_altivec_constant): Likewise.
10034 * config/rs6000/predicates.md (easy_vector_constant_add_self,
10035 easy_vector_constant_msb): Likewise.
10037 2013-06-23 Jakub Jelinek <jakub@redhat.com>
10040 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
10041 add missing return true.
10043 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
10046 * config/sh/predicates.md (general_extend_operand): Invoke
10047 general_movsrc_operand for memory operands.
10048 (general_movsrc_operand): Allow reg+reg addressing, do not use
10049 general_operand for memory operands.
10051 2013-06-23 Sriraman Tallam <tmsriram@google.com>
10053 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
10054 when current target options does not apply.
10055 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
10056 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
10057 * config/i386/bmiintrin.h: Pass appropriate target
10058 attributes to header.
10059 * config/i386/mmintrin.h: Ditto.
10060 * config/i386/nmmintrin.h: Ditto.
10061 * config/i386/avx2intrin.h: Ditto.
10062 * config/i386/fxsrintrin.h: Ditto.
10063 * config/i386/tbmintrin.h: Ditto.
10064 * config/i386/xsaveintrin.h: Ditto.
10065 * config/i386/f16cintrin.h: Ditto.
10066 * config/i386/xtestintrin.h: Ditto.
10067 * config/i386/xsaveoptintrin.h: Ditto.
10068 * config/i386/bmi2intrin.h: Ditto.
10069 * config/i386/lzcntintrin.h: Ditto.
10070 * config/i386/smmintrin.h: Ditto.
10071 * config/i386/wmmintrin.h: Ditto.
10072 * config/i386/x86intrin.h: Remove all header include guards.
10073 * config/i386/prfchwintrin.h: Ditto.
10074 * config/i386/pmmintrin.h: Ditto.
10075 * config/i386/tmmintrin.h: Ditto.
10076 * config/i386/xmmintrin.h: Ditto.
10077 * config/i386/popcntintrin.h: Ditto.
10078 * config/i386/rdseedintrin.h: Ditto.
10079 * config/i386/ammintrin.h: Ditto.
10080 * config/i386/emmintrin.h: Ditto.
10081 * config/i386/immintrin.h: Remove all header include guards.
10082 * config/i386/fma4intrin.h: Ditto.
10083 * config/i386/lwpintrin.h: Ditto.
10084 * config/i386/xopintrin.h: Ditto.
10085 * config/i386/ia32intrin.h: Ditto.
10086 * config/i386/avxintrin.h: Ditto.
10087 * config/i386/rtmintrin.h: Ditto.
10088 * config/i386/fmaintrin.h: Ditto.
10089 * config/i386/mm3dnow.h: Ditto.
10091 2013-06-22 Sriraman Tallam <tmsriram@google.com>
10093 * common/config/i386/i386-common.c: Handle LZCNT.
10095 2013-06-22 Andi Kleen <ak@linux.intel.com>
10097 * doc/extend.texi: Use __atomic_store_n instead of
10098 __atomic_store in HLE example.
10100 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
10102 * config/sh/sh.c: Remove <cstdlib> workaround.
10104 2013-06-21 Andi Kleen <ak@linux.intel.com>
10106 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
10108 2013-06-21 Andi Kleen <ak@linux.intel.com>
10110 * doc/extend.texi: Document that __atomic_clear and
10111 __atomic_test_and_set should only be used with bool.
10113 2013-06-20 Jan Hubicka <jh@suse.cz>
10115 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
10116 types_same_for_odr.
10117 * tree.c (decls_same_for_odr): New function.
10118 (same_for_edr): New function.
10119 (types_same_for_odr): New function.
10120 (get_binfo_at_offset): Use it.
10121 * tree.h (types_same_for_odr): Declare.
10123 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
10124 Jason Merrill <jason@redhat.com>
10126 * system.h: Include <cstdlib> as well as <stdlib.h>.
10128 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
10131 * config/i386/i386.c (construct_container): Report error if
10132 long double is used with disabled x87 float returns.
10134 2013-06-20 Jan Hubicka <jh@suse.cz>
10136 * lto-cgraph.c (input_symtab): Do not set cgraph state.
10138 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
10140 PR rtl-optimization/57425
10141 PR rtl-optimization/57569
10142 * alias.c (write_dependence_p): Remove parameters mem_mode and
10143 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
10144 Changed all callers.
10145 (canon_anti_dependence): Get comments and semantics in sync.
10146 Add parameter mem_canonicalized. Changed all callers.
10147 * rtl.h (canon_anti_dependence): Update prototype.
10149 2013-06-20 Richard Biener <rguenther@suse.de>
10151 * data-streamer-in.c (streamer_read_uhwi): Optimize single
10152 byte case, inline streamer_read_uchar and defer section
10155 2013-06-20 Richard Biener <rguenther@suse.de>
10157 PR tree-optimization/57584
10158 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
10159 SSA names into the expanded expression that take part in
10160 abnormal coalescing.
10162 2013-06-19 Sharad Singhai <singhai@google.com>
10164 * gcov.c (print_usage): Handle new option.
10165 (process_args): Ditto.
10166 (get_gcov_intermediate_filename): New function.
10167 (output_intermediate_file): New function.
10168 (output_gcov_file): New function
10169 (generate_results): Handle new option.
10170 (release_function): Relase demangled name.
10171 (read_graph_file): Handle demangled name.
10172 (output_lines): Ditto.
10173 * doc/gcov.texi: Document gcov intermediate format.
10175 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
10178 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
10179 (lra_emit_add): Use the functions. Add comment about Y as an
10182 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
10185 * collect2.c (collect_atexit): New.
10186 (collect_exit): Delete.
10187 (main): Register collect_atexit with atexit.
10188 (collect_wait): Change collect_exit to exit.
10190 * collect2.h (collect_exit): Delete.
10191 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
10193 2013-06-19 Wei Mi <wmi@google.com>
10195 PR rtl-optimization/57518
10196 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
10197 if regno is used in paradoxical subreg.
10198 (update_equiv_regs): Check pdx_subregs[regno] before
10199 set a reg to be equivalent with a mem.
10201 2013-06-19 Matthias Klose <doko@ubuntu.com>
10204 * file-find.h (find_a_file): Add a mode parameter.
10205 * file-find.c (find_a_file): Likewise.
10206 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
10207 with X_OK for the executables.
10208 * collect2.c (main): Call find_a_file with X_OK.
10210 2013-06-19 Steve Ellcey <sellcey@mips.com>
10213 * config/mips/mips.md (casesi_internal_mips16_<mode>):
10214 Use NEXT_INSN instead of next_real_insn.
10216 2013-06-19 Jan Hubicka <jh@suse.cz>
10218 * cgraph.h (const_value_known_p): Replace by ...
10219 (ctor_for_folding): .. this one.
10220 * cgraphunit.c (process_function_and_variable_attributes): Use it.
10221 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
10222 * expr.c (expand_expr_real_1): Likewise.
10223 (string_constant): Likewise.
10224 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
10225 * ipa.c (process_references): Likewise.
10226 (symtab_remove_unreachable_nodes): Likewise.
10227 * ipa-inline-analysis.c (param_change_prob): Likewise.
10228 * gimple-fold.c (canonicalize_constructor_val): Likewise.
10229 (get_base_constructor): Likwise.
10230 * varpool.c (varpool_remove_node): Likewise.
10231 (varpool_remove_initializer): LIkewise.
10232 (dump_varpool_node): LIkwise.
10233 (const_value_known_p): Rewrite to ...
10234 (ctor_for_folding): ... this one.
10236 2013-06-19 Jakub Jelinek <jakub@redhat.com>
10239 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
10240 PERSONALITY in $PATH derived prefixes.
10242 2013-06-19 Jeff Law <law@redhat.com>
10244 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
10247 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
10248 (simplify_bitwise_binary): Use it to simpify certain binary ops on
10251 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
10253 * config/arm/vfp.md: Move VFP instruction classification documentation
10255 * config/arm/arm.md: ... here. Update instruction classification
10258 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
10260 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
10261 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
10262 pattern. Use more efficient sequences on ARMv5 and Thumb2.
10264 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
10267 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
10268 with NEXT_INSN. Use tablejump_p to check for jump table data
10271 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
10274 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
10275 that now in C++ the value is correct per the C++ standards.
10277 2013-06-19 Richard Biener <rguenther@suse.de>
10279 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
10280 for global context.
10282 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10285 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10288 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
10289 with next_active_insn.
10291 2013-06-18 Sriraman Tallam <tmsriram@google.com>
10293 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
10294 functions are inlined during failures to flag an error.
10295 * tree-inline.c (expand_call_inline): Allow the error to be flagged
10296 in early inline pass.
10298 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
10300 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
10303 2013-06-18 Julian Brown <julian@codesourcery.com>
10305 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
10306 Permit virtual register pre-reload if !strict.
10307 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
10309 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
10311 * config/arm/neon.md (movmisalign<mode>): Use
10312 neon_perm_struct_or_reg_operand instead of
10313 neon_struct_or_register_operand.
10314 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
10315 neon_permissive_struct_operand instead of neon_struct_operand.
10316 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
10317 neon_vector_mem_operand.
10318 * config/arm/predicates.md (neon_struct_operand): Adjust call to
10319 neon_vector_mem_operand.
10320 (neon_permissive_struct_operand): New.
10321 (neon_struct_or_register_operand): Rename to...
10322 (neon_perm_struct_or_reg_operand): This. Adjust call to
10323 neon_vector_mem_operand.
10325 2013-06-18 Richard Biener <rguenther@suse.de>
10327 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
10328 * lto-streamer.h: Include pointer-set.h.
10329 (struct lto_decl_slot): Remove.
10330 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
10331 Remove next_index entry.
10332 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
10333 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
10334 (lto_init_tree_ref_encoder): Adjust.
10335 (lto_destroy_tree_ref_encoder): Likewise.
10336 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
10337 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
10338 (lto_output_decl_index): Adjust.
10339 (lto_new_out_decl_state): Likewise.
10340 (lto_record_function_out_decl_state): Likewise.
10341 * lto-streamer-out.c (copy_function): Likewise.
10343 2013-06-18 Richard Biener <rguenther@suse.de>
10345 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
10346 * cgraphunit.c: Include cfgloop.h.
10347 (init_lowered_empty_function): Initialize the loop tree.
10348 (assemble_thunk): Insert new BBs into loops.
10350 2013-06-18 Richard Biener <rguenther@suse.de>
10352 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
10353 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
10354 the map from cache entry to cache index optional.
10355 (streamer_tree_cache_replace_tree): Adjust accordingly.
10356 (streamer_tree_cache_append): Likewise.
10357 (streamer_tree_cache_delete): Likewise.
10358 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
10359 streamer cache map from cache entry to cache index.
10360 * lto-streamer-out.c (create_output_block): Adjust.
10362 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
10364 * config/arm/arm.md (attribute "insn"): Move multiplication and
10365 division attributes to...
10366 (attribute "type"): ... here. Remove mult.
10367 (attribute "mul32"): New attribute.
10368 (attribute "mul64"): Add umaal.
10369 (*arm_mulsi3): Update attributes.
10370 (*arm_mulsi3_v6): Likewise.
10371 (*thumb_mulsi3): Likewise.
10372 (*thumb_mulsi3_v6): Likewise.
10373 (*mulsi3_compare0): Likewise.
10374 (*mulsi3_compare0_v6): Likewise.
10375 (*mulsi_compare0_scratch): Likewise.
10376 (*mulsi_compare0_scratch_v6): Likewise.
10377 (*mulsi3addsi): Likewise.
10378 (*mulsi3addsi_v6): Likewise.
10379 (*mulsi3addsi_compare0): Likewise.
10380 (*mulsi3addsi_compare0_v6): Likewise.
10381 (*mulsi3addsi_compare0_scratch): Likewise.
10382 (*mulsi3addsi_compare0_scratch_v6): Likewise.
10383 (*mulsi3subsi): Likewise.
10384 (*mulsidi3adddi): Likewise.
10385 (*mulsi3addsi_v6): Likewise.
10386 (*mulsidi3adddi_v6): Likewise.
10387 (*mulsidi3_nov6): Likewise.
10388 (*mulsidi3_v6): Likewise.
10389 (*umulsidi3_nov6): Likewise.
10390 (*umulsidi3_v6): Likewise.
10391 (*umulsidi3adddi): Likewise.
10392 (*umulsidi3adddi_v6): Likewise.
10393 (*smulsi3_highpart_nov6): Likewise.
10394 (*smulsi3_highpart_v6): Likewise.
10395 (*umulsi3_highpart_nov6): Likewise.
10396 (*umulsi3_highpart_v6): Likewise.
10397 (mulhisi3): Likewise.
10398 (*mulhisi3tb): Likewise.
10399 (*mulhisi3bt): Likewise.
10400 (*mulhisi3tt): Likewise.
10401 (maddhisi4): Likewise.
10402 (*maddhisi4tb): Likewise.
10403 (*maddhisi4tt): Likewise.
10404 (maddhidi4): Likewise.
10405 (*maddhidi4tb): Likewise.
10406 (*maddhidi4tt): Likewise.
10407 (divsi3): Likewise.
10408 (udivsi3): Likewise.
10409 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
10410 (thumb2_mulsi_short_compare0): Likewise.
10411 (thumb2_mulsi_short_compare0_scratch): Likewise.
10412 * config/arm/arm1020e.md (1020mult1): Update attribute change.
10413 (1020mult2): Likewise.
10414 (1020mult3): Likewise.
10415 (1020mult4): Likewise.
10416 (1020mult5): Likewise.
10417 (1020mult6): Likewise.
10418 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
10420 (cortex_a15_mult64): Likewise.
10421 (cortex_a15_sdiv): Likewise.
10422 (cortex_a15_udiv): Likewise.
10423 * config/arm/arm1026ejs.md (mult1): Update attribute change.
10429 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
10430 (pj4_ir_div): Likewise.
10431 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
10432 (11_mult2): Likewise.
10433 (11_mult3): Likewise.
10434 (11_mult4): Likewise.
10435 (11_mult5): Likewise.
10436 (11_mult6): Likewise.
10437 (11_mult7): Likewise.
10438 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
10439 (cortex_a8_mla): Likewise.
10440 (cortex_a8_mull): Likewise.
10441 (cortex_a8_smulwy): Likewise.
10442 (cortex_a8_smlald): Likewise.
10443 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
10444 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
10445 (cortex_r4_mul_3): Likewise.
10446 (cortex_r4_mla_4): Likewise.
10447 (cortex_r4_mla_3): Likewise.
10448 (cortex_r4_smlald): Likewise.
10449 (cortex_r4_mull): Likewise.
10450 (cortex_r4_sdiv): Likewise.
10451 (cortex_r4_udiv): Likewise.
10452 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
10453 (cortex_a7_idiv): Likewise.
10454 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
10455 (9_mult2): Likewise.
10456 (9_mult3): Likewise.
10457 (9_mult4): Likewise.
10458 (9_mult5): Likewise.
10459 (9_mult6): Likewise.
10460 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
10461 (cortex_a53_sdiv): Likewise.
10462 (cortex_a53_udiv): Likewise.
10463 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
10464 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
10465 (mp626_mult2): Likewise.
10466 (mp626_mult3): Likewise.
10467 (mp626_mult4): Likewise.
10468 * config/arm/fa526.md (526_mult1): Update attribute change.
10469 (526_mult2): Likewise.
10470 * config/arm/arm-generic.md (mult): Update attribute change.
10471 (mult_ldsched_strongarm): Likewise.
10472 (mult_ldsched): Likewise.
10473 (multi_cycle): Likewise.
10474 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
10475 * config/arm/fa606te.md (606te_mult1): Update attribute change.
10476 (606te_mult2): Likewise.
10477 (606te_mult3): Likewise.
10478 (606te_mult4): Likewise.
10479 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
10480 (cortex_a9_mac16): Likewise.
10481 (cortex_a9_multiply): Likewise.
10482 (cortex_a9_mac): Likewise.
10483 (cortex_a9_multiply_long): Likewise.
10484 * config/arm/fa626te.md (626te_mult1): Update attribute change.
10485 (626te_mult2): Likewise.
10486 (626te_mult3): Likewise.
10487 (626te_mult4): Likewise.
10489 2013-06-18 Richard Biener <rguenther@suse.de>
10492 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
10494 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10497 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
10498 with next_active_insn.
10500 2013-06-18 Alan Modra <amodra@gmail.com>
10502 * config/rs6000/rs6000.h (enum data_align): New.
10503 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
10504 (DATA_ABI_ALIGNMENT): Define.
10505 (CONSTANT_ALIGNMENT): Correct comment.
10506 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
10507 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
10509 2013-06-17 David Malcolm <dmalcolm@redhat.com>
10511 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
10512 ATTRIBUTE_UNUSED marking.
10514 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
10516 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
10517 alternative and update.
10518 (aarch64_dup_lanedi): Delete.
10519 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
10520 * config/aarch64/aarch64-simd-builtins.def: Update.
10522 2013-06-17 Richard Biener <rguenther@suse.de>
10524 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
10525 (lto_input_scc): Declare.
10526 (lto_input_tree_1): Likewise.
10527 (struct lto_stats_d): Add num_tree_bodies_output and
10528 num_pickle_refs_output.
10529 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
10530 (lto_read_tree_1): Split out from ...
10531 (lto_read_tree): ... this.
10532 (lto_input_scc): New function.
10533 (lto_input_tree_1): Split out from ...
10534 (lto_input_tree): ... this. Handle LTO_tree_scc.
10535 (lto_data_in_create): Create the streamer cache without hashes.
10536 * lto-streamer-out.c (create_output_block): Create the streamer
10537 cache with hashes when not doing WPA.
10538 (lto_write_tree_1): Split out from ...
10539 (lto_write_tree): ... this.
10540 (get_symbol_initial_value): New function.
10541 (lto_output_tree_1): Split out from ...
10542 (lto_output_tree): ... this. Write trees as series of SCCs
10543 using a DFS walk via DFS_write_tree.
10544 (struct sccs, struct scc_entry): New types.
10545 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
10546 (DFS_write_tree_body): New function.
10547 (DFS_write_tree): Likewise.
10548 (hash_tree): Likewise.
10549 (scc_entry_compare): Likewise.
10550 (hash_scc): Likewise.
10551 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
10552 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
10553 TREE_CHAIN as regular reference.
10554 (streamer_read_integer_cst): Remove.
10555 (streamer_get_pickled_tree): Adjust.
10556 * tree-streamer-out.c (streamer_write_chain): Disable streaming
10557 of DECL_EXTERNALs in BLOCK_VARS for now.
10558 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
10560 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
10561 Add hash value argument and record that if hashes are recorded
10563 (streamer_tree_cache_insert_1): Adjust.
10564 (streamer_tree_cache_insert): Likewise.
10565 (streamer_tree_cache_insert_at): Rename to ...
10566 (streamer_tree_cache_replace_tree): ... this and adjust.
10567 (streamer_tree_cache_append): Adjust.
10568 (record_common_node): Likewise.
10569 (streamer_tree_cache_create): Add argument whether to
10570 record hash values together with trees.
10571 (streamer_tree_cache_delete): Adjust.
10572 * tree-streamer.h (struct streamer_tree_cache_d): Add
10574 (streamer_read_integer_cst): Remove.
10575 (streamer_tree_cache_insert): Adjust.
10576 (streamer_tree_cache_append): Likewise.
10577 (streamer_tree_cache_insert_at): Rename to ...
10578 (streamer_tree_cache_replace_tree): ... this and adjust.
10579 (streamer_tree_cache_create): Add argument whether to record hashes.
10580 (streamer_tree_cache_get): Rename to ...
10581 (streamer_tree_cache_get_tree): ... this.
10582 (streamer_tree_cache_get_hash): New function.
10583 * tree.c (cache_integer_cst): New function.
10584 * tree.h (cache_integer_cst): Declare.
10585 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
10586 * lto-symtab.c (lto_varpool_replace_node): Only release
10587 DECL_INITIAL of non-prevailing decls.
10588 * varpool.c (varpool_remove_initializer): Do not release
10589 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
10591 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
10593 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
10594 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
10595 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
10596 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
10597 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
10598 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
10599 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
10600 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
10601 instead of TARGET_64BIT.
10602 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
10603 Require ISA_HAS_<D>DIV.
10605 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
10607 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
10608 (mips*-*-linux*): Move default with_llsc setting to where other
10610 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
10612 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
10613 Likewise. Remove default with_tune setting. Move default float
10614 setting to its own block. Handle with_llsc in the same block as above.
10616 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
10618 PR rtl-optimization/57425
10619 PR rtl-optimization/57569
10620 * alias.c (write_dependence_p): Add new parameters mem_mode,
10621 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
10622 Changed all callers.
10623 (canon_anti_dependence): New function.
10624 * cse.c (check_dependence): Use canon_anti_dependence.
10625 * cselib.c (cselib_invalidate_mem): Likewise.
10626 * rtl.h (canon_anti_dependence): Declare.
10628 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
10630 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
10631 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
10632 ".set mips3" for 64-bit targets.
10634 2013-06-15 Dehao Chen <dehao@google.com>
10636 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
10637 * gimple-low.c (gimple_check_call_matching_types): Likewise.
10638 (gimple_check_call_args): Likewise.
10639 * value-prof.c (check_ic_target): Likewise.
10640 * ipa-inline.c (early_inliner): Likewise.
10641 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
10642 * cgraph.c (cgraph_create_edge_1): Likewise.
10643 (cgraph_make_edge_direct): Likewise.
10645 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
10648 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
10649 rs6000_output_move_128bit to handle emitting quad memory
10650 operations. Set attribute length to 8 bytes.
10652 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
10654 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
10656 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
10657 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
10658 (aarch64_<su>mlsl<mode>): Likewise.
10660 2013-06-14 Mike Stump <mikestump@comcast.net>
10662 * Makefile.in (TARGET_H): Add insn-codes.h.
10664 2013-06-14 Alan Modra <amodra@gmail.com>
10666 PR middle-end/57134
10667 PR middle-end/57586
10668 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
10669 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
10670 bitfield expansion when EXPAND_MEMORY.
10671 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
10673 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
10675 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
10676 test for clearing quad memory on 32-bit later.
10678 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
10680 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
10681 (fold_negate_expr): Likewise.
10682 (fold_real_zero_addition_p): Handle vectors.
10683 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
10685 2013-06-14 Alan Modra <amodra@gmail.com>
10687 * varasm.c (force_const_mem): Revert 2013-06-07 change.
10689 2013-06-13 Jan Hubicka <jh@suse.cz>
10691 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
10692 Local comdats are not externally visible.
10693 * symtab.c (dump_symtab_base): Dump externally visible.
10694 (verify_symtab_base): Verify back links in the symtab hash.
10696 2013-06-13 Bin Cheng <bin.cheng@arm.com>
10698 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
10699 CONVERT_EXPR as equal nodes.
10701 2013-06-13 Bin Cheng <bin.cheng@arm.com>
10703 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
10705 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
10707 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
10708 Generalize to complex and vector.
10709 * tree.c (build_all_ones_cst): New function.
10710 * tree.h (build_all_ones_cst): Declare it.
10712 2013-06-13 Alan Modra <amodra@gmail.com>
10714 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
10715 * config/rs6000/rs6000.md (signbittf2): New insn.
10716 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
10717 (abstf2_internal, cmptf_internal2): Likewise.
10718 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
10720 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
10721 Pat Haugen <pthaugen@us.ibm.com>
10722 Peter Bergner <bergner@vnet.ibm.com>
10724 * config/rs6000/rs6000.c (emit_load_locked): Add support for
10725 power8 byte, half-word, and quad-word atomic instructions.
10726 (emit_store_conditional): Likewise.
10727 (rs6000_expand_atomic_compare_and_swap): Likewise.
10728 (rs6000_expand_atomic_op): Likewise.
10730 * config/rs6000/sync.md (larx): Add new modes for power8.
10732 (AINT): New mode iterator to include TImode as well as normal
10733 integer modes on power8.
10734 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
10735 that VSX registers are not considered. Use AINT mode iterator
10736 instead of INT1 to allow inclusion of quad word atomic operations
10738 (load_locked<mode>): Likewise.
10739 (store_conditional<mode>): Likewise.
10740 (atomic_compare_and_swap<mode>): Likewise.
10741 (atomic_exchange<mode>): Likewise.
10742 (atomic_nand<mode>): Likewise.
10743 (atomic_fetch_<fetchop_name><mode>): Likewise.
10744 (atomic_nand_fetch<mode>): Likewise.
10745 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
10747 (ATOMIC): On power8, add QImode, HImode modes.
10748 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
10749 modes that promote to SImode.
10750 (load_lockedti): Convert TImode arguments to PTImode, so that we
10751 get a guaranteed even/odd register pair.
10752 (load_lockedpti): Likewise.
10753 (store_conditionalti): Likewise.
10754 (store_conditionalpti): Likewise.
10756 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
10757 atomic load/store instructions.
10760 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
10762 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
10764 (insn_count): New attribute, with most cases extracted from...
10765 (length): ...here. Redefine most cases in terms of insn_count.
10766 (single_insn): Delete.
10767 (can_delay): Use insn_count to check for single instructions.
10768 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
10769 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
10770 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
10771 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
10772 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
10773 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
10774 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
10775 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
10776 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
10777 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
10778 rather than "length".
10779 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
10780 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
10781 Use "insn_count" rather than "length".
10782 * config/mips/mips-dsp.md
10783 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
10784 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
10787 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
10789 PR tree-optimization/57361
10790 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
10792 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
10794 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
10796 (aarch64_simd_combine<mode>): New instruction expansion.
10797 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
10798 function prototype.
10799 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
10800 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
10802 2013-06-12 Jan Hubicka <jh@suse.cz>
10804 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
10805 decl has when in streaming stage.
10806 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
10807 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
10809 2013-06-12 Roland Stigge <stigge@antcom.de>
10812 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
10814 2013-06-12 Jakub Jelinek <jakub@redhat.com>
10816 PR tree-optimization/57537
10817 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
10818 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
10820 2013-06-12 Richard Biener <rguenther@suse.de>
10822 * data-streamer.h (streamer_write_char_stream): CSE
10823 obs->current_pointer.
10824 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
10825 streamer_write_char_stream manually and optimize the resulting loop.
10826 (streamer_write_hwi_stream): Likewise.
10828 2013-06-12 Jan Hubicka <jh@suse.cz>
10830 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
10831 * cgraph.h (varpool_create_empty_node): Declare.
10832 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
10834 * symtab.c (symtab_unregister_node): Be lax about missin entries
10836 (symtab_get_node): Update comment.
10837 * varpool.c (varpool_create_empty_node): Break out from ...
10838 (varpool_node_for_decl): ... here.
10839 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
10841 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
10843 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
10844 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
10845 part. Use straight-line flow at the end.
10846 <COMPONENT_REF>: Remove superfluous else.
10847 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
10849 2013-06-12 Jakub Jelinek <jakub@redhat.com>
10852 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
10853 target hook even for !TREE_PUBLIC decls. If no resolution info
10854 is available, return false for common and external decls.
10856 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
10858 * config/rl78/constraints.md (U): New constraint.
10859 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
10862 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
10865 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
10866 to allow returning address to AT_PLATFORM name.
10868 2013-06-11 Jan Hubicka <jh@suse.cz>
10870 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
10871 * cgraph.h (symtab_node_base): Add weakref flag.
10872 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
10873 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
10874 (output_weakrefs): Use weakref flag.
10875 * fold-const.c (simple_operand_p): Handle WEAK.
10876 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
10877 * ipa.c (varpool_externally_visible_p): Drop weakref.
10878 (function_and_variable_visibility): Update comment; fix weakref
10879 sanity checks; do not clear DECL_WEAK on them.
10880 * lto-cgraph.c (lto_output_node): update.
10881 (lto_output_varpool_node): Update.
10882 (input_overwrite_node): Update.
10883 (input_node): Update.
10884 (input_varpool_node): Update.
10885 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
10886 (lto_symtab_merge_symbols): Add sanity check.
10887 (lto_symtab_prevailing_decl): Do not special case weakrefs.
10888 * passes.c (rest_of_decl_compilation): Set static flag, too.
10889 * symtab.c (dump_symtab_base): Dump weakref.
10890 (verify_symtab_base): Sanity check weakrefs.
10891 (symtab_make_decl_local): Remove duplicated code.
10892 (symtab_alias_ultimate_target): Simplify.
10893 * varpool.c (varpool_create_variable_alias): Set weakref flag.
10895 2013-06-11 Tom de Vries <tom@codesourcery.com>
10897 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
10898 sequence_vect == NULL.
10900 2013-06-11 DJ Delorie <dj@redhat.com>
10902 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
10903 (rl78_unwind_word_mode): New.
10905 2013-06-11 David Malcolm <dmalcolm@redhat.com>
10907 * final.c (debug_prefix_maps): Make static.
10909 2013-06-11 David Malcolm <dmalcolm@redhat.com>
10911 * function.c (initial_trampoline): Remove stray copy.
10913 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
10915 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
10917 2013-06-11 Martin Jambor <mjambor@suse.cz>
10919 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
10920 within bounds at the beginning of the function.
10922 2013-06-11 Alan Modra <amodra@gmail.com>
10924 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
10926 (get_named_section): Don't NULL !DECL_P decl.
10928 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
10930 * doc/invoke.texi (core-avx2): Document.
10932 (atom): Updated with MOVBE.
10934 2013-06-11 Richard Biener <rguenther@suse.de>
10936 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
10938 2013-06-11 Anton Blanchard <anton@samba.org>
10940 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
10941 correct shift value in little-endian mode.
10943 2013-06-11 Jakub Jelinek <jakub@redhat.com>
10946 * varasm.c (get_variable_align): Move #endif to the right place.
10948 2013-06-10 Cary Coutant <ccoutant@google.com>
10950 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
10951 for hash so that hash table traversal order is deterministic.
10953 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
10954 Pat Haugen <pthaugen@us.ibm.com>
10955 Peter Bergner <bergner@vnet.ibm.com>
10957 * config/rs6000/vector.md (GPR move splitter): Do not split moves
10958 of vectors in GPRS if they are direct moves or quad word load or
10961 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
10963 (direct_move_p): Likewise.
10964 (quad_load_store_p): Likewise.
10966 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
10967 classes into bins based on the physical register type.
10968 (reg_class_to_reg_type): Likewise.
10969 (IS_STD_REG_TYPE): Likewise.
10970 (IS_FP_VECT_REG_TYPE): Likewise.
10971 (reload_fpr_gpr): Arrays to determine what insn to use if we can
10972 use direct move instructions.
10973 (reload_gpr_vsx): Likewise.
10974 (reload_vsx_gpr): Likewise.
10975 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
10976 information that is a simplification of register classes. Also
10977 precalculate direct move reload helpers.
10978 (direct_move_p): New function to return true if the operation can
10979 be done as a direct move instruciton.
10980 (quad_load_store_p): New function to return true if the operation
10981 is a quad memory operation.
10982 (rs6000_legitimize_address): If quad memory, only allow register
10983 indirect for TImode addresses.
10984 (rs6000_legitimate_address_p): Likewise.
10985 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
10986 (rs6000_reload_register_type): Likewise.
10987 (register_to_reg_type): Return register type.
10988 (rs6000_secondary_reload_simple_move): New helper function for
10989 secondary reload and secondary memory needed to identify anything
10990 that is a simple move, and does not need reloading.
10991 (rs6000_secondary_reload_direct_move): New helper function for
10992 secondary reload to identify cases that can be done with several
10993 instructions via the direct move instructions.
10994 (rs6000_secondary_reload_move): New helper function for secondary
10995 reload to identify moves between register types that can be done.
10996 (rs6000_secondary_reload): Add support for quad memory operations
10997 and for direct move.
10998 (rs6000_secondary_memory_needed): Likewise.
10999 (rs6000_debug_secondary_memory_needed): Change argument names.
11000 (rs6000_output_move_128bit): New function to return the move to
11001 use for 128-bit moves, including knowing about the various
11002 limitations of quad memory operations.
11004 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
11005 memory operations. call rs6000_output_move_128bit for the actual
11006 instruciton(s) to generate.
11007 (vsx_movti_64bit): Likewise.
11009 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
11010 (UNSPEC_P8V_MTVSRWZ): Likewise.
11011 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
11012 (UNSPEC_P8V_MTVSRD): Likewise.
11013 (UNSPEC_P8V_XXPERMDI): Likewise.
11014 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
11015 (UNSPEC_FUSION_GPR): Likewise.
11016 (FMOVE128_GPR): New iterator for direct move.
11017 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
11018 (f32_sv): Likewise.
11019 (f32_dm): Likewise.
11020 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
11021 loads and direct move instructions.
11022 (zero_extendsidi2_lfiwzx): Likewise.
11023 (extendsidi2_lfiwax): Likewise.
11024 (extendsidi2_nocell): Likewise.
11025 (floatsi<mode>2_lfiwax): Likewise.
11026 (lfiwax): Likewise.
11027 (floatunssi<mode>2_lfiwzx): Likewise.
11028 (lfiwzx): Likewise.
11029 (fix_trunc<mode>_stfiwx): Likewise.
11030 (fixuns_trunc<mode>_stfiwx): Likewise.
11031 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
11032 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
11033 (parity<mode>2_cmpb): Set length/type attr.
11034 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
11035 for 'mr.' to fast_compare.
11036 (bpermd_<mode>): Change type attr to popcnt.
11037 (p8_fmrgow_<mode>): New insns for power8 direct move support.
11038 (p8_mtvsrwz_1): Likewise.
11039 (p8_mtvsrwz_2): Likewise.
11040 (reload_fpr_from_gpr<mode>): Likewise.
11041 (p8_mtvsrd_1): Likewise.
11042 (p8_mtvsrd_2): Likewise.
11043 (p8_xxpermdi_<mode>): Likewise.
11044 (reload_vsx_from_gpr<mode>): Likewise.
11045 (reload_vsx_from_gprsf): Likewise.
11046 (p8_mfvsrd_3_<mode>): LIkewise.
11047 (reload_gpr_from_vsx<mode>): Likewise.
11048 (reload_gpr_from_vsxsf): Likewise.
11049 (p8_mfvsrd_4_disf): Likewise.
11050 (multi-word GPR splits): Do not split direct moves or quad memory
11053 2013-06-10 David Malcolm <dmalcolm@redhat.com>
11055 * tree-into-ssa.c (interesting_blocks): Make static.
11057 2013-06-10 Jakub Jelinek <jakub@redhat.com>
11060 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
11061 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
11062 Use DATA_ABI_ALIGNMENT for that case instead if defined.
11063 (get_variable_align): New function.
11064 (get_variable_section, emit_bss, emit_common,
11065 assemble_variable_contents, place_block_symbol): Use
11066 get_variable_align instead of DECL_ALIGN.
11067 (assemble_noswitch_variable): Add align argument, use it
11068 instead of DECL_ALIGN.
11069 (assemble_variable): Adjust caller. Use get_variable_align
11070 instead of DECL_ALIGN.
11071 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
11073 (DATA_ABI_ALIGNMENT): Define.
11074 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
11075 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
11076 opt is false, only return the psABI mandated alignment increase.
11077 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
11078 (DATA_ABI_ALIGNMENT): ... this.
11079 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
11080 (DATA_ABI_ALIGNMENT): ... this.
11081 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
11082 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
11083 (DATA_ABI_ALIGNMENT): ... this.
11084 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
11085 * doc/tm.texi: Regenerated.
11087 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
11089 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
11090 cmp_code to construct REG_EQUAL note.
11092 2013-06-09 Jakub Jelinek <jakub@redhat.com>
11095 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
11096 that operands[2] doesn't overlap with operands[0].
11098 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
11099 Jan Hubicka <jh@suse.cz>
11101 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
11102 hack to mark symbols as used.
11104 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
11106 PR rtl-optimization/57559
11107 * lra-constraints.c (process_alt_operands): Don't discourage
11108 memory with known offset for offsetable memory constraint.
11109 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
11111 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
11113 * varasm.c (struct oc_local_state): Reorder fields.
11114 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
11115 and adjust accordingly.
11116 (output_constructor): Reorder initialization code and adjust call to
11117 output_constructor_bitfield.
11119 2013-06-07 Jan Hubicka <jh@suse.cz>
11121 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
11123 2013-06-07 David Malcolm <dmalcolm@redhat.com>
11125 * tree-object-size.c (unknown): Make const.
11127 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11129 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
11130 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
11131 for last alternative in the cpu_facility attribute.
11133 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11136 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
11137 (xordi3): Change operand 2 constraint to arm_xordi_operand.
11138 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
11139 * config/arm/constraints.md (Dg): New constraint.
11140 * config/arm/neon.md (xordi3_neon): Remove.
11141 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
11142 * config/arm/predicates.md (arm_xordi_operand): New predicate.
11144 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11146 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
11147 Clean up alternatives.
11149 2013-06-07 Alan Modra <amodra@gmail.com>
11151 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
11154 2013-06-07 Alan Modra <amodra@gmail.com>
11156 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
11158 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11160 * config/arm/constraints.md (Df): New constraint.
11161 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
11162 Correct length attribute for last two alternatives.
11164 2013-06-07 Alan Modra <amodra@gmail.com>
11166 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
11167 override user -mfp-in-toc.
11168 (offsettable_ok_by_alignment): Consider just the current access
11169 rather than the whole object, unless BLKmode. Handle
11170 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
11171 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
11172 for -mcmodel=medium.
11173 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
11174 override user -mfp-in-toc or -msum-in-toc. Default to
11175 -mno-fp-in-toc for -mcmodel=medium.
11177 2013-06-06 DJ Delorie <dj@redhat.com>
11179 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
11180 TARGET_VALID_POINTER_MODE.
11182 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
11183 Pat Haugen <pthaugen@us.ibm.com>
11184 Peter Bergner <bergner@vnet.ibm.com>
11186 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
11187 Document new power8 builtins.
11189 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
11190 condition code register, to allow 128-bit logical operations to be
11191 done in the VSX or GPR registers.
11192 (nor<mode>3): Use the canonical form for nor.
11193 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
11194 vclz*, and vpopcnt* vector instructions.
11195 (nand<mode>3): Likewise.
11196 (orc<mode>3): Likewise.
11197 (clz<mode>2): LIkewise.
11198 (popcount<mode>2): Likewise.
11200 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
11201 that only the GPRs are recognized.
11203 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
11204 support for new power8 builtins.
11206 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
11208 (xscvdpspn): Likewise.
11214 (vpopcnt): Likewise.
11215 (vpopcntb): Likewise.
11216 (vpopcnth): Likewise.
11217 (vpopcntw): Likewise.
11218 (vpopcntd): Likewise.
11220 (vmrgew): Likewise.
11221 (vmrgow): Likewise.
11223 (eqv_v16qi3): Likewise.
11224 (eqv_v8hi3): Likewise.
11225 (eqv_v4si3): Likewise.
11226 (eqv_v2di3): Likewise.
11227 (eqv_v4sf3): Likewise.
11228 (eqv_v2df3): Likewise.
11230 (nand_v16qi3): Likewise.
11231 (nand_v8hi3): Likewise.
11232 (nand_v4si3): Likewise.
11233 (nand_v2di3): Likewise.
11234 (nand_v4sf3): Likewise.
11235 (nand_v2df3): Likewise.
11237 (orc_v16qi3): Likewise.
11238 (orc_v8hi3): Likewise.
11239 (orc_v4si3): Likewise.
11240 (orc_v2di3): Likewise.
11241 (orc_v4sf3): Likewise.
11242 (orc_v2df3): Likewise.
11244 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
11245 allow power8 quad mode in 64-bit.
11246 (rs6000_builtin_vectorized_function): Add support to vectorize
11247 ISA 2.07 count leading zeros, population count builtins.
11248 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
11249 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
11250 (builtin_function_type): Add vgbbd builtin function which takes an
11252 (altivec_expand_vec_perm_const): Add support for new power8 merge
11255 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
11256 that does not include TImdoe for use with 32-bit.
11257 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
11259 (UNSPEC_VSX_CVDPSPN): Likewise.
11260 (vsx_xscvdpspn): Likewise.
11261 (vsx_xscvspdpn): Likewise.
11262 (vsx_xscvdpspn_scalar): Likewise.
11263 (vsx_xscvspdpn_directmove): Likewise.
11264 (vsx_and<mode>3): Split logical operations into 32-bit and
11265 64-bit. Add support to do logical operations on TImode as well as
11266 VSX vector types. Allow logical operations to be done in either
11267 VSX registers or in general purpose registers in 64-bit mode. Add
11268 splitters if GPRs were used. For AND, add clobber of CCmode to
11269 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
11271 (vsx_and<mode>3_32bit): Likewise.
11272 (vsx_and<mode>3_64bit): Likewise.
11273 (vsx_ior<mode>3): Likewise.
11274 (vsx_ior<mode>3_32bit): Likewise.
11275 (vsx_ior<mode>3_64bit): Likewise.
11276 (vsx_xor<mode>3): Likewise.
11277 (vsx_xor<mode>3_32bit): Likewise.
11278 (vsx_xor<mode>3_64bit): Likewise.
11279 (vsx_one_cmpl<mode>2): Likewise.
11280 (vsx_one_cmpl<mode>2_32bit): Likewise.
11281 (vsx_one_cmpl<mode>2_64bit): Likewise.
11282 (vsx_nor<mode>3): Likewise.
11283 (vsx_nor<mode>3_32bit): Likewise.
11284 (vsx_nor<mode>3_64bit): Likewise.
11285 (vsx_andc<mode>3): Likewise.
11286 (vsx_andc<mode>3_32bit): Likewise.
11287 (vsx_andc<mode>3_64bit): Likewise.
11288 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
11289 and xxlorc instructions.
11290 (vsx_eqv<mode>3_64bit): Likewise.
11291 (vsx_nand<mode>3_32bit): Likewise.
11292 (vsx_nand<mode>3_64bit): Likewise.
11293 (vsx_orc<mode>3_32bit): Likewise.
11294 (vsx_orc<mode>3_64bit): Likewise.
11296 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
11298 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
11300 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
11301 (p8_vmrgow): Likewise.
11302 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
11303 GPRs to be split under VSX.
11304 (p8v_clz<mode>2): Add power8 count leading zero support.
11305 (p8v_popcount<mode>2): Add power8 population count support.
11306 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
11309 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
11312 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
11314 (vec_nand): Likewise.
11315 (vec_vclz): Likewise.
11316 (vec_vclzb): Likewise.
11317 (vec_vclzd): Likewise.
11318 (vec_vclzh): Likewise.
11319 (vec_vclzw): Likewise.
11320 (vec_vgbbd): Likewise.
11321 (vec_vmrgew): Likewise.
11322 (vec_vmrgow): Likewise.
11323 (vec_vpopcnt): Likewise.
11324 (vec_vpopcntb): Likewise.
11325 (vec_vpopcntd): Likewise.
11326 (vec_vpopcnth): Likewise.
11327 (vec_vpopcntw): Likewise.
11329 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
11331 PR rtl-optimization/57468
11332 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
11335 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
11337 PR rtl-optimization/57459
11338 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
11339 type when setting live regs.
11341 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
11343 * config/s390/s390.opt (mlra): New option.
11344 * config/s390/s390.c (s390_decompose_address): Check displacement
11345 for all registers for LRA.
11346 (s390_secondary_reload): Don't used secondary reloads for LRA.
11347 (s390_lra_p): New function.
11348 (TARGET_LRA_P): Define.
11349 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
11350 of attribute cpu_facility to zarch for the last alternative.
11351 (*cmpmem_short): Ditto.
11353 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
11355 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
11356 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
11357 (arm_expand_prologue): Likewise.
11359 2013-06-06 Teresa Johnson <tejohnson@google.com>
11362 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
11363 as this is now done by redirect_edge_and_branch_force.
11364 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
11365 barriers, and fix interaction with splitting.
11366 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
11367 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
11368 reflect changes made in the routine.
11369 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
11370 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
11371 since this is called in cfglayout mode, and replace partition fixup
11372 with assert as that is now done by force_nonfallthru_and_redirect.
11373 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
11374 already be marked with region crossing note.
11375 (insert_section_boundary_note): Make non-static, gate on flag
11376 has_bb_partition, rewrite to also check for multiple partitions.
11377 (rest_of_handle_reorder_blocks): Remove call to
11378 insert_section_boundary_note, now done later during free_cfg.
11379 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
11380 * bb-reorder.h (insert_section_boundary_note): Declare.
11381 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
11382 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
11383 invoke insert_section_boundary_note.
11384 (try_redirect_by_replacing_jump): Remove unnecessary
11385 check for region crossing note.
11386 (fixup_partition_crossing): New function.
11387 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
11388 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
11389 in non-cfglayout mode.
11390 (force_nonfallthru_and_redirect): Fixup partition boundaries,
11391 remove old code that tried to do this. Emit barrier correctly
11392 when we are in cfglayout mode.
11393 (last_bb_in_partition): New function.
11394 (rtl_split_edge): Correctly fixup partition boundaries.
11395 (commit_one_edge_insertion): Remove old code that tried to
11396 fixup region crossing edge since this is now handled in
11397 split_block, and set up insertion point correctly since
11398 block may now end in a jump.
11399 (verify_hot_cold_block_grouping): Guard against checking when not in
11400 linearized RTL mode.
11401 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
11403 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
11404 rtl_verify_flow_info, so not called in cfglayout mode.
11405 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
11406 (fixup_reorder_chain): Remove old code that attempted to fixup region
11407 crossing note as this is now handled in force_nonfallthru_and_redirect.
11408 (duplicate_insn_chain): Don't duplicate switch section notes.
11409 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
11411 * basic-block.h (emit_barrier_after_bb): Declare.
11413 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11415 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
11416 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
11417 arm_usatsihi): Adjust alternatives for arm_restrict_it.
11419 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11421 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
11423 * config/arm/ldmstm.md: Regenerate.
11425 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11427 * config/arm/sync.md (atomic_loaddi_1):
11428 Disable predication for arm_restrict_it.
11429 (arm_load_exclusive<mode>): Likewise.
11430 (arm_load_exclusivesi): Likewise.
11431 (arm_load_exclusivedi): Likewise.
11432 (arm_load_acquire_exclusive<mode>): Likewise.
11433 (arm_load_acquire_exclusivesi): Likewise.
11434 (arm_load_acquire_exclusivedi): Likewise.
11435 (arm_store_exclusive<mode>): Likewise.
11436 (arm_store_exclusive<mode>): Likewise.
11437 (arm_store_release_exclusivedi): Likewise.
11438 (arm_store_release_exclusive<mode>): Likewise.
11440 2013-06-06 Richard Biener <rguenther@suse.de>
11442 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
11444 (lto_tag_is_tree_code_p): Adjust.
11445 (lto_tag_is_gimple_code_p): Likewise.
11446 (lto_gimple_code_to_tag): Likewise.
11447 (lto_tag_to_gimple_code): Likewise.
11448 (lto_tree_code_to_tag): Likewise.
11449 (lto_tag_to_tree_code): Likewise.
11450 * data-streamer.h (streamer_write_hwi_in_range): Use
11451 uhwi streaming to stream the normalized range.
11452 (streamer_read_hwi_in_range): Likewise.
11454 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11456 * config/arm/arm.md (enabled_for_depr_it): New attribute.
11457 (predicable_short_it): Likewise.
11458 (predicated): Likewise.
11459 (enabled): Handle above.
11460 (define_cond_exec): Set predicated attribute to yes.
11462 2013-06-05 Mike Stump <mikestump@comcast.net>
11464 * gdbinit.in (__FUNCTION__): Add.
11466 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
11468 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
11469 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
11471 2013-06-05 Jan Hubicka <jh@suse.cz>
11473 * varasm.c (mark_decl_referenced): Revert the removal until targets
11476 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
11478 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
11479 instead of mark_decl_referenced.
11481 2013-06-05 Jan Hubicka <jh@suse.cz>
11483 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
11484 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
11485 and symtab_used_from_object_file_p.
11486 (cgraph_make_node_local_1): Clear forced_by_abi.
11487 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
11488 * cgraph.h (symtab_node_base): Add forced_by_abi.
11489 (decide_is_variable_needed): Remove.
11490 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
11491 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
11492 (decide_is_symbol_needed): ... this one; handle symbols in general;
11493 always analyze virtuals; honnor forced_by_abi.
11494 (cgraph_finalize_function): Update.
11495 (varpool_finalize_decl): Update.
11496 (symbol_defined_and_needed): Remove.
11497 (analyze_functions): Update.
11498 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
11499 output_refs, input_overwrite_node): Handle forced_by_abi.
11500 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
11501 (address_taken_from_non_vtable_p): ... this one.
11502 (comdat_can_be_unshared_p_1): New function.
11503 (cgraph_comdat_can_be_unshared_p): Rename to ...
11504 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
11505 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
11506 (function_and_variable_visibility): Clear forced_by_abi as needed.
11507 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
11508 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
11509 * symtab.c (dump_symtab_base): Dump forced_by_abi.
11510 * varpool.c (decide_is_variable_needed): Remove.
11512 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11514 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
11515 (arm_option_override): Override arm_restrict_it where appropriate.
11516 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
11517 * config/arm/arm.opt (mrestrict-it): New command-line option.
11518 * doc/invoke.texi: Document -mrestrict-it.
11520 2013-06-05 David Malcolm <dmalcolm@redhat.com>
11522 * tsan.c (tsan_atomic_table): Make const.
11524 2013-06-05 Richard Biener <rguenther@suse.de>
11526 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
11527 index associated with the tree we are supposed to replace.
11528 * tree-streamer-out.c (pack_ts_base_value_fields): Output
11529 TREE_ASM_WRITTEN as zero for everything but SSA names.
11531 2013-06-05 David Malcolm <dmalcolm@redhat.com>
11533 * tree-ssa-structalias.c (call_stmt_vars): Make static.
11535 2013-06-04 Jan Hubicka <jh@suse.cz>
11537 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
11538 (input_node, input_varpool_node): Handle correctly external same
11540 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
11541 nodes at ltrans stage.
11543 2013-06-04 Jan Hubicka <jh@suse.cz>
11545 * ipa-inline.c (update_caller_keys): Fix availability test.
11546 (update_callee_keys): Likewise.
11547 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
11548 to follow ELF standard.
11550 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
11552 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
11553 (mips64r5900el-*-elf*): New configurations.
11554 * config/mips/mips-cpus.def (r5900): New processor.
11555 * config/mips/mips-tables.opt: Regenerate.
11556 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
11557 (mips_issue_rate): Handle PROCESSOR_R5900.
11558 (mips_reorg_process_insns): Force reorder mode for the R5900.
11559 * config/mips/mips.h (TARGET_MIPS5900): Define.
11560 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
11562 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
11564 * config/mips/mips.md (processor): Add r5900.
11565 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
11567 2013-06-04 Ian Bolton <ian.bolton@arm.com>
11569 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
11570 into function to generate MOVI instruction.
11571 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
11572 (aarch64_preferred_simd_mode): Turn into wrapper.
11573 (aarch64_output_scalar_simd_mov_immediate): New function.
11574 * config/aarch64/aarch64-protos.h: Add prototype for above.
11576 2013-06-04 Ian Bolton <ian.bolton@arm.com>
11578 * config/aarch64/aarch64.c (simd_immediate_info): Remove
11579 element_char member.
11580 (sizetochar): Return signed char.
11581 (aarch64_simd_valid_immediate): Remove elchar and other
11582 unnecessary variables.
11583 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
11584 Calculate element_char as required.
11585 * config/aarch64/aarch64-protos.h: Update and move prototype
11586 for aarch64_output_simd_mov_immediate.
11587 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
11590 2013-06-04 Ian Bolton <ian.bolton@arm.com>
11592 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
11593 information completed by aarch64_simd_valid_immediate.
11594 (aarch64_legitimate_constant_p): Update arguments.
11595 (aarch64_simd_valid_immediate): Work with struct rather than many
11597 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
11598 (aarch64_simd_make_constant): Update arguments.
11599 (aarch64_output_simd_mov_immediate): Work with struct rather than
11600 many pointers. Output immediate directly rather than as operand.
11601 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
11603 * config/aarch64/constraints.md (Dn): Update arguments.
11605 2013-06-04 Ian Bolton <ian.bolton@arm.com>
11607 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
11609 (aarch64_simd_immediate_valid_for_move): Remove.
11610 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
11611 (aarch64_simd_make_constant): Update call.
11612 (aarch64_output_simd_mov_immediate): Update call.
11613 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
11615 * config/aarch64/constraints.md (Dn): Update call.
11617 2013-06-04 Ian Bolton <ian.bolton@arm.com>
11619 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
11620 return type to bool for prototype.
11621 (aarch64_legitimate_constant_p): Check for true instead of not -1.
11622 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
11623 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
11625 2013-06-04 Catherine Moore <clm@codesourcery.com>
11627 * config/mips/mips.opt (meva): New.
11628 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
11629 (ASM_SPEC): Handle -meva.
11630 * doc/invoke.texi (meva): Document.
11632 2013-06-04 Alan Modra <amodra@gmail.com>
11634 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
11637 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11639 * rtl.def: Add extra fourth optional field to define_cond_exec.
11640 * gensupport.c (process_one_cond_exec): Process attributes from
11642 * doc/md.texi: Document fourth field in define_cond_exec.
11644 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
11646 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
11647 out the processing order as in store_bit_field_1.
11649 2013-06-04 Jan Hubicka <jh@suse.cz>
11651 PR middle-end/57500
11652 * cgraphunit.c (cgraph_process_same_body_aliases): Create
11653 non-VAR_DECL node if it does not exist yet.
11655 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
11657 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
11658 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
11659 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
11660 target_cpu_default setting.
11662 2013-06-03 Teresa Johnson <tejohnson@google.com>
11664 * dumpfile.c (opt_info_switch_p): Change -fopt-info
11665 default to -fopt-info=optimized instead of all.
11666 * doc/invoke.texi: Ditto.
11667 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
11668 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
11669 (execute_vect_slp): Emit BB vectorization success under
11670 MSG_OPTIMIZED_LOCATIONS.
11671 * tree-vect-slp.c (vect_slp_transform_bb): Change
11672 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
11673 * tree-vect-loop.c (vect_transform_loop): Ditto.
11675 2013-06-03 Jason Merrill <jason@redhat.com>
11678 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
11679 Use TARGET_EXPR for C++.
11681 2013-06-03 Jakub Jelinek <jakub@redhat.com>
11683 PR rtl-optimization/57268
11684 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
11685 if DEBUG_INSN_P (insn).
11688 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
11690 PR rtl-optimization/57268
11691 * sched-deps.c (sched_analyze_2): Flush dependence lists if
11692 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
11694 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
11696 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
11697 (ix86_avoid_lea_for_addr): Likewise.
11698 (exact_dependency_1): Likewise.
11699 (ix86_adjust_cost): Likewise.
11700 (swap_top_of_ready_list): Fix formatting and !reload_completed check
11702 (do_reorder_for_imul): Fix typo, formatting and
11703 !reload_completed check removed.
11704 (ix86_sched_reorder): Fix typo and formatting.
11705 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
11708 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
11710 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
11712 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
11714 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
11715 <STRING_CST>: Likewise.
11716 <VECTOR_CST>: Likewise.
11718 2013-06-01 Janus Weil <janus@gcc.gnu.org>
11719 Mikael Morin <mikael@gcc.gnu.org>
11721 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
11722 * config.in: Regenerated.
11723 * configure: Regenerated.
11725 2013-06-01 Jan Hubicka <jh@suse.cz>
11727 PR middle-end/57366
11728 * cgraphunit.c (compile): When weakref is not supported,
11729 set up transparent aliases before final output pass.
11730 * varasm.c (assemble_alias): Do not try to do it here.
11732 2013-06-01 Jan Hubicka <jh@suse.cz>
11734 PR middle-end/57467
11735 * passes.c (for_per_function): Skip unanalyzed functions.
11737 2013-06-01 Jan Hubicka <jh@suse.cz>
11739 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
11740 (lto_symtab_merge_symbols_1): ... this one.
11741 (lto_symtab_merge_cgraph_nodes): Rename to ...
11742 (lto_symtab_merge_symbols): ... this one; simplify.
11743 * cgraph.c (same_body_aliases_done): Rename to ...
11744 (cpp_implicit_aliases_done): ... this one.
11745 (cgraph_create_function_alias): Update.
11746 (cgraph_same_body_alias): Update.
11747 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
11748 (verify_edge_corresponds_to_fndecl): Simplify.
11749 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
11750 (cgraph_node): Remove same_body_alias.
11751 (varpool_node): Remove alias_of and extra_name_alias.
11752 (same_body_aliases_done): Rename to ..
11753 (cpp_implicit_aliases_done): ... this one.
11754 (symtab_alias_ultimate_target): Add default parameter.
11755 (symtab_resolve_alias): New function.
11756 (fixup_same_cpp_alias_visibility): Declare.
11757 (cgraph_function_node): Add default parameter.
11758 (cgraph_node_asm_name): Likewise.
11759 (cgraph_function_or_thunk_node): Add default parameter; do
11760 not ICE when it is NULL.
11761 (varpool_variable_node): Likewise.
11762 * tree-emutls.c (create_emultls_var): Update.
11763 (ipa_lower_emutls): Update.
11764 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
11765 (cgraph_reset_node): Reset alias info.
11766 (cgraph_finalize_function): Update.
11767 (fixup_same_cpp_alias_visibility): Move to symtab.c.
11768 (analyze_function): Simplify.
11769 (cgraph_process_same_body_aliases): Simplify.
11770 (analyze_functions): Fixup same body aliases.
11771 (handle_alias_pairs): Simplify.
11772 (assemble_thunk): Update.
11773 (assemble_thunks_and_aliases): Update.
11774 (output_weakrefs): Rewrite.
11775 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
11776 (lto_output_varpool_node): Likewise.
11777 (compute_ltrans_boundary): Remve assert.
11778 (get_alias_symbol): New functoin.
11779 (input_node): Rewrite alias handling.
11780 (input_varpool_node): Likewise.
11781 * ipa-pure-const.c (propagate_pure_const): Fix formating.
11782 * ipa.c (process_references): Handle weakrefs correctly.
11783 (symtab_remove_unreachable_nodes): Likewise.
11784 * trans-mem.c (get_cg_data): Update.
11785 (ipa_tm_create_version_alias): Update.
11786 (ipa_tm_execute): Update.
11787 * symtab.c (dump_symtab_base): Dump aliases.
11788 (verify_symtab_base): Verify aliases.
11789 (symtab_node_availability): New function.
11790 (symtab_alias_ultimate_target): Simplify.
11791 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
11792 handle all the fixup cases.
11793 (symtab_resolve_alias): New function.
11794 * passes.c (ipa_write_summaries): Handle weakrefs.
11795 * varpool.c (varpool_analyze_node): Simplify.
11796 (assemble_aliases): Update.
11797 (varpool_create_variable_alias): Simplify.
11798 (varpool_extra_name_alias): Simplify.
11799 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
11800 (lto_symtab_merge_symbols): ... this one.
11802 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
11805 PR rtl-optimization/57268
11806 * sched-deps.c (sched_analyze_2): Flush dependence lists if
11807 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
11809 2013-06-01 Tobias Burnus <burnus@net-b.de>
11811 Partially reverted:
11812 2013-05-31 Tobias Burnus <burnus@net-b.de>
11814 PR middle-end/57073
11815 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
11818 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
11820 PR rtl-optimization/57268
11821 * sched-deps.c (sched_analyze_2): Flush dependence lists if
11822 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
11824 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
11826 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
11827 unordered comparison operators when -fno-trapping-math is in effect
11829 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
11830 and implement unordered comparison operators properly on the e500.
11832 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
11834 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
11835 for constant scalar integers.
11836 (simplify_relational_operation_1): Likewise.
11838 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
11840 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
11841 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
11844 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
11845 Igor Zamyatin <igor.zamyatin@intel.com>
11847 Silvermont (SLM) architecture performance tuning.
11848 * config/i386/i386.h (enum ix86_tune_indices): Add
11849 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
11850 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
11852 * config/i386/i386.c (initial_ix86_tune_features)
11853 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
11854 (ix86_lea_outperforms): Handle Silvermont tuning.
11855 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
11857 (ix86_use_lea_for_mov): Likewise.
11858 (ix86_avoid_lea_for_addr): Likewise.
11859 (ix86_lea_for_add_ok): Likewise.
11860 (exact_dependency_1): New function.
11861 (exact_store_load_dependency): Likewise.
11862 (ix86_adjust_cost): Handle Silvermont tuning.
11863 (do_reoder_for_imul): Likewise.
11864 (swap_top_of_ready_list): New function.
11865 (ix86_sched_reorder): Changed to handle Silvermont tuning.
11867 * config/i386/i386.md (peepholes that split memory operand in fp
11870 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
11872 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
11873 Remove un-necessary braces.
11875 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
11877 * config/aarch64/aarch64.c (aarch64_classify_symbol):
11878 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
11880 2013-05-31 Tobias Burnus <burnus@net-b.de>
11882 PR middle-end/57073
11883 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
11885 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11888 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
11889 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
11890 * config/arm/neon.md (iordi3_neon): Remove.
11891 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
11892 * config/arm/predicates.md (imm_for_neon_logic_operand):
11893 Move to earlier in the file.
11894 (neon_logic_op2): Likewise.
11895 (arm_iordi_operand_neon): New predicate.
11897 2013-05-31 Richard Biener <rguenther@suse.de>
11899 PR tree-optimization/57478
11900 PR tree-optimization/57453
11901 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
11904 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
11906 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
11907 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
11909 2013-05-30 Tobias Burnus <burnus@net-b.de>
11910 Thomas Koenig <tkoenig@gcc.gnu.org>
11912 PR middle-end/57073
11913 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
11914 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
11916 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
11918 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
11920 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
11922 * target.def (register_usage_leveling_p): New hook.
11923 * targhooks.c (default_register_usage_leveling_p): New.
11924 * targhooks.h (default_register_usage_leveling_p): New prototype.
11925 * lra-assigns.c (register_usage_leveling_p): Use the hook.
11926 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
11927 * doc/tm.texi: Update.
11928 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
11930 2013-05-30 Ian Bolton <ian.bolton@arm.com>
11932 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
11933 (*insv_reg<mode>): New define_insn.
11935 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
11937 PR rtl-optimization/57439
11938 * postreload.c (move2add_valid_value_p): Check that we have
11939 a zero subreg_regno_offset when accessing the register in
11940 the requested mode.
11942 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
11943 Igor Zamyatin <igor.zamyatin@intel.com>
11945 Silvermont (SLM) architecture pipeline model, tuning and
11947 * config.gcc: Add slm config options and target.
11949 * config/i386/slm.md: New.
11951 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
11953 * config/i386/i386-c.c (ix86_target_macros_internal): New case
11955 (ix86_target_macros_internal): Likewise.
11957 * config/i386/i386.c (slm_cost): New cost.
11958 (m_SLM): New macro flag.
11959 (initial_ix86_tune_features): Set m_SLM.
11960 (x86_accumulate_outgoing_args): Likewise.
11961 (x86_arch_always_fancy_math_387): Likewise.
11962 (processor_target_table): Add slm cost.
11963 (cpu_names): Add slm cpu name.
11964 (x86_option_override_internal): Set SLM ISA.
11965 (ix86_issue_rate): New case PROCESSOR_SLM.
11966 (ia32_multipass_dfa_lookahead): Likewise.
11967 (fold_builtin_cpu): Add slm.
11969 * config/i386/i386.h (TARGET_SLM): New target macro.
11970 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
11971 (processor_type): Add PROCESSOR_SLM.
11973 * config/i386/i386.md (cpu): Add new value "slm".
11974 (slm.md): Include slm.md.
11976 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
11977 Zhenqiang Chen <zhenqiang.chen@linaro.org>
11979 * config/arm/arm-protos.h: Add and update function protos.
11980 * config/arm/arm.c (use_simple_return_p): New added.
11981 (thumb2_expand_return): Check simple_return flag.
11982 * config/arm/arm.md: Add simple_return and conditional simple_return.
11983 * config/arm/iterators.md: Add iterator for return and simple_return.
11985 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
11987 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
11988 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
11989 (arm_emit_vfp_multi_reg_pop): Likewise.
11990 (thumb2_emit_ldrd_pop): Likewise.
11991 (arm_expand_epilogue): Add misc REG_CFA notes.
11992 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
11994 2013-05-29 Lawrence Crowl <crowl@google.com>
11996 * config/arm/t-arm: Update for below.
11998 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
11999 Change type to hash_table. Update dependent calls and types.
12001 * config/i386/t-cygming: Update for below.
12003 * config/i386/t-interix: Update for below.
12005 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
12006 Change type to hash_table. Update dependent calls and types.
12007 (i386_find_on_wrapper_list::wrappers): Likewise.
12009 * config/ia64/t-ia64: Update for below.
12011 * config/ia64/ia64.c (bundle_state_table):
12012 Change type to hash_table. Update dependent calls and types.
12014 * config/mips/mips.c (mips_reorg_process_insns::htab):
12015 Change type to hash_table. Update dependent calls and types.
12017 * config/sol2.c (solaris_comdat_htab):
12018 Change type to hash_table. Update dependent calls and types.
12020 * config/t-sol2: Update for above.
12022 2013-05-29 Teresa Johnson <tejohnson@google.com>
12024 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
12025 functions are not yet marked as defined.
12027 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
12028 Pat Haugen <pthaugen@us.ibm.com>
12029 Peter Bergner <bergner@vnet.ibm.com>
12031 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
12035 (vrotl<mode>3): Likewise.
12036 (vashl<mode>3): Likewise.
12037 (vlshr<mode>3): Likewise.
12038 (vashr<mode>3): Likewise.
12040 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
12041 support for power8 V2DI builtins.
12043 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
12044 power8 V2DI builtins.
12045 (vupkhsw): Likewise.
12046 (vupklsw): Likewise.
12047 (vaddudm): Likewise.
12048 (vminsd): Likewise.
12049 (vmaxsd): Likewise.
12050 (vminud): Likewise.
12051 (vmaxud): Likewise.
12052 (vpkudum): Likewise.
12053 (vpksdss): Likewise.
12054 (vpkudus): Likewise.
12055 (vpksdus): Likewise.
12060 (vsubudm): Likewise.
12061 (vcmpequd): Likewise.
12062 (vcmpgtsd): Likewise.
12063 (vcmpgtud): Likewise.
12064 (vcmpequd_p): Likewise.
12065 (vcmpgtsd_p): Likewise.
12066 (vcmpgtud_p): Likewise.
12067 (vupkhsw): Likewise.
12068 (vupklsw): Likewise.
12069 (vaddudm): Likewise.
12070 (vmaxsd): Likewise.
12071 (vmaxud): Likewise.
12072 (vminsd): Likewise.
12073 (vminud): Likewise.
12074 (vpksdss): Likewise.
12075 (vpksdus): Likewise.
12076 (vpkudum): Likewise.
12077 (vpkudus): Likewise.
12082 (vsubudm): Likewise.
12084 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
12085 support for power8 V2DI instructions.
12087 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
12088 power8 V2DI instructions. Combine pack and unpack insns to use an
12089 iterator for each mode. Check whether a particular mode supports
12090 Altivec instructions instead of just checking TARGET_ALTIVEC.
12091 (UNSPEC_VPKUWUM): Likewise.
12092 (UNSPEC_VPKSHSS): Likewise.
12093 (UNSPEC_VPKSWSS): Likewise.
12094 (UNSPEC_VPKUHUS): Likewise.
12095 (UNSPEC_VPKSHUS): Likewise.
12096 (UNSPEC_VPKUWUS): Likewise.
12097 (UNSPEC_VPKSWUS): Likewise.
12098 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
12099 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
12100 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
12101 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
12102 (UNSPEC_VUPKHSB): Likewise.
12103 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
12104 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
12105 (UNSPEC_VUPKHSH): Likewise.
12106 (UNSPEC_VUPKLSB): Likewise.
12107 (UNSPEC_VUPKLSH): Likewise.
12109 (VI_char): Likewise.
12110 (VI_scalar): Likewise.
12111 (VI_unit): Likewise.
12113 (VP_small): Likewise.
12114 (VP_small_lc): Likewise.
12115 (VU_char): Likewise.
12116 (add<mode>3): Likewise.
12117 (altivec_vaddcuw): Likewise.
12118 (altivec_vaddu<VI_char>s): Likewise.
12119 (altivec_vadds<VI_char>s): Likewise.
12120 (sub<mode>3): Likewise.
12121 (altivec_vsubcuw): Likewise.
12122 (altivec_vsubu<VI_char>s): Likewise.
12123 (altivec_vsubs<VI_char>s): Likewise.
12124 (altivec_vavgs<VI_char>): Likewise.
12125 (altivec_vcmpbfp): Likewise.
12126 (altivec_eq<mode>): Likewise.
12127 (altivec_gt<mode>): Likewise.
12128 (altivec_gtu<mode>): Likewise.
12129 (umax<mode>3): Likewise.
12130 (smax<mode>3): Likewise.
12131 (umin<mode>3): Likewise.
12132 (smin<mode>3): Likewise.
12133 (altivec_vpkuhum): Likewise.
12134 (altivec_vpkuwum): Likewise.
12135 (altivec_vpkshss): Likewise.
12136 (altivec_vpkswss): Likewise.
12137 (altivec_vpkuhus): Likewise.
12138 (altivec_vpkshus): Likewise.
12139 (altivec_vpkuwus): Likewise.
12140 (altivec_vpkswus): Likewise.
12141 (altivec_vpks<VI_char>ss): Likewise.
12142 (altivec_vpks<VI_char>us): Likewise.
12143 (altivec_vpku<VI_char>us): Likewise.
12144 (altivec_vpku<VI_char>um): Likewise.
12145 (altivec_vrl<VI_char>): Likewise.
12146 (altivec_vsl<VI_char>): Likewise.
12147 (altivec_vsr<VI_char>): Likewise.
12148 (altivec_vsra<VI_char>): Likewise.
12149 (altivec_vsldoi_<mode>): Likewise.
12150 (altivec_vupkhsb): Likewise.
12151 (altivec_vupkhs<VU_char>): Likewise.
12152 (altivec_vupkls<VU_char>): Likewise.
12153 (altivec_vupkhsh): Likewise.
12154 (altivec_vupklsb): Likewise.
12155 (altivec_vupklsh): Likewise.
12156 (altivec_vcmpequ<VI_char>_p): Likewise.
12157 (altivec_vcmpgts<VI_char>_p): Likewise.
12158 (altivec_vcmpgtu<VI_char>_p): Likewise.
12159 (abs<mode>2): Likewise.
12160 (vec_unpacks_hi_v16qi): Likewise.
12161 (vec_unpacks_hi_v8hi): Likewise.
12162 (vec_unpacks_lo_v16qi): Likewise.
12163 (vec_unpacks_hi_<VP_small_lc>): Likewise.
12164 (vec_unpacks_lo_v8hi): Likewise.
12165 (vec_unpacks_lo_<VP_small_lc>): Likewise.
12166 (vec_pack_trunc_v8h): Likewise.
12167 (vec_pack_trunc_v4si): Likewise.
12168 (vec_pack_trunc_<mode>): Likewise.
12170 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
12172 (vec_vmaxsd): Likewise.
12173 (vec_vmaxud): Likewise.
12174 (vec_vminsd): Likewise.
12175 (vec_vminud): Likewise.
12176 (vec_vpksdss): Likewise.
12177 (vec_vpksdus): Likewise.
12178 (vec_vpkudum): Likewise.
12179 (vec_vpkudus): Likewise.
12180 (vec_vrld): Likewise.
12181 (vec_vsld): Likewise.
12182 (vec_vsrad): Likewise.
12183 (vec_vsrd): Likewise.
12184 (vec_vsubudm): Likewise.
12185 (vec_vupkhsw): Likewise.
12186 (vec_vupklsw): Likewise.
12188 2013-05-29 Jan Hubicka <jh@suse.cz>
12190 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
12191 flags; reorder rest of fields in more consistent way.
12192 (varpool_node): Remove analyzed, finalized and alias.
12193 (cgraph_ndoe): Likewise.
12194 (symtab_alias_ultimate_target): New function.
12195 (cgraph_function_node): Move offline.
12196 (cgraph_reset_node): Declare.
12197 (cgraph_comdat_can_be_unshared_p): Remove.
12198 (varpool_remove_initializer): Declare.
12199 (varpool_first_defined_variable, varpool_next_defined_variable
12200 cgraph_first_defined_function, cgraph_next_defined_function): Update.
12201 (cgraph_function_with_gimple_body_p): Update.
12202 (varpool_all_refs_explicit_p): Update.
12203 (symtab_alias_target): New function.
12204 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
12205 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
12206 (cgraph_function_or_thunk_node): Simplify using
12207 symtab_alias_ultimate_target.
12208 (varpool_variable_node): Likewise.
12209 * cgraph.c (cgraph_create_function_alias): Update.
12210 (cgraph_add_thunk): Update.
12211 (cgraph_remove_node): Update.
12212 (dump_cgraph_node): Do not dump removed flags.
12213 (cgraph_function_body_availability): Update.
12214 (cgraph_propagate_frequency): Update.
12215 (verify_cgraph_node): Check sanity of local flag.
12216 (cgraph_function_node): Move here from cgraph.h; revamp for
12217 cgraph_function_or_thunk_node.
12218 * lto-symtab.c (lto_varpool_replace_node): Update.
12219 (lto_symtab_resolve_can_prevail_p): Update.
12220 (lto_symtab_merge_cgraph_nodes): Update.
12221 * ipa-cp.c (determine_versionability, initialize_node_lattices,
12222 propagate_constants_accross_call, devirtualization_time_bonus,
12223 ipcp_propagate_stage): Update.
12224 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
12225 * ipa-inline-transform.c (clone_inlined_nodes,
12226 preserve_function_body_p): Update.
12227 * ipa-reference.c (propagate): Update.
12228 (write_node_summary_p): Update.
12229 * toplev.c (wrapup_global_declaration_2): Update.
12230 * cgraphunit.c (cgraph_analyze_function): Rename to ...
12231 (analyze_function) ... this one.
12232 (cgraph_process_new_functions): Update.
12233 (cgraph_reset_node): Export.
12234 (cgraph_finalize_function): Update.
12235 (cgraph_add_new_function): Update.
12236 (process_function_and_variable_attributes): Update.
12237 (varpool_finalize_decl): Update.
12238 (symbol_finalized): Remove.
12239 (symbol_finalized_and_needed): Rename to ...
12240 (symbol_defined_and_needed): ... update.
12241 (cgraph_analyze_functions): Update.
12242 (handle_alias_pairs): Update.
12243 (mark_functions_to_output): Update.
12244 (assemble_thunk): Update.
12245 (output_in_order): Update.
12246 (output_weakrefs): Update.
12247 (finalize_compilation_unit): Update.
12248 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
12249 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
12250 input_node, input_varpool_node): Update.
12251 * dbxout.c (dbxout_expand_expr): Update.
12252 * cgraphclones.c (cgraph_clone_node): Update.
12253 (cgraph_copy_node_for_versioning): Update.
12254 (cgraph_materialize_clone): Update.
12255 (cgraph_materialize_all_clones): Update.
12256 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
12257 propagate_pure_const, propagate_nothrow): Update.
12258 * lto-streamer-out.c (lto_output, write_symbol): Update.
12259 * ipa-utils.c (ipa_reverse_postorder): Update.
12260 * ipa-inline.c (can_inline_edge_p): Update.
12261 (update_caller_keys, ipa_inline): Update.
12262 * dwarf2out.c (reference_to_unused,
12263 premark_types_used_by_global_vars_helper): Update.
12264 * tree-eh.c (tree_could_trap_p): Update.
12265 * ipa-split.c (consider_split, execute_split_functions): Update.
12266 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
12267 has_addr_references_p): Update; move ahead in file for better
12269 (process_references): Simplify.
12270 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
12271 bodies are removed.
12272 (cgraph_comdat_can_be_unshared_p): Make static.
12273 (cgraph_externally_visible_p): Update.
12274 (varpool_externally_visible_p): Update.
12275 (function_and_variable_visibility): Update.
12276 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
12277 ipa_tm_mark_force_output_node): Update.
12278 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
12279 estimate_edge_devirt_benefit, inline_generate_summary,
12280 inline_write_summary): Update.
12281 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
12282 * ipa-prop.c (ipa_compute_jump_functions): Update.
12283 (ipa_print_node_params, ipa_prop_read_section,
12284 ipa_update_after_lto_read, read_replacements_section): Update.
12285 * varasm.c (mark_decl_referenced): Update.
12286 (assemble_alias, dump_tm_clone_pairs): Update.
12287 * tree-inline.c (copy_bb): Update.
12288 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
12290 * symtab.c (dump_symtab_base): Print new flags.
12291 (verify_symtab_base): Verify new flags.
12292 (symtab_alias_ultimate_target): New function.
12293 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
12294 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
12296 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
12298 * i386.c (ix86_get_function_versions_dispatcher,
12299 ix86_generate_version_dispatcher_body): Update.
12300 (fold_builtin_cpu): Use varpool_add_new_variable.
12301 * varpool.c (varpool_remove_initializer): Break out from ...
12302 (varpool_remove_node): ... this one.
12303 (dump_varpool_node, varpool_node_for_asm,
12304 cgraph_variable_initializer_availability, varpool_analyze_node,
12305 varpool_assemble_decl, varpool_remove_unreferenced_decls,
12306 varpool_finalize_named_section_flags, varpool_create_variable_alias):
12309 2013-05-29 Jan Hubicka <jh@suse.cz>
12311 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
12313 2013-05-29 Easwaran Raman <eraman@google.com>
12315 PR tree-optimization/57442
12316 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
12317 when control exits the main loop.
12319 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
12321 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
12323 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
12324 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
12325 * rx/t-rx: Add rx100 under multi library matches option for nofpu
12328 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12330 PR tree-optimization/57441
12331 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
12332 Don't limit size of incr_vec to number of candidates.
12334 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
12336 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
12337 and mips16 directories.
12338 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
12339 (MULTILIB_DIRNAMES): Ditto.
12340 (MULTILIB_EXCEPTIONS): Add new exceptions.
12341 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
12342 (MULTILIB_DIRNAMES): Ditto.
12343 (MULTILIB_EXCEPTIONS): Add new exceptions.
12345 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
12346 Marcus Shawcroft <marcus.shawcroft@arm.com>
12348 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
12349 SYMBOL_TINY_ABSOLUTE.
12350 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
12351 SYMBOL_TINY_ABSOLUTE.
12352 (aarch64_expand_mov_immediate): Likewise.
12353 (aarch64_classify_symbol): Likewise.
12354 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
12355 Permit SYMBOL_TINY_ABSOLUTE.
12356 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
12358 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
12359 Marcus Shawcroft <marcus.shawcroft@arm.com>
12361 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
12362 Refactor if/switch. Replace gcc_assert with if.
12364 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
12366 * config/i386/i386.c (initial_ix86_tune_features): Enable
12367 FP Reassociation for AMD bdver1 and bdver2.
12369 2013-05-29 Martin Jambor <mjambor@suse.cz>
12371 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
12372 and IMAGPART_EXPR do not occur within other handled_components.
12374 2013-05-29 Richard Biener <rguenther@suse.de>
12376 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
12377 access on whether the use is in the BB we currently try to
12379 (vect_bb_vectorization_profitable_p): Pass the BB we currently
12380 vectorize to vect_bb_slp_scalar_cost.
12382 2013-05-29 Richard Biener <rguenther@suse.de>
12384 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
12385 computing scalar cost offsetted by stmts that are kept live
12387 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
12388 for computation of scalar cost.
12390 2013-05-28 Steve Ellcey <sellcey@mips.com>
12392 * config/mips/mips-cpus.def (mips32r2): Change processor type.
12394 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
12396 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
12397 array notation built-in reduction functions.
12398 * doc/passes.texi (Passes): Added documentation about changes done
12400 * doc/invoke.texi (C Dialect Options): Added documentation about
12401 the -fcilkplus flag.
12402 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
12403 (BUILTINS_DEF): Depend on cilkplus.def.
12404 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
12405 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
12406 * cilkplus.def: New file.
12408 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
12410 PR rtl-optimization/57439
12411 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
12413 2013-05-28 Easwaran Raman <eraman@google.com>
12415 PR tree-optimization/57337
12416 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
12417 (find_insert_point): Correctly identify the insertion point
12418 when two statements with the same UID is compared.
12420 2013-05-28 Richard Biener <rguenther@suse.de>
12422 PR tree-optimization/56787
12423 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
12424 from the list of data references.
12425 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
12427 (vect_analyze_loop_operations): Likewise.
12428 (vect_transform_loop): Remove clobbers.
12430 2013-05-28 Martin Jambor <mjambor@suse.cz>
12432 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
12433 and REALPART_EXPRs have scalar type.
12435 2013-05-28 Richard Biener <rguenther@suse.de>
12437 PR tree-optimization/57411
12438 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
12440 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
12441 virtual operand propagation.
12443 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
12445 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
12446 destination register for bmasksi_vis.
12447 (vector_init_bshuffle): Likewise.
12448 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
12450 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
12452 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
12453 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
12454 mode if the instruction isn't available in the original mode.
12455 * config/sparc/sparc.opt (mfix-ut699): New option.
12456 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
12457 (divdf3): Turn into expander.
12458 (divdf3_nofix): New insn.
12459 (divdf3_fix): Likewise.
12460 (divsf3): Disable if -mfix-ut699.
12461 (sqrtdf2): Turn into expander.
12462 (sqrtdf2_nofix): New insn.
12463 (sqrtdf2_fix): Likewise.
12464 (sqrtsf2): Disable if -mfix-ut699.
12466 2013-05-27 Richard Biener <rguenther@suse.de>
12468 PR middle-end/57412
12469 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
12470 block for the new loop.
12472 2013-05-27 Richard Biener <rguenther@suse.de>
12474 PR tree-optimization/57343
12475 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
12476 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
12477 (number_of_iterations_cond): Do not build the folded tree.
12479 2013-05-27 Richard Biener <rguenther@suse.de>
12482 PR middle-end/57381
12483 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
12484 OEP_CONSTANT_ADDRESS_OF retained.
12486 PR tree-optimization/57417
12487 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
12488 for unchanged base.
12489 (set_ssa_val_to): Compare addresses using
12490 get_addr_base_and_unit_offset.
12492 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
12494 PR rtl-optimization/56833
12495 * postreload.c (move2add_record_mode): New function.
12496 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
12497 (move2add_use_add2_insn): Use move2add_record_sym_value.
12498 (move2add_use_add3_insn): Likewise.
12499 (reload_cse_move2add): Use move2add_valid_value_p and
12500 move2add_record_mode. Invalidate call-clobbered and REG_INC
12501 affected regs by setting reg_mode to VOIDmode.
12502 (move2add_note_store): Don't pretend the inside of a SUBREG is
12503 the actual destination. Invalidate single/leading registers by
12504 setting reg_mode to VOIDmode.
12505 Use move2add_record_sym_value, move2add_valid_value_p and
12506 move2add_record_mode.
12508 2013-05-27 Richard Biener <rguenther@suse.de>
12510 PR tree-optimization/57396
12511 * tree-affine.c (double_int_constant_multiple_p): Properly
12512 return false for val == 0 and div != 0.
12514 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
12516 * config/mips/mips.h: Use #elif in preprocessor conditions.
12518 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
12521 * config/mips/constraints.md (kl): New constraint.
12522 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
12523 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
12524 constraint for operand 0. Split after CSE for MIPS16. Emit a move
12525 from LO for MIPS16.
12526 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
12528 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
12531 * config/mips/mips.c (mips_can_inline_p): New function.
12532 (TARGET_CAN_INLINE_P): Define.
12534 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
12536 * sched-int.h (ds_t, dw_t): Make unsigned int.
12537 Fix documentation that describes how all the ds_t bits are used.
12538 Reserve the last bit for delayed-branch scheduling.
12539 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
12540 (BITS_PER_DEP_WEAK): Fix definition and documentation.
12541 (gen_dep_weak_1): Remove prototype.
12542 * sched-deps.c (get_dep_weak_1): Make static.
12543 * target.def (speculate_insn, needs_block_p, gen_spec_check,
12544 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
12545 * doc/tm.texi: Regenerate.
12546 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
12548 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
12551 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
12553 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
12554 Sandra Loosemore <sandra@codesourcery.com>
12556 * config.gcc (powerpc-*): Allow native for with-cpu.
12558 2013-05-24 Jeff Law <law@redhat.com>
12560 PR tree-optimization/57124
12561 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
12562 conversion feeding a condition if the range has an overflow
12563 if -fstrict-overflow. Add warnings for when we do make the
12566 2013-05-24 Dehao Chen <dehao@google.com>
12568 * tree-cfg.c (locus_discrim_map): Fix the typo.
12569 (locus_discrim_hasher): Likewise.
12570 (locus_discrim_hasher::hash): Likewise.
12571 (locus_discrim_hasher::equal): Likewise.
12573 2013-05-24 Martin Jambor <mjambor@suse.cz>
12575 PR tree-optimization/57294
12576 * cgraph.h (ipa_record_stmt_references): Declare.
12577 * cgraphbuild.c (ipa_record_stmt_references): New function.
12578 (build_cgraph_edges): Use ipa_record_stmt_references.
12579 (rebuild_cgraph_edges): Likewise.
12580 (cgraph_rebuild_references): Likewise.
12581 * ipa-prop.c (ipa_modify_call_arguments): Discard references
12582 associated with the old statement and build references from the
12583 newly built statements.
12584 * ipa-ref.c (ipa_remove_stmt_references): New function.
12585 * ipa-ref.h (ipa_remove_stmt_references): Declare.
12587 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
12589 * lra-constraints.c (emit_spill_move): Use smaller mode for
12591 (check_and_process_move): Consider mem-reg moves for secondary
12593 (curr_insn_transform): Don't lose insns emitted before for
12594 secondary memory moves.
12595 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
12596 reg set up in the current insn.
12598 2013-05-24 Dehao Chen <dehao@google.com>
12600 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
12602 (locus_descrim_hasher::equal): Likewise.
12603 (build_gimple_cfg): New discriminator assignment algorithm.
12604 (make_edges): Likewise.
12605 (next_discriminator_for_locus): Likewise.
12606 (same_line_p): Likewise.
12607 (assign_discriminators): Likewise.
12608 (make_cond_expr_edges): Likewise.
12609 (make_gimple_switch_edges): Likewise.
12610 (make_goto_expr_edges): Likewise.
12611 (make_gimple_asm_edges): Likewise.
12613 2013-05-24 Ian Bolton <ian.bolton@arm.com>
12615 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
12616 X format specifier to only display bottom 16 bits.
12617 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
12618 immediate to match for operand 2, since it will be masked.
12620 2013-05-24 Richard Biener <rguenther@suse.de>
12622 PR tree-optimization/57287
12623 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
12624 all SSA names that occur in abnormal PHIs.
12626 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
12629 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
12630 that index is not negative.
12632 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
12634 PR rtl-optimization/55177
12635 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
12636 (simplify_byte_swapping_operation): New.
12637 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
12638 (simplify_relational_operation_1): Deal with BSWAP.
12640 2013-05-23 Richard Henderson <rth@redhat.com>
12643 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
12644 (ix86_reorg): Call it.
12646 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
12649 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
12650 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
12651 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
12653 2013-05-23 Christian Bruel <christian.bruel@st.com>
12656 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
12658 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
12659 Marcus Shawcroft <marcus.shawcroft@arm.com>
12661 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
12662 * config/aarch64/constraints.md (Usa): Remove.
12663 * doc/md.texi (AArch64 Usa): Remove.
12665 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
12666 Marcus Shawcroft <marcus.shawcroft@arm.com>
12668 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
12669 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
12670 * config/aarch64/predicates.md (aarch64_const_address): Remove.
12671 (aarch64_mov_operand): Use aarch64_mov_operand_p.
12673 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
12675 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
12676 instruction (AdvSIMD).
12677 * config/aarch64/aarch64-builtins.c
12678 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
12679 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
12681 2013-05-23 Martin Jambor <mjambor@suse.cz>
12683 PR middle-end/57347
12684 * tree.h (contains_bitfld_component_ref_p): Declare.
12685 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
12686 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
12688 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
12689 not access a bit-field. Assert all final offsets are byte-aligned.
12691 2013-05-23 Richard Biener <rguenther@suse.de>
12693 PR tree-optimization/57380
12694 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
12695 least one invariant or re-used load.
12696 * passes.c (init_optimization_passes): Move pass_phiprop before
12699 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
12701 * config/aarch64/aarch64-simd.md
12702 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
12704 2013-05-23 Richard Biener <rguenther@suse.de>
12706 PR middle-end/57381
12707 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
12708 OEP_CONSTANT_ADDRESS_OF retained.
12710 2013-05-23 Jakub Jelinek <jakub@redhat.com>
12712 PR middle-end/57344
12713 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
12714 don't lower unit. Handle unit not being always BITS_PER_WORD.
12716 2013-05-23 Richard Biener <rguenther@suse.de>
12718 PR rtl-optimization/57341
12719 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
12720 instead of true_dependence.
12722 2013-05-22 David Malcolm <dmalcolm@redhat.com>
12724 * bb-reorder.c (branch_threshold): Make const.
12725 (exec_threshold): Ditto.
12727 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
12728 Pat Haugen <pthaugen@us.ibm.com>
12729 Peter Bergner <bergner@vnet.ibm.com>
12731 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
12732 documentation for the power8 crypto builtins.
12734 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
12736 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
12737 macros for defining power8 builtin functions.
12738 (BU_P8V_AV_2): Likewise.
12739 (BU_P8V_AV_P): Likewise.
12740 (BU_P8V_VSX_1): Likewise.
12741 (BU_P8V_OVERLOAD_1): Likewise.
12742 (BU_P8V_OVERLOAD_2): Likewise.
12743 (BU_CRYPTO_1): Likewise.
12744 (BU_CRYPTO_2): Likewise.
12745 (BU_CRYPTO_3): Likewise.
12746 (BU_CRYPTO_OVERLOAD_1): Likewise.
12747 (BU_CRYPTO_OVERLOAD_2): Likewise.
12748 (XSCVSPDP): Fix typo, point to the correct instruction.
12749 (VCIPHER): Add power8 crypto builtins.
12750 (VCIPHERLAST): Likewise.
12751 (VNCIPHER): Likewise.
12752 (VNCIPHERLAST): Likewise.
12753 (VPMSUMB): Likewise.
12754 (VPMSUMH): Likewise.
12755 (VPMSUMW): Likewise.
12756 (VPERMXOR_V2DI): Likewise.
12757 (VPERMXOR_V4SI: Likewise.
12758 (VPERMXOR_V8HI: Likewise.
12759 (VPERMXOR_V16QI: Likewise.
12760 (VSHASIGMAW): Likewise.
12761 (VSHASIGMAD): Likewise.
12762 (VPMSUM): Likewise.
12763 (VPERMXOR): Likewise.
12764 (VSHASIGMA): Likewise.
12766 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
12767 __CRYPTO__ if the crypto instructions are available.
12768 (altivec_overloaded_builtins): Add support for overloaded power8
12771 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
12772 support for power8 crypto builtins.
12773 (builtin_function_type): Likewise.
12774 (altivec_init_builtins): Add support for builtins that take vector
12775 long long (V2DI) arguments.
12777 * config/rs6000/crypto.md: New file, define power8 crypto
12780 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
12781 Pat Haugen <pthaugen@us.ibm.com>
12782 Peter Bergner <bergner@vnet.ibm.com>
12784 * doc/invoke.texi (Option Summary): Add power8 options.
12785 (RS/6000 and PowerPC Options): Likewise.
12787 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
12788 constraints.md instead of rs6000.h. Reorder w* constraints. Add
12789 wm, wn, wr documentation.
12791 * config/rs6000/constraints.md (wm): New constraint for VSX
12792 registers if direct move instructions are enabled.
12793 (wn): New constraint for no registers.
12794 (wq): New constraint for quad word even GPR registers.
12795 (wr): New constraint if 64-bit instructions are enabled.
12796 (wv): New constraint if power8 vector instructions are enabled.
12797 (wQ): New constraint for quad word memory locations.
12799 * config/rs6000/predicates.md (const_0_to_15_operand): New
12800 constraint for 0..15 for crypto instructions.
12801 (gpc_reg_operand): If VSX allow registers in VSX registers as well
12802 as GPR and floating point registers.
12803 (int_reg_operand): New predicate to match only GPR registers.
12804 (base_reg_operand): New predicate to match base registers.
12805 (quad_int_reg_operand): New predicate to match even GPR registers
12806 for quad memory operations.
12807 (vsx_reg_or_cint_operand): New predicate to allow vector logical
12808 operations in both GPR and VSX registers.
12809 (quad_memory_operand): New predicate for quad memory operations.
12810 (reg_or_indexed_operand): New predicate for direct move support.
12812 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
12813 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
12814 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
12815 (POWERPC_MASKS): Add power8 options.
12816 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
12819 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
12820 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
12822 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
12823 (-mpower8-fusion): New power8 options.
12824 (-mpower8-fusion-sign): Likewise.
12825 (-mpower8-vector): Likewise.
12826 (-mcrypto): Likewise.
12827 (-mdirect-move): Likewise.
12828 (-mquad-memory): Likewise.
12830 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
12831 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
12833 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
12834 (rs6000_debug_vector_unit): Add p8_vector.
12835 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
12836 definitions. Also print fusion state.
12837 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
12838 (rs6000_builtin_mask_calculate): Add power8 builtin support.
12839 (rs6000_option_override_internal): Add support for power8.
12840 (rs6000_common_init_builtins): Add debugging for skipped builtins
12841 if -mdebug=builtin.
12842 (rs6000_adjust_cost): Add power8 support.
12843 (rs6000_issue_rate): Likewise.
12844 (insn_must_be_first_in_group): Likewise.
12845 (insn_must_be_last_in_group): Likewise.
12846 (force_new_group): Likewise.
12847 (rs6000_register_move_cost): Likewise.
12848 (rs6000_opt_masks): Likewise.
12850 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
12851 power8 capable assembler, default to power7 options.
12852 (TARGET_DIRECT_MOVE): Likewise.
12853 (TARGET_CRYPTO): Likewise.
12854 (TARGET_P8_VECTOR): Likewise.
12855 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
12856 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
12857 (VECTOR_MEM_P8_VECTOR_P): Likewise.
12858 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
12859 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
12860 (TARGET_XSCVDPSPN): Likewise.
12861 (TARGET_XSCVSPDPN): Likewsie.
12862 (TARGET_SYNC_HI_QI): Likewise.
12863 (TARGET_SYNC_TI): Likewise.
12864 (MASK_CRYPTO): Likewise.
12865 (MASK_DIRECT_MOVE): Likewise.
12866 (MASK_P8_FUSION): Likewise.
12867 (MASK_P8_VECTOR): Likewise.
12868 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
12869 temporary used by some of the direct move instructions to get two FP
12870 temporary registers does not force creation of a stack frame.
12871 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
12872 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
12873 that any VSX registers are tieable, even if they are also an
12874 Altivec vector mode.
12875 (r6000_reg_class_enum): Add wm, wr, wv constraints.
12876 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
12877 (RS6000_BTM_CRYPTO): Likewise.
12878 (RS6000_BTM_COMMON): Likewise.
12880 * config/rs6000/rs6000.md (cpu attribute): Add power8.
12881 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
12882 (enum rs6000_vector): Add power8 vector support.
12884 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
12888 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
12889 (any_sibcall_could_use_r3): this and handle indirect calls.
12890 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
12892 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12894 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
12896 2013-05-22 Richard Biener <rguenther@suse.de>
12898 PR middle-end/57349
12899 * profile.c (branch_prob): Do not split blocks that are
12900 abnormally receiving from ECF_RETURNS_TWICE functions.
12902 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
12904 * recog.c (offsettable_address_addr_space_p): Fix calculation of
12905 address mode. Move pointer mode initialization to the same place.
12907 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
12909 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
12910 while it has any effect.
12912 2013-05-21 Easwaran Raman <eraman@google.com>
12914 PR tree-optimization/57322
12915 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
12916 UID of the statement added to the BB to be 1.
12918 2013-05-21 Jakub Jelinek <jakub@redhat.com>
12920 PR tree-optimization/57331
12921 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
12922 of conversion from pointer type to integral type with integer.
12924 2013-05-21 Martin Jambor <mjambor@suse.cz>
12927 * ipa-prop.c (ipa_read_node_info): Process param_used and
12928 controlled_uses in the same order as when writing.
12930 2013-05-21 Magnus Granberg <baldrick@free.fr>
12933 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
12935 2013-05-21 Richard Biener <rguenther@suse.de>
12937 PR tree-optimization/57318
12938 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
12939 estimate stmts with side-effects as likely eliminated.
12941 2013-05-21 Richard Biener <rguenther@suse.de>
12943 PR tree-optimization/57330
12944 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
12945 preserve the call stmts fntype.
12947 2013-05-21 Richard Biener <rguenther@suse.de>
12949 PR tree-optimization/57303
12950 * tree-ssa-sink.c (statement_sink_location): Improve killing
12951 stmt detection and properly handle self-assignments.
12953 2013-05-21 Christian Bruel <christian.bruel@st.com>
12955 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
12956 spanning registers. LEAF_REG_REMAP is supported only for contiguous
12957 registers. Set register size out of the PARALLEL loop.
12959 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
12962 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
12963 (*fmasf4, *fmasf4_media): New insns.
12965 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
12967 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
12968 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
12969 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
12970 (mips_idiv_insns): Update the comments to say that the returned
12971 instruction counts are in units of BASE_INSN_LENGTH.
12972 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
12973 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
12974 using 2 rather than 4 as the length of indirect MIPS16 and
12975 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
12976 length of a NOP. Don't divide MIPS16 lengths by 2.
12977 (mips16_split_long_branches): Assume a branch is long if the
12978 length is greater than 4 rather than 8.
12979 * config/mips/mips.md (length): Give MIPS16 lengths directly,
12980 rather than multiplying them by 2. Multiply instruction counts
12981 by BASE_INSN_LENGTH rather than 4.
12982 (*jump_mips16, tls_get_tp_mips16_<mode>)
12983 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
12985 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
12987 * config/mips/mips.md (extended_mips16): Remove branch case.
12988 (length): Remove duplicated extended_mips16 test.
12990 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
12992 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
12994 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
12996 * recog.h (Recog_data): Rename to...
12997 (recog_data_d): ...this.
12998 (recog_data): Update accordingly.
12999 * recog.c (recog_data): Likewise.
13000 * reload.c (save_recog_data): Likewise.
13001 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
13002 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
13004 2013-05-17 Julian Brown <julian@codesourcery.com>
13006 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
13007 found in a REG_EQUAL note, invalidate it.
13009 2013-05-17 Easwaran Raman <eraman@google.com>
13011 * tree-ssa-reassoc.c (find_insert_point): New function.
13012 (insert_stmt_after): Likewise.
13013 (get_def_stmt): Likewise.
13014 (ensure_ops_are_available): Likewise.
13015 (not_dominated_by): Likewise.
13016 (rewrite_expr_tree): Do not move statements beyond what is
13017 necessary. Remove call to swap_ops_for_binary_stmt...
13018 (reassociate_bb): ... and move it here.
13019 (build_and_add_sum): Assign UIDs for new statements.
13020 (linearize_expr): Likewise.
13021 (do_reassoc): Renumber gimple statement UIDs.
13023 2013-05-17 Jan Hubicka <jh@suse.cz>
13025 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
13027 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
13028 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
13030 (output_weakrefs): Update.
13032 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
13033 Martin Jambor <mjambor@suse.cz>
13035 PR middle-end/57276
13036 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
13037 value that corresponds to the given aggval is found in values vector.
13039 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
13041 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
13042 sse, sse2, sse3, ssse3 and sse4a flags to options.
13044 2013-05-17 David Malcolm <dmalcolm@redhat.com>
13046 * gengtype-state.c: (s_expr_writer): New class, to handle
13047 prettifying of output layout of s-expressions.
13048 (state_writer): New class, to write out gtype.state.
13049 (state_written_type_count): Move this variable into member data of
13051 (s_expr_writer::s_expr_writer): New code: constructor for new class
13052 (state_writer::state_writer(): ditto
13053 (s_expr_writer::write_new_line): New function
13054 (s_expr_writer::write_any_indent): ditto
13055 (s_expr_writer::begin_s_expr): ditto
13056 (s_expr_writer::end_s_expr): ditto
13057 (write_state_fileloc): convert to method of state_writer...
13058 (state_writer:: write_state_fileloc): ...and use methods of
13059 s_expr_writer to write indentation into the gtype.state output file
13060 to visually represent the hierarchical structure of the list
13062 (write_state_fields): ditto, renaming to...
13063 (state_writer::write_state_fields)
13064 (write_state_a_string): ditto, renaming to...
13065 (state_writer::write_state_a_string)
13066 (write_state_string_option): ditto, renaming to...
13067 (state_writer::write_state_string_option)
13068 (write_state_type_option): ditto, renaming to...
13069 (state_writer::write_state_type_option)
13070 (write_state_nested_option): ditto, renaming to...
13071 (state_writer::write_state_nested_option)
13072 (write_state_option): ditto, renaming to...
13073 (state_writer::write_state_option)
13074 (write_state_options): ditto, renaming to...
13075 (state_writer::write_state_options)
13076 (write_state_lang_bitmap): ditto, renaming to...
13077 (state_writer::write_state_lang_bitmap)
13078 (write_state_version): ditto, renaming to...
13079 (state_writer::write_state_version)
13080 (write_state_scalar_type): ditto, renaming to...
13081 (state_writer::write_state_scalar_type)
13082 (write_state_string_type): ditto, renaming to...
13083 (state_writer::write_state_string_type)
13084 (write_state_undefined_type): ditto, renaming to...
13085 (state_writer::write_state_undefined_type)
13086 (write_state_struct_union_type): ditto, renaming to...
13087 (state_writer::write_state_struct_union_type)
13088 (write_state_struct_type): ditto, renaming to...
13089 (state_writer::write_state_struct_type)
13090 (write_state_user_struct_type): ditto, renaming to...
13091 (state_writer::write_state_user_struct_type)
13092 (write_state_lang_struct_type): ditto, renaming to...
13093 (state_writer::write_state_lang_struct_type)
13094 (write_state_param_struct_type): ditto, renaming to...
13095 (state_writer::write_state_param_struct_type)
13096 (write_state_pointer_type): ditto, renaming to...
13097 (state_writer::write_state_pointer_type)
13098 (write_state_array_type): ditto, renaming to...
13099 (state_writer::write_state_array_type)
13100 (write_state_gc_used): ditto, renaming to...
13101 (state_writer::write_state_gc_used)
13102 (write_state_common_type_content): ditto, renaming to...
13103 (state_writer::write_state_common_type_content)
13104 (write_state_type): ditto, renaming to...
13105 (state_writer::write_state_type)
13106 (write_state_pair_list): ditto, renaming to...
13107 (state_writer::write_state_pair_list)
13108 (write_state_pair): ditto, renaming to...
13109 (state_writer::write_state_pair)
13110 (write_state_typedefs): ditto, renaming to...
13111 (state_writer::write_state_typedefs)
13112 (write_state_structures): ditto, renaming to...
13113 (state_writer::write_state_structures)
13114 (write_state_param_structs): ditto, renaming to...
13115 (state_writer::write_state_param_structs)
13116 (write_state_variables): ditto, renaming to...
13117 (state_writer::write_state_variables)
13118 (write_state_srcdir): ditto, renaming to...
13119 (state_writer::write_state_srcdir)
13120 (write_state_files_list): ditto, renaming to...
13121 (state_writer::write_state_files_list)
13122 (write_state_languages): ditto, renaming to...
13123 (state_writer::write_state_languages)
13124 (write_state): create a state_writer instance and use it when
13125 writing out the state file
13127 2013-05-17 Mike Stump <mikestump@comcast.net>
13129 PR rtl-optimization/57304
13130 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
13131 accessing DF_REF_REAL_LOC.
13133 2013-05-17 Jakub Jelinek <jakub@redhat.com>
13135 PR rtl-optimization/57281
13136 PR rtl-optimization/57300
13137 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
13138 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
13139 what the other splitter did if the registers are dead.
13141 2013-05-17 Richard Biener <rguenther@suse.de>
13143 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
13146 2013-05-17 Jakub Jelinek <jakub@redhat.com>
13148 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
13151 2013-05-17 Marek Polacek <polacek@redhat.com>
13153 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
13154 length when doing non-zero store of storing '\0' to '\0'.
13156 2013-05-17 Jakub Jelinek <jakub@redhat.com>
13158 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
13159 vect_external_def oprnd1 with loop_vinfo, try to emit
13160 optional cast, negation and and stmts on the loop preheader
13161 edge instead of into the pattern def seq.
13163 PR tree-optimization/57051
13164 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
13165 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
13167 2013-05-16 Nick Clifton <nickc@redhat.com>
13169 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
13170 (rl78_is_naked_func): New function.
13171 (rl78_expand_prologue): Skip prologue generation for naked functions.
13172 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
13173 * doc/extend.texi (naked): Add RL78 to the list of processors
13174 that supports this attribute.
13176 2013-05-16 Jeff Law <law@redhat.com>
13178 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
13180 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
13182 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
13183 cache parameters using detect_caches_amd also for CYRIX,
13184 NSC and TM2 signatures.
13186 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
13187 Dzianis Kahanovich <mahatma@eu.by>
13191 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
13192 VIA/Centaur processors and determine their cache parameters
13193 using detect_caches_amd.
13195 2013-05-16 Teresa Johnson <tejohnson@google.com>
13197 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
13198 (rtl_verify_edges): New function.
13199 (rtl_verify_bb_insns): Ditto.
13200 (rtl_verify_bb_pointers): Ditto.
13201 (rtl_verify_bb_insn_chain): Ditto.
13202 (rtl_verify_fallthru): Ditto.
13203 (rtl_verify_bb_layout): Ditto.
13204 (rtl_verify_flow_info_1): Outline checks into new functions.
13205 (rtl_verify_flow_info): Ditto.
13207 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
13209 * cfghooks.c (copy_bbs): Add update_dominance argument.
13210 * cfghooks.h (copy_bbs): Update prototype.
13211 * tree-cfg.c (gimple_duplicate_sese_region):
13212 Add update_dominance argument.
13213 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
13214 * tree-ssa-loop-ch.c (copy_loop_headers): Update
13215 gimple_duplicate_sese_region call.
13216 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
13217 Update copy_bbs call.
13218 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
13219 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
13221 2013-05-16 Jakub Jelinek <jakub@redhat.com>
13223 * tree-vectorizer.h (NUM_PATTERNS): Increment.
13224 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
13225 vect_recog_rotate_pattern.
13226 (vect_recog_rotate_pattern): New function.
13228 2013-05-16 Jason Merrill <jason@redhat.com>
13230 * Makefile.in (LLINKER): New variable.
13231 (mostlyclean): Remove link mutex.
13232 * configure.ac: Handle --enable-link-mutex.
13233 * lock-and-run.sh: New script.
13235 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
13238 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
13241 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
13243 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
13245 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
13247 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
13248 * config/arm/arm.c (next_consecutive_mem): New function.
13249 (gen_movmem_ldrd_strd): Likewise.
13250 * config/arm/arm.md (movmemqi): Update condition and code.
13251 (unaligned_loaddi, unaligned_storedi): New patterns.
13253 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
13255 * config.gcc: Obsolete *-*-solaris2.9*.
13256 * doc/install.texi (Specific, *-*-solaris2*): Document it.
13258 2013-05-16 Richard Biener <rguenther@suse.de>
13260 * passes.c (init_optimization_passes): Move pass_parallelize_loops
13261 earlier, after GRAPHITE transforms and IV canonicalization.
13263 2013-05-16 Jakub Jelinek <jakub@redhat.com>
13265 * omp-low.c (extract_omp_for_data): For collapsed loops,
13266 if at least one of the loops is known at compile time to
13267 iterate zero times, set count to 0.
13268 (expand_omp_regimplify_p): New function.
13269 (expand_omp_for_generic): For collapsed loops, if at least
13270 one of the loops isn't known to iterate at least once,
13271 add runtime check with setting count to 0.
13272 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
13273 For unsigned types if it isn't known at compile time that
13274 the loop will iterate at least once, add runtime check to bypass
13275 the whole loop if initial condition isn't true.
13277 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
13279 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
13281 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
13283 PR middle-end/57286
13284 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
13285 transformations to avoid an infinite loop.
13287 2013-05-16 Marek Polacek <polacek@redhat.com>
13289 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
13291 2013-05-15 Leif Ekblad <leif@rdos.net>
13293 * config/i386/i386.c (ix86_decompose_address): Use
13294 DEFAULT_TLS_SEG_REG to access TLS segment register.
13295 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
13296 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
13297 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
13299 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
13302 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
13303 sibling calls to functions that would normally be lazily bound,
13304 unless $gp is call-clobbered.
13306 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
13308 * config/i386/i386.c (ix86_option_override_internal): Update
13309 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
13310 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
13311 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
13312 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
13314 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
13316 2013-05-15 Andreas Schwab <schwab@suse.de>
13318 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
13319 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
13322 2013-05-15 Teresa Johnson <tejohnson@google.com>
13324 * loop-unroll.c (report_unroll_peel): Check decision before
13325 emitting unroll/peel message.
13327 2013-05-15 Teresa Johnson <tejohnson@google.com>
13329 * function.h (has_bb_partition): New rtl_data flag.
13330 (bb_reorder_complete): Ditto.
13331 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
13332 instead of flag_reorder_blocks_and_partition.
13333 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
13334 with some enhancements.
13335 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
13336 * bb-reorder.c (connect_traces): Check for has_bb_partition
13337 instead of flag_reorder_blocks_and_partition.
13338 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
13339 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
13340 verify_hot_cold_block_grouping.
13341 (partition_hot_cold_basic_blocks): Set has_bb_partition.
13343 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
13346 * config/arm/predicates.md (call_insn_operand): New predicate.
13347 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
13348 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
13349 if insn is not a tail call.
13350 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
13352 * config/arm/arm.h (enum reg_class): New caller save register class.
13353 (REG_CLASS_NAMES): Likewise.
13354 (REG_CLASS_CONTENTS): Likewise.
13355 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
13358 2013-05-15 Richard Biener <rguenther@suse.de>
13360 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
13361 of MSG_OPTIMIZED_LOCATIONS.
13362 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
13363 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
13365 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
13366 of MSG_OPTIMIZED_LOCATIONS.
13367 (execute_vect_slp): Likewise.
13368 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
13369 (vect_create_cond_for_alias_checks): Likewise.
13370 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
13371 (vect_recog_widen_mult_pattern): Likewise.
13372 (vect_recog_widen_sum_pattern): Likewise.
13373 (vect_recog_over_widening_pattern): Likewise.
13374 (vect_recog_widen_shift_pattern): Likewise.
13375 (vect_recog_vector_vector_shift_pattern): Likewise.
13376 (vect_recog_divmod_pattern): Likewise.
13377 (vect_recog_mixed_size_cond_pattern): Likewise.
13378 (vect_recog_bool_pattern): Likewise.
13379 (vect_pattern_recog_1): Likewise.
13381 2013-05-15 Martin Jambor <mjambor@suse.cz>
13383 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
13384 non-functions to builtin_unreachable.
13385 * ipa-inline-transform.c (inline_call): Do not assert estimates were
13386 correct when new direct edges were discovered.
13388 2013-05-15 Martin Jambor <mjambor@suse.cz>
13390 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
13391 header, print symbol order instead of node uid, print more information
13392 about indirect edge targets.
13393 (ipa_make_edge_direct_to_target): Print symbol order instead of node
13395 (ipa_make_edge_direct_to_target): Likewise.
13396 (remove_described_reference): Likewise.
13397 (propagate_controlled_uses): Likewise.
13398 (ipa_print_node_params): Also print symbol order.
13399 (ipcp_transform_function): Print symbol order instead of node uids.
13400 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
13401 (cgraph_get_create_real_symbol_node): Likewise.
13402 * ipa-cp.c (print_lattice): Likewise.
13403 (print_all_lattices): Likewise.
13404 (determine_versionability): Likewise.
13405 (initialize_node_lattices): Likewise.
13406 (estimate_local_effects): Likewise.
13407 (update_profiling_info): Likewise.
13408 (create_specialized_node): Likewise.
13409 (perhaps_add_new_callers): Likewise.
13410 (decide_about_value): Likewise.
13411 (decide_whether_version_node): Likewise.
13412 (identify_dead_nodes): Likewise.
13413 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
13414 (dump_inline_summary): Likewise.
13415 (estimate_node_size_and_time): Likewise.
13416 (inline_analyze_function): Likewise.
13417 * ipa-inline.c (report_inline_failed_reason): Likewise.
13418 (want_early_inline_function_p): Likewise.
13419 (edge_badness): Likewise.
13420 (update_edge_key): Likewise.
13421 (inline_small_functions): Likewise. Add dumping of order to two other
13423 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
13424 instead of node uids.
13425 (propagate_pure_const): Likewise.
13426 (propagate_pure_const): Likewise.
13427 * ipa-utils.c (dump_cgraph_node_set): Likewise.
13428 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
13429 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
13431 * tree-pretty-print.c (dump_function_header): Likewise.
13432 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
13433 Print symbol order instead of node uids.
13435 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13437 * config/s390/s390.c (s390_register_move_cost): Don't impose the
13438 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
13440 2013-05-15 Richard Biener <rguenther@suse.de>
13442 PR tree-optimization/57275
13443 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
13444 return value for fail to do runtime alias checks for gather loads.
13446 2013-05-15 Jan Hubicka <jh@suse.cz>
13450 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
13451 weakrefs are not external.
13452 (lto_symtab_merge_decls): Fix thinko when dealing with
13453 non-lto_symtab decls.
13454 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
13455 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
13456 * varpool.c (dump_varpool_node): Dump more flags.
13458 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
13460 * config/i386/i386.c (processor_alias_table): Add instruction
13461 FSGSBASE for AMD bdver3 architecture.
13463 2013-05-14 Jakub Jelinek <jakub@redhat.com>
13465 * tree.c (warn_deprecated_use): Print file:line using locus color.
13466 * diagnostic.c (diagnostic_report_current_module): Print file:line
13467 and file:line:column using locus color.
13469 2013-05-14 Mike Stump <mikestump@comcast.net>
13471 * gdbinit.in: Add __null.
13473 2013-05-14 Mike Stump <mikestump@comcast.net>
13475 * recog.h: Rename struct recog_data to Recog_data.
13476 * recog.c: Likewise.
13477 * reload.c (can_reload_into): Likewise.
13478 * config/picochip/picochip.c: Likewise.
13480 2013-05-14 Mike Stump <mikestump@comcast.net>
13482 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
13484 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
13486 * resource.h (struct resources): Remove unch_memory member.
13487 (CLEAR_RESOURCE): Don't clear unch_memory.
13488 * resource.c (mark_referenced_resources): Don't set it.
13489 (mark_set_resources): Likewise.
13490 (mark_target_live_regs): Don't clear it.
13491 (init_resource_info): Likewise.
13492 * reorg.c (resource_conflicts_p): Don't compare it.
13493 (redundant_insn): Don't set it.
13495 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
13497 * emit-rtl.c (next_label): Remove unused function.
13498 (skip_consecutive_labels, link_cc0_insns): Move to ...
13499 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
13500 only place where these functions are used, and make them static.
13502 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
13504 * fold-const.c (fold_negate_expr): Handle vectors.
13505 (fold_truth_not_expr): Make it static.
13506 (fold_invert_truthvalue): New static function.
13507 (invert_truthvalue_loc): Handle vectors. Do not call
13508 fold_truth_not_expr directly.
13509 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
13510 <TRUTH_NOT_EXPR>: Do not cast to boolean.
13511 (fold_comparison): Handle vector constants.
13512 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
13513 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
13514 * tree.h (fold_truth_not_expr): Remove declaration.
13516 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
13518 * config/aarch64/aarch64-simd.md
13519 (aarch64_vcond_internal<mode>): Rename to...
13520 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
13521 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
13522 float modes. Clarify all iterator modes.
13523 (vcond<mode><mode>): Use new name for vcond expanders.
13524 (vcond<v_cmp_result><mode>): Likewise.
13525 (vcondu<mode><mode>: Likewise.
13526 * config/aarch64/iterators.md (VDQF_COND): New.
13528 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
13531 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
13532 variable for the shift amount. Check that we shift by non-negative
13535 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
13538 * config/arm/arm.h (EPILOGUE_USES): Only return true
13539 for LR_REGNUM after epilogue_completed.
13541 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
13543 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
13544 is error_mark_node.
13546 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
13549 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
13550 and Solaris 11+/x86 with gld.
13551 * configure: Regenerate.
13553 2013-05-14 Jakub Jelinek <jakub@redhat.com>
13555 * expmed.c (expand_shift_1): Canonicalize rotates by
13556 constant bitsize / 2 to bitsize - 1.
13557 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
13558 case ROTATERT>: Likewise.
13561 2013-05-10 Jakub Jelinek <jakub@redhat.com>
13563 * config/i386/i386.md (rotateinv): New code attr.
13564 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
13565 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
13566 roll $31, %eax, etc.
13568 2013-05-14 Richard Biener <rguenther@suse.de>
13570 PR middle-end/57235
13571 * tree-eh.c (sink_clobbers): Give up for successors with
13572 multiple predecessors and no virtual uses.
13574 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
13576 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
13577 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
13579 2013-05-14 Jakub Jelinek <jakub@redhat.com>
13581 PR middle-end/57251
13582 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
13583 the case when both op0 and op1 have VOIDmode.
13585 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
13587 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
13588 in multiply-accumulate mode.
13590 2013-05-13 Guozhi Wei <carrot@google.com>
13592 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
13594 2013-05-13 Kai Tietz <ktietz@redhat.com>
13597 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
13598 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
13599 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
13600 * config/i386/i386.c (ix86_option_override_internal): Likewise.
13601 (ix86_expand_prologue): Likewise.
13602 (ix86_expand_split_stack_prologue): Likewise.
13603 (legitimate_pic_address_disp_p): Likewise.
13604 (legitimize_pic_address): Likewise.
13605 (legitimize_tls_address): Likewise.
13606 (legitimize_pe_coff_symbol): Likewise.
13607 (output_pic_addr_const): Likewise.
13608 (construct_plt_address): Likewise.
13609 (ix86_expand_call): Likewise.
13610 (x86_output_mi_thunk): Likewise.
13611 (x86_function_profiler): Likewise.
13613 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
13615 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
13616 similar switch cases.
13617 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
13618 (aarch64_simd_mov_to_<mode>low): Delete.
13619 (aarch64_simd_mov_to_<mode>high): Delete.
13620 (move_lo_quad_<mode>): Add w<-r alternative.
13621 (aarch64_simd_move_hi_quad_<mode>): Likewise.
13622 (aarch64_simd_mov_from_*): Update type attribute.
13623 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
13626 2013-05-13 Jan Hubicka <jh@suse.cz>
13628 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
13629 * config/i386/i386.c (ix86_compute_frame_layout,
13630 ix86_expand_epilogue, emit_i387_cw_initialization,
13631 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
13632 ix86_local_alignment): Fix use of size/speed predicates.
13634 2013-05-13 Jakub Jelinek <jakub@redhat.com>
13636 PR tree-optimization/45216
13637 PR tree-optimization/57157
13638 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
13639 the (-Y) & (B - 1) variant if OP is |.
13640 * expmed.c (expand_shift_1): For rotations by const0_rtx just
13641 return shifted. Use (-op1) & (prec - 1) as other_amount
13642 instead of prec - op1.
13644 2013-05-13 Martin Jambor <mjambor@suse.cz>
13646 PR middle-end/42371
13647 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
13648 (ipa_constant_data): New type.
13649 (ipa_jump_func): Use ipa_constant_data to hold information about
13650 constant jump functions.
13651 (ipa_get_jf_constant): Adjust to jump function type changes.
13652 (ipa_get_jf_constant_rdesc): New function.
13653 (ipa_param_descriptor): New field controlled_uses.
13654 (ipa_get_controlled_uses): New function.
13655 (ipa_set_controlled_uses): Likewise.
13656 * ipa-ref.h (ipa_find_reference): Declare.
13657 * ipa-prop.c (ipa_cst_ref_desc): New type.
13658 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
13660 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
13661 New parameter cs. Adjust all callers.
13662 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
13663 (remove_described_reference): New function.
13664 (jfunc_rdesc_usable): Likewise.
13665 (try_make_edge_direct_simple_call): Decrement controlled use count,
13666 attempt to remove reference if it hits zero.
13667 (combine_controlled_uses_counters): New function.
13668 (propagate_controlled_uses): Likewise.
13669 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
13670 (ipa_edge_duplication_hook): Duplicate reference descriptions.
13671 (ipa_print_node_params): Print described use counter.
13672 (ipa_write_jump_function): Adjust to jump function type changes.
13673 (ipa_read_jump_function): New parameter CS, pass it to
13674 ipa_set_jf_constant. Adjust caller.
13675 (ipa_write_node_info): Stream controlled use count
13676 (ipa_read_node_info): Likewise.
13677 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
13679 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
13680 count. Remove cloning-added reference if it reaches zero.
13681 * ipa-ref.c (ipa_find_reference): New function.
13683 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
13685 * config/i386/i386.c (processor_target_table): Modified default
13686 alignment values for AMD BD and BT architectures.
13688 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
13690 * tree-vect-generic.c (uniform_vector_p): Move ...
13691 * tree.c (uniform_vector_p): ... here.
13692 * tree.h (uniform_vector_p): Declare it.
13693 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
13696 2013-05-13 Jakub Jelinek <jakub@redhat.com>
13698 PR tree-optimization/57230
13699 * tree-ssa-strlen.c (handle_char_store): Record length for
13700 array store from STRING_CST.
13702 PR tree-optimization/57230
13703 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
13706 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
13708 * config/epiphany/epiphany.c (epiphany_init): Check size of
13709 NUM_MODES_FOR_MODE_SWITCHING.
13710 (epiphany_expand_prologue):
13711 Remove CONFIG_REGNUM initial value handling code.
13712 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
13713 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
13714 (emit_set_fp_mode, epiphany_mode_after): Likewise.
13715 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
13716 Don't return 1 for FP_MODE_NONE.
13717 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
13718 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
13719 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
13720 * config/epiphany/epiphany.md (save_config): New pattern.
13722 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
13724 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
13726 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
13728 * config/i386/i386.md (memory): Handle sseishft1.
13729 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
13730 (*vec_extractv2di_1): Ditto.
13732 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
13734 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
13737 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
13739 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
13740 Add mthumb/march=armv7-a multilib.
13741 Add mthumb/march=armv7-r multilib.
13742 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
13744 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
13746 * config/v850/t-rtems: Add more multilibs.
13748 2013-05-10 Richard Biener <rguenther@suse.de>
13750 PR tree-optimization/57214
13751 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
13752 not propagate from SSA names that occur in abnormal PHI nodes.
13754 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
13756 * stor-layout.c (element_precision): New function.
13757 * machmode.h (element_precision): Declare it.
13758 * tree.c (build_minus_one_cst): New function.
13759 (element_precision): Likewise.
13760 * tree.h (build_minus_one_cst): Declare new function.
13761 (element_precision): Likewise.
13762 * fold-const.c (operand_equal_p): Use element_precision.
13763 (fold_binary_loc): Handle vector types.
13764 * convert.c (convert_to_integer): Use element_precision.
13765 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
13768 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
13770 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
13771 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
13772 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
13773 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
13774 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
13775 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
13776 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
13777 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
13778 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
13779 (Uuw8): New constraints.
13780 (Usb4): Move into alphabetical order.
13781 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
13782 (sd8_operand, ub8_operand, uw8_operand): New predicates.
13783 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
13784 previously unnamed patterns.
13785 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
13786 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
13787 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
13788 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
13789 of set_attr_alternative/if_then_else. Use extended_mips16 instead
13790 of specific lengths.
13792 2013-05-10 Jakub Jelinek <jakub@redhat.com>
13794 * config/i386/i386.md (rotateinv): New code attr.
13795 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
13796 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
13797 roll $31, %eax, etc.
13799 PR tree-optimization/45216
13800 PR tree-optimization/57157
13801 * tree-ssa-forwprop.c (simplify_rotate): New function.
13802 (ssa_forward_propagate_and_combine): Call it.
13804 2013-05-10 Richard Biener <rguenther@suse.de>
13806 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
13807 disable peeling when we version for aliasing.
13808 (vector_alignment_reachable_p): Honor explicit user alignment.
13809 (vect_supportable_dr_alignment): Likewise.
13810 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
13811 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
13812 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
13813 then peeling to arrange for the cost-model check to come first.
13815 2013-05-10 Alan Modra <amodra@gmail.com>
13817 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
13818 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
13819 * configure: Regenerate.
13821 2013-05-10 Alan Modra <amodra@gmail.com>
13824 * varasm.c (default_elf_select_section): Move !DECL_P check..
13825 (get_named_section): ..to here before calling get_section_name.
13827 (default_section_type_flags): Add DECL_P check.
13828 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
13829 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
13831 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
13833 * config/epiphany/epiphany.c (epiphany_expand_prologue):
13834 When using gen_stack_adjust_str with a register offset, add a
13835 REG_FRAME_RELATED_EXPR note.
13837 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
13839 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
13840 (*vec_extractv4si_zext_mem): Ditto.
13841 (*vec_extractv2di): Add 0->x and x->x alternatives.
13842 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
13843 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
13845 2013-05-09 Jason Merrill <jason@redhat.com>
13847 N3639 C++1y VLA support
13848 * gimplify.c (gimplify_vla_decl): Don't touch an existing
13851 * tree.c (build_constructor_va): New.
13852 * tree.h: Declare it.
13854 2013-05-09 Martin Jambor <mjambor@suse.cz>
13857 * gimple-fold.c (canonicalize_constructor_val): Call
13858 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
13860 2013-05-09 Jan Hubicka <jh@suse.cz>
13861 Richard Biener <rguenther@suse.de>
13864 * symtab.c (symtab_make_decl_local): Do not add private names.
13866 2013-05-09 Jan Hubicka <jh@suse.cz>
13869 * symtab.c (insert_to_assembler_name_hash): Handle clones.
13870 (unlink_from_assembler_name_hash): Likewise.
13871 (symtab_prevail_in_asm_name_hash, symtab_register_node,
13872 symtab_unregister_node, symtab_initialize_asm_name_hash,
13873 change_decl_assembler_name): Update.
13875 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
13877 * config/aarch64/aarch64.md: New movtf split.
13878 (*movtf_aarch64): Update.
13879 (aarch64_movdi_tilow): Handle TF modes and rename to
13880 aarch64_movdi_<mode>low.
13881 (aarch64_movdi_tihigh): Handle TF modes and rename to
13882 aarch64_movdi_<mode>high
13883 (aarch64_movtihigh_di): Handle TF modes and rename to
13884 aarch64_mov<mode>high_di
13885 (aarch64_movtilow_di): Handle TF modes and rename to
13886 aarch64_mov<mode>low_di
13887 (aarch64_movtilow_tilow): Remove spurious whitespace.
13888 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
13890 (aarch64_print_operand): Update.
13892 2013-05-09 Alan Modra <amodra@gmail.com>
13894 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
13896 * configure: Regenerate.
13898 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
13900 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
13901 splitter preparation statements.
13902 * config/i386/sse.md (*vec_extract* splitters): Ditto.
13903 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
13906 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13908 * gimple-ssa-strength-reduction.c (count_candidates): Change
13909 return value to int.
13910 (analyze_candidates_and_replace): Change type of length to int.
13912 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
13914 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
13915 (*vec_extract<mode>): Use VI12_128 mode iterator.
13916 (*vec_extract<mode>_mem): Ditto.
13917 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
13920 2013-05-08 Diego Novillo <dnovillo@google.com>
13925 2012-08-17 Diego Novillo <dnovillo@google.com>
13928 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
13929 * config.in: Regenerate.
13930 * configure: Regenerate.
13931 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
13933 2013-05-08 Jan Hubicka <jh@suse.cz>
13936 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
13937 * cgraph.h (symtab_node_base): Add unique_name.
13938 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
13939 input_overwrite_node, input_varpool_node): Stream unique_name.
13940 * cgraphclones.c (cgraph_create_virtual_clone,
13941 cgraph_function_versioning): Set unique_name.
13942 * ipa.c (function_and_variable_visibility): Set unique_name.
13944 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13946 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
13947 (alloc_cand_and_find_basis): Restrict conditional candidate
13948 processing to CAND_MULTs.
13950 2013-05-08 Jan Hubicka <jh@suse.cz>
13953 lto-symtab.c (lto_symtab_symbol_p): New function.
13954 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
13955 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
13956 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
13957 Skip static symbols.
13959 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
13961 PR tree-optimization/57200
13962 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
13963 Only call inform if the preceding warning_at returns true.
13965 2013-05-07 Han Shen <shenhan@google.com>
13967 * cfgexpand.c (record_or_union_type_has_array_p): New function.
13968 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
13969 * common.opt (fstack-protector-strong): New option.
13970 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
13971 * doc/invoke.texi (Optimization Options): Document
13972 "-fstack-protector-strong".
13973 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
13975 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
13977 * config/mips/mips.c (mips_machine_reorg2): Return 0.
13979 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
13981 * ira.c (update_equiv_regs): Add insn having equiv memory even if
13982 it is not lhs of the insn.
13983 (setup_reg_equiv): Remove insn having equiv memory which it is not
13985 * lra-constraints.c (process_address): Try to improve generation
13986 code for address base + disp.
13987 (lra_constraints): Make correct the code for checking insn setting
13988 up backward equivalence. Remove insn only if it is in the init
13990 * lra-eliminations.c (update_reg_eliminate): Change return value.
13991 (lra_eliminate): Use the result.
13993 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
13995 * config/i386/sse.md (ssescalarnummask): New mode attribute.
13996 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
13997 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
13998 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
13999 register target operands.
14000 (*vec_extractv8hi_sse2): New pattern.
14001 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
14002 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
14003 (*vec_extract<mode>_mem): New insn and split pattern.
14005 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
14007 * config/arm/arm.c (arm_asan_shadow_offset): New function.
14008 (TARGET_ASAN_SHADOW_OFFSET): Define.
14009 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
14010 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
14012 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14014 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
14015 (incr_vec_index): Return -1 if increment not found.
14016 (create_add_on_incoming_edge): Assert if increment not found.
14017 (record_increment): Limit number of increments recorded.
14018 (all_phi_incrs_profitable): Return false if an increment not found.
14019 (replace_profitable_candidates): Don't process increments that were
14021 (analyze_candidates_and_replace): Limit size of incr_vec.
14023 2013-05-07 Richard Biener <rguenther@suse.de>
14025 * calls.c (special_function_p): setjmp-like functions are leaf.
14026 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
14027 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
14029 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
14031 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
14032 (aarch64_simd_mov<mode>): New expander.
14033 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
14034 (aarch64_simd_mov_to_<mode>high): Likewise.
14035 (aarch64_simd_mov_from_<mode>low): Likewise.
14036 (aarch64_simd_mov_from_<mode>high): Likewise.
14037 (aarch64_dup_lane<mode>): Update.
14038 (aarch64_dup_lanedi): New instruction pattern.
14039 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
14040 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
14042 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14044 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
14045 (replace_mult_candidate): Remove unnecessary argument; remove
14046 unnecessary parameter from call to introduce_cast_before_cand.
14047 (replace_unconditional_candidate): Remove unnecessary parameter
14048 from call to replace_mult_candidate.
14049 (replace_conditional_candidate): Likewise.
14050 (insert_initializers): Use make_temp_ssa_name.
14051 (introduce_cast_before_cand): Remove unnecessary argument; use
14052 make_temp_ssa_name.
14053 (replace_one_candidate): Remove unnecessary argument; remove
14054 unnecessary parameter from calls to introduce_cast_before_cand.
14055 (replace_profitable_candidates): Remove unnecessary parameters
14056 from calls to replace_one_candidate.
14058 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14060 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
14061 phi def as possibly hiding a basis for a CAND_ADD whose operands
14062 have been commuted in the analysis.
14063 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
14065 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
14067 * config/aarch64/aarch64.md
14068 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
14069 shift value between 0-4.
14071 2013-05-07 Richard Biener <rguenther@suse.de>
14073 * double-int.h (rshift): New overload.
14074 * double-int.c (rshift): New function.
14075 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
14076 (create_reference_ops_from_ref): Remove.
14077 (vn_reference_insert): Use shared ops for constructing the
14078 reference and copy it.
14080 2013-05-07 Richard Biener <rguenther@suse.de>
14082 PR middle-end/57190
14083 * tree-eh.c (sink_clobbers): Properly propagate
14084 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
14086 2013-05-07 Jakub Jelinek <jakub@redhat.com>
14088 PR tree-optimization/57149
14089 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
14090 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
14091 collect_phi_def_edges, execute_late_warn_uninitialized): Use
14092 uninit_undefined_value_p instead of ssa_undefined_value_p.
14095 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
14096 for modifier == EXPAND_INITIALIZER.
14098 2013-05-07 Anton Blanchard <anton@samba.org>
14100 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
14101 for powerpc64 little endian.
14102 * configure: Regenerate.
14104 2013-05-06 Graham Stott <grahams@btinternet.com>
14106 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
14107 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
14108 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
14109 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
14111 2013-05-06 Graham Stott <grahams@btinternet.com>
14113 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
14114 codes which allow non-lvalues.
14116 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
14118 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
14119 components are all 1s.
14120 (integer_minus_onep): New function.
14121 * tree.h (integer_minus_onep): Declare it.
14122 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
14123 integer_minus_onep instead of integer_all_onesp.
14125 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
14128 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
14129 variations of these patterns.
14131 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
14133 * config/i386/i386.md (isa): Add x64_sse4 member.
14134 (enabled): Handle x64_sse4.
14135 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
14136 instruction for 64bit SSE4_1 targets. Update insn attributes.
14137 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
14138 instruction for SSE4_1 targets. Update insn attributes.
14139 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
14140 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
14141 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
14143 (*vec_extractv4si): Rename from *sse4_1_pextrd.
14144 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
14145 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
14147 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
14150 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
14152 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
14154 * final.c (do_assembler_dialects): Don't handle curly braces and
14155 vertical bar escaped by % as dialect delimiters.
14156 (output_asm_insn): Print curly braces and vertical bar if escaped
14157 by % and ASSEMBLER_DIALECT defined.
14158 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
14159 * doc/tm.texi: Regenerated.
14161 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
14163 * config/mips/mips.c: Include tree-pass.h.
14164 (mips_reorg): Split in pre- and post-dbr_schedule parts.
14165 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
14166 (pass_mips_machine_reorg2): New machine specific pass.
14167 (insert_pass_mips_machine_reorg2): New pass plugin definition.
14168 (mips_option_override): Register the new pass.
14169 * rtl.h (cleanup_barriers): Remove prototype.
14170 (dbr_schedule): Likewise.
14171 * jump.c (cleanup_barriers): Make static.
14172 * reorg.c (dbr_schedule): Likewise.
14174 2013-05-06 Richard Biener <rguenther@suse.de>
14176 PR tree-optimization/57185
14177 * tree-parloops.c (add_field_for_reduction): Handle anonymous
14178 SSA names properly.
14180 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
14183 * config/i386/i386.c (add_parameter_dependencies): Add dependence
14184 between "first_arg" and "insn", not "last" and "insn".
14186 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14188 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
14189 (find_candidates_in_block): Re-enable slsr_process_phi.
14190 (create_phi_basis): Fix double counting of candidate adjustment.
14192 2013-05-06 Richard Biener <rguenther@suse.de>
14194 PR middle-end/57147
14195 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
14196 the edge is also fallthru, preserve it and just clear the
14198 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
14199 also complex, preserve that and just clear the fallthru flag.
14200 * tree-inline.c (update_ssa_across_abnormal_edges): Also
14201 update virtual operands.
14203 2013-05-06 Alan Modra <amodra@gmail.com>
14205 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
14206 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
14207 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
14208 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
14209 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
14210 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
14212 2013-05-06 Alan Modra <amodra@gmail.com>
14214 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
14215 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
14216 (DEFAULT_ASM_ENDIAN): Define.
14217 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
14218 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
14219 Update -K PIC clause from sysv4.h.
14220 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
14221 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
14223 2013-05-06 Alan Modra <amodra@gmail.com>
14225 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
14226 twice for little-endian.
14227 (ashrdi3_no_power, ashrdi3): Support little-endian.
14229 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
14232 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
14233 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
14235 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
14236 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
14237 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
14240 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
14241 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
14243 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
14244 * config/pa/pa.opt: Make mbig-switch a no-op.
14245 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
14246 (CASE_VECTOR_MODE): Always return SImode.
14247 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
14248 for the !TARGET_BIG_SWITCH case.
14249 * config/pa/pa-linux.h: Likewise.
14250 * config/pa/pa-openbsd.h: Likewise.
14251 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
14252 * config/pa/pa.md (short_jump): Remove define_insn.
14253 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
14254 (casesi0): Remove define_insn.
14255 (type): Remove btable_branch.
14256 (pa_combine_type): Likewise.
14257 (in_nullified_branch_delay): Likewise.
14258 (in_call_delay): Likewise.
14259 (define_delay): Likewise.
14260 (define_insn_reservation "Z3"): Likewise.
14261 (define_insn_reservation "Z4"): Likewise.
14262 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
14263 (pa_adjust_insn_length): Remove adjustment for btable branches.
14264 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
14267 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
14269 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
14270 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
14271 Add m->r,x alternatives.
14272 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
14273 splitters using SWI48x mode iterator.
14274 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
14275 TARGET_64BIT. Add m->x alternative.
14276 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
14277 Add o->x alternative. Enable for TARGET_SSE.
14278 (sse_storeq): Remove expander.
14279 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
14280 with memory input operand.
14281 (*vec_extractv2di_1 splitter): New.
14282 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
14283 * config/i386/i386.md (ssevecmodelower): New mode attribute.
14285 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
14287 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
14288 (INT_LOWPART): Delete.
14289 (extract_MB): Adjust.
14290 (extract_ME): Adjust.
14291 (print_operand): Adjust.
14293 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
14295 * config/rs6000/predicates.md (reg_or_add_cint_operand,
14296 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
14297 (reg_or_logical_cint_operand, easy_fp_constant,
14298 logical_const_operand): Delete "CONST_DOUBLE" case.
14299 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
14300 "HOST_BITS_PER_WIDE_INT == 64" test.
14301 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
14302 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
14303 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
14304 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
14306 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
14307 CONST_DOUBLE DImode/VOIDmode case.
14308 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
14309 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
14310 CONST_DOUBLE VOIDmode case.
14311 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
14312 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
14313 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
14314 Delete CONST_DOUBLE case.
14315 (splitters for mov FMOVE64 const_double): Delete
14316 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
14317 "HOST_BITS_PER_WIDE_INT >= 64" test.
14318 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
14320 (mov DI const_double): Delete.
14322 2013-05-04 Jakub Jelinek <jakub@redhat.com>
14324 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
14325 on op shows all bits zero in mode of a lowpart subreg, return zero.
14327 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
14330 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
14331 to save TFmode registers and DImode to save TImode registers for
14332 caller save operations.
14333 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
14334 mark being partially clobbered since they only use the first
14337 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
14338 and TDmode only use the upper 64-bits of each VSX register.
14340 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14342 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
14343 (find_candidates_in_block): Disable slsr_process_phi.
14345 2013-05-03 Guozhi Wei <carrot@google.com>
14347 * coverage.c (coverage_obj_init): Move the construction of gcov
14349 (build_init_ctor): ... here.
14351 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14353 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
14354 (slsr_cand_d): Redefine def_phi.
14355 (stride_status, phi_adjust_status, count_phis_status): New enums.
14356 (find_phi_def): New.
14357 (find_basis_for_base_expr): New.
14358 (find_basis_for_candidate): Handle hidden bases.
14359 (alloc_cand_and_find_basis): Handle phi candidates.
14360 (slsr_process_phi): New.
14361 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
14362 (create_mul_imm_cand): Likewise.
14363 (create_add_ssa_cand): Exclude phi base candidates.
14364 (create_add_imm_cand): Likewise.
14365 (slsr_process_cast): Likewise.
14366 (slsr_process_copy): Likewise.
14367 (find_candidates_in_block): Handle phi candidates.
14368 (dump_candidate): Likewise.
14369 (unconditional_cands): Delete.
14370 (unconditional_cands_with_known_stride_p): Delete.
14371 (phi_dependent_cand_p): New.
14372 (cand_increment): Handle phi-dependent candidates.
14373 (replace_dependent): Delete.
14374 (replace_mult_candidate): New.
14375 (replace_unconditional_candidate): New.
14376 (incr_vec_index): Move to avoid forward reference.
14377 (create_add_on_incoming_edge): New.
14378 (create_phi_basis): New.
14379 (replace_dependents): Delete.
14380 (replace_conditional_candidate): New.
14381 (phi_add_costs): New.
14382 (replace_uncond_cands_and_profitable_phis): New.
14383 (record_increment): Handle phi adjustments.
14384 (record_phi_increments): New.
14385 (record_increments): Handle phi adjustments.
14386 (phi_incr_cost): New.
14387 (lowest_cost_path): Handle phis.
14388 (total_savings): Likewise.
14389 (analyze_increments): Likewise.
14390 (ncd_with_phi): New.
14391 (ncd_of_cand_and_phis): New.
14392 (nearest_common_dominator_for_cands): Handle phi increments.
14393 (all_phi_incrs_profitable): New.
14394 (replace_profitable_candidates): Handle phi-dependent candidates.
14395 (analyze_candidates_and_replace): Likewise.
14397 2013-05-03 Teresa Johnson <tejohnson@google.com>
14400 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
14401 do not exceed REG_BR_PROB_BASE.
14403 2013-05-03 Jeff Law <law@redhat.com>
14405 PR tree-optimization/57144
14406 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
14407 operand of the condition will bit into the new type when eliminating
14408 a cast feeding a condition.
14410 2013-05-03 Jakub Jelinek <jakub@redhat.com>
14412 PR rtl-optimization/57130
14413 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
14414 of COMPARE as in_code to the recursive call if needed.
14416 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
14418 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
14419 (enabled): Handle new members.
14420 * config/i386/sse.md (*vec_concatv2si): Merge from
14421 *vec_concatv2si_sse2 and vec_concatv2si_sse.
14422 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
14424 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
14426 PR tree-optimization/57027
14427 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
14428 for fnms opportunity, check we got the prerequisite kind
14429 of tree / gimple before using accessor functions.
14431 2013-05-03 Richard Biener <rguenther@suse.de>
14433 * double-int.h (lshift): New overload without precision
14434 and arith argument.
14435 (operator *=, operator +=, operator -=): Move ...
14436 * double-int.c (operator *=, operator +=, operator -=): ... here
14437 and implement more efficiently.
14438 (mul_double_with_sign): Remove.
14439 (lshift_double): Adjust to take unsinged shift argument, push
14440 dispatching code to callers.
14441 (mul_double_wide_with_sign): Add early out for callers that
14442 are not interested in high parts or overflow.
14443 (lshift): New function.
14444 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
14445 dispatch code here.
14446 (lrotate, rrotate): Use logical shifts.
14447 * expr.c (get_inner_reference): Use lshift.
14448 * fixed-value.c (do_fixed_divide): Likewise.
14449 * tree-dfa.c (get_ref_base_and_extent): Likewise.
14450 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
14451 (indirect_refs_may_alias_p): Likewise.
14452 (stmt_kills_ref_p_1): Likewise.
14454 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
14456 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
14458 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
14460 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
14461 scalar form of FABD instruction.
14463 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
14465 * lra-constraints.c (process_alt_operands): Add checking alt
14466 number to choose the best alternative.
14468 2013-05-02 Richard Biener <rguenther@suse.de>
14470 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
14471 bitmap and its handling.
14472 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
14474 2013-05-02 Richard Biener <rguenther@suse.de>
14476 PR middle-end/57140
14477 * tree-inline.c (copy_loops): Properly handle removed loops.
14478 (copy_cfg_body): Mark destination loops for fixup if source
14479 loops needed fixup.
14481 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
14484 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
14485 generating simple_return for naked functions.
14487 2013-05-02 Martin Jambor <mjambor@suse.cz>
14489 PR middle-end/56988
14490 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
14491 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
14493 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
14494 ipa_agg_replacement_value structures.
14495 (known_aggs_to_agg_replacement_list): Likewise.
14496 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
14497 (read_agg_replacement_chain): Likewise.
14498 (ipcp_transform_function): Also check that by_ref flags match.
14500 2013-05-02 Richard Biener <rguenther@suse.de>
14502 * graphds.h (struct graph): Add obstack member.
14503 * graphds.c (new_graph): Initialize obstack and allocate
14505 (add_edge): Allocate edge from the obstack.
14506 (free_graph): Free the obstack instead of all edges and vertices.
14508 2013-05-02 Teresa Johnson <tejohnson@google.com>
14510 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
14512 * cfg.c (update_bb_profile_for_threading): Ditto.
14513 * tree-inline.c (copy_bb): Ditto.
14514 (copy_edges_for_bb): Ditto.
14515 (initialize_cfun): Ditto.
14516 (copy_cfg_body): Ditto.
14517 (expand_call_inline): Ditto.
14518 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
14519 (estimate_node_size_and_time): Ditto.
14520 (inline_merge_summary): Ditto.
14521 * cgraphclones.c (cgraph_clone_edge): Ditto.
14522 (cgraph_clone_node): Ditto.
14523 * sched-rgn.c (compute_dom_prob_ps): Ditto.
14524 (compute_trg_info): Ditto.
14526 2013-05-02 Ian Bolton <ian.bolton@arm.com>
14528 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
14529 S reg when fp attribute set.
14530 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
14532 2013-05-02 Ian Bolton <ian.bolton@arm.com>
14534 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
14536 (*and_one_cmplsi3_compare0_uxtw): Likewise.
14537 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
14538 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
14540 2013-05-02 Richard Biener <rguenther@suse.de>
14542 * tree-scalar-evolution.c (scev_info_hasher): Remove.
14543 (struct instantiate_cache_entry): New type.
14544 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
14545 (struct instantiate_cache_type): New type.
14546 (set_instantiated_value, get_instantiated_value): Remove.
14547 (get_instantiated_value_entry): New function.
14548 (instantiate_scev_name): Use the new cache and adjust.
14549 (instantiate_scev_poly): Adjust.
14550 (instantiate_scev_binary): Likewise.
14551 (instantiate_array_ref): Likewise.
14552 (instantiate_scev_convert): Likewise.
14553 (instantiate_scev_not): Likewise.
14554 (instantiate_scev_3): Likewise.
14555 (instantiate_scev_2): Likewise.
14556 (instantiate_scev_r): Likewise.
14557 (instantiate_scev): Likewise.
14558 (resolve_mixers): Likewise.
14560 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
14563 * lra-constraints.c (best_small_class_operands_num): Remove.
14564 (process_alt_operands): Remove small_class_operands_num. Take
14565 small classes operands into losers and only if the operand is not
14566 matched. Modify debugging output.
14567 (curr_insn_transform): Remove best_small_class_operands_num.
14570 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14572 * config/aarch64/aarch64-builtins.c
14573 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
14574 * config/aarch64/aarch64-simd-builtins.def
14575 (reduc_splus_): Add new modes.
14576 (reduc_uplus_): New.
14577 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
14578 (reduc_uplus_v4sf): Likewise.
14579 (reduc_splus_v4sf): Likewise.
14580 (aarch64_addv<mode>): Likewise.
14581 (reduc_uplus_<mode>): Likewise.
14582 (reduc_splus_<mode>): Likewise.
14583 (aarch64_addvv2di): Likewise.
14584 (reduc_uplus_v2di): Likewise.
14585 (reduc_splus_v2di): Likewise.
14586 (aarch64_addvv2si): Likewise.
14587 (reduc_uplus_v2si): Likewise.
14588 (reduc_splus_v2si): Likewise.
14589 (reduc_<sur>plus_<mode>): New.
14590 (reduc_<sur>plus_v2di): Likewise.
14591 (reduc_<sur>plus_v2si): Likewise.
14592 (reduc_<sur>plus_v4sf): Likewise.
14593 (aarch64_addpv4sf): Likewise.
14594 * config/aarch64/arm_neon.h
14595 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
14596 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
14597 add UNSPEC_SADDV, UNSPEC_UADDV.
14599 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
14601 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14603 * config/aarch64/arm_neon.h
14604 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
14606 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14608 * config/aarch64/aarch64-builtins
14609 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
14611 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14613 * config/aarch64/aarch64-simd-builtins.def
14614 (reduc_smax_): New.
14615 (reduc_smin_): Likewise.
14616 (reduc_umax_): Likewise.
14617 (reduc_umin_): Likewise.
14618 (reduc_smax_nan_): Likewise.
14619 (reduc_smin_nan_): Likewise.
14622 (smax): Update for V2SF, V4SF and V2DF modes.
14625 (smin_nan): Likewise.
14626 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
14627 (<su><maxmin><mode>3): ...This, refactor.
14628 (s<maxmin><mode>3): New.
14629 (<maxmin_uns><mode>3): Likewise.
14630 (reduc_<maxmin_uns>_<mode>): Refactor.
14631 (reduc_<maxmin_uns>_v4sf): Likewise.
14632 (reduc_<maxmin_uns>_v2si): Likewise.
14633 (aarch64_<fmaxmin><mode>: Remove.
14634 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
14636 (vmin<q>_f<32,64>): Likewise.
14637 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
14639 (su): Add mappings for smax, smin, umax, umin.
14641 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
14642 (FMAXMIN): Rename as...
14643 (FMAXMIN_UNS): ...This.
14645 (fmaxminv): Likewise.
14646 (fmaxmin): Likewise.
14648 (maxmin_uns_op): Likewise.
14650 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14652 * config/aarch64/arm_neon.h
14653 (vac<ge, gt><sd>_f<32, 64>): Rename to...
14654 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
14655 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
14657 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14659 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
14660 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
14662 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14664 * config/aarch64/aarch64-simd.md
14665 (vcond<mode>_internal): Handle special cases for constant masks.
14666 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
14667 (vcondu<mode><mode>): Likewise.
14668 (vcond<v_cmp_result><mode>): New.
14670 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14672 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
14673 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
14674 * config/aarch64/aarch64-simd-builtins.def
14675 (cmeq): Update to BUILTIN_VALLDI.
14680 * config/aarch64/arm_neon.h
14681 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
14682 to builtins or C as appropriate.
14684 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14686 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
14688 (cmhi): Rename to...
14690 * config/aarch64/aarch64-simd.md
14691 (simd_mode): Add SF.
14692 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
14693 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
14694 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
14695 (cstore<mode>_neg): ...This.
14696 * config/aarch64/iterators.md
14698 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
14699 (COMPARISONS): New.
14700 (UCOMPARISONS): Likewise.
14701 (optab): Add missing comparisons.
14707 (VCMP_S): Likewise.
14708 (VCMP_U): Likewise.
14709 (V_cmp_result): Add DF, SF modes.
14710 (v_cmp_result): Likewise.
14712 (vmtype): Likewise.
14713 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
14715 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
14717 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
14718 define_insn to define_insn_and_split.
14719 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
14720 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
14721 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
14722 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
14723 (thumb2_negscc): Likewise.
14725 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
14727 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
14729 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
14731 * config/arm/thumb2.md: Remove trailing whitespaces.
14733 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
14735 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
14736 Use gen_int_mode rather than GEN_INT.
14738 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
14740 * value-prof.c (stream_in_histogram_value): Remove the strayed
14743 2013-04-30 Richard Biener <rguenther@suse.de>
14745 PR middle-end/57122
14746 * cfghooks.c (split_edge): Properly check for the loop latch edge.
14748 2013-04-30 Richard Biener <rguenther@suse.de>
14750 PR middle-end/57107
14751 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
14753 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
14755 PR rtl-optimization/56957
14756 PR rtl-optimization/57105
14757 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
14758 variable. Use just INSN_UID for determining whether an insn
14759 should be only disconnected from the insn stream.
14760 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
14762 2013-04-30 Jakub Jelinek <jakub@redhat.com>
14764 PR tree-optimization/57104
14765 * tsan.c (instrument_expr): Don't instrument accesses to
14766 DECL_HARD_REGISTER VAR_DECLs.
14768 2013-04-30 Richard Biener <rguenther@suse.de>
14770 * function.h (loops_for_fn): New inline function.
14771 (set_loops_for_fn): Likewise.
14772 * cfgloop.h (place_new_loop): Add struct function parameter.
14773 (get_loop): Likewise.
14774 (get_loops): Likewise.
14775 (number_of_loops): Likewise.
14776 (fel_next): Adjust.
14777 (fel_init): Likewise.
14778 * cfg.c (get_loop_copy): Adjust.
14779 * cfgloop.c (flow_loops_dump): Likewise.
14780 (record_loop_exits): Likewise.
14781 (verify_loop_structure): Likewise.
14782 * cfgloopanal.c (mark_irreducible_loops): Likewise.
14783 (estimate_reg_pressure_cost): Likewise.
14784 (mark_loop_exit_edges): Likewise.
14785 * cfgloopmanip.c (place_new_loop): Likewise.
14786 (add_loop): Likewise.
14787 (duplicate_loop): Likewise.
14788 * graph.c (draw_cfg_nodes): Likewise.
14789 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
14790 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
14791 (extract_affine_chrec): Likewise.
14792 (build_scop_iteration_domain): Likewise.
14793 * graphite.c (graphite_initialize): Likewise.
14794 * ira-build.c (create_loop_tree_nodes): Likewise.
14795 (more_one_region_p): Likewise.
14796 (rebuild_regno_allocno_maps): Likewise.
14797 (mark_loops_for_removal): Likewise.
14798 (mark_all_loops_for_removal): Likewise.
14799 (remove_unnecessary_regions): Likewise.
14800 (ira_build): Likewise.
14801 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
14802 * loop-init.c (fix_loop_structure): Likewise.
14803 (gate_rtl_move_loop_invariants): Likewise.
14804 (gate_rtl_unswitch): Likewise.
14805 (gate_rtl_unroll_and_peel_loops): Likewise.
14806 (rtl_doloop): Likewise.
14807 * lto-streamer-in.c (input_cfg): Likewise.
14808 * lto-streamer-out.c (output_cfg): Likewise.
14809 * modulo-sched.c (sms_schedule): Likewise.
14810 * predict.c (tree_estimate_probability): Likewise.
14811 (tree_estimate_probability_driver): Likewise.
14812 (estimate_loops): Likewise.
14813 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
14814 (move_sese_region_to_fn): Likewise.
14815 (debug_loop_num): Likewise.
14816 * tree-chrec.c (chrec_evaluate): Likewise.
14817 (hide_evolution_in_other_loops_than_loop): Likewise.
14818 (chrec_component_in_loop_num): Likewise.
14819 (reset_evolution_in_loop): Likewise.
14820 (evolution_function_is_invariant_rec_p): Likewise.
14821 * tree-if-conv.c (main_tree_if_conversion): Likewise.
14822 * tree-inline.c (copy_loops): Likewise.
14823 (copy_cfg_body): Likewise.
14824 (tree_function_versioning): Likewise.
14825 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
14826 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
14828 (add_to_evolution_1): Likewise.
14829 (scev_const_prop): Likewise.
14830 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
14831 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
14832 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
14833 (tree_ssa_lim_initialize): Likewise.
14834 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
14835 (verify_loop_closed_ssa): Likewise.
14836 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
14837 (tree_ssa_loop_im): Likewise.
14838 (tree_ssa_loop_unswitch): Likewise.
14839 (tree_vectorize): Likewise.
14840 (check_data_deps): Likewise.
14841 (tree_ssa_loop_ivcanon): Likewise.
14842 (tree_ssa_loop_bounds): Likewise.
14843 (tree_complete_unroll): Likewise.
14844 (tree_complete_unroll_inner): Likewise.
14845 (tree_parallelize_loops): Likewise.
14846 (tree_ssa_loop_prefetch): Likewise.
14847 (tree_ssa_loop_ivopts): Likewise.
14848 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
14849 * tree-vectorizer.c (vectorize_loops): Likewise.
14851 2013-04-29 Mike Frysinger <vapier@gentoo.org>
14853 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
14854 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
14855 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
14856 with EABI_LINK_SPEC.
14858 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
14861 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
14864 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
14867 * lra-constraints.c (process_alt_operands): Discourage a bit more
14868 using memory for pseudos. Print cost dump for alternatives.
14869 Modify cost values for conflicts with early clobbers.
14870 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
14872 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
14875 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
14877 2013-04-29 Ian Bolton <ian.bolton@arm.com>
14879 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
14880 from/to S register.
14881 (movdi_aarch64): Support LDR/STR from/to D register.
14883 2013-04-29 Ian Bolton <ian.bolton@arm.com>
14885 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
14886 or higher by default.
14888 2013-04-29 Richard Biener <rguenther@suse.de>
14890 PR middle-end/57075
14891 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
14892 even if not adding abnormal edges for calls that can make
14895 2013-04-29 Richard Biener <rguenther@suse.de>
14897 PR middle-end/57103
14898 * tree-cfg.c (move_stmt_op): Fix condition under which to update
14900 (move_stmt_r): Remove redundant checking.
14902 2013-04-29 Teresa Johnson <tejohnson@google.com>
14905 * basic-block.h (apply_scale): New function.
14906 (apply_probability): Use apply_scale.
14907 * gimple-streamer-in.c (input_bb): Ditto.
14908 * lto-streamer-in.c (input_cfg): Ditto.
14909 * lto-cgraph.c (merge_profile_summaries): Ditto.
14910 * tree-optimize.c (execute_fixup_cfg): Ditto.
14911 * tree-inline.c (copy_bb): Update comment to use apply_scale.
14912 (copy_edges_for_bb): Ditto.
14913 (copy_cfg_body): Ditto.
14915 2013-04-29 Tom de Vries <tom@codesourcery.com>
14917 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
14918 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
14919 (tail_merge_optimize): Handle current_loops == NULL.
14921 2013-04-26 Jeff Law <law@redhat.com>
14923 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
14924 (simplify_cond_using_ranges): Generalize code to simplify
14925 COND_EXPRs where one argument is a constant and the other
14926 is an SSA_NAME created by an integral type conversion.
14928 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14930 * config/arm/arm.md (store_minmaxsi): Use only when
14931 optimize_insn_for_size_p.
14933 2013-04-29 Christian Bruel <christian.bruel@st.com>
14936 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
14938 2013-04-29 Richard Biener <rguenther@suse.de>
14940 PR middle-end/57089
14941 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
14942 loop tree make sure to schedule a fixup for the child as well.
14943 (expand_omp_for_generic): Properly add loops.
14944 (expand_omp_for_static_nochunk): Likewise.
14945 (expand_omp_for_static_chunk): Likewise.
14946 (expand_omp_for): For the degenerate case fixup loops.
14947 (expand_omp_sections): Fix default bb placement in loops.
14948 (expand_omp_atomic_pipeline): Properly add loops.
14950 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14952 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
14954 2013-04-29 Tom de Vries <tom@codesourcery.com>
14956 * tree-ssa-tail-merge.c: Update header comment.
14958 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14960 * config/aarch64/arm_neon.h
14961 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
14962 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
14963 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
14964 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
14965 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
14966 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
14967 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
14968 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
14970 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14972 * config/aarch64/aarch64-simd.md
14973 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
14974 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
14975 fix_trunc, fixuns_trunc.
14976 (ftrunc<VDQF:mode>2): New.
14977 * config/aarch64/iterators.md (optab): Add fix, fixuns.
14978 (fix_trunc_optab): New.
14980 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14982 * config/aarch64/aarch64-builtins.c
14983 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
14984 iceilf, lround, iroundf.
14986 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
14989 * config/i386/i386.h (enum ix86_tune_indices)
14990 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
14991 New, split from X86_TUNE_INTER_UNIT_MOVES.
14992 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
14993 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
14994 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
14995 (TARGET_INTER_UNIT_MOVES): Remove.
14996 * config/i386/i386.c (initial_ix86_tune_features): Update.
14997 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
14998 (ix86_expand_convert_uns_didf_sse): Use
14999 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
15000 (ix86_expand_vector_init_one_nonzero): Ditto.
15001 (ix86_expand_vector_init_interleave): Ditto.
15002 (inline_secondary_memory_needed): Return true for moves from SSE class
15003 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
15004 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
15005 * config/i386/constraints.md (Yi, Ym): Depend on
15006 TARGET_INTER_UNIT_MOVES_TO_VEC.
15007 (Yj, Yn): New constraints.
15008 * config/i386/i386.md (*movdi_internal): Change constraints of
15009 operand 1 from Yi to Yj and from Ym to Yn.
15010 (*movsi_internal): Ditto.
15011 (*movdf_internal): Ditto.
15012 (*movsf_internal): Ditto.
15013 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
15014 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
15015 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
15016 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
15017 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
15018 * config/i386/sse.md (movdi_to_sse): Ditto.
15019 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
15020 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
15021 TARGET_INTER_UNIT_MOVES.
15022 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
15023 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
15024 instead of TARGET_INTER_UNIT_MOVES.
15025 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
15026 operand 1 from Yi to Yj and from Ym to Yn.
15028 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
15030 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
15031 (float_truncate_hi_): Likewise.
15032 (float_extend_lo_): Likewise.
15033 (float_truncate_lo_): Likewise.
15034 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
15035 (aarch64_float_extend_lo_v2df): Likewise.
15036 (vec_unpacks_hi_v4sf): Likewise.
15037 (aarch64_float_truncate_lo_v2sf): Likewise.
15038 (aarch64_float_truncate_hi_v4sf): Likewise.
15039 (vec_pack_trunc_v2df): Likewise.
15040 (vec_pack_trunc_df): Likewise.
15042 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
15044 * config/aarch64/aarch64-builtins.c
15045 (aarch64_fold_builtin): Fold float conversions.
15046 * config/aarch64/aarch64-simd-builtins.def
15047 (floatv2si, floatv4si, floatv2di): New.
15048 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
15049 * config/aarch64/aarch64-simd.md
15050 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
15051 * config/aarch64/iterators.md (FLOATUORS): New.
15052 (optab): Add float, floatuns.
15053 (su_optab): Likewise.
15055 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
15057 * config/aarch64/aarch64-builtins.c
15058 (aarch64_builtin_vectorized_function): Use new names for
15060 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
15061 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
15062 (fcvtzu): Split as...
15063 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
15064 (fcvtas): Split as...
15065 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
15066 (fcvtau): Split as...
15067 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
15068 (fcvtps): Split as...
15069 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
15070 (fcvtpu): Split as...
15071 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
15072 (fcvtms): Split as...
15073 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
15074 (fcvtmu): Split as...
15075 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
15076 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
15077 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
15078 (lfrintnusf, lfrintnudf): Likewise.
15079 * config/aarch64/aarch64-simd.md
15080 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
15082 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
15083 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
15084 (fcvt_pattern): Likewise.
15086 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
15088 * config/aarch64/aarch64-simd.md
15089 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
15090 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
15092 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
15094 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
15095 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
15096 (vrnd<a,m,n,p>_f32): Implement using builtins.
15097 (vrnd<i,x><q>_f<32, 64>): New.
15099 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
15101 * config/aarch64/aarch64-builtins.c
15102 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
15103 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
15104 (frintz): Rename to...
15106 (frintp): Rename to...
15108 (frintm): Rename to...
15110 (frinti): Rename to...
15111 (nearbyint): ...this.
15112 (frintx): Rename to...
15114 (frinta): Rename to...
15116 * config/aarch64/aarch64-simd.md
15117 (aarch64_frint<frint_suffix><mode>): Delete.
15118 (<frint_pattern><mode>2): Convert to insn.
15119 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
15120 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
15121 (frint_pattern): Likewise.
15122 (frint_suffix): Likewise.
15124 2013-04-29 Richard Biener <rguenther@suse.de>
15126 PR tree-optimization/57081
15127 * loop-init.c: Include tree-flow.h.
15128 (loop_optimizer_finalize): Free number of iteration estimates.
15129 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
15131 2013-04-29 Jakub Jelinek <jakub@redhat.com>
15133 PR tree-optimization/57083
15134 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
15135 non-singleton shift count range, zero extend low_bound for uns case.
15137 * config/i386/predicates.md (general_vector_operand): New predicate.
15138 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
15139 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
15140 if they aren't nonimmediate operands. If their original values
15141 satisfy const_vector_equal_evenodd_p, don't shift them.
15142 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
15143 predicates. For the SSE4.1 case force operands[{1,2}] into registers
15144 if not nonimmediate_operand.
15145 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
15146 instead of register_operand.
15147 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
15149 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
15151 * stor-layout.c (finalize_size_functions): Allocate a structure and
15152 reset cfun before dumping the functions.
15154 2013-04-27 Jakub Jelinek <jakub@redhat.com>
15156 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
15159 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
15160 use xop_pmacsdqh if uns_p.
15161 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
15162 the immediate rotate count.
15164 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
15166 * rtl.h (struct rtx_def): Add comment for field jump.
15167 (LRA_SUBREG_P): New macro.
15168 * recog.c (register_operand): Check LRA_SUBREG_P.
15169 * lra.c (lra): Add note at the end of RTL code. Align non-empty
15171 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
15172 (lra_final_code_change): Skip subreg change for operators.
15173 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
15174 if there are no operand changes.
15175 * lra-constraints.c (curr_insn_set): New.
15176 (match_reload): Set LRA_SUBREG_P.
15177 (emit_spill_move): Ditto.
15178 (check_and_process_move): Use curr_insn_set. Process only single
15179 set insns. Don't initialize sec_mem_p and change_p.
15180 (simplify_operand_subreg): Use LRA_SUBREG_P.
15181 (reg_in_class_p): New function.
15182 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
15183 of #ifdef. Add code to remove cycling.
15184 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
15185 non-null disp. Reload inner instead of disp when base and index
15186 are null. Try to put lo_sum into register.
15187 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
15188 (check_and_process_move): Move code for move cost check to
15189 simple_move_p. Remove equiv_substitution.
15190 (simple_move_p): New function.
15191 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
15192 curr_insn_set. Call check_and_process_move only for single set
15193 insns. Use the new function. Move call of check_and_process_move
15194 after operand equiv substitution and address process.
15196 2013-04-26 Jakub Jelinek <jakub@redhat.com>
15199 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
15200 with nonlocal goto receivers or returns twice calls, ignore
15201 unininitialized values from abnormal edges to nl goto receiver
15202 or returns twice call.
15204 2013-04-26 Jakub Jelinek <jakub@redhat.com>
15206 PR tree-optimization/57051
15207 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
15208 and VEC_RSHIFT_EXPR if shift count is a multiple of element
15211 2013-04-26 Richard Biener <rguenther@suse.de>
15213 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
15214 (expand_omp_taskreg): Likewise. Mark loops for fixup.
15215 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
15216 (fixup_loop_arrays_after_move): New function.
15217 (move_sese_region_to_fn): Properly outline the loop tree parts
15218 of the SESE region.
15220 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
15222 * config/i386/i386.md (type, unit): Fix long lines.
15224 2013-04-26 Richard Biener <rguenther@suse.de>
15226 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
15227 (lto-streamer-out.o): Likewise.
15228 * cfgloop.c (init_loops_structure): Export, add struct function
15229 argument and adjust.
15230 (flow_loops_find): Adjust.
15231 * cfgloop.h (enum loop_estimation): Add EST_LAST.
15232 (init_loops_structure): Declare.
15233 * lto-streamer-in.c: Include cfgloop.h.
15234 (input_cfg): Input the loop tree.
15235 * lto-streamer-out.c: Include cfgloop.h.
15236 (output_cfg): Output the loop tree.
15237 (output_struct_function_base): Do not drop PROP_loops.
15239 2013-03-26 Richard Biener <rguenther@suse.de>
15241 * tree-cfg.c (execute_build_cfg): Build the loop tree.
15242 (pass_build_cfg): Provide PROP_loops.
15243 (move_sese_region_to_fn): Remove loops that are outlined into fn
15245 * tree-inline.c: Include cfgloop.h.
15246 (initialize_cfun): Do not drop PROP_loops.
15247 (copy_loops): New function.
15248 (copy_cfg_body): Copy loop structure.
15249 (tree_function_versioning): Initialize destination loop tree.
15250 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
15251 (pass_parallelize_loops): Do IL verification.
15252 * loop-init.c (loop_optimizer_init): Fixup loops if required.
15253 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
15254 the CFG make sure we fixup loops as well.
15255 * tree-ssa-tail-merge.c: Include cfgloop.h.
15256 (replace_block_by): When merging loop latches mark loops for fixup.
15257 * lto-streamer-out.c (output_struct_function_base): Drop
15258 PROP_loops for now.
15259 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
15260 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
15261 * ipa-split.c: Include cfgloop.h.
15262 (split_function): Add the new return block to the loop tree root.
15263 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
15264 whether we have removed the forwarder block.
15265 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
15266 * cfgloop.h (place_new_loop): Declare.
15267 * cfgloopmanip.c (place_new_loop): Export.
15268 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
15269 (tree-switch-conversion.o): Likewise.
15270 (tree-complex.o): Likewise.
15271 (tree-inline.o): Likewise.
15272 (tree-ssa-tailmerge.o): Likewise.
15273 (ipa-split.o): Likewise.
15274 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
15275 (tree-ssa-copy.o): Likewise.
15276 * tree-switch-conversion.c: Include cfgloop.h
15277 (process_switch): If we emit a bit-test cascade, schedule loops
15279 * tree-complex.c: Include cfgloop.h.
15280 (expand_complex_div_wide): Properly add new basic-blocks to loops.
15281 * asan.c: Include cfgloop.h.
15282 (create_cond_insert_point): Properly add new basic-blocks to
15283 loops, schedule loop fixup.
15284 * cfgloop.c (verify_loop_structure): Check that looks are not
15286 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
15288 (expand_omp_for_generic): Likewise.
15289 (expand_omp_sections): Likewise.
15290 (expand_omp_atomic_pipeline): Schedule loops for fixup.
15291 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
15292 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
15293 is initialized, not when loops are present.
15294 * tree-parloops.c (parallelize_loops): Remove checking here.
15295 * passes.c (init_optimization_passes): Schedule a copy-propagation
15296 pass before complete unrolling of inner loops.
15298 2013-04-26 Jakub Jelinek <jakub@redhat.com>
15300 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
15301 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
15302 (colorize_init): Add argument to _WIN32 version.
15303 * toplev.c: Include diagnostic-color.h.
15304 (process_options): Default to -fdiagnostics-color=auto if
15305 GCC_COLORS env var is in the environment.
15306 * common.opt (fdiagnostics-color=): Add Var and Init.
15307 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
15308 env var is in the environment, the default is auto rather than never.
15310 * diagnostic.h (file_name_as_prefix): Add context argument.
15311 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
15312 the string as locus.
15313 * langhooks.c (lhd_print_error_function): Adjust caller.
15315 2013-04-25 Lawrence Crowl <crowl@google.com>
15317 * var-tracking.c (shared_hash_def::htab):
15318 Change type to hash_table. Update dependent calls and types.
15320 2013-04-25 Lawrence Crowl <crowl@google.com>
15322 * Makefile.in: Update as needed below.
15324 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
15325 Move declaration to after the type's method definitons.
15327 * attribs.c (htab_t scoped_attributes::attribute_hash):
15328 Change type to hash_table. Update dependent calls and types.
15330 * bitmap.c (htab_t bitmap_desc_hash):
15331 Change type to hash_table. Update dependent calls and types.
15333 * cselib.c (htab_t cselib_hash_table):
15334 Change type to hash_table. Update dependent calls and types.
15336 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
15337 (hash_string_slot_node): Move implementation into lto-streamer.h
15338 struct string_slot_hasher.
15339 (eq_string_slot_node): Likewise.
15341 * data-streamer-out.c: Update output_block::string_hash_table
15342 dependent calls and types.
15344 * dwarf2cfi.c (htab_t trace_index):
15345 Change type to hash_table. Update dependent calls and types.
15347 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
15348 Change type to hash_table. Update dependent calls and types.
15349 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
15350 (htab_t optimize_external_refs::map): Likewise.
15351 (htab_t output_comp_unit::extern_map): Likewise.
15352 (htab_t output_comdat_type_unit::extern_map): Likewise.
15353 (htab_t output_macinfo::macinfo_htab): Likewise.
15354 (htab_t optimize_location_lists::htab): Likewise.
15355 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
15357 * except.c (htab_t ehspec_hash_type):
15358 Change type to hash_table. Update dependent calls and types.
15359 (assign_filter_values::ttypes): Likewise.
15360 (assign_filter_values::ehspec): Likewise.
15361 (sjlj_assign_call_site_values::ar_hash): Likewise.
15362 (convert_to_eh_region_ranges::ar_hash): Likewise.
15364 * gcse.c (htab_t pre_ldst_table):
15365 Change type to hash_table. Update dependent calls and types.
15367 * ggc-common.c (htab_t saving_htab):
15368 Change type to hash_table. Update dependent calls and types.
15369 (htab_t loc_hash): Likewise.
15370 (htab_t ptr_hash): Likewise.
15371 (call_count): Rename ggc_call_count.
15372 (call_alloc): Rename ggc_call_alloc.
15373 (loc_descriptor): Rename make_loc_descriptor.
15374 (add_statistics): Rename ggc_add_statistics.
15376 * ggc-common.c (saving_htab):
15377 Change type to hash_table. Update dependent calls and types.
15379 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
15380 (push_gimplify_context): Likewise.
15381 (pop_gimplify_context): Likewise.
15382 (struct gimple_temp_hash_elt): Added.
15383 (struct gimplify_hasher): Likewise.
15384 (struct gimplify_ctx.temp_htab):
15385 Change type to hash_table. Update dependent calls and types.
15387 * gimple-fold.c: Include gimplify-ctx.h.
15389 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
15390 Change type to hash_table. Update dependent calls and types.
15391 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
15392 avoid potential global name collision.
15394 * gimplify.c: Include gimplify-ctx.h.
15395 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
15396 (htab_t gimplify_ctx::temp_htab):
15397 Update dependent calls and types for new type hash_table.
15398 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
15399 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
15401 * gimplify-ctx.h: New.
15402 (struct gimple_temp_hash_elt): Move from gimplify.c.
15403 (class gimplify_hasher): New.
15404 (struct gimplify_ctx): Move from gimple.h.
15405 (htab_t gimplify_ctx::temp_htab):
15406 Change type to hash_table. Update dependent calls and types.
15408 * graphite-clast-to-gimple.c: Include graphite-htab.h.
15409 (htab_t ivs_params::newivs_index):
15410 Change type to hash_table. Update dependent calls and types.
15411 (htab_t ivs_params::params_index): Likewise.
15412 (htab_t print_generated_program::params_index): Likewise.
15413 (htab_t gloog::newivs_index): Likewise.
15414 (htab_t gloog::params_index): Likewise.
15416 * graphite.c: Include graphite-htab.h.
15417 4htab_t graphite_transform_loops::bb_pbb_mapping):
15418 Change type to hash_table. Update dependent calls and types.
15420 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
15421 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
15422 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
15424 * graphite-dependences.c: Include graphite-htab.h.
15425 (loop_is_parallel_p): Change hash table type of parameter.
15427 * graphite-htab.h: New.
15428 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
15429 (extern find_pbb_via_hash): Move from graphite-poly.h.
15430 (extern loop_is_parallel_p): Move from graphite-poly.h.
15431 (extern get_loop_body_pbbs): Move from graphite-poly.h.
15433 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
15434 (extern loop_is_parallel_p): Move to graphite-htab.h.
15435 (extern get_loop_body_pbbs): Move to graphite-htab.h.
15437 * haifa-sched.c (htab_t delay_htab):
15438 Change type to hash_table. Update dependent calls and types.
15439 (htab_t delay_htab_i2): Likewise.
15441 * ira-color.c (htab_t allocno_hard_regs_htab):
15442 Change type to hash_table. Update dependent calls and types.
15444 * ira-costs.c (htab_t cost_classes_htab):
15445 Change type to hash_table. Update dependent calls and types.
15447 * loop-invariant.c (htab_t merge_identical_invariants::eq):
15448 Change type to hash_table. Update dependent calls and types.
15450 * loop-iv.c (htab_t bivs):
15451 Change type to hash_table. Update dependent calls and types.
15453 * loop-unroll.c (htab_t opt_info::insns_to_split):
15454 Change type to hash_table. Update dependent calls and types.
15455 (htab_t opt_info::insns_with_var_to_expand): Likewise.
15457 * lto-streamer.h (struct string_slot): Move from data-streamer.h
15458 (struct string_slot_hasher): New.
15459 (htab_t output_block::string_hash_table):
15460 Change type to hash_table. Update dependent calls and types.
15462 * lto-streamer-in.c (freeing_string_slot_hasher): New.
15463 (htab_t file_name_hash_table):
15464 Change type to hash_table. Update dependent calls and types.
15466 * lto-streamer-out.c: Update output_block::string_hash_table dependent
15469 * lto-streamer.c (htab_t tree_htab):
15470 Change type to hash_table. Update dependent calls and types.
15472 * omp-low.c: Include gimplify-ctx.h.
15474 * passes.c (htab_t name_to_pass_map):
15475 Change type to hash_table. Update dependent calls and types.
15476 (pass_traverse): Rename to passes_pass_traverse.
15478 * plugin.c (htab_t event_tab):
15479 Change type to hash_table. Update dependent calls and types.
15481 * postreload-gcse.c (htab_t expr_table):
15482 Change type to hash_table. Update dependent calls and types.
15483 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
15485 * sese.c (debug_rename_map_1): Make extern.
15486 (htab_t copy_bb_and_scalar_dependences::rename_map):
15487 Change type to hash_table. Update dependent calls and types.
15489 * sese.h (extern debug_rename_map): Move to .c file.
15491 * store-motion.c (htab_t store_motion_mems_table):
15492 Change type to hash_table. Update dependent calls and types.
15494 * trans-mem.c (htab_t tm_new_mem_hash):
15495 Change type to hash_table. Update dependent calls and types.
15497 * tree-browser.c (htab_t TB_up_ht):
15498 Change type to hash_table. Update dependent calls and types.
15500 * tree-cfg.c (htab_t discriminator_per_locus):
15501 Change type to hash_table. Update dependent calls and types.
15503 * tree-complex.c: Include tree-hasher.h
15504 (htab_t complex_variable_components):
15505 Change type to hash_table. Update dependent calls and types.
15507 * tree-eh.c (htab_t finally_tree):
15508 Change type to hash_table. Update dependent calls and types.
15510 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
15511 struct int_tree_hasher.
15512 (extern int_tree_map_eq): Likewise.
15513 (uid_decl_map_hash): Removed.
15514 (extern decl_tree_map_eq): Likewise.
15516 * tree-hasher.h: New.
15517 (struct int_tree_hasher): New.
15518 (typedef int_tree_htab_type): New.
15520 * tree-inline.c: Include gimplify-ctx.h.
15522 * tree-mudflap.c: Include gimplify-ctx.h.
15524 * tree-parloops.c: Include tree-hasher.h.
15525 (htab_t eliminate_local_variables_stmt::decl_address):
15526 Change type to hash_table. Update dependent calls and types.
15527 (htab_t separate_decls_in_region::decl_copies): Likewise.
15529 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
15530 Change type to hash_table. Update dependent calls and types.
15532 * tree-sra.c (candidates):
15533 Change type to hash_table. Update dependent calls and types.
15535 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
15537 (int_tree_map_hash): Likewise.
15539 * tree-ssa-dom.c (htab_t avail_exprs):
15540 Change type to hash_table. Update dependent calls and types.
15542 * tree-ssa-live.c (var_map_base_init::tree_to_index):
15543 Change type to hash_table. Update dependent calls and types.
15545 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
15546 Change type to hash_table. Update dependent calls and types.
15548 * tree-ssa-phiopt.c (seen_ssa_names):
15549 Change type to hash_table. Update dependent calls and types.
15551 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
15552 Change type to hash_table. Update dependent calls and types.
15554 * tree-ssa-uncprop.c (equiv):
15555 Change type to hash_table. Update dependent calls and types.
15557 2013-04-25 Jakub Jelinek <jakub@redhat.com>
15559 PR rtl-optimization/57003
15560 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
15561 call note_stores with kill_clobbered_value callback again after
15562 killing regs_invalidated_by_call.
15564 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
15566 * config/aarch64/aarch64-simd.md
15567 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
15568 (aarch64_simd_bsl<mode>): Likewise.
15569 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
15571 2013-04-25 Marek Polacek <polacek@redhat.com>
15573 PR tree-optimization/57066
15574 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
15576 2013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
15578 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
15580 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
15582 * config/aarch64/aarch64-builtins.c
15583 (aarch64_fold_builtin): New.
15584 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
15585 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
15586 * config/aarch64/aarch64-simd-builtins.def (abs): New.
15587 * config/aarch64/arm_neon.h
15588 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
15590 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
15591 Tejas Belagod <tejas.belagod@arm.com>
15593 * config/aarch64/aarch64-builtins.c
15594 (aarch64_gimple_fold_builtin): New.
15595 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
15596 * config/aarch64/aarch64-simd-builtins.def (addv): New.
15597 * config/aarch64/aarch64-simd.md (addpv4sf): New.
15598 (addvv4sf): Update.
15599 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
15601 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
15603 * config/aarch64/aarch64.md
15604 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
15606 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
15608 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
15609 (*ngcsi_uxtw): New pattern.
15611 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15612 Julian Brown <julian@codesourcery.com>
15614 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
15615 (TB_DREG): Add T_V4HF.
15616 (v4hf_UP): New macro.
15617 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
15618 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
15619 Handle initialisation of V4HF. Adjust initialisation of reinterpret
15621 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
15622 (arm_vector_mode_supported_p): Handle V4HF.
15623 (arm_mangle_map): Handle V4HFmode.
15624 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
15625 * config/arm/arm_neon_builtins.def: Add entries for
15626 vcvtv4hfv4sf, vcvtv4sfv4hf.
15627 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
15628 (neon_vcvtv4hfv4sf): Likewise.
15629 * config/arm/neon-gen.ml: Handle half-precision floating point
15631 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
15632 * config/arm/arm_neon.h: Regenerate.
15633 * config/arm/neon.ml (type elts): Add F16.
15634 (type vectype): Add T_float16x4, T_floatHF.
15635 (type vecmode): Add V4HF.
15636 (type features): Add Requires_FP_bit feature.
15637 (elt_width): Handle F16.
15638 (elt_class): Likewise.
15639 (elt_of_class_width): Likewise.
15640 (mode_of_elt): Refactor.
15641 (type_for_elt): Handle F16, fix error messages.
15642 (vectype_size): Handle T_float16x4.
15643 (vcvt_sh): New function.
15644 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
15645 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
15646 (string_of_mode): Handle V4HF.
15647 * doc/arm-neon-intrinsics.texi: Regenerate.
15649 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
15651 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
15652 format specifier in 'X' case.
15654 2013-04-25 Alan Modra <amodra@gmail.com>
15657 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
15658 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
15659 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
15660 Repeat for many other rotate/shift and mask patterns using subregs.
15661 Name lshiftrt insns.
15662 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
15663 on WORDS_BIG_ENDIAN.
15665 2013-04-25 Alan Modra <amodra@gmail.com>
15667 * config.gcc: Support little-endian powerpc-linux targets.
15668 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
15669 (LINK_OS_LINUX_SPEC): Define.
15670 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
15671 Preserve MASK_LITTLE_ENDIAN.
15672 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
15673 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
15674 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
15675 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
15676 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
15677 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
15678 Correct fp word order for little-endian. Don't shift toc entries
15679 smaller than a word for little-endian.
15680 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
15681 (bswapdi2 splits): Correct low-part subreg for little-endian.
15682 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
15683 low/high where such is correct only for be.
15684 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
15685 little-endian for -mcall-aixdesc.
15687 2013-04-25 Alan Modra <amodra@gmail.com>
15689 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
15690 replace_equiv_address_nv.
15692 2013-04-25 Alan Modra <amodra@gmail.com>
15694 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
15696 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
15699 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
15700 * rtl.h (struct rtx_def): ...
15702 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
15704 PR rtl-optimizations/57046
15705 * lra-constraints (split_reg): Set up lra_risky_transformations_p
15706 for multi-reg splits.
15708 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
15710 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
15712 2013-04-24 Sterling Augustine <saugustine@google.com>
15714 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
15715 (comp_dir_string, debug_str_dwo_section): New.
15716 (DEBUG_STR_DWO_SECTION): Rename to ...
15717 (DEBUG_DWO_STR_SECTION): ... this.
15718 (DEBUG_NORM_STR_SECTION): Delete.
15719 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
15720 (DEBUG_STR_DWO_SECTION_FLAGS): New.
15721 (find_AT_string): Move most logic to ...
15722 (find_AT_string_in_table): ... here. New.
15723 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
15724 add_skeleton_AT_string. Delete logic.
15725 (output_skeleton_debug_sections): Remove call to
15726 add_top_level_skeleton_die_attrs.
15727 (add_comp_dir_attribute): Move logic to comp_dir_string.
15728 (dwarf2out_init): Initialize debug_str_dwo_section.
15729 (output_indirect_string): Call find_string_form.
15730 (output_indirect_strings): Rewrite.
15731 (prune_unused_types): Empty skeleton_debug_str_hash.
15732 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
15733 (dwarf2out_finish): Call output_indirect_strings.
15735 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
15737 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
15739 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
15741 * rtl.h (struct rtx_def): Add comment for field jump.
15742 (LRA_SUBREG_P): New macro.
15743 * recog.c (register_operand): Check LRA_SUBREG_P.
15744 * lra.c (lra): Add note at the end of RTL code. Align non-empty
15746 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
15747 (lra_final_code_change): Skip subreg change for operators.
15748 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
15749 if there are no operand changes.
15750 * lra-constraints.c (curr_insn_set): New.
15751 (match_reload): Set LRA_SUBREG_P.
15752 (emit_spill_move): Ditto.
15753 (check_and_process_move): Use curr_insn_set. Process only single
15754 set insns. Don't initialize sec_mem_p and change_p.
15755 (simplify_operand_subreg): Use LRA_SUBREG_P.
15756 (reg_in_class_p): New function.
15757 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
15758 of #ifdef. Add code to remove cycling.
15759 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
15760 non-null disp. Reload inner instead of disp when base and index
15761 are null. Try to put lo_sum into register.
15762 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
15763 (check_and_process_move): Move code for move cost check to
15764 simple_move_p. Remove equiv_substitution.
15765 (simple_move_p): New function.
15766 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
15767 curr_insn_set. Call check_and_process_move only for single set
15768 insns. Use the new function. Move call of check_and_process_move
15769 after operand equiv substitution and address process.
15771 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
15773 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
15774 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
15775 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
15777 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
15779 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
15781 2013-04-24 Marek Polacek <polacek@redhat.com>
15783 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
15784 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
15785 (select_loops_exit_conditions): Likewise.
15786 (number_of_iterations_for_all_loops): Likewise.
15787 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
15788 (scev_analysis): Likewise.
15790 2013-04-02 Catherine Moore <clm@codesourcery.com>
15791 Chao-ying Fu <fu@mips.com>
15793 * config/mips/micromips.md (jraddiusp): New pattern.
15794 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
15795 instruction if possible.
15797 2013-04-24 Alan Modra <amodra@gmail.com>
15799 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
15801 2013-04-24 Julian Brown <julian@codesourcery.com>
15802 Chung-Lin Tang <cltang@codesourcery.com>
15804 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
15805 dependency behavior in enumeration type DIE generation. Add TODO note
15806 to comments about future DW_FORM_sdata/udata re-work of related code.
15808 2013-04-23 Lawrence Crowl <crowl@google.com>
15810 * Makefile.in: Update as needed below.
15812 * hash-table.h (class hash_table):
15813 Correct many methods with parameter types compare_type to the correct
15814 value_type. (Correct code was unlikely to notice the change.)
15815 (hash_table::elements_with_deleted) New.
15816 (class hashtable::iterator): New.
15817 (hashtable::begin()): New.
15818 (hashtable::end()): New.
15819 (FOR_EACH_HASH_TABLE_ELEMENT): New.
15821 * statistics.c (statistics_hashes):
15822 Change type to hash_table. Update dependent calls and types.
15824 * tree-into-ssa.c (var_infos):
15825 Change type to hash_table. Update dependent calls and types.
15827 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
15828 Change type to hash_table. Update dependent calls and types.
15830 * tree-ssa-loop-im.c (struct mem_ref.refs):
15831 Change type to hash_table. Update dependent calls and types.
15833 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
15834 Change type to hash_table. Update dependent calls and types.
15836 * tree-ssa-sccvn.c (vn_tables_s::nary):
15837 Change type to hash_table. Update dependent calls and types.
15838 (vn_tables_s::phis): Likewise.
15839 (vn_tables_s::references): Likewise.
15841 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
15842 (vn_reference_eq): Update parameter and return types.
15844 * tree-ssa-structalias.c (pointer_equiv_class_table):
15845 Change type to hash_table. Update dependent calls and types.
15846 (location_equiv_class_table): Likewise.
15848 * tree-vect-data-refs.c: Consequential changes for making
15849 peeling a hash_table.
15851 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
15852 (destroy_loop_vec_info): Dependent hash_table update.
15854 * tree-vectorizer.h (peeling_htab):
15855 Change type to hash_table. Update dependent calls and types.
15857 2013-04-23 Shiva Chen <shiva0217@gmail.com>
15859 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
15860 to check the register content is equal or not.
15861 * lra-constraints.c (match_reload): Use lra_assign_reg_val
15862 to assign register content record.
15863 * lra-eliminations.c (update_reg_eliminate): Use
15864 lra_update_reg_val_offset to update register content offset.
15865 * lra-int.h (struct lra_reg): Add offset member.
15866 (lra_reg_val_equal_p): New static inline function.
15867 (lra_update_reg_val_offset): New static inline function.
15868 (lra_assign_reg_val): New static inline function.
15869 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
15870 to assign register content record.
15871 (initialize_lra_reg_info_element): Initial offset to zero.
15873 2013-04-23 Catherine Moore <clm@codesourcery.com>
15875 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
15876 operands. Record compression.
15878 2013-04-23 Xinliang David Li <davidxl@google.com>
15880 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
15882 2013-04-23 Richard Biener <rguenther@suse.de>
15884 PR middle-end/57036
15885 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
15886 parameter, only add abnormal goto edges from the copied body
15887 if the call could perform abnormal gotos.
15888 (copy_cfg_body): Adjust.
15890 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
15892 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
15894 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
15896 * coretypes.h (gimple_stmt_iterator): Add struct to make
15899 2013-04-23 Richard Biener <rguenther@suse.de>
15901 PR tree-optimization/57026
15902 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
15903 from SSA names occuring in abnormal PHI nodes.
15905 2013-04-22 Andi Kleen <ak@linux.intel.com>
15907 * lto/lto.c (print_lto_report_1): Fix LTO report names.
15909 2013-04-22 Andi Kleen <ak@linux.intel.com>
15911 * lto/lto.c (print_lto_report_1): Declare early.
15912 (read_cgraph_and_symbols): Call print_lto_report_1 early.
15914 2013-04-22 Andi Kleen <ak@linux.intel.com>
15916 * common.opt (-flto-report-wpa): Add.
15917 * doc/invoke.texi (-flto-report-wpa): Add.
15918 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
15921 2013-04-22 Xinliang David Li <davidxl@google.com>
15923 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
15924 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
15925 * Makefile.in: New dependency
15927 David Daney <ddaney.cavm@gmail.com>
15929 * configure.ac (gcc_cv_as_micromips_support): Use the
15930 --fatal-warnings option.
15931 * configure: Regenerate.
15933 2013-04-22 Marek Polacek <polacek@redhat.com>
15936 * tsan.c (instrument_expr): Don't instrument expression
15937 in case its size is zero.
15939 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
15943 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
15945 * config/alpha/alpha.c (TARGET_LRA_P): New define.
15947 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
15949 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
15950 (gimple_stmt_iterator): New typedef.
15951 * gimple.h (gimple_stmt_iterator): Rename to...
15952 (gimple_stmt_iterator_d): ... This.
15953 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
15954 trees be valid for GIMPLE and GENERIC.
15955 (TARGET_GIMPLE_FOLD_BUILTIN): New.
15956 * gimple-fold.c (gimple_fold_call): Call target hook
15957 gimple_fold_builtin.
15958 * hooks.c (hook_bool_gsiptr_false): New.
15959 * hooks.h (hook_bool_gsiptr_false): New.
15960 * target.def (fold_stmt): New.
15961 * doc/tm.texi: Regenerate.
15963 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
15966 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
15967 a set sp if no stack realignment.
15969 2013-04-22 Nick Clifton <nickc@redhat.com>
15971 * config.gcc (tilegx-linux): Extend extra_objs rather than
15973 (tilepro-linux): Likewise.
15975 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
15977 * config/aarch64/aarch64-builtins.c
15979 (CF0, CF1, CF2, CF3, CF4, CF10): New.
15980 (VAR<1-12>): Add MAP parameter.
15981 (BUILTIN_*): Likewise.
15982 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
15983 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
15984 (aarch64_ushl_n<mode>): Likewise.
15985 (aarch64_sshr_n<mode>): Likewise.
15986 (aarch64_ushr_n<mode>): Likewise.
15987 (aarch64_<maxmin><mode>): Likewise.
15988 (aarch64_sqrt<mode>): Likewise.
15989 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
15990 (vshr<q>_n_*): Likewise.
15992 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
15994 * config/aarch64/aarch64-builtins.c
15995 (aarch64_simd_builtin_type_mode): Handle SF types.
15997 (BUILTIN_GPF): Define.
15998 (aarch64_init_simd_builtins): Handle SF types.
15999 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
16000 (frecps): Likewise.
16001 (frecpx): Likewise.
16002 * config/aarch64/aarch64-simd.md
16003 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
16004 (aarch64_frecpe<mode>): New.
16005 (aarch64_frecps<mode>): Likewise.
16006 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
16007 (v8type): Add frecp<esx>.
16008 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
16009 (aarch64_frecps<mode>): Likewise.
16010 * config/aarch64/iterators.md (FRECP): New.
16011 (frecp_suffix): Likewise.
16012 * config/aarch64/arm_neon.h
16013 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
16015 2013-04-22 Christian Bruel <christian.bruel@st.com>
16018 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
16019 (REG_CLASS_NAMES): Idem.
16020 (REG_CLASS_CONTENTS): Idem.
16021 (REGCLASS_HAS_FP_REG): Idem.
16022 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
16023 (sh_conditional_register_usage): Idem.
16025 2013-04-21 Jeff Law <law@redhat.com>
16027 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
16028 (ssa_forward_propagate_and_combine): Use it.
16030 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
16032 * lra.c: Update the flow chart diagram.
16034 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
16036 PR rtl-optimization/56847
16037 * lra-constraints.c (process_alt_operands): Discourage alternative
16038 with non-matche doffsettable memory constraint fro memory with
16041 2013-04-19 Richard Biener <rguenther@suse.de>
16043 PR tree-optimization/56982
16044 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
16046 * gimplify.c (gimplify_call_expr): Notice special calls.
16047 (gimplify_modify_expr): Likewise.
16048 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
16049 abnormal control flow receivers.
16050 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
16051 in the same way as cfun->has_nonlocal_labels.
16052 (gimple_purge_dead_abnormal_call_edges): Likewise.
16053 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
16054 receivers start a basic-block.
16056 2013-04-19 Richard Biener <rguenther@suse.de>
16058 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
16060 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
16061 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
16062 (SLP_TREE_LOAD_PERMUTATION): Add.
16063 (vect_transform_slp_perm_load): Adjust prototype.
16064 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
16065 (vect_free_slp_instance): Likewise.
16066 (vect_create_new_slp_node): Likewise.
16067 (vect_supported_slp_permutation_p): Remove.
16068 (vect_slp_rearrange_stmts): Adjust.
16069 (vect_supported_load_permutation_p): Likewise. Inline
16070 vect_supported_slp_permutation_p here.
16071 (vect_analyze_slp_instance): Compute load permutations per
16072 slp node instead of per instance.
16073 (vect_get_slp_defs): Adjust.
16074 (vect_transform_slp_perm_load): Likewise.
16075 (vect_schedule_slp_instance): Remove redundant code.
16076 (vect_schedule_slp): Remove hack for PR56270, add it ...
16077 * tree-vect-stmts.c (vectorizable_load): ... here, do not
16078 CSE loads for SLP. Adjust.
16080 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
16082 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
16083 spelling in two comments.
16085 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
16088 * config/arm/arm.c (load_multiple_sequence): Require SP
16089 as base register for loads if SP is in the register list.
16091 2013-04-19 Martin Jambor <mjambor@suse.cz>
16093 PR tree-optimization/56718
16094 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
16095 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
16096 and made public. Adjusted all callers.
16097 (ipa_intraprocedural_devirtualization): New function.
16098 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
16099 (ipa_intraprocedural_devirtualization): Likewise.
16100 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
16102 2013-04-19 Richard Biener <rguenther@suse.de>
16104 PR tree-optimization/57000
16105 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
16107 2013-04-19 Terry Guo <terry.guo@arm.com>
16109 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
16111 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
16112 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
16113 (cortex_m4_fmacs): Use new reservations.
16114 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
16116 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
16118 PR rtl-optimization/56999
16119 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
16121 (lra_coalesce): Remove split_origin_bitmap and related code.
16122 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
16123 ranges if necessary.
16125 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
16127 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
16129 (ix86_expand_call): Remove clobbered_registers array and use
16130 x86_64_ms_sysv_extra_clobbered_registers instead.
16131 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
16133 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
16135 * config/i386/i386.md (*call_rex64_ms_sysv): Use
16136 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
16137 (*call_value_rex64_ms_sysv): Ditto.
16139 2013-04-18 Cary Coutant <ccoutant@google.com>
16141 * dwarf2out.c (output_pubnames): Check die_perennial_p of
16142 parent instead of die_mark.
16144 2013-04-18 Diego Novillo <dnovillo@google.com>
16146 * gimple.c (create_gimple_tmp): New.
16147 (get_expr_type): New.
16148 (build_assign): New.
16149 (build_type_cast): New.
16150 * gimple.h (enum ssa_mode): Define.
16151 (gimple_seq_set_location): New.
16152 * asan.c (build_check_stmt): Change some gimple_build_* calls
16153 to use build_assign and build_type_cast.
16155 2013-04-18 Richard Biener <rguenther@suse.de>
16157 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
16158 handle negative step. Remove redundant checks.
16159 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
16160 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
16161 for negative step and grouped loads fail to vectorize.
16163 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
16165 * emit-rtl.c (reset_insn_used_flags): New function.
16166 (reset_all_used_flags): Use it.
16167 (verify_insn_sharing): New function.
16168 (verify_rtl_sharing): Fix verification for SEQUENCEs.
16170 2013-04-18 Jakub Jelinek <jakub@redhat.com>
16172 PR tree-optimization/56984
16173 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
16174 and (x >> M) >= N don't register any assertion if N << M is the
16177 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
16179 * lower-subreg.c (resolve_simple_move): If called self-recursive,
16180 do not delete_insn insns that have not yet been emitted, only
16181 unlink them with remove_insn.
16182 * df-scan.c (df_insn_delete): Revert r197492.
16184 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
16186 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
16187 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
16189 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
16191 * config/arm/arm.md (movsicc_insn): Convert define_insn into
16192 define_insn_and_split.
16193 (and_scc,ior_scc,negscc): Likewise.
16194 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
16196 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
16198 * config/arm/arm.c (use_return_insn): Return 0 for targets that
16199 can benefit from using a sequence of LDRD instructions in epilogue
16200 instead of a single LDM instruction.
16202 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
16205 * doc/extend.texi: Fix typo.
16207 2013-04-17 Richard Biener <rguenther@suse.de>
16209 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
16210 (vect_build_slp_tree): ... here.
16211 (vect_build_slp_tree_1): Compute which stmts of the SLP group
16212 match. Remove special-casing of mismatched complex loads.
16213 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
16214 re-try the match with swapped commutative operands.
16215 (vect_supported_load_permutation_p): Remove special-casing of
16216 mismatched complex loads.
16217 (vect_analyze_slp_instance): Adjust.
16219 2013-04-17 Richard Biener <rguenther@suse.de>
16221 PR rtl-optimization/56921
16222 * cfgloop.h (struct loop): Add simple_loop_desc member.
16223 (struct niter_desc): Mark with GTY(()).
16224 (simple_loop_desc): Do not use aux field but simple_loop_desc.
16225 * loop-iv.c (get_simple_loop_desc): Likewise.
16226 (free_simple_loop_desc): Likewise.
16229 2013-04-16 Richard Biener <rguenther@suse.de>
16231 PR rtl-optimization/56921
16232 * loop-init.c (pass_rtl_move_loop_invariants): Add
16233 TODO_do_not_ggc_collect to todo_flags_finish.
16234 (pass_rtl_unswitch): Same.
16235 (pass_rtl_unroll_and_peel_loops): Same.
16236 (pass_rtl_doloop): Same.
16238 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
16240 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
16241 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
16242 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
16244 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
16245 * tree-streamer.c (record_common_node): Adjust reference in comment.
16247 2013-04-17 Terry Guo <terry.guo@arm.com>
16249 * config/arm/cortex-m4.md: Add a new bypass.
16251 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16253 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
16255 (*subs_<optab><mode>_multp2): New pattern.
16256 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
16257 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
16259 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16261 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
16262 (*subs_mul_imm_<mode>): New pattern.
16264 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
16267 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
16268 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
16269 (vsx_movti_32bit): Change j->wa to O->wa.
16271 2013-04-16 Richard Biener <rguenther@suse.de>
16273 PR rtl-optimization/56921
16274 * loop-init.c (pass_rtl_move_loop_invariants): Add
16275 TODO_do_not_ggc_collect to todo_flags_finish.
16276 (pass_rtl_unswitch): Same.
16277 (pass_rtl_unroll_and_peel_loops): Same.
16278 (pass_rtl_doloop): Same.
16280 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
16282 * config/arm/arm.c (emit_multi_reg_push): New declaration
16283 for an existing function.
16284 (arm_emit_strd_push): New function.
16285 (arm_expand_prologue): Used here.
16286 (arm_emit_ldrd_pop): New function.
16287 (arm_expand_epilogue): Used here.
16288 (arm_get_frame_offsets): Update condition.
16289 (arm_emit_multi_reg_pop): Add a special case for load of a single
16290 register with writeback.
16292 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
16294 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
16297 2013-04-16 Richard Biener <rguenther@suse.de>
16299 PR tree-optimization/56756
16300 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
16301 (first_mem_ref_loc): New.
16302 (execute_sm): Place the load temporarily before a previous
16303 access instead of in the latch edge to ensure its SSA dependencies
16304 are defined at points dominating the load.
16306 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
16308 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
16309 correct fix by moving header and footer insn to the footer of
16310 the merged basic block. Clear BB_END of the merged-away block.
16312 PR middle-end/43631
16313 * emit-rtl.c (make_note_raw): New function.
16314 (link_insn_into_chain): New static inline function.
16315 (add_insn): Use it.
16316 (add_insn_before, add_insn_after): Factor insn chain linking code...
16317 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
16318 using link_insn_into_chain.
16319 (note_outside_basic_block_p): New helper function for emit_note_after
16320 and emit_note_before.
16321 (emit_note_after): Use nobb variant of add_insn_after if the note
16322 should not be contained in a basic block.
16323 (emit_note_before): Use nobb variant of add_insn_before if the note
16324 should not be contained in a basic block.
16325 (emit_note_copy): Use make_note_raw.
16326 (emit_note): Likewise.
16327 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
16328 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
16329 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
16330 the moved barrier the tail of the basic block it follows.
16331 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
16333 2013-04-15 Jakub Jelinek <jakub@redhat.com>
16335 PR tree-optimization/56962
16336 * gimple-ssa-strength-reduction.c (record_increment): Only set
16337 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
16338 either rhs1 or rhs2 is equal to c->base_expr.
16340 2013-04-15 Richard Biener <rguenther@suse.de>
16342 PR tree-optimization/56933
16343 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
16345 (GROUP_READ_WRITE_DEPENDENCE): Remove.
16346 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
16347 * tree-vect-data-refs.c (vect_analyze_group_access): Move
16348 dependence check ...
16349 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
16351 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
16352 GROUP_READ_WRITE_DEPENDENCE.
16354 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16356 * emit-rtl.c (reset_all_used_flags): New function.
16357 (verify_rtl_sharing): Call reset_all_used_flags before and after
16358 performing the checks.
16360 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16362 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
16363 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
16364 * config/arm/constraints.md (De): New constraint.
16365 * config/arm/neon.md (anddi3_neon): Delete.
16366 (neon_vand<mode>): Expand to standard anddi3 pattern.
16367 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
16368 Move earlier in the file.
16369 (neon_inv_logic_op2): Likewise.
16370 (arm_anddi_operand_neon): New predicate.
16372 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16374 * configure.ac (gcc_cv_ld_as_needed): Set
16375 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
16376 Use -z ignore, -z record on *-*-solaris2*.
16377 (HAVE_LD_AS_NEEDED): Update comment.
16378 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
16379 * configure: Regenerate.
16380 * config.in: Regenerate.
16381 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
16382 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
16383 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
16384 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
16385 equivalents. Fix markup.
16386 * doc/tm.texi: Regenerate.
16388 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
16390 * config/i386/i386.opt: New option mstack-protector-guard=.
16391 * config/i386/i386-opts.h: Add enum stack_protector_guard.
16392 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
16393 TARGET_SSP_TLS_GUARD.
16394 * config/i386/i386.c (ix86_option_override_internal): Set
16395 ix86_stack_protector_guard.
16396 * config/i386/i386.md (stack_protect_set): Enable for
16397 TARGET_SSP_TLS_GUARD only.
16398 (stack_protect_set_<mode>): Ditto.
16399 (stack_protect_test): Ditto.
16400 (stack_protect_test_<mode>): Ditto.
16401 * doc/invoke.texi (i386 Option): Document.
16403 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
16406 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
16407 (S_MODES): Set H_MODE bit.
16408 (SF_MODES): Set only S_MODE and SF_MODE bits.
16409 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
16410 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
16411 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
16412 <MODE_FLOAT>: Likewise.
16414 2013-04-15 Joey Ye <joey.ye@arm.com>
16416 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
16418 2013-04-15 Joey Ye <joey.ye@arm.com>
16420 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
16422 (thumb_far_jump_used_p): Count instruction size and set
16425 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
16427 * reorg.c (fill_simple_delay_slots): Reindent block of code.
16428 * resource.c (mark_target_live_regs): Reformat conditional block.
16430 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
16432 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
16433 notes, they are emitted only just before final.
16434 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
16436 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
16438 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
16439 * cfgrtl.c (delete_insn): Call it here instead.
16440 * lra-spills.c (lra_final_code_change): Use delete_insn.
16441 * haifa-sched.c (sched_remove_insn): Likewise.
16442 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
16443 returning to the nop pool.
16444 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
16445 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
16447 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
16449 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
16450 * doc/tm.texi: Regenerated.
16452 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
16454 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
16457 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
16459 * df-core.c (df_find_def): Compare register numbers.
16460 (df_find_use): Likewise.
16462 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
16465 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
16466 lra_in_progress for return.
16468 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
16470 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
16471 define_insn into define_insn_and_split and emit movsicc patterns.
16473 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
16475 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
16477 2013-04-12 Richard Biener <rguenther@suse.de>
16479 * tree-pass.h (TODO_do_not_ggc_collect): New.
16480 * passes.c (execute_one_ipa_transform_pass): Honor
16481 TODO_do_not_ggc_collect.
16482 (execute_one_pass): Likewise.
16485 2013-04-10 Richard Biener <rguenther@suse.de>
16487 * passes.c (init_optimization_passes): Remove reload pass.
16488 * ira.c (do_reload): Merge into ...
16490 (rest_of_handle_reload): Remove.
16491 (pass_reload): Likewise.
16492 * config/i386/i386.c (ix86_option_override): Refer to ira instead
16493 of reload for vzeroupper pass placement.
16495 2013-04-12 Jakub Jelinek <jakub@redhat.com>
16497 PR tree-optimization/56918
16498 PR tree-optimization/56920
16499 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
16500 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
16501 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
16502 use wide_mul_with_sign method.
16504 2013-04-12 Richard Biener <rguenther@suse.de>
16506 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
16507 not be considered a gimple constant.
16509 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
16511 * fold-const.c (const_binop): Handle vector shifts by a scalar.
16512 (fold_binary_loc): Call const_binop also for mixed vector-scalar
16515 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
16516 Jakub Jelinek <jakub@redhat.com>
16518 * opts.c: Include diagnostic-color.h.
16519 (common_handle_option): Handle OPT_fdiagnostics_color_.
16520 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
16521 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
16522 (diagnostic-color.o): New.
16523 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
16524 (diagnostic_color_rule): New enum.
16525 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
16526 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
16527 the location string.
16528 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
16529 either NULL, or color kind.
16530 * diagnostic-color.c: New file.
16531 * diagnostic-color.h: New file.
16532 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
16534 * doc/invoke.texi (-fdiagnostics-color): Document.
16535 * pretty-print.h (pp_show_color): Define.
16536 (struct pretty_print_info): Add show_color field.
16537 * diagnostic.c: Include diagnostic-color.h.
16538 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
16539 macros. Colorize error:, warning: etc. strings and also the location
16541 (diagnostic_show_locus): Colorize the caret line.
16542 * pretty-print.c: Include diagnostic-color.h.
16543 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
16544 inside of %< %> quotes or quoted through q format modifier.
16546 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16548 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
16550 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16552 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
16553 code in CC_NZ mode.
16554 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
16557 2013-04-11 Marek Polacek <polacek@redhat.com>
16559 PR tree-optimization/48184
16560 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
16562 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
16564 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
16565 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
16566 (skip_simple_arithmetic): Tidy up.
16567 * tree.h (skip_simple_constant_arithmetic): Declare.
16569 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16571 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
16573 2013-04-11 Richard Biener <rguenther@suse.de>
16575 * tree-vect-loop.c (get_initial_def_for_induction): Properly
16576 generate vector constants.
16578 2013-04-11 Richard Biener <rguenther@suse.de>
16580 PR tree-optimization/56878
16581 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
16582 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
16584 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
16585 Prefer to align the DR with the most invariant base address.
16587 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
16589 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
16592 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
16594 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
16595 floating-point vector comparisons against 0.
16597 2013-04-11 Jakub Jelinek <jakub@redhat.com>
16599 PR tree-optimization/56899
16600 * fold-const.c (extract_muldiv_1): Apply distributive law
16601 only if TYPE_OVERFLOW_WRAPS (ctype).
16603 2013-04-11 Bin Cheng <bin.cheng@arm.com>
16606 * ira-costs.c (scan_one_insn): Check whether the source rtx of
16607 loading has side effect.
16609 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
16611 * config/sparc/sparc.c: Include tree-pass.h.
16612 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
16613 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
16614 head of file. Change return type. Split off gate function.
16615 (sparc_gate_work_around_errata): New function.
16616 (pass_work_around_errata): New pass definition.
16617 (insert_pass_work_around_errata) New pass insert definition to
16618 insert pass_work_around_errata just after delayed-branch scheduling.
16619 (sparc_option_override): Insert the pass.
16620 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
16622 2013-04-10 David S. Miller <davem@davemloft.net>
16624 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
16625 or -mcpu=hypersparc.
16627 * target.def (cstore_mode): New hook.
16628 * target.h: Include insn-codes.h
16629 * targhooks.c: Likewise.
16630 (default_cstore_mode): New function.
16631 * targhooks.h: Declare it.
16632 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
16633 * doc/tm.texi: Rebuild.
16634 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
16635 target hook, rather than inspecting the insn_data.
16636 * config/sparc/sparc.c (sparc_cstore_mode): New function.
16637 (TARGET_CSTORE_MODE): Redefine.
16638 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
16640 * config/sparc/predicates.md (cstore_result_operand): New special
16642 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
16643 Use it for operand 0.
16644 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
16645 (*snesi_special): Likewise.
16646 (*snesi_zero): Likewise.
16647 (*seqsi_zero): Likewise.
16648 (*sltu_insn): Likewise.
16649 (*sgeu_insn): Likewise.
16650 (*seqdi_special): Make operand 0 and comparison operation be of
16652 (*snedi_special): Likewise.
16653 (*snedi_special_vis3): Likewise.
16654 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
16655 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
16656 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
16657 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
16658 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
16659 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
16660 (*sltu_extend_sp64): Likewise.
16661 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
16662 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
16663 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
16664 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
16665 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
16667 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
16669 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
16670 (aarch64_start_file): Use the new function.
16672 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
16673 Jason Merrill <jason@redhat.com>
16675 * common.opt: Add -gdwarf.
16676 * opts.c (common_handle_option): Handle it.
16677 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
16679 2013-04-10 Richard Biener <rguenther@suse.de>
16681 * passes.c (execute_todo): Do not call ggc_collect conditional here.
16682 (execute_one_ipa_transform_pass): But unconditionally here.
16683 (execute_one_pass): And here.
16684 (init_optimization_passes): Remove reload pass.
16685 * tree-pass.h (TODO_ggc_collect): Remove.
16686 (pass_reload): Likewise.
16687 * ira.c (do_reload): Merge into ...
16689 (rest_of_handle_reload): Remove.
16690 (pass_reload): Likewise.
16691 * config/i386/i386.c (ix86_option_override): Refer to ira instead
16692 of reload for vzeroupper pass placement.
16693 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
16694 and todo_flags_finish of all passes.
16696 2013-04-10 Richard Biener <rguenther@suse.de>
16698 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
16699 first_const_oprnd field, rename first_def_type to first_op_type.
16700 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
16701 (vect_get_and_check_slp_defs): Always use the type of the
16702 operand. Allow mixed vect_external_def, vect_constant_def types.
16703 (vect_get_constant_vectors): Handle mixed vect_external_def,
16704 vect_constant_def types.
16706 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
16708 PR tree-optimization/55524
16709 * tree-ssa-math-opts.c
16710 (convert_mult_to_fma): Don't use an fms construct
16711 when we don't have an fms operation, but fnma, and it looks
16712 likely that we'll be able to use the latter.
16714 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
16716 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
16718 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
16719 inline fail caused by overwritable functions.
16721 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
16723 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
16724 unnecessary bits in the constant power of two case.
16726 2013-04-10 Richard Biener <rguenther@suse.de>
16728 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
16729 broken code swapping operands.
16730 (vect_build_slp_tree): Do not compute load permutations here.
16731 (vect_analyze_slp_instance): Compute load permutations here,
16732 after building the SLP tree.
16734 2013-04-09 Christian Bruel <christian.bruel@st.com>
16736 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
16737 of next/prev_real_insn.
16739 2013-04-09 Jan Hubicka <jh@suse.cz>
16741 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
16742 Drop aliased parameter.
16743 (function_and_variable_visibility): Do not handle alias pairs.
16744 * cgraph.c (varpool_externally_visible_p): Update prototype.
16745 * varpool.c (varpool_add_new_variable): Update.
16747 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16749 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
16751 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
16753 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
16755 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
16757 2013-04-09 Marek Polacek <polacek@redhat.com>
16759 PR tree-optimization/48762
16760 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
16762 2013-04-09 Richard Biener <rguenther@suse.de>
16764 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
16766 (vect_build_slp_tree): Likewise.
16767 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
16768 calculating the cost of a SLP instance.
16769 (vect_analyze_slp_instance): Use it from here, after building
16772 2013-04-09 Jakub Jelinek <jakub@redhat.com>
16774 PR middle-end/56883
16775 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
16776 expand_omp_for_static_chunk): Use simple_p = true in
16777 force_gimple_operand_gsi calls when assigning to addressable decls.
16779 2013-04-09 Jeff Law <law@redhat.com>
16781 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
16782 when the boolean was created by converting a wider object which
16783 had a boolean range.
16785 2013-04-09 Richard Biener <rguenther@suse.de>
16787 * tree-vectorizer.h (slp_void_p): Remove.
16788 (slp_tree): Typedef before _slp_tree declaration.
16789 (struct _slp_tree): Use a vector of slp_tree as children.
16790 (vect_get_place_in_interleaving_chain): Remove.
16791 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
16793 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
16795 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
16796 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
16797 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
16798 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
16799 Use slp_node instead of slp_void_p and adjust.
16801 2013-04-09 Richard Biener <rguenther@suse.de>
16803 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
16804 work that is not necessary.
16806 2013-04-09 Jakub Jelinek <jakub@redhat.com>
16808 PR tree-optimization/56854
16809 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
16810 forward into clobber stmts if it would change MEM_REF lhs into
16813 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
16815 * tree.c (type_hash_lookup, type_hash_add): Make static.
16816 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
16818 2013-04-09 Richard Biener <rguenther@suse.de>
16820 * tree.h (unsave_expr_now): Remove.
16821 * tree-inline.c (mark_local_for_remap_r): Remove.
16822 (unsave_expr_1): Likewise.
16823 (unsave_r): Likewise.
16824 (unsave_expr_now): Likewise.
16825 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
16826 (propagate_tree_value): Likewise.
16828 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
16830 * doc/rtl.texi (sequence): Rewrite documentation to match the
16831 current use of SEQUENCE rtl objects.
16832 * rtl.def (SEQUENCE): Likewise.
16834 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
16835 Update documentation.
16836 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
16837 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
16839 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
16841 2013-04-08 Teresa Johnson <tejohnson@google.com>
16843 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
16844 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
16846 (estimate_edge_size_and_time): Add comment to suggest using rounding
16848 (estimate_node_size_and_time): Ditto.
16849 (remap_edge_change_prob): Use helper rounding divide methods.
16850 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
16851 (gimple_mod_pow2_value_transform): Ditto.
16852 (gimple_mod_subtract_transform): Ditto.
16853 (gimple_ic_transform): Ditto.
16854 (gimple_stringops_transform): Ditto.
16855 * stmt.c (conditional_probability): Ditto.
16856 (emit_case_dispatch_table): Ditto.
16857 * lto-cgraph.c (merge_profile_summaries): Ditto.
16858 * tree-optimize.c (execute_fixup_cfg): Ditto.
16859 * cfgcleanup.c (try_forward_edges): Ditto.
16860 * cfgloopmanip.c (scale_loop_profile): Ditto.
16862 (duplicate_loop_to_header_edge): Ditto.
16863 (lv_adjust_loop_entry_edge): Ditto.
16864 * tree-vect-loop.c (vect_transform_loop): Ditto.
16865 * profile.c (compute_branch_probabilities): Ditto.
16866 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
16867 * lto-streamer-in.c (input_cfg): Ditto.
16868 * gimple-streamer-in.c (input_bb): Ditto.
16869 * ipa-cp.c (update_profiling_info): Ditto.
16870 (update_specialized_profile): Ditto.
16871 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
16872 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
16874 * sched-rgn.c (compute_dom_prob_ps): Ditto.
16875 (compute_trg_info): Ditto.
16876 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
16877 (purge_dead_edges): Ditto.
16878 * loop-unswitch.c (unswitch_loop): Ditto.
16879 * cgraphclones.c (cgraph_clone_edge): Ditto.
16880 (cgraph_clone_node): Ditto.
16881 * tree-inline.c (copy_bb): Ditto.
16882 (copy_edges_for_bb): Ditto.
16883 (initialize_cfun): Ditto.
16884 (copy_cfg_body): Ditto.
16885 (expand_call_inline): Ditto.
16887 2013-04-08 Kai Tietz <ktietz@redhat.com>
16889 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
16890 TARGET_CYGWIN64 by TARGET_64BIT.
16892 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
16894 * config/epiphany/epiphany.md (GPR_1): New constant.
16895 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
16896 * config/epiphany/epiphany.c (gen_compare_reg):
16897 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
16898 is already in place.
16899 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
16900 Don't require being called during rtl expansion; If y operlaps r0,
16902 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
16903 (epiphany_expand_epilogue): Likewise.
16905 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
16906 Don't use CC_FPmode for ORDERED / UNORDERED.
16907 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
16909 * config/epiphany/constraints.md (CnL): New constraint.
16910 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
16911 * config/epiphany/predicates.md (add_operand): Allow 1024.
16913 * config/epiphany/epiphany.md (logical_op): New code iterator.
16914 (op_mnc): New code attribute.
16915 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
16916 (mov_f+1, mov_f+2): New peephole2 patterns.
16918 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
16919 (cstoresi4): Also allow re-use of zero result when doing a NE
16920 comparison to a non-zero operand.
16921 Use (clobber (scratch)) for first insn if the gpr output is not needed.
16923 * config/epiphany/epiphany.md (<insn_opname>v2si3):
16924 Use gen_addsi3_i / gen_subsi3_i.
16926 2013-04-08 Jakub Jelinek <jakub@redhat.com>
16930 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
16931 contain anything but clobbers, at most one __builtin_stack_restore,
16932 optionally debug stmts and final resx, and if it has at least one
16933 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
16934 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
16935 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
16936 which isn't defaut definition, remove them.
16937 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
16938 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
16939 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
16940 with MEM_REF LHS with SSA_NAME address.
16942 2013-04-08 Jeff Law <law@redhat.com>
16944 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
16946 2013-04-08 Richard Biener <rguenther@suse.de>
16948 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
16950 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
16951 determined vector type.
16952 (vect_analyze_data_refs): Likewise.
16953 (vect_get_new_vect_var): Adjust.
16954 (vect_create_destination_var): Preserve SSA name versions.
16955 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
16956 not dump anything here.
16958 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
16960 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
16961 Add member lr_slot_known.
16962 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
16964 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
16965 Remove code that sets lr_slot_offset according to what a previous
16966 version of epiphany_emit_save_restore used to do.
16967 (epiphany_emit_save_restore): When doing an lr save or restore,
16968 set/verify lr_slot_known and lr_slot_offset.
16970 2013-04-08 Xinyu Qi <xyqi@marvell.com>
16973 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
16976 2013-04-08 Richard Biener <rguenther@suse.de>
16978 * alias.c (find_base_term): Fix thinko in previous change.
16980 2013-04-08 Jakub Jelinek <jakub@redhat.com>
16982 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
16983 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
16984 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
16985 if possible to compute val.
16986 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
16987 For QImode integers don't require anything about precision. Use
16988 const_with_all_bytes_same to find out if the constant doesn't have
16989 repeated bytes in it.
16991 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16993 * config/s390/s390.c (s390_expand_insv): Only accept insertions
16996 2013-04-08 Marek Polacek <polacek@redhat.com>
16998 PR rtl-optimization/48182
16999 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
17002 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17005 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
17006 nuses, make sure we have a label.
17008 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17011 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
17012 (rs6000_emit_swdiv_low_precision): Remove.
17013 (rs6000_emit_swdiv): Rewrite to handle between one and four
17014 iterations of Newton-Raphson generally; modify required number of
17015 iterations for some cases.
17016 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
17018 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
17020 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
17021 set-but-unused variable.
17023 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
17024 basic blocks of released function bodies garbage-collectable.
17026 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
17027 (struct rtl_opt_pass): Add TODO_df_finish.
17029 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
17031 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17033 * config/arm/constraints.md (q): New constraint.
17034 * config/arm/ldrdstrd.md: New file.
17035 * config/arm/arm.md (ldrdstrd.md) New include.
17036 (arm_movdi): Use "q" instead of "r" constraint
17037 for double-word memory access.
17038 (movdf_soft_insn): Likewise.
17039 * config/arm/vfp.md (movdi_vfp): Likewise.
17040 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
17041 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
17042 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
17043 (mem_ok_for_ldrd_strd): Likewise.
17044 (output_move_double): Update assertion.
17046 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17048 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
17050 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17052 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
17053 define_insn_and_split.
17054 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
17056 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17058 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
17059 define_insn_and_split.
17060 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
17061 (shiftsi3_compare): New pattern.
17062 (rrx): New pattern.
17063 * config/arm/unspecs.md (UNSPEC_RRX): New.
17065 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17067 * config/arm/arm.md (negdi_extendsidi): New pattern.
17068 (negdi_zero_extendsidi): Likewise.
17070 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17072 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
17073 define_insn_and_split.
17074 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
17075 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
17077 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17079 * config/arm/arm.md (arm_subdi3): Convert define_insn into
17080 define_insn_and_split.
17081 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
17082 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
17084 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17086 * config/arm/arm.md (subsi3_carryin): New pattern.
17087 (subsi3_carryin_const): Likewise.
17088 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
17089 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
17091 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17093 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
17095 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17097 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
17098 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
17100 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17102 * config/arm/arm.c (arm_expand_builtin): Change fcode
17103 type to unsigned int.
17105 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17107 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
17109 2013-04-04 Ian Lance Taylor <iant@google.com>
17111 * doc/standards.texi (Standards): The Go frontend supports the Go 1
17114 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
17116 PR middle-end/56729
17117 * df-scan.c (df_insn_delete): Disable failing assert.
17119 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17121 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
17122 New function prototype.
17123 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
17124 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
17125 (arm_builtin_vectorized_function): New function.
17127 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17129 * config/arm/arm_neon_builtins.def: New file.
17130 * config/arm/arm.c (neon_builtin_data): Move contents to
17131 arm_neon_builtins.def.
17132 (enum arm_builtins): Include neon builtin definitions.
17133 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
17134 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
17136 2013-04-04 Marek Polacek <polacek@redhat.com>
17138 PR tree-optimization/48186
17139 * predict.c (maybe_hot_frequency_p): Return false if
17140 HOT_BB_FREQUENCY_FRACTION is 0.
17141 (cgraph_maybe_hot_edge_p): Likewise.
17143 2013-04-04 Richard Biener <rguenther@suse.de>
17145 PR tree-optimization/56826
17146 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
17149 2013-04-04 Richard Biener <rguenther@suse.de>
17151 PR tree-optimization/56213
17152 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
17153 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
17155 2013-04-04 Richard Biener <rguenther@suse.de>
17157 PR tree-optimization/56837
17158 * tree-loop-distribution.c (classify_partition): For non-zero
17159 values require that the value has the same precision as its
17160 mode to be useful as memset value.
17162 2013-04-03 Nick Clifton <nickc@redhat.com>
17164 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
17165 (fmssf4): Use fmsf.s on E3V5 architectures.
17166 (fnmasf4): Use fnmaf.s on E3V5 architectures.
17167 (fnmssf4): Use fnmsf.s on E3V5 architectures.
17169 2013-04-03 Jeff Law <law@redhat.com>
17171 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
17172 (lra-eliminations.o): Likewise.
17174 2013-04-03 Teresa Johnson <tejohnson@google.com>
17176 * gcov-io.c (compute_working_sets): Moved most of body of old
17177 compute_working_sets here from profile.c.
17178 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
17179 (gcov_working_set_t): Moved typedef here from basic-block.h
17180 (compute_working_set): Declare.
17181 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
17182 (get_working_sets): Renamed from compute_working_set,
17183 replace most of body with call to new compute_working_sets.
17184 (get_exec_counts): Replace call to compute_working_sets
17185 to get_working_sets.
17186 * profile.h (get_working_sets): Renamed from compute_working_set.
17187 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
17188 to get_working_sets.
17189 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
17190 * gcov-dump.c (dump_working_sets): New function.
17192 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
17194 * hwint.c (sext_hwi, zext_hwi): New functions.
17195 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
17196 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
17197 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
17198 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
17199 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
17200 (sext_hwi, zext_hwi): New functions.
17202 2013-04-03 Jeff Law <law@redhat.com>
17204 PR tree-optimization/56799
17205 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
17206 back test for widening conversion erroneously dropped in prior change.
17208 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17211 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
17212 instead of next_real_insn.
17214 2013-04-03 Marek Polacek <polacek@redhat.com>
17217 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
17219 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17222 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
17224 (thumb1_output_casesi): Likewise.
17225 (thumb2_output_casesi): Likewise.
17227 2013-04-03 Richard Biener <rguenther@suse.de>
17229 PR tree-optimization/56817
17230 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
17232 (tree_unroll_loops_completely_1): ... new function to manually
17233 walk the loop tree, properly defering outer loops of unrolled
17234 loops to later iterations.
17236 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
17238 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
17239 (vectorizable_load): Likewise.
17240 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
17241 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
17243 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
17245 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
17248 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
17250 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
17252 2013-04-03 Bin Cheng <bin.cheng@arm.com>
17254 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
17256 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
17258 PR tree-optimization/56790
17259 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
17262 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
17264 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
17266 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
17267 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
17270 2013-04-03 Jakub Jelinek <jakub@redhat.com>
17273 * tree.h (force_folding_builtin_constant_p): New decl.
17274 * builtins.c (force_folding_builtin_constant_p): New variable.
17275 (fold_builtin_constant_p): Fold immediately also if
17276 force_folding_builtin_constant_p.
17278 2013-04-03 Richard Biener <rguenther@suse.de>
17280 PR tree-optimization/56812
17281 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
17282 DRs of the same interleaving chain are independent.
17284 2013-04-02 Jason Merrill <jason@redhat.com>
17286 * gdbinit.in (pbb): Use debug fn.
17288 2013-04-02 Lawrence Crowl <crowl@google.com>
17290 * sese.h (struct ivtype_map_elt_s): Remove unused.
17291 (extern debug_ivtype_map): Remove unused.
17292 (extern eq_ivtype_map_elts): Remove unused.
17293 * sese.c (debug_ivtype_map): Removed unused.
17294 (debug_ivtype_map_1): Removed unused.
17295 (debug_ivtype_elt): Remove unused.
17296 (eq_ivtype_map_elts): Remove unused.
17299 2013-04-02 Kai Tietz <ktietz@redhat.com>
17302 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
17303 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
17304 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
17306 (legitimize_pe_coff_symbol): Likewise.
17307 (is_imported_p): New helper-function.
17308 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
17309 for Windows x64 targets.
17310 (ix86_expand_prologue): Optimize for pe-coff targets.
17311 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
17312 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
17313 medium/large code-model.
17314 (legitimize_pic_address): Likewise.
17315 (legitimize_tls_address): Likewise.
17316 (ix86_expand_call): Likewise.
17317 (x86_output_mi_thunk): Likewise.
17318 (get_dllimport_decl): Add new beimport argument.
17319 (construct_plt_address): Don't assert for x64 pe-coff targets.
17320 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
17322 (SYMBOL_FLAG_STUBVAR): New macro.
17323 (SYMBOL_REF_STUBVAR_P): Likewise.
17324 * config/i386/winnt.c (stub_list): New structure.
17325 (stub_head): New local variable.
17326 (i386_pe_record_stub): New function.
17327 (i386_pe_file_end): Emit refptr-stubs.
17329 2013-04-02 Jakub Jelinek <jakub@redhat.com>
17331 PR rtl-optimization/56745
17332 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
17333 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
17336 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
17337 and both of them are MEM_REFs, just compare first argument for
17338 equality and attempt to deal even with differing offsets.
17341 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
17342 of gimple_clobber_p to be MEM_REF.
17343 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
17344 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
17345 after gimplification.
17346 * asan.c (get_mem_ref_of_assignment): Don't instrument
17347 gimple_clobber_p stmts.
17348 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
17349 gimple_clobber_p stmt if they have MEM_REF lhs and
17350 are dead because of another gimple_clobber_p stmt.
17351 * tree-ssa-live.c (clear_unused_block_pointer): Treat
17352 gimple_clobber_p stmts like debug stmts.
17353 (remove_unused_locals): Remove clobbers with MEM_REF lhs
17354 that refer to unused VAR_DECLs or uninitialized values.
17355 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
17356 gimple_clobber_p stmts if they refer to removed parameters.
17357 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
17360 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
17362 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
17363 using SWI48 mode attribute.
17365 2013-04-02 Wei Mi <wmi@google.com>
17367 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
17368 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
17369 *<rotate_insn><mode>3_mask in i386.md.
17371 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
17373 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
17375 2013-04-02 Richard Biener <rguenther@suse.de>
17377 PR tree-optimization/56778
17378 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
17379 Runtime alias tests are not supported for gather loads.
17380 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
17381 stmts referenced from SSA operands before updating SSA form.
17383 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
17384 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17386 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
17387 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
17388 * config/arm/cortex-a53.md: New file.
17389 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
17390 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
17391 * config/arm/arm.c (arm_issue_rate): Likewise.
17392 * config/arm/arm-tune.md: Regenerate
17393 * config/arm/arm-tables.opt: Regenerate.
17394 * config/arm/arm-cores.def: Add cortex-a53.
17396 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
17398 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
17401 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
17403 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
17404 scalar load/store operations using B/H registers.
17405 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
17407 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
17409 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
17411 * config/aarch64/aarch64.c
17412 (aarch64_simd_scalar_immediate_valid_for_move): New.
17413 * config/aarch64/aarch64-protos.h
17414 (aarch64_simd_scalar_immediate_valid_for_move): New.
17415 * config/aarch64/constraints.md (Dh, Dq): New.
17416 * config/aarch64/iterators.md (hq): New.
17418 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
17420 * reorg.c (get_branch_condition): Deal with conditional returns.
17421 (fill_simple_delay_slots): Remove dead code dealing with jumps.
17423 2013-04-01 Wei Mi <wmi@google.com>
17425 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
17426 Truncate operand 2 using %b asm operand modifier.
17427 (*<shift_insn><mode>3_mask): Ditto.
17428 (*<rotate_insn><mode>3_mask): Ditto.
17430 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
17432 PR middle-end/56798
17433 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
17435 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
17437 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
17439 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
17441 2013-03-30 Lawrence Crowl <crowl@google.com>
17443 * dse.c (clear_alias_sets): Remove never set.
17444 (disqualified_clear_alias_sets): Remove never set.
17445 (clear_alias_mode_pool): Remove never set.
17446 (dse_step0): Remove condition that is never true.
17447 (canon_address): Remove condition that is never true.
17448 (dse_step7): Remove condition that is never true.
17449 (rest_of_handle_dse): Remove condition that is never true.
17450 (rest_of_handle_dse::did_global): Remove never read from above.
17451 (dse_step2_spill): Remove never called from above.
17452 (dse_step5_spill): Remove never called from above.
17454 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
17456 * doc/md.texi (Standard Names) <casesi>: Update documentation for
17457 JUMP_TABLE_DATA changes.
17458 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
17459 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
17460 (Insns) <jump_table_data>: New entry.
17461 * doc/tm.texi: Regenerate.
17463 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
17465 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
17466 for table jump at the end of a basic block using tablejump_p.
17467 * targhooks.c (default_invalid_within_doloop): Likewise.
17468 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
17469 target hook implementation that is identical to the default hook.
17470 (rs6000_invalid_within_doloop): Remove.
17472 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
17473 unused variable from tablejump_p call.
17475 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
17476 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
17477 (INSN_DELETED_P): Likewise.
17478 (emit_jump_table_data): New prototype.
17479 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
17480 after 4th as unused.
17481 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
17482 * sched-vis.c (print_insn): Likewise.
17483 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
17484 insn for compatibility with back ends that use next_active_insn to
17485 identify jump table data.
17486 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
17487 (remove_insn): Likewise.
17488 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
17490 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
17491 (emit_jump_table_data): New function.
17493 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
17494 basic block, a JUMP_TABLE_DATA never is.
17495 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
17496 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
17497 off from code handling real insns.
17498 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
17499 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
17500 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
17501 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
17502 is not a NONDEBUG_INSN_P.
17503 * ira-costs.c (scan_one_insn): Likewise.
17504 * jump.c (mark_all_labels): Likewise.
17505 (mark_jump_label_1): Likewise.
17506 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
17507 * lra.c (get_insn_freq): Expect all insns reaching here to be in
17509 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
17510 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
17511 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
17512 JUMP_TABLE_DATA_P insns.
17513 (calculate_elim_costs_all_insns): Likewise.
17514 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
17515 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
17516 (delete_output_reload): Code style fixups.
17517 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
17518 insn flags on this non-insn.
17519 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
17520 as scheduling barriers, for pre-change compatibility.
17521 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
17522 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
17524 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
17525 redundant JUMP_TABLE_DATA_P test.
17526 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
17527 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
17528 (frv_for_each_packet): Likewise.
17529 * config/i386/i386.c (min_insn_size): Likewise.
17530 (ix86_avoid_jump_mispredicts): Likewise.
17531 * config/m32r/m32r.c (m32r_is_insn): Likewise.
17532 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
17533 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
17534 (mips16_insn_length): Robustify.
17535 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
17536 (mips16_split_long_branches): Likewise.
17537 * config/pa/pa.c (pa_combine_instructions): Likewise.
17538 * config/rs6000/rs6000.c (get_next_active_insn): Treat
17539 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
17540 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
17541 as contributing to pool range lengths.
17542 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
17543 Remove redundant JUMP_TABLE_DATA_P test.
17544 (sh_loop_align): Likewise.
17545 (split_branches): Likewise.
17546 (sh_insn_length_adjustment): Likewise.
17547 * config/spu/spu.c (get_branch_target): Likewise.
17549 2013-03-29 Jan Hubicka <jh@suse.cz>
17551 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
17552 gcov streaming; stream hot bb threshold to ltrans.
17553 * predict.c (get_hot_bb_threshold): Break out from ....
17554 (maybe_hot_count_p): ... here.
17555 (set_hot_bb_threshold): New function.
17556 * lto-section-in.c (lto_section_name): Add profile.
17557 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
17558 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
17559 and data-streamer.h
17560 (histogram_entry): New structure.
17561 (histogram, histogram_pool): New global vars.
17562 (histogram_hash): New structure.
17563 (histogram_hash::hash): New method.
17564 (histogram_hash::equal): Likewise.
17565 (account_time_size): New function.
17566 (cmp_counts): New function.
17567 (dump_histogram): New function.
17568 (ipa_profile_generate_summary): New function.
17569 (ipa_profile_write_summary): New function.
17570 (ipa_profile_read_summary): New function.
17571 (ipa_profile): Decide on threshold.
17572 (pass_ipa_profile): Add ipa_profile_write_summary and
17573 ipa_profile_read_summary.
17574 * Makefile.in (ipa.o): Update dependencies.
17575 * lto-streamer.h (LTO_section_ipa_profile): New section.
17577 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
17579 * tree.h (VAR_P): New.
17581 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
17584 * doc/invoke.texi ([-fwhole-program]): Fix typo.
17586 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
17588 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
17589 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
17590 (control_flow_insn_p): Likewise.
17591 * cfgrtl.c (duplicate_insn_chain): Likewise.
17592 * final.c (get_attr_length_1): Likewise.
17593 (shorten_branches): Likewise.
17594 (final_scan_insn): Likewise.
17595 * function.c (instantiate_virtual_regs): Likewise.
17596 * gcse.c (insert_insn_end_basic_block): Likewise.
17597 * ira-costs.c (scan_one_insn): Likewise.
17598 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
17599 * lra.c (check_rtl): Likewise.
17600 * reload1.c (elimination_costs_in_insn): Likewise.
17601 * reorg.c (follow_jumps): Likewise.
17603 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
17604 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
17605 (thumb_far_jump_used_p): Likewise.
17606 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
17607 (workaround_speculation): Likewise.
17608 (add_sched_insns_for_speculation): Likewise.
17609 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
17610 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
17611 (frv_for_each_packet): Likewise.
17612 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
17613 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
17614 (final_emit_insn_group_barriers): Likewise.
17615 * config/m32r/m32r.c (m32r_is_insn): Likewise.
17616 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
17617 (mips16_insn_length): Likewise.
17618 * config/pa/pa.c (pa_reorg): Likewise.
17619 (pa_combine_instructions): Likewise.
17620 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
17621 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
17622 (sh_reorg): Likewise.
17623 (split_branches): Likewise.
17624 * config/spu/spu.c (get_branch_target): Likewise.
17626 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
17629 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
17631 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
17632 Fix declaration name.
17634 2013-03-28 Lawrence Crowl <crowl@google.com>
17636 * graphds.h (struct graph.indicies): Remove unused.
17637 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
17638 (SCOP_ORIGINAL_PDDRS): Remove unused.
17639 * sese.h (extern insert_loop_close_phis): Removed unused.
17640 (extern insert_guard_phis): Removed unused.
17641 (extern ivtype_map_elt_info): Removed unused.
17642 (new_ivtype_map_elt): Removed unused.
17643 * sese.c (ivtype_map_elt_info): Removed unused.
17645 2013-03-28 Lawrence Crowl <crowl@google.com>
17647 * Makefile.in: Add several missing include dependences.
17649 (test-dump.o): New. This object is not added to any executable,
17650 but is present for ad-hoc testing.
17652 (debug (const bitmap_head_def &)): New.
17653 (debug (const bitmap_head_def *)): New.
17655 (extern debug (const bitmap_head_def &)): New.
17656 (extern debug (const bitmap_head_def *)): New.
17658 (debug (edge_def &)): New.
17659 (debug (edge_def *)): New.
17661 (debug (basic_block_def &)): New.
17662 (debug (basic_block_def *)): New.
17664 (dump_node (const_tree, int, FILE *)): Correct source file.
17666 (debug (die_struct &)): New.
17667 (debug (die_struct *)): New.
17669 (extern debug (die_struct &)): New.
17670 (extern debug (die_struct *)): New.
17671 * gimple-pretty-print.c
17672 (debug (gimple_statement_d &)): New.
17673 (debug (gimple_statement_d *)): New.
17674 * gimple-pretty-print.h
17675 (extern debug (gimple_statement_d &)): New.
17676 (extern debug (gimple_statement_d *)): New.
17678 (debug (ira_allocno_copy &)): New.
17679 (debug (ira_allocno_copy *)): New.
17680 (debug (ira_allocno &)): New.
17681 (debug (ira_allocno *)): New.
17683 (extern debug (ira_allocno_copy &)): New.
17684 (extern debug (ira_allocno_copy *)): New.
17685 (extern debug (ira_allocno &)): New.
17686 (extern debug (ira_allocno *)): New.
17688 (debug (live_range &)): New.
17689 (debug (live_range *)): New.
17691 (debug (lra_live_range &)): New.
17692 (debug (lra_live_range *)): New.
17694 (debug (lra_live_range &)): New.
17695 (debug (lra_live_range *)): New.
17697 (debug (omega_pb_d &)): New.
17698 (debug (omega_pb_d *)): New.
17700 (extern debug (omega_pb_d &)): New.
17701 (extern debug (omega_pb_d *)): New.
17703 (debug (const rtx_def &)): New.
17704 (debug (const rtx_def *)): New.
17706 (debug_tree (tree): Move within file.
17707 (debug_raw (const tree_node &)): New.
17708 (debug_raw (const tree_node *)): New.
17709 (dump_tree_via_hooks (const tree_node *, int)): New.
17710 (debug (const tree_node &)): New.
17711 (debug (const tree_node *)): New.
17712 (debug_verbose (const tree_node &)): New.
17713 (debug_verbose (const tree_node *)): New.
17714 (debug_head (const tree_node &)): New.
17715 (debug_head (const tree_node *)): New.
17716 (debug_body (const tree_node &)): New.
17717 (debug_body (const tree_node *)): New.
17718 (debug_vec_tree (tree): Move and reimplement in terms of dump.
17719 (debug (vec<tree, va_gc> &)): New.
17720 (debug (vec<tree, va_gc> *)): New.
17722 (extern debug (const rtx_def &)): New.
17723 (extern debug (const rtx_def *)): New.
17725 (debug_raw (simple_bitmap_def &)): New.
17726 (debug_raw (simple_bitmap_def *)): New.
17727 (debug (simple_bitmap_def &)): New.
17728 (debug (simple_bitmap_def *)): New.
17730 (extern debug (simple_bitmap_def &)): New.
17731 (extern debug (simple_bitmap_def *)): New.
17732 (extern debug_raw (simple_bitmap_def &)): New.
17733 (extern debug_raw (simple_bitmap_def *)): New.
17735 (debug (vinsn_def &)): New.
17736 (debug (vinsn_def *)): New.
17737 (debug_verbose (vinsn_def &)): New.
17738 (debug_verbose (vinsn_def *)): New.
17739 (debug (expr_def &)): New.
17740 (debug (expr_def *)): New.
17741 (debug_verbose (expr_def &)): New.
17742 (debug_verbose (expr_def *)): New.
17743 (debug (vec<rtx> &)): New.
17744 (debug (vec<rtx> *)): New.
17746 (extern debug (vinsn_def &)): New.
17747 (extern debug (vinsn_def *)): New.
17748 (extern debug_verbose (vinsn_def &)): New.
17749 (extern debug_verbose (vinsn_def *)): New.
17750 (extern debug (expr_def &)): New.
17751 (extern debug (expr_def *)): New.
17752 (extern debug_verbose (expr_def &)): New.
17753 (extern debug_verbose (expr_def *)): New.
17754 (extern debug (vec<rtx> &)): New.
17755 (extern debug (vec<rtx> *)): New.
17757 (_list_iter_cond_expr): Make inline instead of static.
17759 (debug (sreal &)): New.
17760 (debug (sreal *)): New.
17762 (extern debug (sreal &)): New.
17763 (extern debug (sreal *)): New.
17765 (extern debug_raw (const tree_node &)): New.
17766 (extern debug_raw (const tree_node *)): New.
17767 (extern debug (const tree_node &)): New.
17768 (extern debug (const tree_node *)): New.
17769 (extern debug_verbose (const tree_node &)): New.
17770 (extern debug_verbose (const tree_node *)): New.
17771 (extern debug_head (const tree_node &)): New.
17772 (extern debug_head (const tree_node *)): New.
17773 (extern debug_body (const tree_node &)): New.
17774 (extern debug_body (const tree_node *)): New.
17775 (extern debug (vec<tree, va_gc> &)): New.
17776 (extern debug (vec<tree, va_gc> *)): New.
17778 (debug (struct loop &)): New.
17779 (debug (struct loop *)): New.
17780 (debug_verbose (struct loop &)): New.
17781 (debug_verbose (struct loop *)): New.
17782 * tree-dump.c: Add header dependence.
17784 (extern debug (struct loop &)): New.
17785 (extern debug (struct loop *)): New.
17786 (extern debug_verbose (struct loop &)): New.
17787 (extern debug_verbose (struct loop *)): New.
17789 (debug (data_reference &)): New.
17790 (debug (data_reference *)): New.
17791 (debug (vec<data_reference_p> &)): New.
17792 (debug (vec<data_reference_p> *)): New.
17793 (debug (vec<ddr_p> &)): New.
17794 (debug (vec<ddr_p> *)): New.
17796 (extern debug (data_reference &)): New.
17797 (extern debug (data_reference *)): New.
17798 (extern debug (vec<data_reference_p> &)): New.
17799 (extern debug (vec<data_reference_p> *)): New.
17800 (extern debug (vec<ddr_p> &)): New.
17801 (extern debug (vec<ddr_p> *)): New.
17803 (debug (pt_solution &)): New.
17804 (debug (pt_solution *)): New.
17806 (extern debug (pt_solution &)): New.
17807 (extern debug (pt_solution *)): New.
17809 (debug (_var_map &)): New.
17810 (debug (_var_map *)): New.
17811 (debug (tree_live_info_d &)): New.
17812 (debug (tree_live_info_d *)): New.
17814 (extern debug (_var_map &)): New.
17815 (extern debug (_var_map *)): New.
17816 (extern debug (tree_live_info_d &)): New.
17817 (extern debug (tree_live_info_d *)): New.
17819 2013-03-28 Jan Hubicka <jh@suse.cz>
17821 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
17823 2013-03-28 Ian Bolton <ian.bolton@arm.com>
17825 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
17826 record only when desired or required.
17828 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
17830 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
17831 *vec_extractv2di_1_rex64. Use x64 isa attribute.
17833 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17835 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
17836 (*andsi3_compare0_uxtw): New pattern.
17837 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
17838 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
17840 2013-03-28 Jan Hubicka <jh@suse.cz>
17842 * data-streamer-in.c (streamer_read_gcov_count): New function.
17843 * gimple-streamer-out.c: Include value-prof.h.
17844 (output_gimple_stmt): Output histogram.
17845 (output_bb): Use streamer_write_gcov_count.
17846 * value-prof.c: Include data-streamer.h
17847 (dump_histogram_value): Add HIST_TYPE_MAX.
17848 (stream_out_histogram_value): New function.
17849 (stream_in_histogram_value): New function.
17850 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
17851 (stream_out_histogram_value, stream_in_histogram_value): Declare.
17852 * data-streamer-out.c (streamer_write_gcov_count): New function.
17853 (streamer_write_gcov_count_stream): New function.
17854 * lto-cgraph.c (lto_output_edge): Update counter streaming.
17855 (lto_output_node): Likewise.
17856 (input_node, input_edge): Likewise.
17857 * lto-streamer-out.c (output_cfg): Update streaming.
17858 * lto-streamer-in.c (input_cfg): Likewise.
17859 * data-streamer.h (streamer_write_gcov_count,
17860 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
17861 * gimple-streamer-in.c: Include value-prof.h
17862 (input_gimple_stmt): Input histograms.
17863 (input_bb): Update profile streaming.
17865 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
17867 * genmodes.c (emit_max_int): New function.
17868 (emit_insn_modes_h): Added call to emit_max_function.
17869 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
17871 * machmode.def: Fixed comment.
17873 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
17875 * combine.c (try_combine): Removed useless assert.
17876 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
17878 2013-03-28 Marek Polacek <polacek@redhat.com>
17879 Richard Biener <rguenther@suse.de>
17881 PR tree-optimization/56695
17882 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
17883 build signed result of a vector comparison.
17884 * tree-cfg.c (verify_gimple_comparison): Check that a result
17885 of a vector comparison has signed type.
17887 2013-03-28 Richard Biener <rguenther@suse.de>
17889 PR tree-optimization/37021
17890 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
17891 do not restrict gaps between groups.
17892 * tree-vect-stmts.c (vectorizable_load): Properly account for
17893 a gap between groups.
17895 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
17897 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
17898 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
17901 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
17903 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
17904 * doc/extend.texi (Named Address Spaces): Ditto.
17905 (Variable Attributes): Ditto.
17907 2013-03-27 Kai Tietz <ktietz@redhat.com>
17909 * config.build: Add support for cygwin x64 target.
17910 * config.gcc: Likewise.
17911 * config.host: Likewise.
17912 * configure.ac: Likewise
17913 * configure: Regenerated.
17915 2013-03-27 Kai Tietz <ktietz@redhat.com>
17917 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
17918 * config/i386/t-cygwin-w64: New file.
17919 * config/i386/cygwin-w64.h: New file.
17920 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
17921 and add support for x64-cygwin target.
17922 (CPP_SPEC): Likewise.
17923 (CXX_WRAP_SPEC_LIST): Undefine before define.
17924 (LIBGCJ_SONAME): Use 15 as version.
17926 2013-03-27 Richard Biener <rguenther@suse.de>
17928 PR tree-optimization/56716
17929 * tree-ssa-structalias.c (perform_var_substitution): Adjust
17930 dumping for ref nodes.
17932 2013-03-27 Martin Jambor <mjambor@suse.cz>
17934 PR tree-optimization/55334
17935 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
17936 restricted pointers to arrays.
17938 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
17940 * Makefile.in (.SUFFIXES): Add .cc.
17941 (.c.o): Apply same recipe for implicit rule .cc.o.
17943 2013-03-27 Richard Biener <rguenther@suse.de>
17945 PR tree-optimization/37021
17946 * tree-vect-data-refs.c (vect_check_strided_load): Allow
17947 REALPART/IMAGPART_EXPRs around the supported refs.
17948 * tree-ssa-structalias.c (find_func_aliases): Assume that
17949 floating-point values are not used to transfer pointers.
17951 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
17953 * target.def (TARGET_HAS_IFUNC_P): New target hook.
17954 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
17955 * doc/tm.texi: Regenerate.
17956 * targhooks.h (default_has_ifunc_p): New.
17957 * targhooks.c (default_has_ifunc_p): Ditto.
17958 * config/linux-protos.h: New file.
17959 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
17960 hook for linux which disables support of indirect functions in android.
17961 * config/linux-android.c: New file.
17962 * config/t-linux-android.c: Ditto.
17963 * config.gcc: Added new object file linux-android.o.
17964 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
17965 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
17966 * varasm.c (do_assemble_alias): Likewise.
17967 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
17968 doesn't support indirect functions.
17969 * configure: Regenerate.
17971 2013-03-27 Bin Cheng <bin.cheng@arm.com>
17974 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
17975 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
17978 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17980 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
17982 2013-03-27 Terry Guo <terry.guo@arm.com>
17984 * config/arm/arm-cores.def: Added core cortex-r7.
17985 * config/arm/arm-tune.md: Regenerated.
17986 * config/arm/arm-tables.opt: Regenerated.
17987 * doc/invoke.texi: Added entry for core cortex-r7.
17989 2013-03-27 Walter Lee <walt@tilera.com>
17991 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
17992 double-decrement of next_scratch_regno.
17994 2013-03-27 Walter Lee <walt@tilera.com>
17996 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
17998 (insn_v1mulus): Ditto.
17999 (insn_v2muls): Ditto.
18001 2013-03-27 Walter Lee <walt@tilera.com>
18003 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
18004 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
18006 2013-03-27 Walter Lee <walt@tilera.com>
18008 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
18009 (*sibcall_value): Ditto.
18011 2013-03-27 Walter Lee <walt@tilera.com>
18013 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
18014 (insn_mnz_v8qi): ... this ...
18015 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
18017 (insn_v<n>mnz): Replaced by ...
18018 (insn_v1mnz): ... this ...
18019 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
18021 (insn_mz_<mode>): Replaced by ...
18022 (insn_mz_v8qi): ... this ...
18023 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
18025 (insn_v<n>mz): Replaced by ...
18026 (insn_v1mz): ... this ...
18027 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
18030 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
18032 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
18034 2013-03-26 Roland McGrath <mcgrathr@google.com>
18036 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
18037 than fprintf with a non-constant, non-format string.
18039 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
18041 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
18042 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
18043 operand 0 predicate.
18044 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
18045 attribute. Use general_x64nomem_operand as operand 1 predicate.
18046 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
18047 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
18048 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
18049 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
18050 (mov<mode>_insv_1): Remove expander. Merge insn with
18051 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
18052 Use general_x64nomem_operand as operand 1 predicate.
18053 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
18054 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
18055 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
18056 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
18057 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
18058 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
18059 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
18060 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
18061 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
18062 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
18063 (general_x64nomem_operand): Ditto.
18065 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
18067 * config/rtems.opt: Add -pthread option.
18069 2013-03-26 Richard Biener <rguenther@suse.de>
18071 * alias.c (find_base_term): Avoid redundant and not used recursion.
18072 (base_alias_check): Get the initial base term from the caller.
18073 (true_dependence_1): Compute and pass base terms to base_alias_check.
18074 (write_dependence_p): Likewise.
18075 (may_alias_p): Likewise.
18077 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
18079 * config/aarch64/aarch64.c (aarch64_classify_address): Support
18080 PC-relative load in SI modes and above only.
18082 2013-03-26 Xinyu Qi <xyqi@marvell.com>
18084 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
18085 * config/arm/iwmmxt.md (WCGR0): Update.
18086 (WCGR1, WCGR2, WCGR3): Likewise.
18088 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
18090 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
18091 Use x64 and nox64 isa attributes.
18093 2013-03-26 Richard Biener <rguenther@suse.de>
18095 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
18096 alignment computations and rely on get_object_alignment_1
18097 for the !TYPE_P case.
18098 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
18100 2013-03-26 Walter Lee <walt@tilera.com>
18102 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
18103 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
18105 2013-03-25 Jeff Law <law@redhat.com>
18107 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
18108 check for INTEGRAL_TYPE_P that was missing due to checking in
18109 wrong version of prior patch.
18111 2013-03-25 Walter Lee <walt@tilera.com>
18113 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
18114 TILEGX_INSN_SHUFFLEBYTES1.
18115 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
18117 (tilegx_builtins): Ditto.
18118 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
18120 2013-03-25 Walter Lee <walt@tilera.com>
18122 * config/tilegx/tilegx.md (floatsisf2): New pattern.
18123 (floatunssisf2): New pattern.
18124 (floatsidf2): New pattern.
18125 (floatunssidf2): New pattern.
18127 2013-03-25 Walter Lee <walt@tilera.com>
18129 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
18130 tests for constraint J, K, N, P.
18132 2013-03-25 Walter Lee <walt@tilera.com>
18134 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
18135 Use indirect/pcrel encoding.
18136 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
18139 2013-03-25 Steve Ellcey <sellcey@mips.com>
18141 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
18142 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
18143 * config/mips/mips.c (mips_option_override): Set IMADD default.
18144 * config/mips/mips.h (PTF_AVOID_IMADD): New.
18145 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
18146 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
18147 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
18148 * doc/invoke.texi (-mimadd/-mno-imadd): New.
18150 2013-03-25 Jeff Law <law@redhat.com>
18152 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
18153 slightly to avoid creating and folding useless trees. Simplify
18154 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
18156 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
18158 * config/i386/i386.md (*zero_extendsidi2): Merge with
18159 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
18160 * config/i386/predicates.md (x86_64_zext_operand): Rename from
18161 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
18162 targets. Clarify comment.
18164 2013-03-25 Martin Jambor <mjambor@suse.cz>
18166 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
18167 pass-through jump functions differently.
18168 (ipa_read_jump_function): Likewise. Also use setter functions to set
18171 2013-03-25 Martin Jambor <mjambor@suse.cz>
18173 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
18174 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
18176 (ipa_get_indirect_edge_target): New function.
18177 (devirtualization_time_bonus): New parameter known_aggs, pass it to
18178 ipa_get_indirect_edge_target. Update all callers.
18179 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
18180 ipa_get_indirect_edge_target_1 instead of calling
18181 ipa_get_indirect_edge_target.
18182 (create_specialized_node): Pass aggvlas to
18183 ipcp_discover_new_direct_edges.
18185 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18187 * config/arm/arm.md (f_sels, f_seld): New types.
18188 (*cmov<mode>): New pattern.
18189 * config/arm/predicates.md (arm_vsel_comparison_operator): New
18192 2013-03-25 Kai Tietz <ktietz@redhat.com>
18194 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
18195 POSIX-printf for mingw-hosted builds.
18197 2013-03-25 Richard Biener <rguenther@suse.de>
18199 PR middle-end/56694
18200 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
18201 must-not-throw stmt location.
18203 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18205 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
18206 Emit load-acquire versions when acq is true.
18207 (arm_emit_store_exclusive): Add rel parameter.
18208 Emit store-release versions when rel is true.
18209 (arm_split_compare_and_swap): Use acquire-release instructions
18211 of barriers when appropriate.
18212 (arm_split_atomic_op): Likewise.
18213 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
18214 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
18215 (VUNSPEC_SLX): Likewise.
18216 (VUNSPEC_LDA): Likewise.
18217 (VUNSPEC_STL): Likewise.
18218 * config/arm/sync.md (atomic_load<mode>): New pattern.
18219 (atomic_store<mode>): Likewise.
18220 (arm_load_acquire_exclusive<mode>): Likewise.
18221 (arm_load_acquire_exclusivesi): Likewise.
18222 (arm_load_acquire_exclusivedi): Likewise.
18223 (arm_store_release_exclusive<mode>): Likewise.
18225 2013-03-25 Catherine Moore <clm@codesourcery.com>
18227 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
18228 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
18229 * config/mip/predicates.md (lwsp_swsp_operand,
18230 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
18231 sb16_operand, db4_operand, db7_operand, ib3_operand,
18232 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
18233 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
18234 andi16_operand): New predicates.
18235 * config/mips/mips.md (compression): New attribute.
18236 (enabled): New attribute.
18237 (length): Consider compression in computing length.
18238 (shift_compression): New code attribute.
18239 (*add<mode>3): New operands. Record compression.
18240 (sub<mode>3): Likewise.
18241 (one_cmpl<mode>2): Likewise.
18242 (*and<mode>3): Likewise.
18243 (*ior<mode>3): Likewise.
18244 (unnamed pattern for xor): Likewise.
18245 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
18246 (*<optab><mode>3): Likewise.
18247 (*mov<mode>_internal: Likewise.
18248 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
18249 (mips_unsigned_immediate_p): New.
18250 (umips_lwsp_swsp_address_p): New.
18251 (m16_based_address_p): New.
18252 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
18253 (mips_unsigned_immediate_p): New prototype.
18254 (lwsp_swsp_address_p): New prototype.
18255 (m16_based_address_p): New prototype.
18256 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
18257 (mips_signed_immediate_p): New function.
18258 (m16_based_address_p): New function.
18259 (lwsp_swsp_address_p): New function.
18260 (mips_print_operand_punctuation): Recognize short delay slot insns
18261 for microMIPS.add<mode>3"
18263 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18266 * config/arm/iterators.md (v_cmp_result): New mode attribute.
18267 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
18269 2013-03-25 Richard Biener <rguenther@suse.de>
18271 PR tree-optimization/56689
18272 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
18275 2013-03-25 Richard Biener <rguenther@suse.de>
18277 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
18279 (memory_references): Likewise.
18280 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
18281 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
18282 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
18283 (gather_mem_refs_in_loops): Fold into ...
18284 (analyze_memory_references): ... this. Move initialization
18285 to tree_ssa_lim_initialize.
18286 (fill_always_executed_in): Rename to ...
18287 (fill_always_executed_in_1): ... this.
18288 (fill_always_executed_in): Move contains_call computation to
18289 this new function from ...
18290 (tree_ssa_lim_initialize): ... here.
18291 (tree_ssa_lim): Call fill_always_executed_in.
18293 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
18295 * postreload.c (reload_combine): Fix code detecting returns.
18297 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
18299 * function.c (emit_use_return_register_into_block): On cc0 targets,
18300 do not emit the sequence between cc0 setter and user.
18302 2013-03-25 Kai Tietz <ktietz@redhat.com>
18304 * config/i386/predicates.md (local_symbolic_operand): Interpret
18305 dll-imported symbols as none-local.
18307 2013-03-25 Richard Biener <rguenther@suse.de>
18309 * tree-ssa-loop-im.c (struct depend): Remove.
18310 (struct lim_aux_data): Make depends a vec of gimples.
18311 (free_lim_aux_data): Adjust.
18312 (add_dependency): Likewise.
18313 (set_level): Likewise.
18315 2013-03-25 Richard Biener <rguenther@suse.de>
18317 PR middle-end/56434
18318 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
18319 the pointer returned by calls with ECF_MALLOC set.
18321 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
18323 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
18325 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
18327 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
18328 using MMXMODE mode iterator.
18329 (*move<mode>_internal): Merge with *movv2sf_internal and
18330 *movv2sf_internal_rex64 using MMXMODE mode iterator.
18332 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
18334 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
18335 (record_last_mem_set_info): Likewise.
18337 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
18338 of XNEWVEC followed by memset.
18339 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
18341 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
18343 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
18344 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
18345 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
18346 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
18347 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
18348 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
18349 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
18350 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
18351 BARRIER_P instead of GET_CODE.
18353 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
18355 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
18356 inaccuracy in the probing code.
18358 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
18359 (ctrapdi4): Likewise.
18361 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
18363 * calls.c (expand_call): Add missing guard to code handling return
18364 of non-BLKmode structures in MSB.
18365 * function.c (expand_function_end): Likewise.
18367 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
18369 * combine.c (try_combine): Adjust comment. Do not add the set of
18370 insn #0 if the destination indirectly is set or dies in insn #2.
18371 Tidy up code to distribute a new note.
18373 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
18375 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
18376 also for alternatives 16 and 17.
18378 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
18380 * config/i386/sse.md (*mov<mode>_internal): Merge with
18381 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
18382 Emit insn template depending on type attribute. Use
18383 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
18384 movd instead of movq mnemonic for interunit moves. Rewrite mode
18385 attribute calculation. Remove unit attribute calculation.
18386 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
18387 Set prefix_data16 attribute for DImode ssemov types.
18388 Use Ym instead of y for SSE-MMX conversion alternatives.
18389 Reorder operand constraints.
18391 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
18393 * df.h (df_insn_delete): Adjust prototype.
18394 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
18395 and let it decide whether mark the basic block dirty.
18396 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
18397 * df-scan.c (df_insn_info_delete): New helper function, split
18398 off from df_insn_delete.
18399 (df_scan_free_bb_info): Use it.
18400 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
18402 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
18403 that the insn is actually an insn and it has a non-NULL basic block.
18404 Do not mark basic block dirty if only deleting a DEBUG_INSN.
18406 2013-03-22 Richard Biener <rguenther@suse.de>
18408 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
18410 (mem_ref_alloc): Do not allocate them.
18411 (refs_independent_p): Do not query or maintain a cache.
18413 2013-03-22 Richard Biener <rguenther@suse.de>
18415 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
18416 (gather_mem_refs_in_loops): Do not compute it.
18417 (analyze_memory_references): Do not allocate it.
18418 (tree_ssa_lim_finalize): Do not free it.
18419 (for_all_locs_in_loop): Do not query all_refs_in_loop.
18421 2013-03-22 Richard Biener <rguenther@suse.de>
18423 * is-a.h (as_a): Use gcc_checking_assert.
18425 2013-03-22 Ian Bolton <ian.bolton@arm.com>
18427 * config/aarch64/aarch64.c (aarch64_print_operand): New
18428 format specifier for printing a constant in hex.
18429 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
18430 format specifier for printing second operand.
18432 2013-03-22 Richard Biener <rguenther@suse.de>
18434 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
18436 (gather_mem_refs_in_loops): Perform store accumulation here.
18437 (create_vop_ref_mapping_loop): Remove.
18438 (create_vop_ref_mapping): Likewise.
18439 (analyze_memory_references): Initialize refs_stored_in_loop.
18440 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
18441 (record_indep_loop): Remove.
18442 (record_dep_loop): New function.
18443 (ref_indep_loop_p_1): Adjust to only walk over references
18444 in the loop, not its subloops.
18445 (ref_indep_loop_p): Rename to ...
18446 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
18447 maintaining a more fine-grained cache.
18448 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
18449 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
18451 2013-03-22 Richard Biener <rguenther@suse.de>
18453 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
18454 (struct mem_ref): Make accesses_in_loop a vec of a vec of
18455 aggregate mem_ref_loc.
18456 (free_mem_ref_locs): Inline into ...
18457 (memref_free): ... this and adjust.
18458 (mem_ref_alloc): Adjust.
18459 (mem_ref_locs_alloc): Remove.
18460 (record_mem_ref_loc): Adjust.
18461 (get_all_locs_in_loop): Rewrite into ...
18462 (for_all_locs_in_loop): ... this iterator.
18463 (rewrite_mem_ref_loc): New functor.
18464 (rewrite_mem_refs): Use for_all_locs_in_loop.
18465 (sm_set_flag_if_changed): New functor.
18466 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
18467 (ref_always_accessed): New functor.
18468 (ref_always_accessed_p): Use for_all_locs_in_loop.
18470 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
18472 * tree-pass.h (PROP_gimple_lvec): New.
18473 * passes.c (dump_properties): Handle PROP_gimple_lvec.
18474 (init_optimization_passes): Move pass_lower_vector.
18475 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
18477 (pass_lower_vector): Provide PROP_gimple_lvec.
18478 (pass_lower_vector_ssa): Likewise.
18479 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
18481 2013-03-21 Mark Wielaard <mjw@redhat.com>
18483 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
18485 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
18487 * config/i386/i386.md (*movdi_internal): Disparage slightly
18488 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
18489 conversion alternatives.
18491 2013-03-21 Jakub Jelinek <jakub@redhat.com>
18493 PR middle-end/48087
18494 * diagnostic.def (DK_WERROR): New kind.
18495 * diagnostic.h (werrorcount): Define.
18496 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
18497 promoted to DK_ERROR, increment DK_WERROR counter instead of
18499 * toplev.c (toplev_main): Call print_ignored_options even if
18500 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
18501 even if just werrorcount is non-zero.
18504 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
18506 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
18507 (string_cst_pool_decl): New function.
18508 (optimize_one_addr_into_implicit_ptr): New function.
18509 (resolve_addr_in_expr): Optimize DWARF location expression
18510 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
18511 which doesn't live in memory, but has DW_AT_location or
18512 DW_AT_const_value, or refers to a string literal, into
18513 DW_OP_GNU_implicit_pointer.
18514 (optimize_location_into_implicit_ptr): New function.
18515 (resolve_addr): If removing DW_AT_location of a variable because
18516 it was DW_OP_addr of address of the variable, but the variable doesn't
18517 live in memory, try to emit const value attribute for the initializer.
18519 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
18521 * tree.h (VECTOR_TYPE_P): New macro.
18522 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
18523 TYPE_MODE): Use it.
18524 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
18525 VEC_COND_EXPR cannot be lvalues.
18526 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
18528 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
18530 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
18531 Restrict the transformation to equal modes.
18533 2013-03-21 Richard Biener <rguenther@suse.de>
18535 PR tree-optimization/39326
18536 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
18537 (MEM_ANALYZABLE): Adjust.
18538 (record_mem_ref_loc): Move bitmap ops ...
18539 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
18540 unanalyzable refs, do not record locations for it.
18541 (analyze_memory_references): Allocate ref zero as shared
18543 (refs_independent_p): Do not test for unanalyzed mems here.
18544 (ref_indep_loop_p_1): Special-case disambiguation against
18545 the unanalyzed ref.
18546 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
18548 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
18550 * config/arm/arm-protos.h (tune_params): Add
18551 prefer_neon_for_64bits field.
18552 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
18553 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
18554 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
18555 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
18556 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
18557 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
18558 (arm_option_override): Handle -mneon-for-64bits new option.
18559 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
18560 (prefer_neon_for_64bits): Declare new variable.
18561 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
18562 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
18563 (arch_enabled): Handle new arch types. Remove support for onlya8
18565 (one_cmpldi2): Use new arch names.
18566 (zero_extend<mode>di2, extend<mode>di2): Ditto.
18567 * config/arm/arm.opt (mneon-for-64bits): Add option.
18568 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
18569 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
18570 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
18572 * doc/invoke.texi (-mneon-for-64bits): Document.
18574 2013-03-21 Richard Biener <rguenther@suse.de>
18576 PR tree-optimization/39326
18577 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
18578 (sort_bbs_in_loop_postorder_cmp): New function.
18579 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
18581 2013-03-21 Richard Biener <rguenther@suse.de>
18583 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
18584 (vect_insert_into_interleaving_chain): Likewise.
18585 (vect_drs_dependent_in_basic_block): Inline ...
18586 (vect_slp_analyze_data_ref_dependence): ... here. New function,
18588 (vect_analyze_data_ref_dependence): ... here. Simplify.
18589 (vect_check_interleaving): Simplify.
18590 (vect_analyze_data_ref_dependences): Likewise. Split out ...
18591 (vect_slp_analyze_data_ref_dependences): ... this new function.
18592 (dr_group_sort_cmp): New function.
18593 (vect_analyze_data_ref_accesses): Compute data-reference groups
18594 here instead of in vect_analyze_data_ref_dependence. Use
18595 a more efficient algorithm.
18596 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
18597 vect_slp_analyze_data_ref_dependences. Call
18598 vect_analyze_data_ref_accesses earlier.
18599 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
18600 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
18601 (vect_slp_analyze_data_ref_dependences): New prototype.
18603 2013-03-21 Richard Biener <rguenther@suse.de>
18605 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
18606 ref is stored in the loop.
18607 (find_refs_for_sm): Walk only over all stores.
18608 (store_motion_loop): Allocate from lim_bitmap_obstack.
18609 (store_motion): Likewise.
18611 2013-03-21 Richard Biener <rguenther@suse.de>
18613 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
18614 Update virtual SSA form.
18616 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
18618 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
18619 * configure: Regenerate.
18620 * config.in: Regenerate.
18621 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
18622 if !HAVE_LD_EH_FRAME_CIEV3.
18624 2013-03-21 Richard Biener <rguenther@suse.de>
18626 * tree-cfg.c (verify_expr_no_block): New function.
18627 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
18628 nor DECL_VALUE_EXPR have locations with associated blocks.
18629 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
18630 (clear_unused_block_pointer): Remove code dealing with
18631 blocks in DECL_DEBUG_EXPR locations.
18633 2013-03-21 Richard Biener <rguenther@suse.de>
18635 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
18636 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
18637 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
18638 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
18639 instead of DECL_DEBUG_EXPR_IS_FROM.
18640 * gimplify.c (gimplify_modify_expr): Likewise.
18641 * tree-cfg.c (verify_expr_location_1): Likewise.
18642 * tree-complex.c (create_one_component_var): Likewise.
18643 * tree-sra.c (create_access_replacement): Likewise.
18644 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
18645 (clear_unused_block_pointer): Likewise.
18646 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
18647 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
18648 * var-tracking.c (var_debug_decl): Likewise.
18649 (track_expr_p): Likewise.
18650 * tree-inline.c (add_local_variables): Likewise. Set
18651 DECL_HAS_DEBUG_EXPR_P after copying it.
18652 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
18653 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
18655 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
18658 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
18659 * configure: Regenerate.
18660 * config.in: Regenerate.
18661 * config/i386/i386.md (*movdf_internal): Use
18662 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
18663 movd instead of movq mnemonic for interunit moves.
18664 (*movdi_internal): Ditto.
18666 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
18668 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
18669 (abd<mode>_3): New pattern.
18670 (aba<mode>_3): New pattern.
18671 (fabd<mode>_3): New pattern.
18673 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
18675 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
18676 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
18677 occurrence of REGISTER_PREFIX as its empty string.
18679 2013-03-20 Jeff Law <law@redhat.com>
18681 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
18682 addititional equivalences for equality comparisons between an SSA_NAME
18683 and a constant where the SSA_NAME was set from a widening conversion.
18685 2013-03-20 Walter Lee <walt@tilera.com>
18687 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
18689 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
18691 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
18692 depending on type attribute.
18693 (*movti_internal): Ditto.
18694 (*movtf_internal): Ditto.
18695 (*movxf_internal): Ditto.
18696 (*movdf_internal): Ditto.
18697 (*movsf_internal): Ditto.
18699 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
18701 * config/i386/i386.md (*movti_internal): Set prefix attribute to
18702 maybe_vex for sselog1 and ssemov types.
18703 (*movdi_internal): Reorder operand constraints.
18704 (*movsi_internal): Ditto. Set prefix attribute to
18705 maybe_vex for sselog1 and ssemov types.
18706 (*movtf_internal): Set prefix attribute to maybe_vex
18707 for sselog1 and ssemov types.
18708 (*movdf_internal): Ditto. Set prefix_data16 attribute for
18709 DImode ssemov types. Reorder operand constraints.
18710 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
18711 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
18712 attribute for SImode ssemov types. Reorder operand constraints.
18714 2013-03-20 Martin Jambor <mjambor@suse.cz>
18716 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
18717 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
18719 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
18721 * config/rs6000/predicates.md (indexed_address, update_address_mem
18722 update_indexed_address_mem): New predicates.
18723 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
18724 attribute for load/store instructions.
18725 * config/rs6000/dfp.md (movsd_store): Likewise.
18726 (movsd_load): Likewise.
18727 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
18728 (unnamed HI->DI extend define_insn): Likewise.
18729 (unnamed SI->DI extend define_insn): Likewise.
18730 (unnamed QI->SI extend define_insn): Likewise.
18731 (unnamed QI->HI extend define_insn): Likewise.
18732 (unnamed HI->SI extend define_insn): Likewise.
18733 (unnamed HI->SI extend define_insn): Likewise.
18734 (extendsfdf2_fpr): Likewise.
18735 (movsi_internal1): Likewise.
18736 (movsi_internal1_single): Likewise.
18737 (movhi_internal): Likewise.
18738 (movqi_internal): Likewise.
18739 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
18740 attribute for load/store instructions.
18741 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
18743 (mov<mode>_softfloat): Likewise.
18744 (mov<mode>_hardfloat32): Likewise.
18745 (mov<mode>_hardfloat64): Likewise.
18746 (mov<mode>_softfloat64): Likewise.
18747 (movdi_internal32): Likewise.
18748 (movdi_internal64): Likewise.
18749 (probe_stack_<mode>): Likewise.
18751 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
18753 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
18754 floating point, and decimal floating point to reload iterator.
18756 * config/rs6000/constraints.md (wl constraint): New constraints to
18757 return FLOAT_REGS if certain options are used to reduce the number
18758 of separate patterns that exist in the file.
18759 (wx constraint): Likewise.
18760 (wz constraint): Likewise.
18762 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
18763 -mdebug=reg, print wg, wl, wx, and wz constraints.
18764 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
18765 Initialize the reload functions for 64-bit binary/decimal floating
18767 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
18768 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
18769 create the buffer on the stack to overcome not having a 32-bit
18771 (rs6000_emit_move): Likewise.
18772 (rs6000_secondary_memory_needed_rtx): Likewise.
18773 (rs6000_alloc_sdmode_stack_slot): Likewise.
18774 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
18775 via xxlxor, just like DFmode 0.0.
18777 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
18778 define as 1 if we are running on a power7 or newer.
18779 (enum r6000_reg_class_enum): Add new constraints.
18781 * config/rs6000/dfp.md (movsd): Delete, combine with binary
18782 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
18783 with other moves by using conditional constraits (wg). Use LFIWZX
18784 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
18785 (movsd splitter): Likewise.
18786 (movsd_hardfloat): Likewise.
18787 (movsd_softfloat): Likewise.
18789 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
18790 binary and decimal floating point moves.
18791 (fmove_ok): New attributes to combine binary and decimal floating
18792 point moves, and to combine power6x (mfpgpr) moves along normal
18794 (real_value_to_target): Likewise.
18795 (f32_lr): Likewise.
18796 (f32_lm): Likewise.
18797 (f32_li): Likewise.
18798 (f32_sr): Likewise.
18799 (f32_sm): Likewise.
18800 (f32_si): Likewise.
18801 (movsf): Combine binary and decimal floating point moves. Combine
18802 power6x (mfpgpr) moves with other moves by using conditional
18803 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
18804 (mov<mode> for SFmode/SDmode); Likewise.
18805 (SFmode/SDmode splitters): Likewise.
18806 (movsf_hardfloat): Likewise.
18807 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
18808 (movsf_softfloat): Likewise.
18809 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
18811 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
18812 wx and wz constraints.
18814 * config/rs6000/constraints.md (wg constraint): New constraint to
18815 return FLOAT_REGS if -mmfpgpr (power6x) was used.
18817 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
18820 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
18821 -mdebug=reg, print wg, wl, wx, and wz constraints.
18822 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
18823 Initialize the reload functions for 64-bit binary/decimal floating
18825 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
18826 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
18827 create the buffer on the stack to overcome not having a 32-bit
18829 (rs6000_emit_move): Likewise.
18830 (rs6000_secondary_memory_needed_rtx): Likewise.
18831 (rs6000_alloc_sdmode_stack_slot): Likewise.
18832 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
18833 via xxlxor, just like DFmode 0.0.
18835 * config/rs6000/dfp.md (movdd): Delete, combine with binary
18836 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
18837 with other moves by using conditional constraits (wg). Use LFIWZX
18838 and STFIWX for loading SDmode on power7.
18839 (movdd splitters): Likewise.
18840 (movdd_hardfloat32): Likewise.
18841 (movdd_softfloat32): Likewise.
18842 (movdd_hardfloat64_mfpgpr): Likewise.
18843 (movdd_hardfloat64): Likewise.
18844 (movdd_softfloat64): Likewise.
18846 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
18847 64-bit binary and decimal floating point moves.
18848 (FMOVE64X): Likewise.
18849 (movdf): Combine 64-bit binary and decimal floating point moves.
18850 Combine power6x (mfpgpr) moves with other moves by using
18851 conditional constraits (wg).
18852 (mov<mode> for DFmode/DDmode): Likewise.
18853 (DFmode/DDmode splitters): Likewise.
18854 (movdf_hardfloat32): Likewise.
18855 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
18856 (movdf_softfloat32): Likewise.
18857 (movdf_hardfloat64_mfpgpr): Likewise.
18858 (movdf_hardfloat64): Likewise.
18859 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
18860 (movdf_softfloat64): Likewise.
18861 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
18862 (reload_<mode>_load): Move to later in the file so they aren't in
18863 the middle of the floating point move insns.
18864 (reload_<mode>_store): Likewise.
18866 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
18869 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
18870 constraint if -mdebug=reg.
18871 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
18872 Enable using dd reload support if needed.
18874 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
18875 binary and decimal floating point moves in rs6000.md.
18876 (movtd_internal): Likewise.
18878 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
18879 decimal floating point moves.
18881 (movtf_internal): Likewise.
18882 (mov<mode>_internal, TDmode/TFmode): Likewise.
18883 (movtf_softfloat): Likewise.
18884 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
18886 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
18887 movdi_internal64, using wg constraint for move direct operations.
18888 (movdi_internal64): Likewise.
18890 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
18891 MODES_TIEABLE_P for selected modes. Print the numerical value of
18892 the various virtual registers. Use GPR/FPR first/last values,
18893 instead of hard coding the register numbers. Print which modes
18894 have reload functions registered.
18895 (rs6000_option_override_internal): If -mdebug=reg, trace the options
18896 settings before/after setting cpu, target and subtarget settings.
18897 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
18898 and for secondary reload failures in rs6000_secondary_reload_inner.
18899 (rs6000_secondary_reload_fail): Likewise.
18900 (rs6000_secondary_reload_inner): Likewise.
18902 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
18903 macros for first/last GPR and FPR registers.
18904 (LAST_GPR_REGNO): Likewise.
18905 (FIRST_FPR_REGNO): Likewise.
18906 (LAST_FPR_REGNO): Likewise.
18908 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
18909 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
18910 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
18911 (vcond<mode><mode>): Likewise.
18912 (vcondu<mode><mode>): Likewise.
18913 (vector_gtu<mode>): Likewise.
18914 (vector_gte<mode>): Likewise.
18915 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
18916 to prevent the compiler from converting DImode operations to TImode.
18917 (ior<mode>3): Likewise.
18918 (and<mode>3): Likewise.
18919 (one_cmpl<mode>2): Likewise.
18920 (nor<mode>3): Likewise.
18921 (andc<mode>3): Likewise.
18923 * config/rs6000/constraints.md (wt constraint): New constraint
18924 that returns VSX_REGS if TImode is allowed in VSX registers.
18926 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
18927 constant under VSX.
18929 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
18930 similar to TImode, but it is restricted to being in the GPRs.
18932 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
18933 TImode to occupy a single VSX register.
18935 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
18936 -mvsx-timode for power7/power8.
18937 (power7 cpu): Likewise.
18938 (power8 cpu): Likewise.
18940 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
18941 sure that TFmode/TDmode take up two registers if they are ever
18942 allowed in the upper VSX registers.
18943 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
18945 (rs6000_init_hard_regno_mode_ok): Likewise.
18946 (rs6000_debug_reg_global): Add debugging for PTImode and wt
18947 constraint. Print if LRA is turned on.
18948 (rs6000_option_override_internal): Give an error if -mvsx-timode
18949 and VSX is not enabled.
18950 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
18951 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
18952 to reg+offset addressing. Use PTImode when checking offset
18953 addresses for validity.
18954 (reg_offset_addressing_ok_p): Likewise.
18955 (rs6000_legitimate_offset_address_p): Likewise.
18956 (rs6000_legitimize_address): Likewise.
18957 (rs6000_legitimize_reload_address): Likewise.
18958 (rs6000_legitimate_address_p): Likewise.
18959 (rs6000_eliminate_indexed_memrefs): Likewise.
18960 (rs6000_emit_move): Likewise.
18961 (rs6000_secondary_reload): Likewise.
18962 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
18963 reloads to fpr registers to continue to use reg+offset addressing,
18964 but 64-bit reloads to altivec registers need reg+reg addressing.
18965 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
18966 it. Treat LO_SUM like a PLUS operation.
18967 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
18968 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
18969 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
18970 registers to share a register with a smaller sized type, since VSX
18971 puts scalars in the upper 64-bits.
18972 (print_operand): Add support for PTImode.
18973 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
18974 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
18975 registers, but don't have arithmetic support.
18976 (rs6000_memory_move_cost): Add test for VSX.
18977 (rs6000_opt_masks): Add -mvsx-timode.
18979 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
18982 (VSr): Use wt constraint for TImode.
18983 (VSv): Drop TImode support.
18984 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
18985 (vsx_movti_64bit): Likewise.
18986 (vsx_movti_32bit): Likewise.
18987 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
18988 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
18989 one '?' on the appropriate output constraint. Do not allow TImode
18990 logical operations on 32-bit systems.
18991 (vsx_ior<mode>3): Likewise.
18992 (vsx_xor<mode>3): Likewise.
18993 (vsx_one_cmpl<mode>2): Likewise.
18994 (vsx_nor<mode>3): Likewise.
18995 (vsx_andc<mode>3): Likewise.
18996 (vsx_concat_<mode>): Likewise.
18997 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
18999 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
19000 OPTION_MASK_VSX_TIMODE.
19001 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
19002 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
19004 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
19005 (TI2 iterator): New iterator for TImode, PTImode.
19006 (wd mode attribute): Add values for vector types.
19007 (movti_string): Replace TI move operations with operations for TImode
19008 and PTImode. Add support for TImode being allowed in VSX registers.
19009 (mov<mode>_string, TImode/PTImode): Likewise.
19010 (movti_ppc64): Likewise.
19011 (mov<mode>_ppc64, TImode/PTImode): Likewise.
19012 (TI mode splitters): Likewise.
19014 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
19017 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
19019 PR tree-optimization/56355
19020 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
19021 Also handle integers with undefined overflow.
19023 2013-03-20 Catherine Moore <clm@codesourcery.com>
19024 Maciej W. Rozycki <macro@codesourcery.com>
19025 Tom de Vries <tom@codesourcery.com>
19026 Nathan Sidwell <nathan@codesourcery.com>
19027 Iain Sandoe <iain@codesourcery.com>
19028 Nathan Froyd <froydnj@codesourcery.com>
19029 Chao-ying Fu <fu@mips.com>
19031 * doc/extend.texi: (micromips, nomicromips, nocompression):
19032 Document new function attributes.
19033 * doc/invoke.texi (minterlink-compressed, mmicromips,
19034 m14k, m14ke, m14kec): Document new options.
19035 (minterlink-mips16): Update documentation.
19036 * doc/md.texi (ZC, ZD): Document new constraints.
19037 * configure.ac (gcc_cv_as_micromips): Check if linker
19038 supports the .set micromips directive.
19039 * configure: Regenerate.
19040 * config.in: Regenerate.
19041 * config/mips/mips-tables.opt: Regenerate.
19042 * config/mips/micromips.md: New file.
19043 * constraints.md (ZC, ZD): New constraints.
19044 * config/mips/predicates.md (movep_src_register): New predicate.
19045 (movep_src_operand): New predicate.
19046 (non_volatile_mem_operand): New predicate.
19047 * config/mips/mips.md (multimem): New type.
19048 (length): Differentiate between 17-bit and 18-bit branch offsets.
19049 (MOVEP1, MOVEP2): New mode iterator.
19050 (mov_<load>l): Use ZC constraint.
19051 (mov_<load>r): Likewise.
19052 (mov_<store>l): Likewise.
19053 (mov_<store>r): Likewise.
19054 (*branch_equality<mode>_inverted): Add microMIPS support.
19055 (*branch_equality<mode>): Likewise.
19056 (*jump_absolute): Likewise.
19057 (indirect_jump_<mode>): Likewise.
19058 (tablejump_<mode>): Likewise.
19059 (<optab>_internal): Likewise.
19060 (sibcall_internal): Likewise.
19061 (sibcall_value_internal): Likewise.
19062 (prefetch): Use constraint ZD.
19063 * config/mips/mips.opt (minterlink-compressed): New option.
19064 (minterlink-mips16): Now an alias for minterlink-compressed.
19065 (mmicromips): New option.
19066 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
19067 (compare_and_swap_12): Likewise.
19068 (sync_add<mode>): Likewise.
19069 (sync_<optab>_12): Likewise.
19070 (sync_old_<optab>_12): Likewise.
19071 (sync_new_<optab>_12): Likewise.
19072 (sync_nand_12): Likewise.
19073 (sync_old_nand_12): Likewise.
19074 (sync_new_nand_12): Likewise.
19075 (sync_sub<mode>): Likewise.
19076 (sync_old_add<mode>): Likewise.
19077 (sync_old_sub<mode>): Likewise.
19078 (sync_new_add<mode>): Likewise.
19079 (sync_new_sub<mode>): Likewise.
19080 (sync_<optab><mode>): Likewise.
19081 (sync_old_<optab><mode>): Likewise.
19082 (sync_new_<optab><mode>): Likewise.
19083 (sync_nand<mode>): Likewise.
19084 (sync_old_nand<mode>): Likewise.
19085 (sync_new_nand<mode>): Likewise.
19086 (sync_lock_test_and_set<mode>): Likewise.
19087 (test_and_set_12): Likewise.
19088 (atomic_compare_and_swap<mode>): Likewise.
19089 (atomic_exchange<mode>_llsc): Likewise.
19090 (atomic_fetch_add<mode>_llsc): Likewise.
19091 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
19092 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
19093 (umips_save_restore_pattern_p): Likewise.
19094 (umips_load_store_pair_p): Likewise.
19095 (umips_output_load_store_pair): Likewise.
19096 (umips_movep_target_p): Likewise.
19097 (umips_12bit_offset_address_p): Likewise.
19098 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
19099 (mips_base_mips16): Rename this...
19100 (mips_base_compression_flags): ...to this. Update all uses.
19101 (mips_attribute_table): Add micromips, nomicromips and nocompression.
19102 (mips_mips16_decl_p): Delete.
19103 (mips_nomips16_decl_p): Delete.
19104 (mips_get_compress_on_flags): New function.
19105 (mips_get_compress_off_flags): New function.
19106 (mips_get_compress_mode): New function.
19107 (mips_get_compress_on_name): New function.
19108 (mips_get_compress_off_name): New function.
19109 (mips_insert_attributes): Support multiple compression types.
19110 (mips_merge_decl_attributes): Likewise.
19111 (umips_12bit_offset_address_p): New function.
19112 (mips_start_function_definition): Emit .set micromips directive.
19113 (mips_call_may_need_jalx_p): New function.
19114 (mips_function_ok_for_sibcall): Add microMIPS support.
19115 (mips_print_operand_punctuation): Support short delay slots and
19117 (umips_swm_mask, umips_swm_encoding): New.
19118 (umips_build_save_restore): New function.
19119 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
19120 (was_mips16_p): Remove.
19121 (old_compression_mode): New.
19122 (mips_set_compression_mode): New function.
19123 (mips_set_current_function): Add microMIPS support.
19124 (mips_option_override): Likewise.
19125 (umips_save_restore_pattern_p): New function.
19126 (umips_output_save_restore): New function.
19127 (umips_load_store_pair_p_1): New function.
19128 (umips_load_store_pair_p): New function.
19129 (umips_output_load_store_pair_1): New function.
19130 (umips_output_load_store_pair): New function.
19131 (umips_movep_target_p) New function.
19132 (mips_prepare_pch_save): Add microMIPS support.
19133 * config/mips/mips.h (TARGET_COMPRESSION): New.
19134 (TARGET_CPU_CPP_BUILTINS): Update macro
19135 to use new compression flags and to support microMIPS.
19136 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
19137 (MIPS_ARCH_FLOAT_SPEC): Likewise.
19138 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
19139 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
19140 (ASM_SPEC): Support mmicromips and mno-micromips.
19141 (M16STORE_REG_P): New macro.
19142 (MIPS_CALL): Support TARGET_MICROMIPS.
19143 (MICROMIPS_J): New macro.
19144 (mips_base_mips16): Rename this...
19145 (mips_base_compression_flags): ...to this.
19146 (UMIPS_12BIT_OFFSET_P): New macro.
19147 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
19148 (MULTILIB_DIRNAMES): Likewise.
19149 2013-03-20 Richard Biener <rguenther@suse.de>
19151 PR tree-optimization/56661
19152 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
19153 the result does not have to be distinct.
19155 2013-03-20 Richard Biener <rguenther@suse.de>
19157 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
19160 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
19161 Steven Bosscher <steven@gcc.gnu.org>
19163 PR rtl-optimization/56605
19164 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
19166 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
19169 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
19170 that require movd instead of movq.
19172 2013-03-20 Richard Biener <rguenther@suse.de>
19174 * tree-ssa-structalias.c (struct variable_info): Add pointer
19175 to the first field of an aggregate with sub-vars. Make
19176 this and the pointer to the next subfield its ID.
19177 (vi_next): New function.
19178 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
19179 storedanything_id, integer_id): Increment by one.
19180 (new_var_info, get_call_vi, lookup_call_clobber_vi,
19181 get_call_clobber_vi): Adjust.
19182 (solution_set_expand): Simplify and speedup.
19183 (solution_set_add): Inline into ...
19184 (set_union_with_increment): ... this. Adjust accordingly.
19185 (do_sd_constraint): Likewise.
19186 (do_ds_constraint): Likewise.
19187 (do_complex_constraint): Simplify.
19188 (build_pred_graph): Adjust.
19189 (solve_graph): Likewise. Simplify and speedup.
19190 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
19191 get_constraint_for_component_ref, get_constraint_for_1,
19192 first_vi_for_offset, first_or_preceding_vi_for_offset,
19193 create_function_info_for, create_variable_info_for_1,
19194 create_variable_info_for, intra_create_variable_infos): Adjust.
19195 (init_base_vars): Push NULL for ID zero.
19196 (compute_points_to_sets): Adjust.
19198 2013-03-20 Richard Biener <rguenther@suse.de>
19200 * cfgloop.c (verify_loop_structure): Streamline and avoid
19201 ICEing on corrupt loop tree.
19202 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
19205 2013-03-20 Richard Biener <rguenther@suse.de>
19207 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
19208 check whether an SSA update is needed.
19210 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
19212 * config/mips/constraints.md (T): Rename to...
19216 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
19217 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
19219 2013-03-19 Ian Bolton <ian.bolton@arm.com>
19221 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
19222 (*subsi3_carryin_uxtw): Likewise.
19224 2013-03-19 Ian Bolton <ian.bolton@arm.com>
19226 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
19227 (*rorsi3_insn_uxtw): Likewise.
19229 2013-03-19 Ian Bolton <ian.bolton@arm.com>
19231 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
19232 (*extrsi5_insn_uxtw): Likewise.
19234 2013-03-19 Richard Biener <rguenther@suse.de>
19236 PR tree-optimization/56273
19237 * passes.c (init_optimization_passes): Move second VRP after DOM.
19239 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
19241 * config/i386/i386.md (*movti_internal): Merge from
19242 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
19243 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
19244 nox64 isa attributes.
19246 2013-03-18 Richard Biener <rguenther@suse.de>
19248 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
19250 (merge_node_constraints): Likewise.
19251 (build_succ_graph): Likewise.
19252 (valid_graph_edge): Inline into single caller.
19253 (unify_nodes): Likewise. Use bitmap_set_bit return value
19255 (scc_visit): Fix formatting and variable use.
19256 (do_sd_constraint): Use gcc_checking_assert.
19257 (do_ds_constraint): Likewise.
19258 (do_complex_constraint): Likewise.
19259 (condense_visit): Likewise. Cleanup.
19260 (dump_pred_graph): New function.
19261 (perform_var_substitution): Dump the pred-graph before
19262 variable substitution.
19263 (find_equivalent_node): Use gcc_checking_assert.
19264 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
19266 2013-03-18 Richard Biener <rguenther@suse.de>
19268 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
19269 Remove cond_expr_stmt_list argument and do not gimplify the
19271 (vect_loop_versioning): Adjust.
19272 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
19273 Cleanup to use less temporaries.
19274 (vect_create_data_ref_ptr): Cleanup.
19276 2013-03-18 Jakub Jelinek <jakub@redhat.com>
19278 PR tree-optimization/56635
19279 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
19280 require types_compatible_p types.
19282 2013-03-18 Nick Clifton <nickc@redhat.com>
19284 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
19285 spurious backslash.
19287 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
19288 Add missing line to comment describing function.
19290 2013-03-18 Richard Biener <rguenther@suse.de>
19292 PR tree-optimization/56210
19293 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
19294 Handle string / character search functions.
19295 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
19297 2013-03-18 Richard Biener <rguenther@suse.de>
19299 PR middle-end/56483
19300 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
19301 and implement properly.
19302 * gimple.h (gimple_cond_single_var_p): Remove.
19304 2013-03-18 Richard Biener <rguenther@suse.de>
19306 * tree-data-ref.h (find_data_references_in_loop): Declare.
19307 * tree-data-ref.c (get_references_in_stmt): Use a stack
19308 vector pre-allocated in the callers.
19309 (find_data_references_in_stmt): Adjust.
19310 (graphite_find_data_references_in_stmt): Likewise.
19311 (create_rdg_vertices): Likewise.
19312 (find_data_references_in_loop): Export.
19313 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
19314 Compute dependences here...
19315 (vect_analyze_data_refs): ...not here. When we encounter
19316 a non-vectorizable data reference in basic-block vectorization
19317 truncate the data reference vector. Do not bother to
19318 fixup data-dependence information for gather loads.
19319 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
19320 of data references, as reported.
19322 2013-03-18 Richard Biener <rguenther@suse.de>
19324 PR tree-optimization/3713
19325 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
19326 has_constants and expr.
19327 (stmt_has_constants): Properly valueize SSA names when deciding
19328 whether the stmt has constants.
19330 2013-03-18 Richard Biener <rguenther@suse.de>
19332 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
19333 whole function when there is nothing to do.
19334 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
19335 * tree-vectorizer.c (vectorize_loops): Update virtual and
19336 loop-closed SSA once.
19337 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
19339 2013-03-18 Richard Biener <rguenther@suse.de>
19341 PR middle-end/56113
19342 * domwalk.c (bb_postorder): New global static.
19343 (cmp_bb_postorder): New function.
19344 (walk_dominator_tree): Replace scheme imposing an order for
19345 visiting dominator sons by one sorting them at the time they
19346 are pushed on the stack.
19348 2013-03-18 Richard Biener <rguenther@suse.de>
19350 PR tree-optimization/39326
19351 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
19352 (struct mem_ref): Replace mem member with ao_ref typed member.
19353 (MEM_ANALYZABLE): Adjust.
19354 (memref_eq): Likewise.
19355 (mem_ref_alloc): Likewise.
19356 (gather_mem_refs_stmt): Likewise.
19357 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
19358 (execute_sm_if_changed_flag_set): Adjust.
19359 (execute_sm): Likewise.
19360 (ref_always_accessed_p): Likewise.
19361 (refs_independent_p): Likewise.
19362 (can_sm_ref_p): Likewise.
19364 2013-03-18 Jakub Jelinek <jakub@redhat.com>
19367 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
19368 return 1 even for !unsignedp.
19370 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
19372 * config/i386/i386.md (isa): Add x64 and nox64.
19373 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
19374 (*pushtf): Enable *roF alternative for x64 isa only.
19375 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
19376 mode attribute of integer alternatives to DImode for TARGET_64BIT.
19377 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
19378 (*movtf_internal): Merge from *movtf_internal_rex64 and
19379 *movtf_internal_sse. Use x64 and nox64 isa attributes.
19380 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
19381 nox64 isa attributes.
19382 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
19383 nox64 isa attributes.
19384 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
19386 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
19388 * config/alpha/alpha.c (TARGET_LRA_P): New define.
19390 2013-03-17 Jakub Jelinek <jakub@redhat.com>
19393 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
19394 class names. Remove trailing comma after "ALL_REGS".
19396 2013-03-16 Jan Hubicka <jh@suse.cz>
19398 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
19399 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
19400 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
19401 of cgraph_get_create_node.
19402 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
19404 2013-03-16 Jason Merrill <jason@redhat.com>
19407 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
19408 with DW_AT_default_value.
19410 2013-03-16 Jakub Jelinek <jakub@redhat.com>
19412 * BASE-VER: Set to 4.9.0.
19414 2013-03-14 Andi Kleen <ak@linux.intel.com>
19417 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
19418 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
19419 Document _x* TSX intrinsics.
19421 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
19422 David Holsgrove <david.holsgrove@xilinx.com>
19424 * configure.ac: Add MicroBlaze TLS support detection.
19425 * configure: Regenerate.
19426 * config/microblaze/microblaze-protos.h
19427 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
19428 symbol_mentioned_p, label_mentioned_p): Add prototypes.
19429 * config/microblaze/microblaze.c (microblaze_address_type): Add
19430 ADDRESS_TLS and tls_reloc address types.
19431 (microblaze_address_info): Add tls_reloc.
19432 (TARGET_HAVE_TLS): Define.
19433 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
19434 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
19435 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
19436 load_tls_operand, microblaze_call_tls_get_addr,
19437 microblaze_legitimize_tls_address): New functions.
19438 (microblaze_classify_unspec): Handle UNSPEC_TLS.
19439 (get_base_reg): Use microblaze_tls_symbol_p.
19440 (microblaze_classify_address): Handle TLS.
19441 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
19442 label_mentioned_p and microblaze_tls_referenced_p.
19443 (microblaze_legitimize_address): Handle TLS.
19444 (microblaze_address_insns): Handle ADDRESS_TLS.
19445 (pic_address_needs_scratch): Handle TLS.
19446 (print_operand_address): Handle TLS.
19447 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
19448 (microblaze_expand_move): Handle TLS.
19449 (microblaze_legitimate_constant_p): Check
19450 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
19451 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
19452 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
19453 (PIC_OFFSET_TABLE_REGNUM): Set.
19454 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
19455 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
19456 (addsi3, movsi_internal2, movdf_internal): Update constraints
19457 * config/microblaze/predicates.md (arith_plus_operand): Define
19458 (move_operand): Redefine as move_src_operand,
19459 check microblaze_tls_referenced_p.
19461 2013-03-14 Ian Bolton <ian.bolton@arm.com>
19463 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
19464 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
19466 2013-03-14 Ian Bolton <ian.bolton@arm.com>
19468 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
19471 2013-03-14 Jakub Jelinek <jakub@redhat.com>
19473 PR tree-optimization/53265
19474 * common.opt (Waggressive-loop-optimizations): New option.
19475 * tree-ssa-loop-niter.c: Include tree-pass.h.
19476 (do_warn_aggressive_loop_optimizations): New function.
19477 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
19478 if number_of_latch_executions returned constant.
19479 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
19480 early. If number_of_latch_executions returned constant, set
19481 nb_iterations_upper_bound back to it.
19482 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
19484 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
19485 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
19487 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
19488 (MULTILIB_OSDIRNAMES): Set.
19489 * genmultilib: If defaultosdirname doesn't start with :: , set
19490 defaultosdirname2 instead, clear it and emit two . multilib_raw
19491 entries instead of just one.
19493 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
19495 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
19496 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
19497 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
19498 (SUBTARGET_OVERRIDE_OPTIONS): New.
19500 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
19503 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
19504 (musermode): Convert to Var(TARGET_USERMODE).
19505 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
19506 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
19507 * config/sh/sh.c (sh_option_override): Use
19508 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
19509 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
19511 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
19513 (udivsi3_i4_single, divsi3_i4_single): Use
19514 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
19516 2013-03-13 Dave Korn <dave.korn.cygwin@....>
19518 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
19521 2013-03-13 Richard Biener <rguenther@suse.de>
19523 PR tree-optimization/56608
19524 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
19525 calls when vectorizing basic-blocks.
19527 2013-03-13 Jakub Jelinek <jakub@redhat.com>
19530 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
19533 2013-03-12 Jakub Jelinek <jakub@redhat.com>
19535 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
19537 2013-03-11 Jan Hubicka <jh@suse.cz>
19540 * lto-streamer-out.c (output_symbol_p): Skip references from
19541 constructors of external variables.
19543 2013-03-11 Jan Hubicka <jh@suse.cz>
19545 PR middle-end/56571
19546 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
19548 * emit-rtl.c (verify_rtx_sharing): Likewise.
19549 (copy_insn_1): Likewise.
19550 * rtl.c (copy_rtx): Likewise.
19552 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
19555 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
19556 output_operand_lossage message.
19558 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
19561 * arm.c (shift_op): Validate RTL pattern on the fly.
19562 (arm_print_operand, case 'S'): Don't use shift_operator to validate
19565 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
19568 * config/pa/pa.md (call_value): Check for calls to powf and direct to
19569 new call patterns that clobber %fr12.
19570 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
19571 split and postreload patterns.
19572 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
19573 registers %fr12 and %fr12R as call used.
19575 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
19577 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
19578 (canon_address, record_store, replace_read, check_mem_read_rtx,
19579 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
19580 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
19581 rest_of_handle_dse): Likewise.
19583 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
19585 PR middle-end/56524
19586 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
19588 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
19589 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
19590 (save_optabs_if_changed): Replace with...
19591 (init_tree_optimization_optabs): ...this.
19592 * optabs.c (save_optabs_if_changed): Rename to...
19593 (init_tree_optimization_optabs): ...this. Take the optimization node
19594 as argument. Do nothing if the base optabs are already correct.
19595 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
19596 to recompute optabs.
19597 * function.h (function): Remove optabs field.
19598 * function.c (invoke_set_current_function_hook): Call
19599 init_tree_optimization_optabs. Use the result to initialize
19602 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
19604 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
19605 if GTMA_HAS_NO_INSTRUMENTATION.
19606 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
19607 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
19608 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
19609 * gimple-pretty-print.c (dump_gimple_transaction): Handle
19610 GTMA_HAS_NO_INSTRUMENTATION.
19612 2013-03-08 Jakub Jelinek <jakub@redhat.com>
19614 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
19617 2013-03-08 Marek Polacek <polacek@redhat.com>
19618 Jakub Jelinek <jakub@redhat.com>
19620 PR tree-optimization/56478
19621 * predict.c (is_comparison_with_loop_invariant_p): Change the
19622 type of loop_step to tree.
19623 (predict_loops): Adjust.
19624 (predict_iv_comparison): Perform the computations on double_ints.
19626 2013-03-08 Richard Biener <rguenther@suse.de>
19628 PR tree-optimization/56570
19629 * tree-cfg.c (verify_expr_location_1): Verify locations for
19631 * tree-sra.c (create_access_replacement): Strip locations
19632 from DECL_DEBUG_EXPRs.
19634 2013-03-08 Richard Biener <rguenther@suse.de>
19636 * tree-inline.c (expand_call_inline): Do not associate
19637 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
19638 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
19640 2013-03-08 Richard Biener <rguenther@suse.de>
19642 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
19643 or block changes with -Og. Fix for location / block encoding
19644 changes and PHI arguments with locations.
19646 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
19648 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
19650 (struct output_info): Likewise.
19651 (register_overhead): Remove bad gcc_assert.
19652 (bitmap_find_bit): If there is only a single bitmap element, do not
19653 count a miss as a search.
19654 (print_statistics): Update for counter type changes.
19655 (dump_bitmap_statistics): Likewise. Print headers such that they
19656 are properly lined up with the printed counters.
19658 2013-03-07 Jakub Jelinek <jakub@redhat.com>
19660 PR tree-optimization/56559
19661 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
19662 check that it has only a single use.
19664 2013-03-07 Richard Biener <rguenther@suse.de>
19666 * doc/invoke.texi (fwhole-program): Discourage use in combination
19669 2013-03-06 Jakub Jelinek <jakub@redhat.com>
19671 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
19673 PR tree-optimization/56539
19674 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
19675 instead of GSI_CONTINUE_LINKING as last argument to
19676 force_gimple_operand_gsi. Adjust function comment.
19678 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
19681 PR middle-end/56548
19682 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
19683 promoted mode, convert the result back to the original mode.
19685 2013-03-06 Richard Biener <rguenther@suse.de>
19687 PR middle-end/56294
19688 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
19689 (insert_updated_phi_nodes_compare_uids): New function.
19690 (update_ssa): Sort symbols_to_rename after UID before
19691 traversing it to insert PHI nodes.
19693 2013-03-06 Richard Biener <rguenther@suse.de>
19695 PR middle-end/50494
19696 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
19697 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
19700 2013-02-13 Richard Biener <rguenther@suse.de>
19703 * varasm.c (output_constant_def_1): Get the decl representing
19704 the constant as argument.
19705 (output_constant_def): Wrap output_constant_def_1.
19706 (make_decl_rtl): Use output_constant_def_1 with the decl
19707 representing the constant.
19708 (build_constant_desc): Optionally re-use a decl already
19709 representing the constant.
19710 (tree_output_constant_def): Adjust.
19712 2013-03-06 Joey Ye <joey.ye@arm.com>
19715 * gcc.c (convert_white_space): New function.
19716 (main): Handles white space in function name.
19718 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
19721 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
19722 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
19723 to SH_DIV_CALL_TABLE for TARGET_SH2.
19724 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
19726 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
19727 call-table options.
19729 2013-03-05 Sterling Augustine <saugustine@google.com>
19730 Cary Coutant <ccoutant@google.com>
19733 * dwarf2out.c (resolve_addr): Don't call
19734 remove_loc_list_addr_table_entries a second time for the same
19737 2013-03-05 Jakub Jelinek <jakub@redhat.com>
19740 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
19741 (avoid_complex_debug_insns): New function.
19742 (expand_debug_locations): Call it.
19744 PR rtl-optimization/56484
19745 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
19746 lifetimes of hard registers on small register class machines.
19748 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
19750 * config/microblaze/microblaze-protos.h: Rename
19751 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
19752 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
19754 (microblaze_fast_interrupt_function_p): New function.
19755 (microblaze_is_interrupt_handler): Rename to
19756 microblaze_is_interrupt_variant and add fast_interrupt check.
19757 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
19758 (save_restore_insns): Likewise.
19759 (compute_frame_size): Likewise.
19760 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
19761 (microblaze_globalize_label): Likewise.
19762 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
19763 * config/microblaze/microblaze.md: Use wrapper
19764 microblaze_is_interrupt_variant.
19766 2013-03-05 Kai Tietz <ktietz@redhat.com>
19768 * sdbout.c (sdbout_one_type): Switch to current function's section
19769 supporting cold/hot.
19771 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
19773 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
19776 2013-03-05 Jakub Jelinek <jakub@redhat.com>
19778 PR middle-end/56461
19779 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
19780 if VALGRIND_GET_VBITS is defined, temporarily make object
19781 memory all defined, and restore previous valgrind addressability
19782 and definability afterwards. Free this_object at the end.
19784 PR middle-end/56461
19785 * lra.c (lra): Call lra_clear_live_ranges if live_p,
19786 right before calling lra_create_live_ranges, also call it
19787 when clearing live_p. Only call lra_clear_live_ranges
19788 at the end if live_p.
19790 PR middle-end/56461
19791 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
19793 2013-03-05 Richard Biener <rguenther@suse.de>
19795 PR tree-optimization/56521
19796 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
19799 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
19802 * except.h (remove_unreachable_eh_regions): New prototype.
19803 * except.c (remove_eh_handler_splicer): New function, split out
19804 of remove_eh_handler.
19805 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
19806 warning about running it on many EH regions one at a time.
19807 (remove_unreachable_eh_regions_worker): New function, walk the
19808 EH tree in depth-first order and remove non-marked regions.
19809 (remove_unreachable_eh_regions): New function.
19810 * tree-eh.c (mark_reachable_handlers): New function, split out
19811 from remove_unreachable_handlers.
19812 (remove_unreachable_handlers): Use mark_reachable_handlers and
19813 remove_unreachable_eh_regions.
19814 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
19815 and remove_unreachable_eh_regions.
19817 2013-03-05 Richard Biener <rguenther@suse.de>
19819 PR middle-end/56525
19820 * loop-init.c (fix_loop_structure): Remove loops in two stages,
19821 not freeing them until the end.
19823 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19825 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
19827 2013-03-05 Richard Biener <rguenther@suse.de>
19829 PR tree-optimization/56270
19830 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
19831 of loads after scheduling an SLP instance.
19833 2013-03-05 Jakub Jelinek <jakub@redhat.com>
19835 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
19837 (check_gcc_parallelize): Run guality.exp as a separate job from
19838 vect.exp with unsorted.exp and $(dg_target_exps) separately from
19839 struct-layout-1.exp with stackalign.exp.
19841 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
19843 PR middle-end/56461
19844 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
19845 load_index sbitmap even if some bit in it isn't set.
19847 PR middle-end/56461
19848 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
19849 (discover_iteration_bound_by_body_walk): Change queues to
19850 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
19851 spelling in comment. Call safe_push on queues[bound_index] directly.
19852 Release queues[queue_index] in every iteration unconditionally.
19853 Release bounds vector.
19855 PR middle-end/56461
19856 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
19857 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
19858 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
19861 2013-03-05 Richard Biener <rguenther@suse.de>
19864 * tree-inline.c (remap_blocks_to_null): New function.
19865 (expand_call_inline): When expanding a call stmt without
19866 an associated block inline remap all callee blocks to NULL.
19868 2013-03-05 Jakub Jelinek <jakub@redhat.com>
19870 PR rtl-optimization/56494
19871 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
19872 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
19873 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
19875 PR middle-end/56461
19876 * sel-sched-ir.c (free_sched_pools): Release
19877 succs_info_pool.stack[succs_info_pool.max_top] vectors too
19878 if succs_info_pool.max_top isn't -1.
19881 * opts.c (opts_obstack, opts_concat): Moved to...
19882 * opts-common.c (opts_obstack, opts_concat): ... here.
19884 2013-03-04 Jakub Jelinek <jakub@redhat.com>
19886 PR middle-end/56461
19887 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
19889 2013-03-04 Martin Jambor <mjambor@suse.cz>
19891 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
19892 all appropriate places.
19894 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
19896 PR tree-optimization/56424
19897 * ipa-split.c (split_function): Do not set the RSO flag if result is
19898 not by reference and its type is a register type.
19900 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
19902 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
19903 (microblaze_legitimate_pic_operand): Likewise
19904 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
19905 new function microblaze_legitimate_pic_operand
19906 * config/microblaze/microblaze-protos.h
19907 (microblaze_legitimate_pic_operand): Declare.
19909 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
19911 * config/microblaze/predicates.md (call_insn_simple_operand):
19912 New predicate for supported rtx code types.
19913 * config/microblaze/microblaze.md (call_internal1): Use
19914 call_insn_simple_operand predicate.
19916 2013-03-04 Jakub Jelinek <jakub@redhat.com>
19918 PR middle-end/56461
19919 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
19920 partitions.ordered_remove.
19922 PR middle-end/56461
19923 * tree-vect-stmts.c (vectorizable_conversion): Don't call
19924 vec_oprnds0.create (1) for modifier == NONE.
19926 PR middle-end/56461
19927 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
19928 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
19929 vec_oprnds1 right before pushing anything to it for
19932 PR middle-end/56461
19933 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
19934 set nbbs to 0 instead of having separate code path.
19935 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
19936 instead of false as last argument if returning NULL.
19938 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
19940 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
19941 the attribute is now called "target" instead of "option".
19942 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
19943 * doc/tm.texi.in (Target Attributes): Likewise document the correct
19944 attribute/pragma name for TARGET_OPTION_VALID_P and
19945 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
19946 * doc/tm.texi: Regenerated.
19948 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
19950 * config/microblaze/microblaze.c:
19951 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
19952 * config/microblaze/microblaze.h: Add -mxl-reorder to
19954 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
19955 instructions emitted if TARGET_REORDER.
19956 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
19957 or 0 for -m/-mno case, but initialises as 2 to detect default use case
19960 2013-03-01 Xinliang David Li <davidxl@google.com>
19962 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
19965 2013-03-01 Jakub Jelinek <jakub@redhat.com>
19967 PR middle-end/56461
19968 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
19969 vector even when returning true. Fix up function comment formatting.
19971 PR middle-end/56461
19972 * ira-build.c (ira_loop_nodes_count): New variable.
19973 (create_loop_tree_nodes): Initialize it.
19974 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
19976 PR middle-end/56461
19977 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
19978 method on dr_chain and result_chain.
19979 * tree-vect-stmts.c (vectorizable_store): Only call
19980 result_chain.create if j == 0.
19982 PR middle-end/56461
19983 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
19984 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
19985 before overwriting it.
19987 2013-03-01 Tobias Burnus <burnus@net-b.de>
19989 * doc/extended.texi (C Extensions): Change order in @menu
19991 (Other MIPS Built-in Functions): Move last MIPS entry before
19992 "picoChip Built-in Functions".
19993 (SH Built-in Functions): Move after RX Built-in Functions.
19994 * doc/gcc.texi (Introduction): Change order in @menu
19996 * doc/md.texi (Constraints): Ditto.
19997 * gty.texi (Type Information): Ditto.
19998 (User-provided marking routines for template types): Make
20000 * doc/invoke.texi (AArch64 Options): Move before
20001 "Adapteva Epiphany Options".
20003 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
20004 Jakub Jelinek <jakub@redhat.com>
20007 * asan.c (gate_asan): Lookup no_sanitize_address instead of
20008 no_address_safety_analysis attribute.
20009 * doc/extend.texi (no_address_safety_attribute): Rename to
20010 no_sanitize_address attribute, mention no_address_safety_analysis
20011 attribute as deprecated alias.
20013 2013-02-28 Jakub Jelinek <jakub@redhat.com>
20015 PR middle-end/56461
20016 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
20017 type to vec<vec<tree> > *.
20018 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
20019 to be vec<tree> instead of vec<tree> *, set vec_defs
20020 to vNULL and call vec_defs.create (number_of_vects), adjust other
20022 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
20023 vectorizable_condition): Adjust vect_get_slp_defs callers.
20025 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
20027 * config/aarch64/aarch64.c
20028 (aarch64_float_const_representable): Remove unused variable.
20030 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
20032 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
20034 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
20036 * config/aarch64/aarch64-builtins.c
20037 (aarch64_init_simd_builtins): Make static.
20039 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
20041 * config/aarch64/aarch64.c
20042 (aarch64_simd_make_constant): Make static.
20044 2013-02-28 Martin Jambor <mjambor@suse.cz>
20046 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
20047 with no initialization to the RHS of debug statements.
20049 2013-02-28 Martin Jambor <mjambor@suse.cz>
20051 PR tree-optimization/56294
20052 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
20054 (get_access_replacement): Do not call create_access_replacement.
20055 Assert a replacement exists.
20056 (get_repl_default_def_ssa_name): Create the replacement declaration
20059 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20061 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
20062 final_end_function.
20064 2013-02-28 Marek Polacek <polacek@redhat.com>
20066 PR rtl-optimization/56466
20067 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
20068 if we're changing a loop.
20069 (peel_loops_completely): Likewise.
20071 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
20074 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
20076 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
20079 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
20080 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
20081 INTX_FTYPE_FX, FX_FTYPE_INTX.
20082 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
20084 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
20086 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
20087 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
20088 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
20089 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
20090 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
20091 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
20092 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
20093 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
20094 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
20095 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
20096 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
20097 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
20098 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
20099 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
20100 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
20101 (avrxmega6): Increase max flash segments from 5 to 6.
20102 * config/avr/t-multilib: Regenerate.
20103 * config/avr/avr-tables.opt: Regenerate.
20104 * doc/avr-mmcu.texi: Regenerate.
20106 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
20108 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
20109 (avr_device_to_arch): Rename to avr_device_to_ld.
20110 (avr_device_to_as): New prototype.
20111 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
20112 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
20113 * config/avr/driver-avr.c (avr_device_to_as): New.
20114 (avr_device_to_arch): Rename to avr_device_to_ld.
20116 2013-02-27 Jakub Jelinek <jakub@redhat.com>
20118 PR middle-end/56461
20119 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
20120 method on dr_chain and result_chain.
20122 PR middle-end/56461
20123 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
20124 pointer_set_destroy on not_executed_last_iteration.
20126 PR middle-end/56461
20127 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
20129 PR middle-end/56461
20130 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
20131 FOR_EACH_DEFINED_FUNCTION when freeing state.
20133 PR middle-end/56461
20134 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
20136 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
20139 PR middle-end/56461
20140 * ipa-cp.c (decide_whether_version_node): Call vec_free on
20141 known_aggs[i].items and release known_aggs vector.
20143 PR middle-end/56461
20144 * ipa-reference.c (propagate): Free node_info even for alias nodes.
20146 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20148 * config/microblaze/microblaze.c (microblaze_emit_compare):
20149 Use xor for EQ/NE comparisions.
20150 * config/microblaze/microblaze.md (cstoresf4): Add constraints
20151 (cbranchsf4): Adjust operator to comparison_operator.
20153 2013-02-27 Jakub Jelinek <jakub@redhat.com>
20155 PR middle-end/56461
20156 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
20158 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
20159 vec_safe_push, always update *slot.
20160 (redirect_edge_var_map_clear): Use vec_free.
20161 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
20162 (free_var_map_entry): Use vec_free.
20163 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
20164 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
20166 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
20168 PR middle-end/45472
20169 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
20170 when the may_trap_p bit of the exprs being merged differs.
20171 Reorder tests for speculativeness in the logical and operator.
20173 2013-02-27 Jakub Jelinek <jakub@redhat.com>
20175 * incpath.c (add_standard_paths): Use reconcat instead of concat
20176 where appropriate and avoid leaking memory.
20178 * opts.h: Include obstack.h.
20179 (opts_concat): New prototype.
20180 (opts_obstack): New declaration.
20181 * opts.c (opts_concat): New function.
20182 (opts_obstack): New variable.
20183 (init_options_struct): Call gcc_init_obstack on opts_obstack.
20184 (finish_options): Use opts_concat instead of concat
20185 and XOBNEWVEC instead of XNEWVEC.
20186 * opts-common.c (generate_canonical_option, decode_cmdline_option,
20187 generate_option): Likewise.
20188 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
20189 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
20192 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
20193 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
20195 2013-02-26 Jakub Jelinek <jakub@redhat.com>
20197 PR middle-end/56461
20198 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
20200 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
20202 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
20203 (arm_block_move_unaligned_straight): Likewise.
20204 (arm_adjust_block_mem): Likewise.
20206 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
20209 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
20210 temp, cond and label.
20211 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
20214 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
20215 * config/c6x/c6x.h (dbx_register_map): Update declaration.
20218 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
20219 of prologue/epilogue functions.
20222 * config/tilegx/tilegx.c (tilegx_expand_prologue):
20223 Remove unused variable cfa_offset.
20224 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
20227 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
20228 type promotion to unsigned.
20231 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
20232 for HOST_WIDE_INT of 32 bit / same size as int.
20233 (arm_block_move_unaligned_straight): Likewise.
20234 (arm_adjust_block_mem): Likewise.
20237 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
20240 2013-02-26 Marek Polacek <polacek@redhat.com>
20242 PR tree-optimization/56426
20243 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
20245 2013-02-26 Richard Biener <rguenther@suse.de>
20248 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
20249 unused variable loops.
20251 2013-02-26 Jakub Jelinek <jakub@redhat.com>
20253 PR tree-optimization/56448
20254 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
20255 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
20256 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
20257 later operands of the references, or even first operand for
20258 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
20260 PR tree-optimization/56443
20261 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
20262 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
20263 to type_for_mode langhook.
20265 2013-02-25 Matt Turner <mattst88@gmail.com>
20267 * doc/invoke.texi: Document r4700.
20269 2013-02-25 Richard Biener <rguenther@suse.de>
20271 PR tree-optimization/56175
20272 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
20274 (simplify_bitwise_binary): ... here. Also guard the conversion
20275 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
20277 2013-02-25 Catherine Moore <clm@codesourcery.com>
20280 2013-02-24 Catherine Moore <clm@codesourcery.com>
20281 Maciej W. Rozycki <macro@codesourcery.com>
20282 Tom de Vries <tom@codesourcery.com>
20283 Nathan Sidwell <nathan@codesourcery.com>
20284 Iain Sandoe <iain@codesourcery.com>
20285 Nathan Froyd <froydnj@codesourcery.com>
20286 Chao-ying Fu <fu@mips.com>
20288 * doc/extend.texi: (micromips, nomicromips, nocompression):
20289 Document new function attributes.
20290 * doc/invoke.texi (minterlink-compressed, mmicromips,
20291 m14k, m14ke, m14kec): Document new options.
20292 (minterlink-mips16): Update documentation.
20293 * doc/md.texi (ZC, ZD): Document new constraints.
20294 * configure.ac (gcc_cv_as_micromips): Check if linker
20295 supports the .set micromips directive.
20296 * configure: Regenerate.
20297 * config.in: Regenerate.
20298 * config/mips/mips-tables.opt: Regenerate.
20299 * config/mips/micromips.md: New file.
20300 * constraints.md (ZC, AD): New constraints.
20301 * config/mips/predicates.md (movep_src_register): New predicate.
20302 (movep_src_operand): New predicate.
20303 (non_volatile_mem_operand): New predicate.
20304 * config/mips/mips.md (multimem): New type.
20305 (length): Differentiate between 17-bit and 18-bit branch offsets.
20306 (MOVEP1, MOVEP2): New mode iterator.
20307 (mov_<load>l): Use ZC constraint.
20308 (mov_<load>r): Likewise.
20309 (mov_<store>l): Likewise.
20310 (mov_<store>r): Likewise.
20311 (*branch_equality<mode>_inverted): Add microMIPS support.
20312 (*branch_equality<mode>): Likewise.
20313 (*jump_absolute): Likewise.
20314 (indirect_jump_<mode>): Likewise.
20315 (tablejump_<mode>): Likewise.
20316 (<optab>_internal): Likewise.
20317 (sibcall_internal): Likewise.
20318 (sibcall_value_internal): Likewise.
20319 (prefetch): Use constraint ZD.
20320 * config/mips/mips.opt (minterlink-compressed): New option.
20321 (minterlink-mips16): Now an alias for minterlink-compressed.
20322 (mmicromips): New option.
20323 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
20324 (compare_and_swap_12): Likewise.
20325 (sync_add<mode>): Likewise.
20326 (sync_<optab>_12): Likewise.
20327 (sync_old_<optab>_12): Likewise.
20328 (sync_new_<optab>_12): Likewise.
20329 (sync_nand_12): Likewise.
20330 (sync_old_nand_12): Likewise.
20331 (sync_new_nand_12): Likewise.
20332 (sync_sub<mode>): Likewise.
20333 (sync_old_add<mode>): Likewise.
20334 (sync_old_sub<mode>): Likewise.
20335 (sync_new_add<mode>): Likewise.
20336 (sync_new_sub<mode>): Likewise.
20337 (sync_<optab><mode>): Likewise.
20338 (sync_old_<optab><mode>): Likewise.
20339 (sync_new_<optab><mode>): Likewise.
20340 (sync_nand<mode>): Likewise.
20341 (sync_old_nand<mode>): Likewise.
20342 (sync_new_nand<mode>): Likewise.
20343 (sync_lock_test_and_set<mode>): Likewise.
20344 (test_and_set_12): Likewise.
20345 (atomic_compare_and_swap<mode>): Likewise.
20346 (atomic_exchange<mode>_llsc): Likewise.
20347 (atomic_fetch_add<mode>_llsc): Likewise.
20348 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
20349 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
20350 (umips_save_restore_pattern_p): Likewise.
20351 (umips_load_store_pair_p): Likewise.
20352 (umips_output_load_store_pair): Likewise.
20353 (umips_movep_target_p): Likewise.
20354 (umips_12bit_offset_address_p): Likewise.
20355 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
20356 (mips_base_mips16): Rename this...
20357 (mips_base_compression_flags): ...to this. Update all uses.
20358 (mips_attribute_table): Add micromips, nomicromips and nocompression.
20359 (mips_mips16_decl_p): Delete.
20360 (mips_nomips16_decl_p): Delete.
20361 (mips_get_compress_on_flags): New function.
20362 (mips_get_compress_off_flags): New function.
20363 (mips_get_compress_mode): New function.
20364 (mips_get_compress_on_name): New function.
20365 (mips_get_compress_off_name): New function.
20366 (mips_insert_attributes): Support multiple compression types.
20367 (mips_merge_decl_attributes): Likewise.
20368 (umips_12bit_offset_address_p): New function.
20369 (mips_start_function_definition): Emit .set micromips directive.
20370 (mips_call_may_need_jalx_p): New function.
20371 (mips_function_ok_for_sibcall): Add microMIPS support.
20372 (mips_print_operand_punctuation): Support short delay slots and
20374 (umips_swm_mask, umips_swm_encoding): New.
20375 (umips_build_save_restore): New function.
20376 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
20377 (was_mips16_p): Remove.
20378 (old_compression_mode): New.
20379 (mips_set_compression_mode): New function.
20380 (mips_set_current_function): Add microMIPS support.
20381 (mips_option_override): Likewise.
20382 (umips_save_restore_pattern_p): New function.
20383 (umips_output_save_restore): New function.
20384 (umips_load_store_pair_p_1): New function.
20385 (umips_load_store_pair_p): New function.
20386 (umips_output_load_store_pair_1): New function.
20387 (umips_output_load_store_pair): New function.
20388 (umips_movep_target_p) New function.
20389 (mips_prepare_pch_save): Add microMIPS support.
20390 * config/mips/mips.h (TARGET_COMPRESSION): New.
20391 (TARGET_CPU_CPP_BUILTINS): Update macro
20392 to use new compression flags and to support microMIPS.
20393 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
20394 (MIPS_ARCH_FLOAT_SPEC): Likewise.
20395 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
20396 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
20397 (ASM_SPEC): Support mmicromips and mno-micromips.
20398 (M16STORE_REG_P): New macro.
20399 (MIPS_CALL): Support TARGET_MICROMIPS.
20400 (MICROMIPS_J): New macro.
20401 (mips_base_mips16): Rename this...
20402 (mips_base_compression_flags): ...to this.
20403 (UMIPS_12BIT_OFFSET_P): New macro.
20404 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
20405 (MULTILIB_DIRNAMES): Likewise.
20407 2013-02-25 Tom de Vries <tom@codesourcery.com>
20409 PR rtl-optimization/56131
20410 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
20411 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
20412 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
20414 2013-02-25 Tobias Burnus <burnus@net-b.de>
20416 * doc/invoke.texi (-fsanitize=): Move from optimization
20417 to debugging options.
20419 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
20421 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
20423 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
20424 Alexander Monakov <amonakov@ispras.ru>
20426 PR middle-end/56077
20427 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
20428 flush pending lists also on non-jumps. Adjust comment.
20430 2013-02-24 Catherine Moore <clm@codesourcery.com>
20431 Maciej W. Rozycki <macro@codesourcery.com>
20432 Tom de Vries <tom@codesourcery.com>
20433 Nathan Sidwell <nathan@codesourcery.com>
20434 Iain Sandoe <iain@codesourcery.com>
20435 Nathan Froyd <froydnj@codesourcery.com>
20436 Chao-ying Fu <fu@mips.com>
20438 * doc/extend.texi: (micromips, nomicromips, nocompression):
20439 Document new function attributes.
20440 * doc/invoke.texi (minterlink-compressed, mmicromips,
20441 m14k, m14ke, m14kec): Document new options.
20442 (minterlink-mips16): Update documentation.
20443 * doc/md.texi (ZC, ZD): Document new constraints.
20444 * configure.ac (gcc_cv_as_micromips): Check if linker
20445 supports the .set micromips directive.
20446 * configure: Regenerate.
20447 * config.in: Regenerate.
20448 * config/mips/mips-tables.opt: Regenerate.
20449 * config/mips/micromips.md: New file.
20450 * constraints.md (ZC, AD): New constraints.
20451 * config/mips/predicates.md (movep_src_register): New predicate.
20452 (movep_src_operand): New predicate.
20453 (non_volatile_mem_operand): New predicate.
20454 * config/mips/mips.md (multimem): New type.
20455 (length): Differentiate between 17-bit and 18-bit branch offsets.
20456 (MOVEP1, MOVEP2): New mode iterator.
20457 (mov_<load>l): Use ZC constraint.
20458 (mov_<load>r): Likewise.
20459 (mov_<store>l): Likewise.
20460 (mov_<store>r): Likewise.
20461 (*branch_equality<mode>_inverted): Add microMIPS support.
20462 (*branch_equality<mode>): Likewise.
20463 (*jump_absolute): Likewise.
20464 (indirect_jump_<mode>): Likewise.
20465 (tablejump_<mode>): Likewise.
20466 (<optab>_internal): Likewise.
20467 (sibcall_internal): Likewise.
20468 (sibcall_value_internal): Likewise.
20469 (prefetch): Use constraint ZD.
20470 * config/mips/mips.opt (minterlink-compressed): New option.
20471 (minterlink-mips16): Now an alias for minterlink-compressed.
20472 (mmicromips): New option.
20473 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
20474 (compare_and_swap_12): Likewise.
20475 (sync_add<mode>): Likewise.
20476 (sync_<optab>_12): Likewise.
20477 (sync_old_<optab>_12): Likewise.
20478 (sync_new_<optab>_12): Likewise.
20479 (sync_nand_12): Likewise.
20480 (sync_old_nand_12): Likewise.
20481 (sync_new_nand_12): Likewise.
20482 (sync_sub<mode>): Likewise.
20483 (sync_old_add<mode>): Likewise.
20484 (sync_old_sub<mode>): Likewise.
20485 (sync_new_add<mode>): Likewise.
20486 (sync_new_sub<mode>): Likewise.
20487 (sync_<optab><mode>): Likewise.
20488 (sync_old_<optab><mode>): Likewise.
20489 (sync_new_<optab><mode>): Likewise.
20490 (sync_nand<mode>): Likewise.
20491 (sync_old_nand<mode>): Likewise.
20492 (sync_new_nand<mode>): Likewise.
20493 (sync_lock_test_and_set<mode>): Likewise.
20494 (test_and_set_12): Likewise.
20495 (atomic_compare_and_swap<mode>): Likewise.
20496 (atomic_exchange<mode>_llsc): Likewise.
20497 (atomic_fetch_add<mode>_llsc): Likewise.
20498 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
20499 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
20500 (umips_save_restore_pattern_p): Likewise.
20501 (umips_load_store_pair_p): Likewise.
20502 (umips_output_load_store_pair): Likewise.
20503 (umips_movep_target_p): Likewise.
20504 (umips_12bit_offset_address_p): Likewise.
20505 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
20506 (mips_base_mips16): Rename this...
20507 (mips_base_compression_flags): ...to this. Update all uses.
20508 (mips_attribute_table): Add micromips, nomicromips and nocompression.
20509 (mips_mips16_decl_p): Delete.
20510 (mips_nomips16_decl_p): Delete.
20511 (mips_get_compress_on_flags): New function.
20512 (mips_get_compress_off_flags): New function.
20513 (mips_get_compress_mode): New function.
20514 (mips_get_compress_on_name): New function.
20515 (mips_get_compress_off_name): New function.
20516 (mips_insert_attributes): Support multiple compression types.
20517 (mips_merge_decl_attributes): Likewise.
20518 (umips_12bit_offset_address_p): New function.
20519 (mips_start_function_definition): Emit .set micromips directive.
20520 (mips_call_may_need_jalx_p): New function.
20521 (mips_function_ok_for_sibcall): Add microMIPS support.
20522 (mips_print_operand_punctuation): Support short delay slots and
20524 (umips_swm_mask, umips_swm_encoding): New.
20525 (umips_build_save_restore): New function.
20526 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
20527 (was_mips16_p): Remove.
20528 (old_compression_mode): New.
20529 (mips_set_compression_mode): New function.
20530 (mips_set_current_function): Add microMIPS support.
20531 (mips_option_override): Likewise.
20532 (umips_save_restore_pattern_p): New function.
20533 (umips_output_save_restore): New function.
20534 (umips_load_store_pair_p_1): New function.
20535 (umips_load_store_pair_p): New function.
20536 (umips_output_load_store_pair_1): New function.
20537 (umips_output_load_store_pair): New function.
20538 (umips_movep_target_p) New function.
20539 (mips_prepare_pch_save): Add microMIPS support.
20540 * config/mips/mips.h (TARGET_COMPRESSION): New.
20541 (TARGET_CPU_CPP_BUILTINS): Update macro
20542 to use new compression flags and to support microMIPS.
20543 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
20544 (MIPS_ARCH_FLOAT_SPEC): Likewise.
20545 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
20546 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
20547 (ASM_SPEC): Support mmicromips and mno-micromips.
20548 (M16STORE_REG_P): New macro.
20549 (MIPS_CALL): Support TARGET_MICROMIPS.
20550 (MICROMIPS_J): New macro.
20551 (mips_base_mips16): Rename this...
20552 (mips_base_compression_flags): ...to this.
20553 (UMIPS_12BIT_OFFSET_P): New macro.
20554 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
20555 (MULTILIB_DIRNAMES): Likewise.
20557 2013-02-24 Jakub Jelinek <jakub@redhat.com>
20560 * target-globals.c (save_target_globals): For init_reg_sets and
20561 target_reinit remporarily set this_fn_optabs to this_target_optabs.
20563 2013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
20565 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
20566 * config/aarch64/t-aarch64
20567 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
20569 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
20571 PR inline-asm/56148
20572 * lra-constraints.c (process_alt_operands): Reload operand
20573 conflicting with earlier clobber only if no more other conflicting
20576 2013-02-22 Jakub Jelinek <jakub@redhat.com>
20579 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
20580 if not linking a shared library.
20582 2013-02-22 Seth LaForge <sethml@google.com>
20584 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
20586 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
20588 * config/arm/arm.md (split for extendsidi): Update condition.
20589 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
20590 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
20591 (qhs_zextenddi_cstr): Likewise.
20593 2013-02-21 Jakub Jelinek <jakub@redhat.com>
20595 PR middle-end/56420
20596 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
20597 avoid signed wrapping.
20598 (expand_mult): Handle properly multiplication by
20599 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
20600 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
20601 in the compiler if coeff is HOST_WIDE_INT_MIN.
20602 (expand_divmod): Don't make ext_op1 static, change it's type to
20603 uhwi. Avoid undefined behavior in -INTVAL (op1).
20605 PR rtl-optimization/50339
20606 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
20608 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
20609 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
20610 into splitting_ashiftrt field.
20611 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
20613 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
20616 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
20618 PR middle-end/56108
20619 * trans-mem.c (execute_tm_mark): Do not expand transactions that
20620 are sure to go irrevocable.
20622 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
20624 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
20625 scalars are valid operands.
20627 2013-02-21 Martin Jambor <mjambor@suse.cz>
20629 PR tree-optimization/56310
20630 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
20631 only matching indices and non-negative final offsets.
20632 (intersect_aggregates_with_edge): Pass src_idx to
20633 agg_replacements_to_vector. Pass src_idx insstead of index to
20634 intersect_with_agg_replacements.
20636 2013-02-21 Martin Jambor <mjambor@suse.cz>
20638 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
20639 instead of hard-wired defaults.
20641 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
20643 * doc/invoke.texi (MIPS Options): Update documentation of the
20644 floating-point multiply-accumulate instruction restrictions.
20646 2013-02-21 Kostya Serebryany <kcc@google.com>
20648 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
20649 asan_shadow_offset on x86_64 linux.
20651 2013-02-21 Richard Biener <rguenther@suse.de>
20653 PR tree-optimization/56415
20655 2013-02-11 Richard Biener <rguenther@suse.de>
20657 PR tree-optimization/56273
20658 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
20661 2013-02-21 Jakub Jelinek <jakub@redhat.com>
20664 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
20667 PR inline-asm/56405
20668 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
20669 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
20671 2013-02-20 Jan Hubicka <jh@suse.cz>
20673 PR tree-optimization/56265
20674 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
20675 when target is referenced for first time.
20677 2013-02-20 Richard Biener <rguenther@suse.de>
20679 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
20680 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
20681 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
20682 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
20683 not return anything.
20684 (rename_ssa_copies): Do not remove unused locals.
20685 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
20686 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
20687 * passes.c (execute_function_todo): Do not schedule unused locals
20688 removal if cleanup_tree_cfg did something.
20689 * tree-ssa-live.c (remove_unused_locals): Dump statistics
20690 about the number of removed locals.
20692 2013-02-20 Richard Biener <rguenther@suse.de>
20694 PR tree-optimization/56398
20695 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
20697 2013-02-20 Martin Jambor <mjambor@suse.cz>
20699 PR tree-optimization/55334
20700 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
20701 restricted pointers to arrays.
20703 2013-02-20 Richard Biener <rguenther@suse.de>
20704 Jakub Jelinek <jakub@redhat.com>
20706 PR tree-optimization/56396
20707 * tree-ssa-ccp.c (n_const_val): New static variable.
20708 (get_value): Return NULL for SSA names we don't have a lattice
20710 (ccp_initialize): Initialize n_const_val.
20711 * tree-ssa-copy.c (n_copy_of): New static variable.
20712 (init_copy_prop): Initialize n_copy_of.
20713 (get_value): Return NULL_TREE for SSA names we don't have a
20716 2013-02-20 Martin Jambor <mjambor@suse.cz>
20718 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
20720 2013-02-20 Richard Biener <rguenther@suse.de>
20722 * genpreds.c (write_lookup_constraint): Do not compare first
20723 letter of the constraint again.
20725 2013-02-20 Richard Biener <rguenther@suse.de>
20727 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
20729 (get_use_iv_cost): Terminate hashtable walk when coming across
20732 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
20734 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
20735 reassociation for avx2 targets.
20737 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20739 * config/microblaze/microblaze.c: microblaze_has_clz = 0
20740 Add version check for v8.10.a to enable microblaze_has_clz
20741 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
20742 version and TARGET_PATTERN_COMPARE check
20743 * config/microblaze/microblaze.md: New clzsi2 instruction
20745 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20747 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
20748 function before branching.
20750 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
20752 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
20754 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
20756 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
20758 PR middle-end/55889
20759 * sel-sched.c: Include ira.h.
20760 (implicit_clobber_conflict_p): New function.
20761 (moveup_expr): Use it.
20762 * Makefile.in (sel-sched.o): Depend on ira.h.
20764 2013-02-19 Richard Biener <rguenther@suse.de>
20766 PR tree-optimization/56384
20767 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
20768 (vn_hash_type): Split out from ...
20769 (vn_hash_constant_with_type): ... here.
20770 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
20771 (vn_phi_eq): Compare types from vn_phi_s structure.
20772 (vn_phi_lookup): Populate vn_phi_s type.
20773 (vn_phi_insert): Likewise.
20775 2013-02-19 Jakub Jelinek <jakub@redhat.com>
20777 PR tree-optimization/56350
20778 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
20779 if haven't found reduction or nested cycle operand, rather than
20780 asserting we must find it.
20782 PR tree-optimization/56381
20783 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
20786 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
20787 Jakub Jelinek <jakub@redhat.com>
20790 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
20791 (swap_optab_enable): Same.
20792 (init_all_optabs): Use argument instead of global.
20793 * tree.h (struct tree_optimization_option): New field target_optabs.
20794 * expr.h (init_all_optabs): Add argument to prototype.
20795 (TREE_OPTIMIZATION_OPTABS): New.
20796 (save_optabs_if_changed): Protoize.
20797 * optabs.h: Declare this_fn_optabs.
20798 * optabs.c (save_optabs_if_changed): New.
20799 Declare this_fn_optabs.
20800 (init_optabs): Add argument to init_all_optabs() call.
20801 * function.c (invoke_set_current_function_hook): Handle per
20803 * function.h (struct function): New field optabs.
20804 * config/mips/mips.c (mips_set_mips16_mode): Handle when
20805 optimization_current_node has changed.
20806 * target-globals.h (save_target_globals_default_opts): Protoize.
20807 * target-globals.c (save_target_globals_default_opts): New.
20809 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20812 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
20813 registers %fr12 and %fr12R as call used.
20816 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
20817 and HImode, require all displacements to be an integer multiple of
20819 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
20820 only allow QImode and HImode when reload is in progress and strict is
20821 true. Likewise for symbolic addresses. Use base14_operand to check
20822 displacements in REG+BASE addresses.
20824 2013-02-18 Richard Biener <rguenther@suse.de>
20826 PR tree-optimization/56366
20827 * tree-vect-loop.c (get_initial_def_for_induction): Properly
20828 handle sign-conversion of outer-loop initial induction value.
20830 2013-02-18 Richard Biener <rguenther@suse.de>
20832 PR middle-end/56349
20833 * cfghooks.c (merge_blocks): If we merge a latch into another
20834 block adjust references to it.
20835 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
20836 (verify_loop_structure): Verify that a recorded latch is in fact
20839 2013-02-18 Richard Biener <rguenther@suse.de>
20841 PR tree-optimization/56321
20842 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
20843 order SSA name release and virtual operand unlinking.
20845 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20847 * config/microblaze/microblaze.md (save_stack_block): Define.
20848 (restore_stack_block): Likewise.
20850 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20852 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
20853 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
20854 * config/microblaze/microblaze.c (microblaze_option_override):
20855 Bail out early for PIC modes when target does not support PIC.
20857 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20859 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
20860 Replace with a microblaze version.
20861 (microblaze_trampoline_init): Adapt for microblaze.
20862 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
20865 2013-02-16 Jakub Jelinek <jakub@redhat.com>
20866 Dodji Seketeli <dodji@redhat.com>
20869 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
20870 (instrument_mem_region_access): Do not forget to always put
20871 instrumentation of the of 'base' and 'base + len' in a "if (len !=
20872 0) statement, even for cases where either 'base' or 'base + len'
20873 are not instrumented -- because they have been previously
20874 instrumented. Simplify the logic by putting all the statements
20875 instrument 'base + len' inside a sequence, and then insert that
20876 sequence right before the current insertion point. Then, to
20877 instrument 'base + len', just get an iterator on that statement.
20878 And do not forget to update the pointer to iterator the function
20879 received as argument.
20881 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
20883 PR rtl-optimization/56348
20884 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
20886 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
20888 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
20889 (clean_graph_dump_file): Pass base to start_graph_dump.
20891 2013-02-14 Richard Henderson <rth@redhat.com>
20894 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
20896 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
20898 * collect2-aix.h: Define F_LOADONLY.
20900 2013-02-14 Richard Biener <rguenther@suse.de>
20903 * varasm.c (output_constant_def_1): Get the decl representing
20904 the constant as argument.
20905 (output_constant_def): Wrap output_constant_def_1.
20906 (make_decl_rtl): Use output_constant_def_1 with the decl
20907 representing the constant.
20908 (build_constant_desc): Optionally re-use a decl already
20909 representing the constant.
20910 (tree_output_constant_def): Adjust.
20912 2013-02-14 Dodji Seketeli <dodji@redhat.com>
20915 * asan.c (instrument_builtin_call): Really put the length of the
20916 second source argument into src1_len.
20918 2013-02-13 Jakub Jelinek <jakub@redhat.com>
20920 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
20921 argument. If it is false, don't create edge from then_bb to
20923 (insert_if_then_before_iter): Pass true to it.
20924 (build_check_stmt): Pass false to it.
20925 (transform_statements): Flush hash table only on extended basic
20926 block boundaries, rather than at the beginning of every bb.
20927 Don't flush hash table on nonfreeing_call_p calls.
20928 * tree-flow.h (nonfreeing_call_p): New prototype.
20929 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
20931 2013-02-13 David S. Miller <davem@davemloft.net>
20933 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
20935 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
20938 * ira.c (max_regno_before_ira): Move from ...
20940 (fix_reg_equiv_init): Use max_regno_before_ira instead of
20943 2013-02-13 Jakub Jelinek <jakub@redhat.com>
20945 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
20947 2013-02-13 Richard Biener <rguenther@suse.de>
20950 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
20951 globals in MEM_REFs.
20953 2013-02-13 Richard Biener <rguenther@suse.de>
20955 * loop-init.c (loop_optimizer_init): Clear loop state when
20956 re-initializing preserved loops.
20957 * loop-unswitch.c (unswitch_single_loop): Return whether
20958 we unswitched the loop. Do not verify loop state here.
20959 (unswitch_loops): When we unswitched a loop discover new loops.
20961 2013-02-13 Kostya Serebryany <kcc@google.com>
20963 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
20965 * sanitizer.def: Rename __asan_init to __asan_init_v1.
20967 2013-02-12 Dodji Seketeli <dodji@redhat.com>
20969 Avoid instrumenting duplicated memory access in the same basic block
20970 * Makefile.in (asan.o): Add new dependency on hash-table.h
20971 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
20972 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
20973 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
20974 (free_mem_ref_resources, has_mem_ref_been_instrumented)
20975 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
20976 (get_mem_ref_of_assignment): New functions.
20977 (get_mem_refs_of_builtin_call): Extract from
20978 instrument_builtin_call and tweak a little bit to make it fit with
20980 (instrument_builtin_call): Use the new
20981 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
20982 of is_gimple_builtin_call.
20983 (instrument_derefs, instrument_mem_region_access): Insert the
20984 instrumented memory reference into the hash table.
20985 (maybe_instrument_assignment): Renamed instrument_assignment into
20986 this, and change it to advance the iterator when instrumentation
20987 actually happened and return true in that case. This makes it
20988 homogeneous with maybe_instrument_assignment, and thus give a
20989 chance to callers to be more 'regular'.
20990 (transform_statements): Clear the memory reference hash table
20991 whenever we enter a new BB, when we cross a function call, or when
20992 we are done transforming statements. Use
20993 maybe_instrument_assignment instead of instrumentation. No more
20994 need to special case maybe_instrument_assignment and advance the
20995 iterator after calling it; it's now handled just like
20996 maybe_instrument_call. Update comment.
20998 2013-02-13 Richard Biener <rguenther@suse.de>
21000 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
21001 Fix loop discovery code.
21003 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
21005 PR inline-asm/56148
21006 * lra-constraints.c (process_alt_operands): Match early clobber
21007 operand with itself. Check conflicts with earlyclobber only if
21008 the operand is not reloaded. Prefer to reload conflicting operand
21009 if earlyclobber and matching operands are the same.
21011 2013-02-12 Richard Biener <rguenther@suse.de>
21014 * lto-streamer-out.c (write_symbol): Do not output symbols
21015 for hard register variables.
21017 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
21020 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
21021 (umulsidi3_insn, mulsidi3_insn): New insns.
21023 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
21025 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
21026 (struct tune_params): Add vec_costs field.
21027 * config/arm/arm.c (arm_builtin_vectorization_cost)
21028 (arm_add_stmt_cost): New functions.
21029 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
21030 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
21031 (arm_default_vec_cost): New struct of type cpu_vec_costs.
21032 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
21033 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
21034 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
21035 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
21037 2013-02-12 Richard Biener <rguenther@suse.de>
21040 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
21041 decls again if possible.
21043 2013-02-12 Richard Biener <rguenther@suse.de>
21045 PR middle-end/56288
21046 * tree-ssa.c (verify_ssa_name): Fix check, move
21047 SSA_NAME_IN_FREE_LIST check up.
21049 2013-02-12 Jakub Jelinek <jakub@redhat.com>
21050 Steven Bosscher <steven@gcc.gnu.org>
21052 PR rtl-optimization/56151
21053 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
21054 equal to op0 or op1, and last_insn pattern is CODE operation
21055 with MEM dest and one of the operands matches that MEM.
21057 2013-02-11 Sriraman Tallam <tmsriramgoogle.com>
21059 * doc/extend.texi: Document Function Multiversioning and "default"
21060 parameter string to target attribute.
21061 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
21062 target attribute parameter is "default".
21063 (ix86_compare_version_priority): Remove checks for target attribute.
21064 (ix86_mangle_function_version_assembler_name): Change error to sorry.
21065 Remove check for target attribute equal to NULL. Add assert.
21066 (ix86_generate_version_dispatcher_body): Change error to sorry.
21068 2013-02-11 Iain Sandoe <iain@codesourcery.com>
21069 Jack Howarth <howarth@bromo.med.uc.edu>
21070 Patrick Marlier <patrick.marlier@gmail.com>
21073 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
21074 define ENDFILE_SPEC as TM_DESTRUCTOR.
21075 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
21077 2013-02-11 Alexander Potapenko <glider@google.com>
21078 Jack Howarth <howarth@bromo.med.uc.edu>
21079 Jakub Jelinek <jakub@redhat.com>
21082 * config/darwin.c (cdtor_record): Rename ctor_record.
21083 (sort_cdtor_records): Rename sort_ctor_records.
21084 (finalize_dtors): New routine to sort destructors by
21085 priority before use in assemble_integer.
21086 (machopic_asm_out_destructor): Use finalize_dtors if needed.
21088 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
21090 PR rtl-optimization/56275
21091 * simplify-rtx.c (avoid_constant_pool_reference): Check that
21092 offset is non-negative and less than cmode size before
21093 calling simplify_subreg.
21095 2013-02-11 Richard Biener <rguenther@suse.de>
21097 PR tree-optimization/56264
21098 * cfgloop.h (fix_loop_structure): Adjust prototype.
21099 * loop-init.c (fix_loop_structure): Return the number of
21100 newly discovered loops.
21101 * tree-cfgcleanup.c (repair_loop_structures): When new loops
21102 are discovered, do a full loop-closed SSA rewrite.
21104 2013-02-11 Richard Biener <rguenther@suse.de>
21106 PR tree-optimization/56273
21107 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
21109 (check_array_ref): Fix missing newline in dumps.
21110 (search_for_addr_array): Likewise.
21112 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
21114 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
21116 2013-02-09 Jakub Jelinek <jakub@redhat.com>
21119 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
21121 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
21123 PR rtl-optimization/56246
21124 * lra-constraints.c (simplify_operand_subreg): Try to reuse
21126 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
21127 constraints are satisfied.
21129 2013-02-08 Jeff Law <law@redhat.com>
21132 * emit-rtl.c (reg_is_parm_p): New function.
21133 * regs.h (reg_is_parm_p): New prototype.
21134 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
21135 callee-clobbered registers.
21137 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
21140 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
21141 If there is no implicit builtin declaration, just return NULL.
21143 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
21145 * config/i386/sse.md (FMAMODEM): New mode iterator.
21146 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
21147 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
21149 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
21151 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
21152 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
21153 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
21155 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
21157 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
21158 (microblaze*-*-elf): Likewise.
21159 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
21161 * config/microblaze/microblaze-c.c: Add builtin defines for
21162 _LITTLE_ENDIAN and _BIG_ENDIAN.
21163 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
21164 add to TARGET_DEFAULT flags.
21165 Expand ASM_SPEC and LINK_SPEC.
21166 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
21167 * config/microblaze/microblaze.md: Update extendsidi2 and
21168 movdi_internal instructions to use low-order / high-order reg
21170 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
21171 options and inversemask / mask of LITTLE_ENDIAN.
21172 * config/microblaze/t-microblaze: Expand multilib options to
21173 include mlittle-endian (le) and update exceptions patterns.
21175 2013-02-08 Jakub Jelinek <jakub@redhat.com>
21177 PR rtl-optimization/56195
21178 * lra-constraints.c (get_reload_reg): Don't reuse regs
21179 if they have smaller mode than requested, if they have
21180 wider mode than requested, try to return a SUBREG.
21182 PR tree-optimization/56250
21183 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
21184 if type is unsigned and code isn't MULT_EXPR.
21186 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
21188 PR tree-optimization/56064
21189 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
21190 bits according to mode.
21191 * fixed-value.h (fixed_from_double_int)
21192 (const_fixed_from_double_int): Adjust comments.
21194 2013-02-08 Richard Biener <rguenther@suse.de>
21197 * lto-streamer.h (struct data_in): Remove current_file, current_line
21198 and current_col members.
21199 * lto-streamer-out.c (lto_output_location): Stream changed bits
21200 en-block for efficiency.
21201 * lto-streamer-in.c (clear_line_info): Remove.
21202 (lto_input_location): Cache current file, line and column
21203 globally via local statics. Read changed bits en-block.
21204 (input_function): Do not call clear_line_info.
21205 (lto_read_body): Likewise.
21206 (lto_input_toplevel_asms): Likewise.
21208 2013-02-08 Michael Matz <matz@suse.de>
21210 PR tree-optimization/52448
21211 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
21212 (nt_call_phase): New static.
21213 (add_or_mark_expr): Only mark accesses with newer phase than any
21215 (nonfreeing_call_p): New.
21216 (nt_init_block): Update nt_call_phase, mark blocks as visited.
21217 (nt_fini_block): Keep blocks marked as visited.
21218 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
21220 2013-02-08 Richard Biener <rguenther@suse.de>
21222 * ira.c (ira): Free broken dominator information.
21224 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
21226 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
21228 2013-02-08 Marek Polacek <polacek@redhat.com>
21230 * cfgloop.c (verify_loop_structure): Add more checking of headers.
21232 2013-02-08 Richard Biener <rguenther@suse.de>
21234 PR middle-end/56181
21235 * cfgloop.h (flow_loops_find): Adjust.
21236 (bb_loop_header_p): Declare.
21237 * cfgloop.c (bb_loop_header_p): New function split out from ...
21238 (flow_loops_find): ... here. Adjust function signature,
21239 support incremental loop structure update.
21240 (verify_loop_structure): Cleanup. Verify a loop is a loop.
21241 * cfgloopmanip.c (fix_loop_structure): Move ...
21242 * loop-init.c (fix_loop_structure): ... here.
21243 (apply_loop_flags): Split out from ...
21244 (loop_optimizer_init): ... here.
21245 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
21246 in incremental mode, only remove dead loops here.
21248 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
21251 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
21252 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
21253 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
21254 (*round<mode>3.libgcc): New insns for fixed-modes.
21255 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
21256 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
21257 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
21258 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
21259 implementations. Define to __builtin_avr_absFX,
21260 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
21261 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
21262 __builtin_avr_countlsFX, respectively.
21263 * config/avr/avr-c.c (target.h): Include it.
21264 (enum avr_builtin_id): New enum.
21265 (avr_resolve_overloaded_builtin): New static function.
21266 (avr_register_target_pragmas): Use it to set
21267 targetm.resolve_overloaded_builtin.
21268 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
21269 tree nodes used by DEF_BUILTIN.
21270 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
21271 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
21272 <AVR_BUILTIN_xxBITS>: Same.
21274 2013-02-08 Richard Biener <rguenther@suse.de>
21276 * cfgloop.c (verify_loop_structure): Properly handle
21277 a loop exiting to another loop header.
21278 * ira-int.h (ira_loops): Remove.
21279 * ira.c (ira_loops): Remove.
21280 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
21281 (do_reload): Use loop_optimizer_finalize.
21282 * ira-build.c (create_loop_tree_nodes): Use get_loops and
21283 number_of_loops to access the loop tree.
21284 (more_one_region_p): Likewise.
21285 (finish_loop_tree_nodes): Likewise.
21286 (rebuild_regno_allocno_maps): Likewise.
21287 (mark_loops_for_removal): Likewise.
21288 (mark_all_loops_for_removal): Likewise.
21289 (remove_unnecessary_regions): Likewise.
21290 (ira_build): Likewise.
21291 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
21293 2013-02-08 Richard Biener <rguenther@suse.de>
21295 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
21296 * ipa-pure-const.c (analyze_function): Avoid calling
21297 mark_irreducible_loops twice.
21298 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
21300 2013-02-07 David S. Miller <davem@davemloft.net>
21302 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
21304 * var-tracking.c (vt_add_function_parameter): Test the presence of
21305 HAVE_window_save properly and do not remap argument registers when
21306 we have a leaf function.
21308 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
21311 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
21313 * config/i386/i386.c (ix86_print_operand): Ditto.
21315 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
21317 * lra-constraints.c (process_alt_operands): Fix recently added comment.
21319 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
21321 PR rtl-optimization/56225
21322 * lra-constraints.c (process_alt_operands): Check that reload hard
21323 reg can hold value for strict_low_part.
21325 2013-02-07 Jakub Jelinek <jakub@redhat.com>
21328 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
21329 dwarf2out_end_function.
21330 (in_first_function_p, maybe_at_text_label_p,
21331 first_loclabel_num_not_at_text_label): New variables.
21332 (dwarf2out_var_location): In the first function find out
21333 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
21334 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
21337 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
21339 PR rtl-optimization/56178
21340 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
21341 SUBREG of a register. Tidy up related block of code.
21342 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
21343 note if the source is a register or a SUBREG of a register.
21345 2013-02-07 Jakub Jelinek <jakub@redhat.com>
21348 * config/rs6000/rs6000.md (ptrm): New mode attr.
21349 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
21350 call_value_indirect_aix<pttrsize>,
21351 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
21354 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
21356 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
21357 if -bnortl. Convert to strcmp and strncmp.
21359 2013-02-07 Alan Modra <amodra@gmail.com>
21362 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
21363 addresses won't wrap when offsetting.
21364 (rs6000_secondary_reload): Provide secondary reloads needed for
21365 wrapping LO_SUM addresses.
21367 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
21369 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
21370 MACH, just __MACH__.
21372 2013-02-06 Richard Biener <rguenther@suse.de>
21374 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
21375 instead of calling fix_loop_structure.
21377 2013-02-06 Jakub Jelinek <jakub@redhat.com>
21379 PR middle-end/56217
21380 * omp-low.c (use_pointer_for_field): Return false if
21381 lower_send_shared_vars doesn't generate any copy-out code.
21383 2013-02-06 Tom de Vries <tom@codesourcery.com>
21385 PR rtl-optimization/56131
21386 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
21387 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
21388 of the label is NULL. Add comment.
21390 2013-02-05 Jakub Jelinek <jakub@redhat.com>
21392 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
21395 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
21396 (STATIC_LIBTSAN_LIBS): Likewise.
21397 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
21398 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
21399 is defined, don't add anything else beyond that.
21400 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
21401 (LINK_COMMAND_SPEC): Use them.
21403 PR tree-optimization/56205
21404 * tree-stdarg.c (check_all_va_list_escapes): Return true if
21405 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
21406 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
21408 2013-02-05 Richard Biener <rguenther@suse.de>
21410 PR tree-optimization/53342
21411 PR tree-optimization/53185
21412 * tree-vectorizer.h (vect_check_strided_load): Remove.
21413 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
21414 not disallow peeling for vectorized strided loads.
21415 (vect_check_strided_load): Make static and simplify.
21416 (vect_analyze_data_refs): Adjust.
21417 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
21418 correctly when vectorizing strided loads.
21420 2013-02-05 Richard Biener <rguenther@suse.de>
21422 * doc/install.texi: Refer to ISL, not PPL.
21424 2013-02-05 Jan Hubicka <jh@suse.cz>
21426 PR tree-optimization/55789
21427 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
21429 2013-02-05 Jan Hubicka <jh@suse.cz>
21431 PR tree-optimization/55789
21432 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
21433 the dead call anyway.
21435 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
21438 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
21440 2013-02-04 Alexander Potapenko <glider@google.com>
21441 Jack Howarth <howarth@bromo.med.uc.edu>
21442 Jakub Jelinek <jakub@redhat.com>
21445 * config/darwin.c (sort_ctor_records): Stabilized qsort
21446 on constructor priority by using original position.
21447 (finalize_ctors): New routine to sort constructors by
21448 priority before use in assemble_integer.
21449 (machopic_asm_out_constructor): Use finalize_ctors if needed.
21451 2013-02-04 Jakub Jelinek <jakub@redhat.com>
21454 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
21455 about visibility on artificial decls.
21456 * config/sol2.c (solaris_assemble_visibility): Likewise.
21458 2013-02-04 Kai Tietz <ktietz@redhat.com>
21461 * config/i386/i386.c (function_value_ms_64): Add additional valtype
21462 argument and improve checking of return-argument types for 16-byte
21464 (ix86_function_value_1): Add additional valtype argument on call
21465 of function_value_64.
21466 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
21467 handling infunction_value_64 function.
21469 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
21471 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
21473 2013-02-04 Richard Biener <rguenther@suse.de>
21475 PR tree-optimization/56188
21476 * tree-ssa-structalias.c (label_visit): Consider case with
21477 initially non-empty points-to set.
21478 (perform_var_substitution): Dump node mapping and clean up.
21480 2013-02-04 Richard Guenther <rguenther@suse.de>
21483 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
21484 node prevail as last resort.
21485 (lto_symtab_merge_decls): Remove guard on LTRANS here.
21486 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
21488 2013-02-04 Richard Biener <rguenther@suse.de>
21490 PR tree-optimization/56113
21491 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
21493 (equiv_class_lookup_or_add): ... this.
21494 (label_visit): Adjust and fix error in previous patch.
21495 (perform_var_substitution): Adjust.
21497 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
21499 * config/sh/divtab.c: Fix formatting and comments throughout the file.
21500 * config/sh/sh4-300.md: Likewise.
21501 * config/sh/sh4a.md: Likewise.
21502 * config/sh/constraints.md: Likewise.
21503 * config/sh/sh.md: Likewise.
21504 * config/sh/netbsd-elf.h: Likewise.
21505 * config/sh/predicates.md: Likewise.
21506 * config/sh/sh-protos.h: Likewise.
21507 * config/sh/ushmedia.h: Likewise.
21508 * config/sh/linux.h: Likewise.
21509 * config/sh/sh.c: Likewise.
21510 * config/sh/superh.h: Likewise.
21511 * config/sh/elf.h: Likewise.
21512 * config/sh/sh4.md: Likewise.
21513 * config/sh/sh.h: Likewise.
21515 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21517 * config/pa/constraints.md: Adjust unused letters. Change "T"
21518 constraint to match_test floating_point_store_memory_operand().
21519 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
21520 (base14_operand): New.
21521 (floating_point_store_memory_operand): New.
21522 (integer_store_memory_operand): Revise to use base14_operand and
21523 reg_plus_base_memory_operand.
21524 (move_dest_operand): Allow symbolic_memory_operands.
21525 (symbolic_memory_operand): Check for LO_SOM.
21526 (symbolic_operand): Change default case to break.
21527 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
21528 CONST_DOUBLE values to be reloaded by putting them into memory when
21529 the destination is a floating point register.
21530 (movdf): Remove code to handle CONST_DOUBLE.
21532 (reload_indf_r1): New.
21533 (reload_insf_r1): New.
21534 Consistently use "Q" and "T" constraints with integer and floating
21535 point move instructions, respectively.
21536 (movdi): Remove FAIL.
21537 Change predicate for source operand unamed DImode move from
21538 general_operand to move_src_operand.
21539 (umulsidi3): Change predicate for destination operand to
21541 Likewise for similar unamed patterns.
21542 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
21543 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
21544 (hppa_legitimize_address): Simplify mask calculation.
21545 (pa_emit_move_sequence): Revised handling of secondary reloads from
21546 REG+D addresses for floating point loads and stores. Directly handle
21547 loading CONST0_RTX (mode) to a floating point register.
21548 (pa_secondary_reload): Handle reloading DF and SFmode constant values
21549 to floating point registers. Don't restrict secondary reloads to
21550 floating point registers to integer modes. Revise some comments and
21552 (TARGET_LEGITIMATE_ADDRESS_P): Define.
21553 (pa_legitimate_address_p): New.
21554 (pa_legitimize_reload_address): New.
21555 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
21556 (STRICT_REG_OK_FOR_BASE_P): New.
21557 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
21558 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
21560 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
21561 Andrew Dixie <andrewd@gentrack.com>
21563 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
21566 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
21568 * expmed.c (extract_bit_field_1): Pass the full width of the
21569 structure to get_best_reg_extraction_insn.
21571 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
21574 * configure.ac (use_cxa_atexit): Add AIX.
21575 * configure: Regenerate.
21577 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
21579 2013-02-01 Jakub Jelinek <jakub@redhat.com>
21582 * final.c (need_profile_function): New variable.
21583 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
21584 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
21585 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
21586 notes, targetm.asm_out.function_prologue doesn't emit anything,
21587 HAVE_prologue and profiler should be emitted before prologue,
21588 set need_profile_function instead of emitting it.
21589 (final_scan_insn): If need_profile_function, emit
21590 profile_function on the first NOTE_INSN_BASIC_BLOCK or
21591 NOTE_INSN_FUNCTION_BEG note.
21593 2013-02-01 Richard Henderson <rth@redhat.com>
21595 * config/rs6000/rs6000.md (smulditi3): New.
21598 * config/alpha/alpha.md (umulditi3): New.
21600 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
21602 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
21603 (ASM_OUTPUT_ALIGNED_LOCAL): New.
21605 2013-02-01 Richard Biener <rguenther@suse.de>
21607 PR tree-optimization/56113
21608 * tree-ssa-structalias.c (label_visit): Reduce work for
21609 single-predecessor nodes.
21611 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
21613 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
21614 range isn't testing for zero.
21616 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
21618 PR middle-end/56113
21619 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
21621 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
21622 Nick Clifton <nickc@redhat.com>
21624 * config/v850/constraints.md (Q): Define as a memory constraint.
21625 * config/v850/predicates.md (label_ref_operand): New predicate.
21626 (e3v5_shift_operand): New predicate.
21627 (ior_operator): New predicate.
21628 * config/v850/t-v850: Add e3v5 multilib.
21629 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
21630 (v850_gen_movdi): Prototype.
21631 * config/v850/v850.c: Add support for e3v5 architecture.
21632 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
21634 (construct_save_jarl): Add e3v5 long JARL support.
21635 (v850_adjust_insn_length): New function. Adjust length of call
21636 insns when using e3v5 instructions.
21637 (v850_gen_movdi): New function: Generate instructions to move a
21639 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
21640 (CPP_SPEC): Define __v850e3v5__ as appropriate.
21641 (TARGET_USE_FPU): Enable for e3v5.
21642 (CONST_OK_FOR_W): New macro.
21643 (ADJUST_INSN_LENGTH): Define.
21644 * config/v850/v850.md (UNSPEC_LOOP): Define.
21645 (attr cpu): Add v850e3v5.
21646 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
21647 (movdi): New pattern.
21648 (movdi_internal): New pattern.
21649 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
21650 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
21651 (cstoresf4): Likewise.
21652 (cstoredf4): Likewise.
21653 (insv): New pattern.
21654 (rotlso3_a): New pattern.
21655 (rotlsi3_b): New pattern
21656 (rotlsi3_v850e3v5): New pattern.
21657 (doloop_begin): New pattern.
21658 (fix_loop_counter): New pattern.
21659 (doloop_end): New pattern.
21660 (branch_normal): Add e3v5 long branch support.
21661 (branch_invert): Likewise.
21662 (branch_z_normal): Likewise.
21663 (branch_z_invert): Likewise.
21664 (branch_nz_normal): Likewise.
21665 (branch_nz_invert): Likewise.
21666 (call_internal_short): Add e3v5 register-indirect JARL support.
21667 (call_internal_long): Likewise.
21668 (call_value_internal_short): Likewise.
21669 (call_value_internal_long): Likewise.
21670 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
21671 (mloop): New option.
21672 * config.gcc: Add support for configuring v840e3v5 target.
21673 * doc/invoke.texi: Document new v850 specific command line options.
21675 2013-01-31 Paul Koning <ni1d@arrl.net>
21679 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
21680 children if parent is a class.
21681 (prune_unused_types_prune): Don't add DW_AT_declaration.
21683 2013-01-31 Richard Biener <rguenther@suse.de>
21685 PR tree-optimization/56157
21686 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
21687 match up operand with SLP child.
21689 2013-01-31 Jason Merrill <jason@redhat.com>
21692 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
21693 parameters the first time.
21694 (gen_scheduled_generic_parms_dies): Check completeness here.
21696 2013-01-31 Richard Biener <rguenther@suse.de>
21698 PR middle-end/53073
21699 * common.opt (faggressive-loop-optimizations): New flag,
21700 enabled by default.
21701 * doc/invoke.texi (faggressive-loop-optimizations): Document.
21702 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
21703 infer_loop_bounds_from_undefined by it.
21705 2013-01-31 Richard Biener <rguenther@suse.de>
21707 PR tree-optimization/56150
21708 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
21709 visit virtual operands.
21710 (find_uses_to_rename_bb): Likewise.
21712 2013-01-31 Richard Biener <rguenther@suse.de>
21714 PR tree-optimization/56150
21715 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
21716 mixed store non-store stmts.
21718 2013-01-30 Jakub Jelinek <jakub@redhat.com>
21721 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
21722 LIBASAN_EARLY_SPEC is defined.
21723 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
21724 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
21726 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
21729 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
21730 invalid args instead of ICEing on it.
21731 (ix86_valid_target_attribute_tree): Return error_mark_node if
21732 ix86_valid_target_attribute_inner_p failed.
21733 (ix86_valid_target_attribute_p): Return false only if
21734 ix86_valid_target_attribute_tree returned error_mark_node. Allow
21735 target("default") attribute.
21736 (sorted_attr_string): Change argument from const char * to tree,
21737 merge in all target attribute arguments rather than just one.
21738 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
21739 instead of free. Avoid using strcat.
21740 (ix86_mangle_function_version_assembler_name): Mangle
21741 target("default") as if no target attribute is present. Adjust
21742 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
21743 instead of xmalloc and XDELETEVEC instead of free.
21744 (ix86_function_versions): Don't return true if one of the decls
21745 doesn't have target attribute. If they don't and one of the decls
21746 is DECL_FUNCTION_VERSIONED, report an error. Adjust
21747 sorted_attr_string caller. Use XDELETEVEC instead of free.
21748 (ix86_supports_function_versions): Remove.
21749 (make_name): Fix up formatting.
21750 (make_dispatcher_decl): Remove resolver_name and its initialization.
21751 Avoid leaking memory.
21752 (is_function_default_version): Return true if there is
21753 target("default") attribute rather than no target attribute at all.
21754 (make_resolver_func): Avoid leaking memory.
21755 (ix86_generate_version_dispatcher_body): Likewise.
21756 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
21757 * target.def (supports_function_versions): Remove.
21758 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
21759 * doc/tm.texi: Regenerated.
21761 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
21763 PR rtl-optimization/56144
21764 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
21765 for values with side effects.
21767 2013-01-30 Richard Biener <rguenther@suse.de>
21769 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
21770 (sparseset_pop): Likewise.
21771 * cfganal.c (compute_idf): Likewise. Increase work-stack size
21772 to be able to use quick_push in the worker loop.
21774 2013-01-30 Marek Polacek <polacek@redhat.com>
21776 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
21778 2013-01-30 Richard Biener <rguenther@suse.de>
21781 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
21783 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
21785 PR tree-optimization/56064
21786 * fixed-value.c (fixed_from_double_int): New function.
21787 * fixed-value.h (fixed_from_double_int): New prototype.
21788 (const_fixed_from_double_int): New static inline function.
21789 * fold-const.c (native_interpret_fixed): New static function.
21790 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
21791 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
21792 (native_encode_fixed): New static function.
21793 (native_encode_expr) <FIXED_CST>: Use it.
21794 (native_interpret_int): Move double_int worker code to...
21795 * double-int.c (double_int::from_buffer): ...this new static method.
21796 * double-int.h (double_int::from_buffer): Prototype it.
21798 2013-01-30 Richard Biener <rguenther@suse.de>
21800 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
21801 New pointer-map and obstack.
21802 (init_alias_vars): Allocate pointer-map and obstack.
21803 (delete_points_to_sets): Free them.
21804 (find_what_var_points_to): Cache result.
21805 (find_what_p_points_to): Adjust for changed interface of
21806 find_what_var_points_to.
21807 (compute_points_to_sets): Likewise.
21808 (ipa_pta_execute): Likewise.
21810 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
21812 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
21813 * configure: Regenerate.
21814 * config.in: Regenerate.
21815 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
21816 #nobits/#progbits if supported.
21818 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
21821 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
21822 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
21823 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
21825 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
21827 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
21828 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
21830 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
21832 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
21833 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
21835 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
21837 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
21839 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
21840 * config/arm/cortex-a7.md: New bypasses using
21841 arm_mac_accumulator_is_result.
21843 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
21845 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
21846 (cortex_a7_neon_mla): Likewise.
21847 (cortex_a7_fpfmad): New reservation.
21848 (cortex_a7_fpmacs): Use ffmas and update required units.
21849 (cortex_a7_fpmuld): Update required units and latency.
21850 (cortex_a7_fpmacd): Likewise.
21851 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
21852 (cortex_a7_neon). Likewise.
21853 (bypass) Update participating units.
21855 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
21857 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
21858 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
21860 * config/arm/vfp11.md (vfp_farith): Use ffmas.
21861 (vfp_fmul): Use ffmad.
21862 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
21863 (cortex_r4_fmacd): Use ffmad.
21864 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
21865 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
21866 (cortex_a9_fmacd): Use ffmad.
21867 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
21868 (cortex_a8_vfp_macd): Use ffmad.
21869 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
21870 (cortex_a5_fpmacd): Use ffmad.
21871 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
21872 (cortex_a15_vfp_macd): Use ffmad.
21873 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
21875 2013-01-29 Jason Merrill <jason@redhat.com>
21878 * varasm.c (default_assemble_visibility): Don't warn about
21879 visibility on artificial decls.
21881 2013-01-29 Richard Biener <rguenther@suse.de>
21883 PR tree-optimization/56113
21884 * tree-ssa-structalias.c (equiv_class_lookup): Also return
21886 (label_visit): Free duplicate bitmaps and record the leader instead.
21887 (perform_var_substitution): Adjust.
21889 2013-01-29 Richard Biener <rguenther@suse.de>
21891 PR tree-optimization/55270
21892 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
21893 the CFG, schedule loops for fixup.
21895 2013-01-29 Nick Clifton <nickc@redhat.com>
21897 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
21900 2013-01-28 Leif Ekblad <leif@rdos.net>
21902 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
21903 * config/i386/i386.h (TARGET_RDOS): New macro.
21904 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
21905 * config/i386/i386.c (ix86_option_override_internal): For 64bit
21906 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
21907 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
21908 DEFAULT_LARGE_SECTION_THRESHOLD.
21909 * config/i386/i386.md (R14_REG, R15_REG): New constants.
21910 * config/i386/rdos.h: New file.
21911 * config/i386/rdos64.h: New file.
21913 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
21916 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
21919 2013-01-28 Jakub Jelinek <jakub@redhat.com>
21921 PR rtl-optimization/56117
21922 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
21923 call cselib_lookup_from_insn on the MEM before calling
21924 add_insn_mem_dependence.
21926 2013-01-28 Richard Biener <rguenther@suse.de>
21928 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
21929 to a stmt that didn't have one.
21930 (copy_phis_for_bb): Likewise for PHI arguments.
21931 (copy_debug_stmt): Likewise for debug stmts.
21933 2013-01-28 Richard Biener <rguenther@suse.de>
21935 PR tree-optimization/56034
21936 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
21937 (partition_builtin_p): Adjust.
21938 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
21939 it is the last partition.
21940 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
21941 up the vertex for the definition.
21942 (classify_partition): Classify whether a partition is a
21943 PKIND_REDUCTION, thus has uses outside of the loop.
21944 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
21945 Merge all PKIND_REDUCTION partitions into the last partition.
21946 (tree_loop_distribution): Seed partitions from reductions as well.
21948 2013-01-28 Jakub Jelinek <jakub@redhat.com>
21950 PR tree-optimization/56125
21951 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
21952 pow(x,c) into sqrt(x) * powi(x, n/2) or
21953 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
21954 optimizing for size.
21955 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
21956 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
21959 PR tree-optimization/56094
21960 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
21961 to UNKNOWN_LOCATION while gimplifying expr.
21963 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
21966 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
21967 operand 0 in movabs insn template for -masm=intel asm alternative.
21968 (*movabs<mode>_2): Ditto for operand 1.
21970 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
21973 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
21976 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
21978 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
21981 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
21983 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
21984 Undef to avoid warning.
21986 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
21988 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
21989 * configure: Regenerate.
21991 2013-01-25 Jakub Jelinek <jakub@redhat.com>
21993 PR tree-optimization/56098
21994 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
21995 for stmts with volatile ops.
21996 (cond_store_replacement): Don't optimize if assign has volatile ops.
21997 (cond_if_else_store_replacement_1): Don't optimize if either
21998 then_assign or else_assign have volatile ops.
21999 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
22002 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
22004 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
22006 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
22008 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
22009 missing ':' in asm example.
22011 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
22013 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
22014 entries into lane and laneq entries.
22015 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
22016 Remove AdvSIMD scalar modes.
22017 (aarch64_sq<r>dmulh_laneq<mode>): New.
22018 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
22020 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
22021 builtin implementations to relfect changes in RTL in aarch64-simd.md.
22022 * config/aarch64/iterators.md (VCOND): New.
22025 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
22028 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
22029 Add NULL LIBNAME argument to existing definitions.
22030 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
22031 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
22032 * config/avr/avr.c (DEF_BUILTIN): Same.
22033 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
22034 (avr_expand_builtin): Expand to a vanilla call if a libgcc
22035 implementation is available (DECL_ASSEMBLER_NAME is set).
22036 (avr_fold_absfx): New static function.
22037 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
22038 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
22039 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
22040 AVR_BUILTIN_ABSLLK.
22041 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
22042 (abshk, absk, abslk, absllk): Provide as static inline functions.
22044 2013-01-25 Marek Polacek <polacek@redhat.com>
22046 PR tree-optimization/56035
22047 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
22049 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
22051 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
22052 (*movtf_internal_rex64): Add (!o,C) alternative
22053 (*movxf_internal_rex64): Ditto.
22054 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
22056 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
22058 * doc/invoke.texi: fix typo.
22059 * doc/objc.texi: fix typo.
22061 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
22063 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
22064 for the first two alternatives.
22066 2013-01-24 Diego Novillo <dnovillo@google.com>
22068 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
22069 (ggc-zone.o): Remove.
22070 * configure.ac: Remove option --with-gc.
22071 * configure: Re-generate.
22072 * doc/install.texi: Remove documentation for --with-gc.
22073 * gengtype.c (write_enum_defn): Remove. Update all users.
22074 (write_Types_process_field): Remove generation of gt_e_* argument.
22075 (output_type_enum): Remove. Update all users.
22076 (write_enum_defn): Remove. Update all users.
22077 (enum alloc_zone): Remove. Update all users.
22078 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
22079 * ggc-common.c (ggc_splay_alloc): Remove first argument.
22080 Update all callers.
22081 (struct ptr_data): Remove field TYPE. Update all users.
22082 (gt_pch_note_object): Remove argument TYPE. Update all users.
22083 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
22085 * ggc-none.c (ggc_alloc_typed_stat): Remove.
22086 (struct alloc_zone): Remove.
22087 (ggc_internal_alloc_zone_stat): Remove.
22088 (ggc_internal_cleared_alloc_zone_stat): Remove.
22089 * ggc-page.c (ggc_alloc_typed_stat): Remove.
22090 (ggc_pch_count_object): Remove last argument. Update all users.
22091 (ggc_pch_alloc_object): Remove last argument. Update all users.
22092 (struct alloc_zone): Remove.
22093 * ggc-zone.c: Remove.
22094 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
22095 (struct alloc_zone): Remove.
22096 (ggc_alloc_typed_stat): Remove.
22097 (ggc_alloc_typed): Remove.
22098 (ggc_splay_alloc): Remove first argument.
22099 (rtl_zone): Remove. Update all users.
22100 (tree_zone): Remove. Update all users.
22101 (tree_id_zone): Remove. Update all users.
22102 (ggc_internal_zone_alloc_stat): Remove. Update all users.
22103 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
22104 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
22105 * tree-ssanames.c: Remove references to zone allocator in comments.
22107 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
22109 * config/avr/avr.c (avr_out_fract): Make register numbers that
22110 might be outside of source operand signed.
22112 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
22114 * config/i386/constraints.md (Yf): New constraint.
22115 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
22116 of f constraint to conditionaly disable x87 register preferences.
22117 (*movdf_internal): Ditto.
22118 (*movsf_internal): Ditto.
22120 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
22122 PR inline-asm/55934
22123 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
22124 that have operands with impossible constraints.
22125 Add a FIXME for a speed-up opportunity.
22126 * lra-constraints.c (process_alt_operands): Verify that a class
22127 selected from constraints on asms is valid for the operand mode.
22128 (curr_insn_transform): Remove incorrect comment.
22130 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
22132 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
22133 TOC operand is a valid symbol ref in the constant pool.
22135 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
22137 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
22139 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
22142 * config/avr/stdfix.h: New file.
22143 * t-avr (stdfix-gcc.h): New rule to build it.
22144 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
22146 2013-01-23 Kostya Serebryany <kcc@google.com>
22148 * config/darwin.h: remove dependency on
22149 CoreFoundation (asan on Mac OS).
22151 2013-01-23 Jakub Jelinek <jakub@redhat.com>
22154 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
22155 instead of cmpdi_operand for first comparison operand.
22156 Don't assert that comparison operands aren't both constants.
22158 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
22160 * doc/install.texi (Downloading the Source): Update references to
22161 downloading separate components.
22163 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
22165 * doc/extend.texi (__int128): Improve grammar.
22167 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
22170 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
22171 alternative to (o,r).
22172 (*movdi_internal_rex64): Remove (!o,n) alternative.
22173 (DImode immediate->memory splitter): Remove.
22174 (DImode immediate->memory peephole2): Remove.
22175 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
22176 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
22177 alternative to (!o,*r).
22178 (*movtf_internal_sse): New pattern.
22179 (*movxf_internal_rex64): New pattern.
22180 (*movxf_internal): Disable for TARGET_64BIT.
22181 (*movdf_internal_rex64): Remove (!o,F) alternative.
22183 2013-01-22 Jakub Jelinek <jakub@redhat.com>
22185 PR middle-end/56074
22186 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
22187 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
22188 * tree-vect-loop-manip.c (find_loop_location): Also ignore
22189 stmt locations where LOCATION_LOCUS of the stmt location is
22190 UNKNOWN_LOCATION or BUILTINS_LOCATION.
22193 * config/i386/i386.md (UNSPEC_STOS): New.
22194 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
22195 *strsetqi_1): Add UNSPEC_STOS.
22197 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
22200 * doc/invoke.texi: Remove left over -Wsynth example.
22202 2013-01-21 Jakub Jelinek <jakub@redhat.com>
22204 PR tree-optimization/56051
22205 * fold-const.c (fold_binary_loc): Don't fold
22206 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
22207 a narrowing conversion, or widening conversion from signed
22210 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
22212 PR rtl-optimization/56023
22213 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
22214 dependent on debug instruction.
22216 2013-01-21 Martin Jambor <mjambor@suse.cz>
22218 PR middle-end/56022
22219 * function.c (allocate_struct_function): Call
22220 invoke_set_current_function_hook earlier.
22222 2013-01-21 Jakub Jelinek <jakub@redhat.com>
22224 * reload1.c (init_reload): Only initialize reload_obstack
22225 during the first call.
22227 2013-01-21 Marek Polacek <polacek@redhat.com>
22229 * cfgloop.c (verify_loop_structure): Fix up grammar.
22231 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
22233 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
22234 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
22236 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
22239 * config/arm/marvell-pj4.md: Update copyright year.
22240 Fix up use of alu to alu_reg and simple_alu_imm.
22242 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
22244 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
22246 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
22249 * lra-constraints.c (curr_insn_transform): Don't reuse original
22250 insn for secondary memory move when memory mode should be different.
22252 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
22254 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
22255 atomic_storedi_1): New patterns.
22257 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
22259 btver2 pipeline descriptions.
22260 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
22262 * config/i386/i386.md (btver2_decode): New type attributes.
22263 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
22265 * config/i386/btver2.md: New file describing btver2 pipelines.
22267 2013-01-19 Andrew Pinski <apinski@cavium.com>
22269 PR tree-optimization/52631
22270 * tree-ssa-sccvn (visit_use): Before looking up the original
22271 statement, try looking up the simplified expression.
22273 2013-01-19 Anthony Green <green@moxielogic.com>
22275 * config/moxie/moxie.c (moxie_expand_prologue): Set
22276 current_function_static_stack_size.
22278 2013-01-18 Jakub Jelinek <jakub@redhat.com>
22280 PR tree-optimization/56029
22281 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
22282 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
22284 2013-01-18 Sharad Singhai <singhai@google.com>
22286 PR tree-optimization/55995
22287 * dumpfile.c (dump_loc): Print location only if available.
22288 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
22290 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
22293 * lra-constraints.c (curr_insn_transform): Reuse original insn for
22294 secondary memory move.
22295 (inherit_reload_reg): Use rclass instead of cl for
22296 check_secondary_memory_needed_p.
22298 2013-01-18 Jakub Jelinek <jakub@redhat.com>
22300 PR middle-end/56015
22301 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
22302 the case where writing real complex part of target modifies op1.
22304 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
22306 * config/aarch64/aarch64-simd.md
22307 (aarch64_vcond_internal<mode>): Handle unordered cases.
22308 * config/aarch64/iterators.md (v_cmp_result): New.
22310 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
22311 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
22313 * config/arm/marvell-pj4.md: New file.
22314 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
22315 * config/arm/arm.md (generic_sched): Add marvell_pj4.
22316 (generic_vfp): Likewise.
22317 * config/arm/arm-cores.def: Add marvell-pj4.
22318 * config/arm/arm-tune.md: Regenerate.
22319 * config/arm/arm-tables.opt: Regenerate.
22320 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
22321 * doc/invoke.texi: Document marvell-pj4.
22323 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
22325 * config/aarch64/arm_neon.h: Map scalar types to standard types.
22327 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
22332 * var-tracking.c (negative_power_of_two_p): New.
22333 (global_get_addr_cache, local_get_addr_cache): New.
22334 (get_addr_from_global_cache, get_addr_from_local_cache): New.
22335 (vt_canonicalize_addr): Rewrite using the above. Adjust the
22337 (vt_stack_offset_p): Remove.
22338 (vt_canon_true_dep): Always canonicalize loc's address.
22339 (clobber_overlapping_mems): Make sure we have a MEM.
22340 (local_get_addr_clear_given_value): New.
22341 (val_reset): Clear local cached entries.
22342 (compute_bb_dataflow): Create and release the local cache.
22343 Disable duplicate MEMs clobbering.
22344 (emit_notes_in_bb): Clobber MEMs likewise.
22345 (vt_emit_notes): Create and release the local cache.
22346 (vt_initialize, vt_finalize): Create and release the global
22347 cache, respectively.
22348 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
22350 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
22352 PR libmudflap/53359
22353 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
22354 not found in the symtab.
22356 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
22359 PR rtl-optimization/55547
22360 PR rtl-optimization/53827
22363 * alias.c (offset_overlap_p): New, factored out of...
22364 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
22365 the conservative special case for symbolic constants. Don't
22366 adjust zero sizes on alignment.
22368 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
22370 PR rtl-optimization/52573
22371 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
22372 REG_UNUSED for the same register.
22374 2013-01-17 Richard Biener <rguenther@suse.de>
22375 Marek Polacek <polacek@redhat.com>
22377 PR rtl-optimization/55833
22378 * loop-unswitch.c (unswitch_loops): Move loop verification...
22379 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
22380 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
22381 Set it to true when we're removing a loop from hierarchy tree in
22382 an irreducible region.
22383 (fix_bb_placements): Adjust caller.
22384 (fix_loop_placements): Likewise.
22386 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
22388 * config/avr/builtins.def (DEF_BUILTIN): Factor out
22389 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
22390 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
22391 Remove ID. Adjust comments.
22392 * config/avr/avr-c.c (avr_builtin_name): Remove.
22393 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
22394 * config/avr/avr.c (avr_tolower): New static function.
22395 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
22396 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
22397 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
22400 2013-01-17 Jan Hubicka <jh@suse.cz>
22402 PR tree-optimization/55273
22403 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
22405 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
22408 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
22409 store through atomic_store<mode>_1.
22410 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
22412 2013-01-17 Martin Jambor <mjambor@suse.cz>
22414 PR tree-optimizations/55264
22415 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
22416 for virtual methods.
22417 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
22418 virtual methods before inlining is over.
22419 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
22421 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
22424 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
22426 PR rtl-optimization/56005
22427 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
22428 pending reads for prefetch.
22430 2013-01-16 Ian Bolton <ian.bolton@arm.com>
22432 * config/aarch64/aarch64.md
22433 (*cstoresi_neg_uxtw): New pattern.
22434 (*cmovsi_insn_uxtw): New pattern.
22435 (*<optab>si3_uxtw): New pattern.
22436 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
22437 (*<optab>si3_insn_uxtw): New pattern.
22438 (*bswapsi2_uxtw): New pattern.
22440 2013-01-16 Richard Biener <rguenther@suse.de>
22442 * tree-inline.c (tree_function_versioning): Remove set but
22443 never used variable.
22445 2013-01-16 Richard Biener <rguenther@suse.de>
22447 PR tree-optimization/55964
22448 * tree-flow.h (rename_variables_in_loop): Remove.
22449 (rename_variables_in_bb): Likewise.
22450 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
22451 (copy_loop_before): Adjust and delete update-ssa status.
22452 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
22453 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
22454 (rename_variables_in_loop): Remove.
22455 (slpeel_update_phis_for_duplicate_loop): Likewise.
22456 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
22457 use available cfg machinery instead of duplicating it.
22458 Update PHI nodes and perform poor-mans SSA update here.
22459 (slpeel_tree_peel_loop_to_edge): Adjust.
22461 2013-01-16 Richard Biener <rguenther@suse.de>
22463 PR tree-optimization/54767
22464 PR tree-optimization/53465
22465 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
22466 (vrp_visit_phi_node): For PHI arguments coming via backedges
22467 drop all symbolical range information.
22468 (execute_vrp): Compute backedges.
22470 2013-01-16 Richard Biener <rguenther@suse.de>
22472 * doc/install.texi: Update CLooG and ISL requirements to
22475 2013-01-16 Christian Bruel <christian.bruel@st.com>
22478 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
22479 (broken_move): Handle UNSPECV_SP_SWITCH_B.
22480 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
22482 2013-01-16 DJ Delorie <dj@redhat.com>
22484 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
22485 (UNSPECV_SP_SWITCH_E): New.
22486 (sp_switch_1): Change to an unspec.
22487 (sp_switch_2): Change to an unspec. Don't use post-inc when we
22490 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
22492 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
22493 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
22494 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
22495 (expand_mem_thread_fence): Ditto.
22496 (expand_mem_signal_fence): Ditto.
22497 (expand_atomic_load): Ditto.
22498 (expand_atomic_store): Ditto.
22500 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
22502 PR rtl-optimization/55547
22503 PR rtl-optimization/53827
22506 * alias.c (memrefs_conflict_p): Set sizes to negative after
22509 2013-01-15 Jakub Jelinek <jakub@redhat.com>
22512 * function.c (thread_prologue_and_epilogue_insns): Always
22513 add crtl->drap_reg to set_up_by_prologue.set, even if
22514 stack_realign_drap is false.
22516 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
22518 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
22519 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
22520 *call): Fix indention.
22522 2013-01-15 Tom de Vries <tom@codesourcery.com>
22525 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
22528 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
22530 PR rtl-optimization/55153
22531 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
22533 2013-01-15 Martin Jambor <mjambor@suse.cz>
22535 PR tree-optimization/55920
22536 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
22537 accesses as grp_to_be_debug_replaced.
22539 2013-01-15 Jakub Jelinek <jakub@redhat.com>
22541 PR tree-optimization/55920
22542 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
22543 there is non-useless type conversion needed from debug rhs to lhs,
22544 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
22546 2013-01-15 Joseph Myers <joseph@codesourcery.com>
22547 Mikael Pettersson <mikpe@it.uu.se>
22550 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
22552 (ASM_OUTPUT_CASE_LABEL): Remove.
22553 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
22554 * final.c (shorten_branches): Update alignment of labels before
22555 jump tables if CASE_VECTOR_SHORTEN_MODE.
22557 2013-01-15 Richard Biener <rguenther@suse.de>
22560 * system.h: Do not include gmp.h for building host tools.
22562 2013-01-15 Richard Biener <rguenther@suse.de>
22564 PR middle-end/55882
22565 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
22566 account for bitpos when computing alignment.
22568 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
22570 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
22571 (ix86_target_macros_internal): Likewise.
22573 * config/i386/i386.c (m_CORE2I7): Removed.
22574 (m_CORE_HASWELL): New macro.
22575 (m_CORE_ALL): Likewise.
22576 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
22577 (initial_ix86_arch_features): Likewise.
22578 (processor_target_table): Initializations for Core avx2.
22579 (cpu_names): New names "core-avx2".
22580 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
22581 PROCESSOR_CORE_HASWELL.
22582 (ix86_issue_rate): New case.
22583 (ia32_multipass_dfa_lookahead): Likewise.
22584 (ix86_sched_init_global): Likewise.
22586 * config/i386/i386.h (TARGET_HASWELL): New macro.
22587 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
22588 (processor_type): New PROCESSOR_HASWELL.
22590 2013-01-15 Jakub Jelinek <jakub@redhat.com>
22592 PR tree-optimization/55955
22593 * tree-vect-loop.c (vectorizable_reduction): Give up early on
22594 *SHIFT_EXPR and *ROTATE_EXPR codes.
22596 PR tree-optimization/48766
22597 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
22598 -ftrapv disable -fwrapv.
22600 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
22603 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
22604 etc. to 1 and not to __flash.
22605 Use LL suffix for __INT24_MAX__ with -mint8.
22606 Use ULL suffix for __UINT24_MAX__ with -mint8.
22608 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
22610 * config/avr/avr-arch.h
22611 (struct base_arch_s): Use typedef avr_arch_t instead.
22612 (struct arch_info_s): Use typedef avr_arch_info_t instead.
22613 (struct mcu_type_s): Use typedef avr_mcu_t instead.
22614 * config/avr/avr.c: Same.
22615 * config/avr/avr-devices.c: Same.
22616 * config/avr/driver-avr.c: Same.
22617 * config/avr/gen-avr-mmcu-texi.c: Same.
22618 * config/avr/avr-mcus.def: Adjust comment.
22620 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
22622 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
22623 * config/aarch64/iterators.md (VALLDI): New.
22625 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
22626 Andi Kleen <ak@linux.intel.com>
22629 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
22630 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
22633 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
22635 * config/avr/avr-stdint.h: Remove trailing blanks.
22636 * config/avr/avr-log.h: Same.
22637 * config/avr/avr-arch.h: Same.
22638 * config/avr/avr-devices.c: Same.
22639 * config/avr/avr-dimode.md: Same.
22640 * config/avr/predicates.md: Same.
22641 * config/avr/avr-c.c: Same. And fix typo.
22643 * config/avr/avr-protos.h: Same. And:
22644 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
22645 (init_cumulative_args): Rename to avr_init_cumulative_args.
22646 (expand_prologue): Rename to avr_expand_prologue.
22647 (expand_epilogue): Rename to avr_expand_epilogue.
22648 (adjust_insn_length): Rename to avr_adjust_insn_length.
22649 (notice_update_cc): Rename to avr_notice_update_cc.
22650 (final_prescan_insn): Rename to avr_final_prescan_insn.
22651 * config/avr/avr.c: Same.
22652 * config/avr/avr.h: Same.
22653 * config/avr/avr.md: Remove trailing blanks.
22654 (prologue): Use avr_expand_prologue.
22655 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
22657 2013-01-14 Richard Biener <rguenther@suse.de>
22659 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
22660 verify_location, collect_subblocks): New functions.
22661 (verify_gimple_in_cfg): Verify that locations only reference
22662 BLOCKs in the functions BLOCK tree.
22664 2013-01-14 Richard Biener <rguenther@suse.de>
22666 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
22668 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
22670 (insert_out_of_ssa_copy_on_edge): Likewise.
22671 (rewrite_close_phi_out_of_ssa): Likewise.
22672 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
22674 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
22675 propagated constants.
22676 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
22679 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
22681 * config/avr/avr-modes.def: Add GPL copyright notice.
22683 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
22685 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
22686 MEMMODEL_MASK to determine memory model.
22687 (atomic_store<mode>): Ditto from operands[2].
22688 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
22690 2013-01-13 Jakub Jelinek <jakub@redhat.com>
22693 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
22694 (fold_gimple_assign): Don't call unshare_expr here.
22695 (fold_ctor_reference): Call unshare_expr.
22697 2013-01-13 Terry Guo <terry.guo@arm.com>
22699 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
22700 * doc/fragments.texi: Document MULTILIB_REUSE.
22701 * gcc.c (multilib_reuse): New internal spec.
22702 (set_multilib_dir): Also search multilib from multilib_reuse.
22703 * genmultilib (tmpmultilib3): Refactor code.
22704 (tmpmultilib4): Ditto.
22705 (multilib_reuse): New multilib argument.
22707 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
22709 * Makefile.in: Update copyright.
22711 2013-01-12 Tom de Vries <tom@codesourcery.com>
22713 PR middle-end/55890
22714 * calls.c (expand_call): Check if arg_nr is valid.
22716 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
22718 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
22719 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
22720 documentation. Add missing '__' in front of
22721 __builtin_ia32_packssdw256.
22723 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
22726 * config/s390/s390.c (s390_preferred_reload_class): Do not return
22727 NO_REGS for larl operands.
22728 (s390_reload_larl_operand): Use s390_load_address instead of
22731 2013-01-11 Richard Biener <rguenther@suse.de>
22733 * tree-cfg.c (verify_node_sharing_1): Split out from ...
22734 (verify_node_sharing): ... here.
22735 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
22737 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
22739 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
22740 Substitute TREECHECKING.
22741 * configure: Regenerate.
22742 * Makefile.in (TREECHECKING): New.
22744 2013-01-11 Richard Guenther <rguenther@suse.de>
22746 PR tree-optimization/44061
22747 * tree-vrp.c (extract_range_basic): Compute zero as
22748 value-range for __builtin_constant_p of function parameters.
22750 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
22752 Update copyright years.
22754 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
22756 PR rtl-optimization/55672
22757 * lra-eliminations.c (mark_not_eliminable): Permit addition with
22758 const to be eliminable.
22760 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
22762 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
22763 * configure: Regenerate.
22765 2013-01-10 Richard Biener <rguenther@suse.de>
22767 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
22769 2013-01-10 Richard Biener <rguenther@suse.de>
22772 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
22773 locations for virtual PHI arguments.
22774 (rewrite_update_phi_arguments): Likewise.
22776 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
22778 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
22781 2013-01-10 Jakub Jelinek <jakub@redhat.com>
22783 PR tree-optimization/55921
22784 * tree-complex.c (expand_complex_asm): New function.
22785 (expand_complex_operations_1): Call it for GIMPLE_ASM.
22787 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
22790 * config/s390/s390.c (s390_symref_operand_p)
22791 (s390_loadrelative_operand_p): Merge the two functions.
22792 (s390_check_qrst_address, print_operand_address): Add parameters
22793 to s390_loadrelative_operand_p invokation.
22794 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
22795 (s390_reload_larl_operand, s390_secondary_reload): Use
22796 s390_loadrelative_operand_p instead of s390_symref_operand_p.
22797 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
22799 2013-01-09 Mike Stump <mikestump@comcast.net>
22801 * dse.c (record_store): Remove unnecessary assert.
22803 2013-01-09 Jan Hubicka <jh@suse.cz>
22805 PR tree-optimization/55569
22806 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
22807 * cfgloop.h (scale_loop_profile): Likewise.
22809 2013-01-09 Jan Hubicka <jh@suse.cz>
22812 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
22814 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
22816 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
22818 PR middle-end/55114
22819 * expr.h (maybe_emit_group_store): Declare.
22820 * expr.c (maybe_emit_group_store): New function.
22821 * builtins.c (expand_builtin_int_roundingfn): Call it.
22822 (expand_builtin_int_roundingfn_2): Likewise.
22824 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
22826 PR rtl-optimization/55829
22827 * lra-constraints.c (match_reload): Add code for absent output.
22828 (curr_insn_transform): Add code for reloads of matched inputs
22831 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
22833 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
22834 attribute of movddup insn to DF.
22835 (*vec_interleave_lowv2df): Ditto.
22836 (vec_dupv2df): Ditto.
22838 2013-01-09 Jan Hubicka <jh@suse.cz>
22840 PR tree-optimiation/55875
22841 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
22842 EVERY_ITERATION parameter.
22843 (number_of_iterations_exit): Check if exit is executed every iteration.
22844 (idx_infer_loop_bounds): Similarly here.
22845 (n_of_executions_at_most): Simplify
22846 to only test for cases where statement is dominated by the
22847 particular bound; handle correctly the "postdominance" test.
22848 (scev_probably_wraps_p): Use max loop iterations info
22849 as a global bound first.
22851 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
22852 Nick Clifton <nickc@redhat.com>
22854 * config/v850/v850.md (cbranchsf4): New pattern.
22855 (cstoresf4): New pattern.
22856 (cbranchdf4): New pattern.
22857 (cstoredf4): New pattern.
22858 (movsicc): Disallow floating point comparisons.
22859 (cmpsf_le_insn): Fix order of operators.
22860 (cmpsf_lt_insn): Likewise.
22861 (cmpsf_eq_insn): Likewise.
22862 (cmpdf_le_insn): Likewise.
22863 (cmpdf_lt_insn): Likewise.
22864 (cmpdf_eq_insn): Likewise.
22865 (cmpsf_ge_insn): Use LE comparison.
22866 (cmpdf_ge_insn): Likewise.
22867 (cmpsf_gt_insn): Use LT comparison.
22868 (cmpdf_gt_insn): Likewise.
22869 (cmpsf_ne_insn): Delete pattern.
22870 (cmpdf_ne_insn): Delete pattern.
22871 * config/v850/v850.c (v850_gen_float_compare): Use
22872 gen_cmpdf_eq_insn for NE comparison.
22873 (v850_float_z_comparison_operator)
22874 (v850_float_nz_comparison_operator): Move from here ...
22875 * config/v850/predicates.md: ... to here. Move GT and GE
22876 comparisons into v850_float_z_comparison_operator.
22877 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
22879 (v850_float_nz_comparison_operator): Likewise.
22881 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
22883 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
22884 with calls to gen_insvsi/gen_insvdi.
22886 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
22888 * config/i386/i386.c (initial_ix86_tune_features): Set up
22889 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
22891 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
22892 Jakub Jelinek <jakub@redhat.com>
22894 PR tree-optimization/48189
22895 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
22896 If nitercst is 0, don't predict the exit edge.
22898 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22900 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
22901 in asm_fprintf with reg_names.
22902 (aarch64_print_operand_address): Likewise.
22903 (aarch64_return_addr): Likewise.
22904 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
22906 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
22908 * config/pa/pa.h (VAL_U6_BITS_P): Define.
22909 (INT_U6_BITS): Likewise.
22910 * config/pa/predicates.md (uint6_operand): New predicate.
22911 (shift5_operand, shift6_operand): Likewise.
22912 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
22914 (lshrdi3): Use shift6_operand.
22915 (shrpsi4, shrpdi4): New insn patterns.
22916 (extzv): Delete expander.
22917 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
22918 predicates in unamed zero extract patterns. Tighten common constraint.
22919 (extv): Delete expander.
22920 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
22921 predicates in unamed sign extract patterns. Tighten common constraint.
22922 (insv): Delete expander.
22923 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
22924 predicates in unamed insert patterns. Tighten common constraint.
22925 Change uint32_operand predicate to uint6_operand predicate in unamed
22926 DImode pattern to insert constant values of type 1...1xxxx.
22928 2013-01-04 Jan Hubicka <jh@suse.cz>
22930 PR tree-optimization/55823
22931 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
22934 2013-01-08 Jakub Jelinek <jakub@redhat.com>
22935 Uros Bizjak <ubizjak@gmail.com>
22937 PR rtl-optimization/55845
22938 * df-problems.c (can_move_insns_across): Stop scanning at
22939 volatile_insn_p source instruction or give up if
22940 across_from .. across_to range contains any volatile_insn_p
22943 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
22945 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
22946 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
22948 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
22949 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
22951 2013-01-08 Jakub Jelinek <jakub@redhat.com>
22954 * asan.c (asan_clear_shadow): New function.
22955 (asan_emit_stack_protection): Use it.
22957 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
22959 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
22960 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
22961 with tab instead of space.
22963 2013-01-08 Nick Clifton <nickc@redhat.com>
22965 * config/rl78/rl78.c (rl78_expand_prologue): Always select
22966 register bank 0 at the start of an interrupt handler.
22967 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
22970 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
22972 * config/aarch64/aarch64-simd.md
22973 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
22974 (aarch64_simd_bsl): Likewise.
22975 (aarch64_vcond_internal<mode>): Likewise.
22976 (vcond<mode><mode>): Likewise.
22977 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
22978 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
22980 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
22982 * config/aarch64/aarch64-builtins.c
22983 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
22985 2013-01-08 Martin Jambor <mjambor@suse.cz>
22988 * tree-sra.c (analyze_access_subtree): Return true also after
22989 potentially creating a debug-only replacement.
22991 2013-01-08 Jakub Jelinek <jakub@redhat.com>
22993 PR middle-end/55890
22994 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
22996 PR tree-optimization/54120
22997 * tree-vrp.c (range_fits_type_p): Don't allow
22998 src_precision < precision from signed vr to unsigned_p
22999 if vr->min or vr->max is negative.
23000 (simplify_float_conversion_using_ranges): Test can_float_p
23001 against CODE_FOR_nothing.
23003 2013-01-08 Jakub Jelinek <jakub@redhat.com>
23004 Richard Biener <rguenther@suse.de>
23006 PR middle-end/55851
23007 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
23008 types instead of just INTEGER_TYPE types.
23010 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
23012 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
23015 2013-01-07 Steve Ellcey <sellcey@mips.com>
23018 * config/mips/mips.opt: Change mad to mmad to match documentation.
23020 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
23023 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
23024 .progmemx.data now.
23026 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
23029 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
23030 (avr_addrspace_t): Add .section_name field.
23031 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
23033 (avr_addrspace): Same. Initialize .section_name. Remove last
23034 NULL entry. Put __memx into .progmemx.data.
23035 (progmem_section_prefix): Remove.
23036 (avr_asm_init_sections): No need to initialize progmem_section.
23037 (avr_asm_named_section): Use avr_addrspace[].section_name to get
23038 section name prefix.
23039 (avr_asm_select_section): Ditto. And use get_unnamed_section to
23040 retrieve the progmem section.
23041 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
23042 boundary to run over avr_addrspace[].
23043 (avr_register_target_pragmas): Ditto.
23045 2013-01-06 Jakub Jelinek <jakub@redhat.com>
23047 * varasm.c (output_constant_def_contents): For asan_protect_global
23048 protected strings, adjust DECL_ALIGN if needed, before testing for
23050 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
23051 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
23053 (output_object_block): For asan protected decls, emit asan padding
23054 after their contents.
23055 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
23056 (asan_finish_file): Test it here instead.
23058 2013-01-07 Nick Clifton <nickc@redhat.com>
23059 Matthias Klose <doko@debian.org>
23060 Doug Kwan <dougkwan@google.com>
23061 H.J. Lu <hongjiu.lu@intel.com>
23064 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
23066 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
23068 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
23070 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
23072 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
23074 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
23077 * doc/install.texi (Cross-Compiler-Specific Options): Document
23080 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
23082 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
23083 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
23084 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
23085 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
23086 vqmovun_high_s64): Fix source operand number and update copyright.
23088 2013-01-07 Richard Biener <rguenther@suse.de>
23090 PR middle-end/55890
23091 * gimple.h (gimple_call_builtin_p): New overload.
23092 * gimple.c (validate_call): New function.
23093 (gimple_call_builtin_p): Likewise.
23094 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
23095 Use gimple_call_builtin_p.
23096 (find_func_clobbers): Likewise.
23097 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
23098 (strlen_optimize_stmt): Likewise.
23100 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
23102 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
23103 (vld1q_dup_*): Likewise.
23104 (vld1_*): Likewise.
23105 (vld1q_*): Likewise.
23106 (vld1_lane_*): Likewise.
23107 (vld1q_lane_*): Likewise.
23109 2013-01-07 Richard Biener <rguenther@suse.de>
23111 * lto-streamer.h (LTO_minor_version): Bump to 2.
23113 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
23115 * config/aarch64/aarch64-protos.h
23116 (aarch64_const_double_zero_rtx_p): Rename to...
23117 (aarch64_float_const_zero_rtx_p): ...this.
23118 (aarch64_float_const_representable_p): New.
23119 (aarch64_output_simd_mov_immediate): Likewise.
23120 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
23121 move immediate case.
23122 * config/aarch64/aarch64.c
23123 (aarch64_const_double_zero_rtx_p): Rename to...
23124 (aarch64_float_const_zero_rtx_p): ...this.
23125 (aarch64_print_operand): Allow printing of new constants.
23126 (aarch64_valid_floating_const): New.
23127 (aarch64_legitimate_constant_p): Check for valid floating-point
23129 (aarch64_simd_valid_immediate): Likewise.
23130 (aarch64_vect_float_const_representable_p): New.
23131 (aarch64_float_const_representable_p): Likewise.
23132 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
23133 (aarch64_output_simd_mov_immediate): New.
23134 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
23135 (*movdf_aarch64): Likewise.
23136 * config/aarch64/constraints.md (Ufc): New.
23137 (Y): call aarch64_float_const_zero_rtx.
23138 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
23140 2013-01-07 Richard Biener <rguenther@suse.de>
23142 PR tree-optimization/55888
23143 PR tree-optimization/55862
23144 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
23145 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
23146 not if it is contained therein.
23148 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
23150 * config/avr/t-avr: Typo.
23152 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
23155 * config/avr/t-avr: Don't automatically rebuild
23156 $(srcdir)/config/avr/t-multilib
23157 $(srcdir)/config/avr/avr-tables.opt
23158 $(srcdir)/doc/avr-mmcu.texi
23159 (avr-mcus): New phony target to build them on request.
23160 (s-avr-mlib, s-avr-mmcu-texi): Remove.
23161 * avr/avr-mcus.def: Adjust comments.
23163 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
23165 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
23167 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
23169 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
23171 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
23173 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
23175 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
23177 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
23178 to generate profiling.
23179 * config/rs6000/aix64.h (LIB_SPEC): Same.
23181 2013-01-04 Andrew Pinski <apinski@cavium.com>
23183 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
23185 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
23187 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
23189 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
23191 (ix86_expand_move): Ditto.
23192 (ix86_zero_extend_to_Pmode): Ditto.
23193 (ix86_expand_call): Ditto.
23194 (ix86_expand_special_args_builtin): Ditto.
23195 (ix86_expand_builtin): Ditto.
23197 2013-01-04 Richard Biener <rguenther@suse.de>
23199 PR tree-optimization/55862
23200 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
23201 translating them through PHI nodes.
23203 2013-01-04 Martin Jambor <mjambor@suse.cz>
23205 PR tree-optimization/55755
23206 * tree-sra.c (sra_modify_assign): Do not check that an access has no
23207 children when trying to avoid producing a VIEW_CONVERT_EXPR.
23209 2013-01-04 Marek Polacek <polacek@redhat.com>
23211 PR middle-end/55859
23212 * opts.c (default_options_optimization): Clarify error message.
23214 2013-01-04 Richard Biener <rguenther@suse.de>
23216 PR middle-end/55863
23217 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
23220 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23223 * config/pa/pa.md (movsi): Revert previous change.
23224 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
23227 2013-01-03 Richard Henderson <rth@redhat.com>
23229 * config/i386/i386.c (ix86_expand_move): Always assign to op1
23230 after eliminating TLS symbols.
23232 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
23235 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
23236 * graphite-poly.c (debug_gmp_value): Likewise.
23238 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
23241 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
23242 selected code model, define __code_mode_small__, __code_model_medium__,
23243 __code_model_large__, __code_model_32__ or __code_model_kernel__.
23244 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
23245 xchg temporary register with %k. Declare temporary register as
23247 [__x86_64__]: For medium and large code models, preserve %rbx register.
23249 2013-01-03 Richard Biener <rguenther@suse.de>
23251 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
23252 (dump_subscript): Adjust.
23253 (finalize_ddr_dependent): Do not dump redundant info.
23254 (analyze_siv_subscript): Adjust.
23255 (subscript_dependence_tester): Likewise.
23256 (compute_affine_dependence): Likewise.
23258 2013-01-03 Richard Biener <rguenther@suse.de>
23261 2013-01-03 Richard Biener <rguenther@suse.de>
23263 PR tree-optimization/55857
23264 * tree-vect-stmts.c (vectorizable_load): Do not setup
23265 re-alignment for invariant loads.
23267 2013-01-02 Richard Biener <rguenther@suse.de>
23269 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
23270 invariant load do not generate a vector load from the scalar location.
23272 2013-01-03 Richard Biener <rguenther@suse.de>
23274 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
23275 for not vectorizing.
23276 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
23277 not build INDIRECT_REFs, call get_name once only.
23278 (vect_create_data_ref_ptr): Likewise. Dump base object kind
23279 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
23281 2013-01-03 Richard Biener <rguenther@suse.de>
23283 PR tree-optimization/55857
23284 * tree-vect-stmts.c (vectorizable_load): Do not setup
23285 re-alignment for invariant loads.
23287 2013-01-03 Richard Biener <rguenther@suse.de>
23290 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
23291 prefer a built-in decl.
23293 2013-01-03 Jakub Jelinek <jakub@redhat.com>
23295 * gcc.c (process_command): Update copyright notice dates.
23296 * gcov.c (print_version): Likewise.
23297 * gcov-dump.c (print_version): Likewise.
23299 PR rtl-optimization/55838
23300 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
23301 iv0.step, iv1.step and step.
23303 2013-01-03 Jakub Jelinek <jakub@redhat.com>
23304 Marc Glisse <marc.glisse@inria.fr>
23306 PR tree-optimization/55832
23307 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
23308 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
23309 integer_{one,zero}_node.
23311 2013-01-03 Jakub Jelinek <jakub@redhat.com>
23314 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
23315 * var-tracking.c (reverse_op): Don't add reverse ops to
23316 VALUEs that have already
23317 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
23319 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
23321 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
23323 2013-01-02 Teresa Johnson <tejohnson@google.com>
23325 * dumpfile.c (dump_loc): Print filename with location.
23326 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
23327 new location_t parameter to emit complete unroll message with
23328 new dump framework.
23329 (canonicalize_loop_induction_variables): Compute loops location
23330 and pass to try_unroll_loop_completely.
23331 * loop-unroll.c (report_unroll_peel): New function.
23332 (peel_loops_completely): Use new dump format with location
23333 for main dumpfile message, and invoke report_unroll_peel on success.
23334 (decide_unrolling_and_peeling): Ditto.
23335 (decide_peel_once_rolling): Remove old dumpfile message subsumed
23336 by report_unroll_peel.
23337 (decide_peel_completely): Ditto.
23338 (decide_unroll_constant_iterations): Ditto.
23339 (decide_unroll_runtime_iterations): Ditto.
23340 (decide_peel_simple): Ditto.
23341 (decide_unroll_stupid): Ditto.
23342 * cfgloop.c (get_loop_location): New function.
23343 * cfgloop.h (get_loop_location): Declare.
23345 2013-01-02 Sriraman Tallam <tmsriram@google.com>
23347 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
23350 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23352 PR middle-end/55198
23353 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
23354 BLKmode objects when EXPAND_MEMORY is specified.
23356 2013-01-02 Sriraman Tallam <tmsriram@google.com>
23358 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
23360 (fold_builtin_cpu): Do not share cpu model decls across statements.
23362 2013-01-02 Jason Merrill <jason@redhat.com>
23365 * tree.c (build_array_type_1): Revert earlier change.
23367 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
23369 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
23371 * config/aarch64/aarch64-tune.md: Re-generate.
23373 2013-01-02 Richard Biener <rguenther@suse.de>
23375 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
23376 invariant load do not generate a vector load from the scalar location.
23378 2013-01-02 Richard Biener <rguenther@suse.de>
23381 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
23382 * configure: Regenerate.
23384 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
23386 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
23387 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
23388 (expand_builtin_int_roundingfn_2): Keep the original target around
23389 for the fallback case.
23391 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
23393 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
23394 to be clear for sign changes.
23396 2013-01-01 Jan Hubicka <jh@suse.cz>
23398 * ipa-inline-analysis.c: Fix formatting.
23400 2013-01-01 Jakub Jelinek <jakub@redhat.com>
23402 PR tree-optimization/55831
23403 * tree-vect-loop.c (get_initial_def_for_induction): Use
23404 gsi_after_labels instead of gsi_start_bb.
23406 Copyright (C) 2013 Free Software Foundation, Inc.
23408 Copying and distribution of this file, with or without modification,
23409 are permitted in any medium without royalty provided the copyright
23410 notice and this notice are preserved.