1 2013-09-27 Diego Novillo <dnovillo@google.com>
3 * gimple.h (enum ssa_mode): Remove.
5 2013-09-27 Paulo Matos <pmatos@broadcom.com>
8 * gcc.dg/pr58463.c: New test.
10 2013-09-27 Paulo Matos <pmatos@broadcom.com>
12 * cfgloop.h (number_of_loops): Fix typo in check for null.
14 2013-09-27 Jakub Jelinek <jakub@redhat.com>
17 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
18 are children of outermost saved_cfun's loop, and set it up to
19 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
20 if loop != loop0 and SESE region contains bbs that belong to loop0.
22 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
24 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
25 (binary_scale_code_p, get_base_term, get_index_term): New functions.
26 (set_address_segment, set_address_base, set_address_index)
27 (set_address_disp): Accept the argument unconditionally.
28 (baseness): Remove must_be_base_p and must_be_index_p checks.
29 (decompose_normal_address): Classify as much as possible in the
32 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
34 * cse.c (count_reg_usage): Handle INT_LIST.
35 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
36 * reginfo.c (reg_scan_mark_refs): Likewise.
37 * reload1.c (eliminate_regs_1): Likewise.
39 2013-09-27 Iain Sandoe <iain@codesourcery.com>
42 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
45 2013-09-27 Richard Biener <rguenther@suse.de>
47 PR tree-optimization/58459
48 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
49 restriction not propagating into loops.
51 2013-09-26 Florian Weimer <fw@deneb.enyo.de>
53 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
54 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
55 * doc/tree-ssa.texi (Walking use-def chains): Delete.
57 2013-09-26 Richard Biener <rguenther@suse.de>
59 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names
62 2013-09-26 Richard Biener <rguenther@suse.de>
64 * alias.h (component_uses_parent_alias_set): Rename to ...
65 (component_uses_parent_alias_set_from): ... this.
66 * alias.c (component_uses_parent_alias_set): Rename to ...
67 (component_uses_parent_alias_set_from): ... this and return
69 (reference_alias_ptr_type_1): Use the result from
70 component_uses_parent_alias_set_from instead of stripping
71 components one at a time.
72 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
74 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
76 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs): Move
78 * tree-ssa-ter.h: New File. Move prototypes here.
79 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
80 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
81 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
82 from is_replaceable_p.
83 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
84 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
85 (process_replaceable): Use ter_is_replaceable_p.
86 (find_replaceable_in_bb): Use ter_is_replaceable_p.
87 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
88 newly refactored ssa_is_replaceable_p.
89 * cfgexpand.c: Include tree-outof-ssa.h.
90 * ssaexpand.h: Delete.
92 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
94 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
95 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
97 (create_gimple_tmp): Delete.
98 (get_expr_type, build_assign, build_type_cast): Move to...
99 * gimple-builder.c: New File.
100 (get_expr_type): Relocate from gimple.c.
101 (build_assign, build_type_cast): Change to only create ssanames.
102 * gimple.h: Move prototypes to...
103 * gimple-builder.h: New File. Here.
104 * tree-ssa.h: And here.
105 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
106 count_uses_and_derefs): Relocate from gimple.c.
107 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
108 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
109 * tree-ssa-math-opts (execute_cse_reciprocals): Use
110 gimple_replace_ssa_lhs.
111 * asan.c: Include gimple-builder.h.
112 * Makefile.in: Add gimple-builder.o.
114 2013-09-26 Richard Biener <rguenther@suse.de>
116 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
117 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
118 (loe_visit_block): Use gcc_checking_assert.
119 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
120 gimple_assign_ssa_name_copy_p.
121 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
124 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
126 * config/rs6000/t-rs6000 (rs6000.o): Remove.
127 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
129 2013-09-26 Richard Biener <rguenther@suse.de>
131 PR tree-optimization/58539
132 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
133 the fact that debug statements are not taking part in loop-closed
136 2013-09-26 Nick Clifton <nickc@redhat.com>
138 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
139 time warning message.
140 (msp430_print_operand_raw): Delete unused letter parameter.
141 (TARGET_PRINT_OPERAND_ADDRESS): Define.
142 (msp430_print_operand_address): New function.
143 (msp430_print_operand): Move address printing code from here to
145 * config/msp430/msp430.md (movsipsi2): Add comment in generated
147 (zero_extendpsisi2): Likewise.
148 (extendpsisi2): New pattern.
149 (andneghi3): New pattern.
151 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
153 * config/aarch64/aarch64.opt (mlra): New option.
154 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
155 (TARGET_LRA_P): Define.
157 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
159 * expr.c (expand_assignment): Remove obsolete comment.
161 2013-09-25 Jeff Law <law@redhat.com>
163 * tree-flow.h (thread_through_all_blocks): Prototype moved into
164 tree-ssa-threadupdate.h.
165 (register_jump_thread): Similarly.
166 * tree-ssa-threadupdate.h: New header file.
167 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
168 * tree-vrp.c: Likewise.
169 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
170 (thread_around_empty_blocks): Change type of path vector argument to
171 an edge,type pair from just an edge. Initialize both elements when
172 appending to a jump threading path. Tweak references to elements
174 (thread_across_edge): Similarly. Release memory for the elements
176 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
177 (dump_jump_thread_path): New function broken out from
178 register_jump_thread.
179 (register_jump_thread): Use dump_jump_thread_path. Change type of
180 path vector entries. Search the path for NULL edges and dump
181 the path if one is found. Tweak the conversion of path to 3-edge
182 form to use the block copy type information embedded in the path.
184 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
186 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
188 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
189 Vladimir Makarov <vmakarov@redhat.com>
191 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
192 from the least significant bit.
193 (strip_address_mutations): Add bitfield operations handling.
194 (must_be_index_p): Add shifting and rotate operations handling.
195 (set_address_base): Use must_be_base_p predicate.
196 (set_address_index): Use must_be_index_p predicate.
198 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
199 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
200 Sergey Lega <sergey.s.lega@intel.com>
201 Anna Tikhonova <anna.tikhonova@intel.com>
202 Ilya Tocar <ilya.tocar@intel.com>
203 Andrey Turetskiy <andrey.turetskiy@intel.com>
204 Ilya Verbin <ilya.verbin@intel.com>
205 Kirill Yukhin <kirill.yukhin@intel.com>
206 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
208 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
210 (ix86_expand_vector_move_misalign): Support new unaligned load and
211 stores and use new names.
212 (CODE_FOR_sse2_storedqu): Rename to ...
213 (CODE_FOR_sse2_storedquv16qi): ... this.
214 (CODE_FOR_sse2_loaddqu): Rename to ...
215 (CODE_FOR_sse2_loaddquv16qi): ... this.
216 (CODE_FOR_avx_loaddqu256): Rename to ...
217 (CODE_FOR_avx_loaddquv32qi): ... this.
218 (CODE_FOR_avx_storedqu256): Rename to ...
219 (CODE_FOR_avx_storedquv32qi): ... this.
220 * config/i386/i386.md (fpint_logic): New.
221 * config/i386/sse.md (VMOVE): Extend for AVX512.
225 (VI_UNALIGNED_LOADSTORE): Ditto.
226 (sse2_avx_avx512f): Ditto.
227 (sse2_avx2): Extend for AVX512.
228 (sse4_1_avx2): Ditto.
230 (sse): Extend for AVX512.
233 (avxsizesuffix): Ditto.
234 (sseintvecmode): Ditto.
236 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
237 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
238 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
239 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
240 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
241 (<sse2_avx_avx512f>_storedqu<mode): ... this.
242 (<sse>_movnt<mode>): Replace constraint "x" with "v".
243 (STORENT_MODE): Extend for AVX512.
244 (*absneg<mode>2): Replace constraint "x" with "v".
245 (*mul<mode>3): Ditto.
246 (*ieee_smin<mode>3): Ditto.
247 (*ieee_smax<mode>3): Ditto.
248 (avx_cmp<mode>3): Replace VF with VF_128_256.
249 (*<sse>_maskcmp<mode>3_comm): Ditto.
250 (<sse>_maskcmp<mode>3): Ditto.
251 (<sse>_andnot<mode>3): Extend for AVX512.
252 (<code><mode>3, anylogic): Replace VF with VF_128_256.
253 (<code><mode>3, fpint_logic): New.
254 (*<code><mode>3): Extend for AVX512.
255 (avx512flogicsuff): New.
256 (avx512f_<logic><mode>): Ditto.
257 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
259 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
260 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
261 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
262 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
263 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
264 (xop_vpermil2<mode>3): Ditto.
265 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
266 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
267 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
268 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
270 2013-09-25 Tom Tromey <tromey@redhat.com>
272 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
273 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
274 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
275 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
276 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
277 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
278 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
279 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
280 (GRAPHITE_HTAB_H): Remove.
282 2013-09-25 Tom Tromey <tromey@redhat.com>
284 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
286 2013-09-25 Tom Tromey <tromey@redhat.com>
288 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
290 2013-09-25 Tom Tromey <tromey@redhat.com>
292 * config/i386/t-i386 (i386.o): Remove.
293 (i386-c.o): Use COMPILE and POSTCOMPILE.
295 2013-09-25 Tom Tromey <tromey@redhat.com>
297 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
299 2013-09-25 Tom Tromey <tromey@redhat.com>
301 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
302 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
303 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
304 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
305 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
306 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
307 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
308 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
309 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
310 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
311 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
312 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
313 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
314 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
315 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
316 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
317 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
318 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
319 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
320 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
321 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
322 (tree-ssa-pre.o, tree-ssa-sccvn.o)
323 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
324 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
325 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
326 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
327 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
328 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
329 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
330 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
331 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
332 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
333 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
334 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
335 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
336 (graphite-clast-to-gimple.o, graphite-dependences.o)
337 (graphite-interchange.o, graphite-poly.o)
338 (graphite-scop-detection.o, graphite-sese-to-poly.o)
339 (graphite-optimize-isl.o, tree-vect-loop.o)
340 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
341 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
342 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
343 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
344 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
345 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
346 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
347 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
348 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
349 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
350 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
351 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
352 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
353 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
354 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
355 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
356 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
357 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
358 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
359 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
360 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
361 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
362 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
363 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
364 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
365 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
366 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
367 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
368 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
369 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
370 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
371 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
372 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
373 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
374 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
375 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
376 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
377 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
378 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
379 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
380 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
381 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
382 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
383 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
384 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
385 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
386 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
387 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
388 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
389 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
390 (gcov-dump.o): Remove.
391 (default-c.o): Use COMPILE and POSTCOMPILE.
392 (CFLAGS-gcc.o): New variable.
393 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
395 2013-09-25 Tom Tromey <tromey@redhat.com>
397 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
398 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
399 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
400 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
401 (c-family/c-pragma.o, c-family/c-pretty-print.o)
402 (c-family/c-semantics.o, c-family/c-ada-spec.o)
403 (c-family/array-notation-common.o, c-family/stub-objc.o)
404 (c-family/c-ubsan.o): Remove.
406 2013-09-25 Tom Tromey <tromey@redhat.com>
408 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
410 2013-09-25 Tom Tromey <tromey@redhat.com>
412 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
413 to add -DENABLE_SHARED_LIBGCC.
414 (gcc.o): Don't use subshell.
416 2013-09-25 Tom Tromey <tromey@redhat.com>
418 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
419 * configure.ac: Don't invoke AM_PROG_CC_C_O.
420 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
421 * configure, config.in: Rebuild.
423 2013-09-25 Tom Tromey <tromey@redhat.com>
425 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
426 (COMPILE, POSTCOMPILE): New variables.
427 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
428 (DEPFILES): New variable.
430 * configure.ac: Add checks for dependency checking.
431 * configure, aclocal.m4: Regenerate.
433 2013-09-25 Tom Tromey <tromey@redhat.com>
435 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
436 ($(ALL_HOST_OBJS)): Move order-only dependency to end
439 2013-09-25 Tom Tromey <tromey@redhat.com>
441 * Makefile.in (generated_files): Add options.h,
442 target-hooks-def.h, insn-opinit.h,
443 common/common-target-hooks-def.h, pass-instances.def,
444 c-family/c-target-hooks-def.h.
446 2013-09-25 Jeff Law <law@redhat.com>
448 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
449 than foo[foo.length () - 1] to access last member in a vec.
450 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
452 2013-09-25 Richard Biener <rguenther@suse.de>
455 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
457 2013-09-25 Jan Hubicka <jh@suse.cz>
459 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
462 2013-09-25 Marek Polacek <polacek@redhat.com>
465 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
466 when determining the type name.
468 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
470 * config/sh/sh.md: Fix formatting.
472 2013-09-24 Xinliang David Li <davidxl@google.com>
474 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
475 Check max peel iterations parameter.
476 * param.def: New parameter.
477 * doc/invoke.texi: Document New parameter.
479 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
481 * gimple-pretty-print.c: Various whitespace tweaks.
482 * tree-core.h: Likewise.
483 * tree-pretty-print.c: Likewise.
484 * tree-ssa-alias.c: Likewise.
485 * tree-ssa-copy.c: Likewise.
486 * tree-ssanames.c: Likewise.
487 * tree-ssanames.h: Likewise.
488 * tree-vrp.c: Likewise.
490 2013-09-24 Alan Modra <amodra@gmail.com>
494 * stmt.c (expand_asm_operands): Call expand_expr with
495 EXPAND_MEMORY for output operands that disallow regs. Don't
496 use EXPAND_WRITE on inout operands.
498 2013-09-24 Richard Biener <rguenther@suse.de>
501 * tree.c (reference_alias_ptr_type): Move ...
502 * alias.c (reference_alias_ptr_type): ... here and implement
503 in terms of the new reference_alias_ptr_type_1.
504 (ref_all_alias_ptr_type_p): New helper.
505 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
506 use ref_all_alias_ptr_type_p.
507 (get_deref_alias_set): Add flag_strict_aliasing check here.
508 (reference_alias_ptr_type_1): New function, split out from ...
509 (get_alias_set): ... here.
510 (alias_ptr_types_compatible_p): New function.
511 * alias.h (reference_alias_ptr_type): Declare.
512 (alias_ptr_types_compatible_p): Likewise.
513 * tree.h (reference_alias_ptr_type): Remove.
514 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
515 to compare MEM_REF alias types.
517 2013-09-24 Richard Biener <rguenther@suse.de>
519 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
521 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
523 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
524 reload helper function arrays into a single array reg_addr.
525 (reload_fpr_gpr): Likewise.
526 (reload_gpr_vsx): Likewise.
527 (reload_vsx_gpr): Likewise.
528 (struct rs6000_reg_addr): Likewise.
529 (reg_addr): Likewise.
530 (rs6000_debug_reg_global): Change rs6000_vector_reload,
531 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
532 (rs6000_init_hard_regno_mode_ok): Likewise.
533 (rs6000_secondary_reload_direct_move): Likewise.
534 (rs6000_secondary_reload): Likewise.
536 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
537 constraints: wu, ww, and wy. Repurpose wv constraint added during
538 power8 changes. Put wg constraint in alphabetical order.
540 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
541 for future work to add ISA 2.07 VSX single precision support.
542 (-mvsx-scalar-double): Change default from -1 to 1, update
543 documentation comment.
544 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
545 (-mupper-regs-df): New debug switch to control whether DF values
546 can go in the traditional Altivec registers.
547 (-mupper-regs-sf): New debug switch to control whether SF values
548 can go in the traditional Altivec registers.
550 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
552 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
553 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
554 Add new constraints, wu/ww/wy. Repurpose wv constraint.
555 (rs6000_debug_legitimate_address_p): Print if we are running
556 before, during, or after reload.
557 (rs6000_secondary_reload): Add a comment.
558 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
560 * config/rs6000/constraints.md (wa constraint): Sort w<x>
561 constraints. Update documentation string.
562 (wd constraint): Likewise.
563 (wf constraint): Likewise.
564 (wg constraint): Likewise.
565 (wn constraint): Likewise.
566 (ws constraint): Likewise.
567 (wt constraint): Likewise.
568 (wx constraint): Likewise.
569 (wz constraint): Likewise.
570 (wu constraint): New constraint for ISA 2.07 SFmode scalar
572 (ww constraint): Likewise.
573 (wy constraint): Likewise.
574 (wv constraint): Repurpose ISA 2.07 constraint that we not used in
575 the previous submissions.
576 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
578 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
580 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
581 (REG_BR_PROB): Say that the probability is stored in an int_list.
582 * reg-notes.def: Update commentary to mention INT_LIST.
583 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
585 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
586 * rtlanal.c (int_reg_note_p): New function.
587 (alloc_reg_note): Assert that the note does not have an int argument.
588 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
589 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
590 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
591 rather than an INSN_LIST. Handle INT_LIST.
592 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
593 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
594 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
595 Manipulate them as ints rather than rtxes.
596 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
597 * regmove.c (copy_src_to_dest): Likewise.
598 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
600 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
601 into the cases that need it.
602 * config/arm/arm.c (arm_unwind_emit): Likewise.
604 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
605 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
606 * loop-doloop.c (add_test, doloop_modify): Likewise.
607 * loop-unswitch.c (compare_and_jump_seq): Likewise.
608 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
609 * predict.c (combine_predictions_for_insn): Likewise.
610 * print-rtl.c (print_rtx): Handle INT_LIST.
611 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
612 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
613 * config/arm/arm.c (emit_unlikely_jump): Likewise.
614 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
615 (ix86_split_fp_branch, predict_jump): Likewise.
616 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
617 * config/sh/sh.c (expand_cbranchsi4): Likewise.
618 * config/spu/spu.c (ea_load_store_inline): Likewise.
620 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
621 value of a REG_BR_PROB note.
622 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
623 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
624 * emit-rtl.c (try_split): Likewise.
625 * predict.c (br_prob_note_reliable_p): Likewise.
626 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
627 * reorg.c (mostly_true_jump): Likewise.
628 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
629 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
630 * config/i386/i386.c (ix86_print_operand): Likewise.
631 * config/ia64/ia64.c (ia64_print_operand): Likewise.
632 * config/mmix/mmix.c (mmix_print_operand): Likewise.
633 * config/rs6000/rs6000.c (output_cbranch): Likewise.
634 * config/s390/s390.c (s390_expand_tbegin): Likewise.
635 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
636 * config/sparc/sparc.c (output_cbranch): Likewise.
637 * config/spu/spu.c (get_branch_target): Likewise.
638 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
639 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
641 2013-09-23 Jan Hubicka <jh@suse.cz>
643 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
645 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
646 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
648 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
649 (gimple_fold_call): Dump inconsistent devirtualizations; add
650 sanity check for type based devirtualizations.
651 * ipa-prop.c: Include ipa-utils.h
652 (ipa_intraprocedural_devirtualization): Add sanity check.
653 (try_make_edge_direct_virtual_call): Likewise.
655 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
657 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
658 assignment statements.
660 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
662 * gimple-pretty-print.c (dump_ssaname_info): New function.
663 (dump_gimple_phi): Call it.
664 (pp_gimple_stmt_1): Likewise.
665 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
666 (range_info_def): Declare.
667 * tree-pretty-print.c (pp_double_int): New function.
668 (dump_generic_node): Call it.
669 * tree-pretty-print.h (pp_double_int): Declare.
670 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
671 * tree-ssanames.h (range_info_def): New structure.
672 (value_range_type): Move definition here.
673 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
675 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
677 (set_range_info): New function.
678 (get_range_info): Likewise.
679 (duplicate_ssa_name_range_info): Likewise.
680 (duplicate_ssa_name_fn): Check pointer type and call
681 duplicate_ssa_name_range_info.
682 * tree-ssa-copy.c (fini_copy_prop): Likewise.
683 * tree-vrp.c (value_range_type): Remove definition, now in
685 (vrp_finalize): Call set_range_info to update value range of
687 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
688 (SSA_NAME_RANGE_INFO): New macro.
690 2013-09-23 Richard Biener <rguenther@suse.de>
692 PR tree-optimization/58464
693 * tree-ssa-pre.c (phi_trans_lookup): Remove.
694 (phi_trans_add): Change to add conditionally on being not
696 (phi_translate_1): Remove recursion detection here.
697 (phi_translate): Pre-seed the cache with NULL to catch
698 recursion here in a more generic way.
699 (bitmap_find_leader): Adjust comment.
700 (get_representative_for): Dump value-numbers.
701 (create_expression_by_pieces): Likewise.
702 (insert_into_preds_of_block): Likewise.
704 2013-09-23 Christian Bruel <christian.bruel@st.com>
707 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
708 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
710 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
713 2013-09-20 Renlin Li <renlin.li@arm.com>
715 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
716 (aarch64_expand_epilogue): Likewise.
717 (aarch64_legitimize_reload_address): Likewise.
719 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
721 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
722 NULL_TREE before pushing them onto the vector. Likewise for labels.
724 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
726 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
727 comparison operators when -fno-trapping-math is in effect.
728 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
729 comparison operators in TFmode and assert that unsupported operators
731 (ia64_print_operand): Likewise.
733 2013-09-21 Jan Hubicka <jh@suse.cz>
735 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
736 (sse_typeless_stores): Enable for core
737 (sse_load0_by_pxor): Likewise.
738 (four_jump_limit): Disable for core.
739 (pad_returns): Likewise.
740 (avoid_vector_decode): Likewise.
741 (fuse_cmp_and_branch): Enable for cores.
742 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
744 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
747 * config/pa/pa.c (pa_option_override): Disable auto increment and
748 decrement instructions until reload is completed.
750 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
751 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
752 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
754 2013-09-20 DJ Delorie <dj@redhat.com>
755 Nick Clifton <nickc@redhat.com>
757 * config/rl78/rl78.c: Various whitespace and comment tweaks.
758 (need_to_save): Save bank 0 on interrupts.
759 (characterize_address): Strip far address wrappers.
760 (rl78_as_legitimate_address): Likewise.
761 (transcode_memory_rtx): Likewise.
762 (rl78_peep_movhi_p): Disable this peephole after devirt.
763 (rl78_propogate_register_origins): Forget all origins when a
765 * config/rl78/rl78-virt.md: Various whitespace tweaks.
766 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
768 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
769 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
770 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
772 * config/rl78/constraints.md: Various whitespace and paren tweaks.
774 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
776 * config/pa/pa.md: In "scc" insn patterns, change output template to
777 handle const0_rtx in reg_or_0_operand operands.
779 2013-09-20 Martin Husemann <martin@NetBSD.org>
782 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
783 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
785 2013-09-20 Richard Biener <rguenther@suse.de>
788 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
789 remembering SSA name version and block index.
790 (new_scev_info_str): Adjust.
791 (hash_scev_info): Likewise. Also hash the block index.
792 (eq_scev_info): Adjust.
793 (find_var_scev_info): Likewise.
794 (struct instantiate_cache_entry): Remove.
795 (struct instantiate_cache_type): Use a htab to map name, block
797 (instantiate_cache_type::~instantiate_cache_type): Adjust.
798 (get_instantiated_value_entry): Likewise.
799 (hash_idx_scev_info, eq_idx_scev_info): New functions.
800 (instantiate_scev_name): Adjust.
802 2013-09-20 Jeff Law <law@redhat.com>
804 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
806 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
808 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
809 Call aarch64_simd_expand_args to update op[argc].
811 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
813 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
816 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
818 * gengtype.c (file_rules): Added rule for *.cc files.
819 (get_output_file_with_visibility): Give fatal message when no
822 2013-09-20 Renlin Li <renlin.li@arm.com>
824 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
825 (aarch64_expand_epilogue): Likewise.
826 (aarch64_legitimize_reload_address): Likewise.
828 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
831 * expr.c (expand_assignment): Remove misalignp code path.
833 2013-09-20 Marek Polacek <polacek@redhat.com>
836 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
837 TYPE_PRECISION. Add asserts.
839 2013-09-20 Richard Biener <rguenther@suse.de>
841 PR tree-optimization/58453
842 * tree-loop-distribution.c (distribute_loop): Apply the cost
843 model for -ftree-loop-distribute-patterns, too.
845 2013-09-20 Richard Biener <rguenther@suse.de>
848 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
849 make type comparison less strict.
851 2013-09-20 Alan Modra <amodra@gmail.com>
853 * configure: Regenerate.
854 * aclocal.m4: Regenerate.
856 2013-09-20 Marek Polacek <polacek@redhat.com>
859 * doc/extend.texi: Document that attribute used is meant to be used
860 on variables with static storage duration.
862 2013-09-19 Jakub Jelinek <jakub@redhat.com>
864 PR tree-optimization/58472
865 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
866 simd_lane_access set inv_p = false.
867 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
868 the simduid magic VAR_DECL.
870 2013-09-19 Jan Hubicka <jh@suse.cz>
872 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
874 2013-09-17 Jeff Law <law@redhat.com>
876 * tree-ssa-dom.c (record_temporary_equivalences): New function
877 split out of dom_opt_dom_walker::after_dom_children.
878 (dom_opt_dom_walker::thread_across_edge): Move common code
879 in here from dom_opt_dom_walker::after_dom_children.
880 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
882 2013-09-19 Jan Hubicka <jh@suse.cz>
884 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
885 (TARGET_GENERIC): Use PROCESOR_GENERIC
886 (enum processor_type): Unify generic32 and 64.
887 * i386.md (cpu): Likewise.
888 * x86-tune.def (use_leave): Enable for generic32.
889 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
890 * athlon.md: Change generic64 to generic in all occurences.
891 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
892 (ix86_target_macros_internal): Likewise.
893 * driver-i386.c (host_detect_local_cpu): Likewise.
894 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename to ..
895 (generic_memcpy, generic_memset, generic_cost): This one.
896 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
897 (m_GENERIC32, m_GENERIC64): Remove.
898 (m_GENERIC): Turn into one flag.
899 (processor_target): Unify generic tunnings.
900 (ix86_option_override_internal): Replace generic32/64 by generic.
901 (ix86_issue_rate): Likewise.
902 (ix86_adjust_cost): Likewise.
904 2013-09-19 Jan Hubicka <jh@suse.cz>
906 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
909 2013-09-19 Jakub Jelinek <jakub@redhat.com>
911 * omp-low.c (expand_omp_sections): Always pass len - 1 to
912 GOMP_sections_start, even if !exit_reachable.
914 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
916 * lra-constraints.c (need_for_all_save_p): Use macro
917 HARD_REGNO_CALL_PART_CLOBBERED.
918 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
919 set up pseudo conflict hard regs.
921 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
924 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
927 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
929 PR rtl-optimization/58438
930 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
931 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
932 from previous subpasses.
934 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
936 * arm.c (arm_get_frame_offsets): Validate architecture supports
937 LDRD/STRD before accepting the tuning preference.
938 (arm_expand_prologue): Likewise.
939 (arm_expand_epilogue): Likewise.
941 2013-09-18 Richard Biener <rguenther@suse.de>
943 PR tree-optimization/58417
944 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
945 have chrecs with symbols defined in the loop as operands.
946 (chrec_fold_multiply): Likewise.
947 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
948 parameters before folding binary operations.
949 (struct instantiate_cache_entry_hasher): Remove.
950 (struct instantiate_cache_type): Use a pointer-map.
951 (instantiate_cache_type::instantiate_cache_type): New function.
952 (instantiate_cache_type::get): Likewise.
953 (instantiate_cache_type::set): Likewise.
954 (instantiate_cache_type::~instantiate_cache_type): Adjust.
955 (get_instantiated_value_entry): Likewise.
956 (global_cache): New global.
957 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
958 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
959 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
960 (instantiate_scev_name): Adjust.
961 (instantiate_scev): Construct global instead of local cache.
962 (resolve_mixers): Likewise.
964 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
965 Paolo Carlini <paolo.carlini@oracle.com>
968 * doc/implement-cxx.texi: Fix references to the C++ standards.
970 2013-09-18 Jakub Jelinek <jakub@redhat.com>
972 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
973 * tree-vect-data-refs.c (vect_analyze_data_refs): For
974 simd_lane_access drs, update also DR_ALIGNED_TO.
976 2013-09-18 Marek Polacek <polacek@redhat.com>
979 * doc/extend.texi: Document no_sanitize_undefined attribute.
980 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
981 no_sanitize_undefined attribute.
983 2013-09-18 Nick Clifton <nickc@redhat.com>
985 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
986 (ASM_DECLARE_FUNCTION_NAME): Define.
988 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
990 * compare-elim.c (find_comparison_dom_walker): New class
991 (find_comparisons_in_bb): Rename to
992 find_comparison_dom_walker::before_dom_children
993 (find_comparisons): Adjust
994 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
996 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
997 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
998 (init_walk_dominator_tree): Remove declaration.
999 (fini_walk_dominator_tree): Remove declaration.
1000 * fwprop.c (single_def_use_dom_walker): New class
1001 (single_def_use_enter_block): Convert to
1002 single_def_use_dom_walker::before_dom_children.
1003 (single_def_use_leave_block): Convert to
1004 single_def_use_dom_walker::after_dom_children.
1005 (build_single_def_use_links): Adjust.
1006 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
1008 (find_candidates_in_block): Convert to
1009 find_candidates_dom_walker::before_dom_children.
1010 (execute_strength_reduction): Adjust.
1011 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
1012 (sese_dom_walker): New class.
1013 (sese_dom_walker::sese_dom_walker): New constructor.
1014 (sese_dom_walker::~sese_dom_walker): New destructor.
1015 (build_sese_conditions_before): Convert to
1016 sese_dom_walker::before_dom_children.
1017 (build_sese_conditions_after): Convert to
1018 sese_dom_walker::after_dom_children.
1019 (build_poly_scop): Adjust
1020 * tree-into-ssa.c (rewrite_dom_walker): New class
1021 (rewrite_enter_block): Convert to
1022 rewrite_dom_walker::before_dom_children.
1023 (rewrite_leave_block): Convert to
1024 rewrite_dom_walker::after_dom_children.
1025 (rewrite_update_dom_walker): New class.
1026 (rewrite_update_enter_block): Convert to
1027 rewrite_update_dom_walker::before_dom_children.
1028 (rewrite_update_leave_block): Convert to
1029 rewrite_update_dom_walker::after_dom_children.
1030 (rewrite_blocks, rewrite_into_ssa): Adjust.
1031 (mark_def_dom_walker): New class.
1032 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
1033 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
1034 (mark_def_sites_blocks): Convert to
1035 mark_def_dom_walker::before_dom_children.
1036 (mark_def_site_blocks): Remove.
1037 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
1038 (tree_ssa_dominator_optimize): Adjust.
1039 (dom_thread_across_edge): Convert to method
1040 dom_opt_dom_walker::thread_across_edge.
1041 (dom_opt_enter_block): Convert to member function
1042 dom_opt_dom_walker::before_dom_children.
1043 (dom_opt_leave_block): Convert to member function
1044 dom_opt_dom_walker::after_dom_children.
1045 * tree-ssa-dse.c (dse_dom_walker): New class.
1046 (dse_enter_block): Convert to member function
1047 dse_dom_walker::before_dom_children.
1048 (tree_ssa_dse): Adjust.
1049 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
1050 (determine_invariantness_stmt): Convert to method
1051 invariantness_dom_walker::before_dom_children.
1052 (determine_invariantness): Remove
1053 (move_computations_dom_walker): New class.
1054 (move_computations_stmt): Convert to method
1055 move_computations_dom_walker::before_dom_children.
1056 (move_computations, tree_ssa_lim): Adjust.
1057 * tree-ssa-phiopt.c (nontrapping_dom_walker): new class
1058 (nt_init_block): Make method
1059 notrappping_dom_walker::before_dom_children.
1060 (nt_fini_block): Make
1061 method nontrapping_dom_walker::after_dom_children.
1062 (get_non_trapping): Adjust.
1063 * tree-ssa-pre.c (eliminate_dom_walker): New class.
1064 (eliminate_bb): Make method eliminate_dom_walker::before_dom_children.
1065 (eliminate_leave_block): Make method.
1066 eliminate_dom_walker::after_dom_children.
1068 * tree-ssa-strlen.c (strlen_dom_walker): New class.
1069 (strlen_enter_block): Make method
1070 strlen_dom_walker::before_dom_children.
1071 (strlen_leave_block): Make
1072 method strlen_dom_walker::after_dom_children.
1073 (tree_ssa_strlen): Adjust.
1074 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
1075 (tree_ssa_uncprop): Adjust.
1076 (uncprop_leave_block): Make method
1077 uncprop_dom_walker::after_dom_children.
1078 (uncprop_leave_block): Make method
1079 uncprop_dom_walker::before_dom_children.
1081 2013-09-18 Bin Cheng <bin.cheng@arm.com>
1083 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn
1084 before branch in same basic block. Check both src and dest of
1087 2013-09-17 Nick Clifton <nickc@redhat.com>
1089 * config/rl78/rl78-real.md (bf): New pattern.
1091 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
1092 (rl78_print_operand): Do not put a # before a %B.
1093 * config/rl78/rl78.opt: Tweak doc strings.
1095 2013-09-17 DJ Delorie <dj@redhat.com>
1097 * config/rl78/constraints.md (Wcv): Allow up to $r31.
1098 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
1099 (rl78_option_override): Likewise, if -mallregs.
1100 (is_virtual_register): Likewise.
1101 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
1102 (REGNO_OK_FOR_BASE_P): Likewise.
1103 * config/rl78/rl78.opt (-mallregs): New.
1105 2013-09-17 Nick Clifton <nickc@redhat.com>
1107 * config/rl78/rl78.c (need_to_save): Change return type to bool.
1108 For interrupt functions: save all call clobbered registers if the
1109 interrupt handler is not a leaf function.
1110 (rl78_expand_prologue): Always recompute the frame information.
1111 For interrupt functions: only select bank 0 if one of the bank 0
1112 registers is going to be psuhed.
1114 2013-09-17 DJ Delorie <dj@redhat.com>
1116 * config/rl78/constraints.md: For each W* constraint, rename to C*
1117 and create a W* constraint that checks for an optional ES: prefix
1119 * config/rl78/rl78.md (UNS_ES_ADDR): New.
1120 (es_addr): New. Used to wrap far addresses.
1121 * config/rl78/rl78-protos.h (rl78_es_addr): New.
1122 (rl78_es_base): New.
1123 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
1124 wrapped far addresses.
1125 (rl78_print_operand_1): Unwrap far addresses before processing.
1126 (rl78_lo16): Wrap far addresses in unspecs.
1127 (rl78_es_addr): New.
1128 (rl78_es_base): New.
1129 (insn_ok_now): Check for not-yet-wrapped far addresses.
1130 (transcode_memory_rtx): Properly re-wrap far addresses.
1132 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
1134 * config/sparc/t-rtems: Add leon3 multilibs.
1136 2013-09-17 Cong Hou <congh@google.com>
1138 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
1139 when checking the dot production pattern. The type of rhs operand
1140 of multiply is now checked correctly.
1142 2013-09-17 Jeff Law <law@redhat.com>
1144 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
1145 edge implied equivalences into successor phis.
1146 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
1147 here from tree-ssa-threadedge.c.
1148 (mark_threaded_blocks): When threading through a joiner, if both
1149 successors of the joiner's clone reach the same block, verify the
1150 PHI arguments are equal. If not, cancel the jump threading request.
1151 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
1152 tree-ssa-threadupdate.c
1153 (thread_across_edge): Don't check PHI argument equality when
1154 threading through joiner block here.
1156 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
1158 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
1159 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
1160 (warn_uninit, warn_uninitialized_vars, execute_early_warn_uninitialized,
1161 make_pass_early_warn_uninitialized): Move to tree-ssa-uninit.c.
1162 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c
1163 (has_undefined_value_p): New. Pass dependant parts of
1164 ssa_undefined_value_p.
1165 (uninit_undefined_value_p): Use has_undefined_value_p.
1166 (warn_uninit, warn_uninitialized_vars, execute_early_warn_uninitialized,
1167 make_pass_early_warn_uninitialized): Move from tree-ssa.c
1168 * tree-ssa.h: Adjust prototypes
1170 2013-09-17 Jan Hubicka <jh@suse.cz>
1173 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
1174 * ipa-inline.c (can_inline_edge_p): Do not downgrade
1175 FUNCTION_NOT_OPTIMIZED.
1176 * ipa-inline-analysis.c (compute_inline_parameters): Function
1177 not optimized is not inlinable unless it is alwaysinline.
1178 (inline_analyze_function): Force calls in not optimized
1179 function not inlinable.
1181 2013-09-17 Jan Hubicka <jh@suse.cz>
1184 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
1186 * ipa.c (function_and_variable_visibility): Likewise.
1187 * ipa-profile.c (ipa_profile): Likewise.
1189 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
1192 * cgraph.c (cgraph_function_body_availability): Check for ifunc
1193 attribute, and don't inline the resolver in this case.
1195 2013-09-17 Teresa Johnson <tejohnson@google.com>
1197 * coverage.c (get_coverage_counts): Add missing newline.
1199 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1201 PR tree-optimization/58088
1202 * fold-const.c (mask_with_trailing_zeros): New function.
1203 (fold_binary_loc): Make sure we don't recurse infinitely
1204 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
1205 Use mask_with_trailing_zeros where appropriate.
1207 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
1209 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
1210 of 'prev' var to get better distance estimation.
1212 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
1214 * tree-inline.h (struct copy_body_data): Add transform_parameter.
1215 * tree-inline.c (is_parameter_of): New predicate.
1216 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
1217 a parameter has been remapped.
1218 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
1219 (optimize_inline_calls): Initialize transform_parameter.
1220 (copy_gimple_seq_and_replace_locals): Likewise.
1221 (tree_function_versioning): Likewise.
1222 (maybe_inline_call_in_expr): Likewise.
1224 2013-09-17 Nick Clifton <nickc@redhat.com>
1226 * config/msp430/msp430-protos.h: Add prototypes for new functions.
1227 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
1229 (is_attr_func): New function.
1230 (msp430_is_interrupt_func): New function.
1231 (is_naked_func): New function.
1232 (is_reentrant_func): New function.
1233 (is_critical_func): New function.
1234 (msp430_start_function): Add annotations for function attributes.
1235 (msp430_attr): New function.
1236 (msp430_attribute_table): New.
1237 (msp430_function_section): New function.
1238 (TARGET_ASM_FUNCTION_SECTION): Define.
1239 (msp430_builtin): New enum.
1240 (msp430_init_builtins): New function.
1241 (msp430_builtin_devl): New function.
1242 (msp430_expand_builtin): New function.
1243 (TARGET_INIT_BUILTINS): Define.
1244 (TARGET_EXPAND_BUILTINS): Define.
1245 (TARGET_BUILTIN_DECL): Define.
1246 (msp430_expand_prologue): Add support for naked, interrupt,
1247 critical and reentrant functions.
1248 (msp430_expand_epilogue): Likewise.
1249 (msp430_print_operand): Handle 'O' character.
1250 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
1252 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
1253 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
1254 (pushm): Use a 'n' rather than an 'i' constraint.
1255 (msp_return): Add generation of the interrupt return instruction.
1256 (disable_interrupts): New pattern.
1257 (enable_interrupts): New pattern.
1258 (push_intr_state): New pattern.
1259 (pop_intr_state): New pattern.
1260 (bic_SR): New pattern.
1261 (bis_SR): New pattern.
1262 * doc/extend.texi: Document MSP430 function attributes and builtin
1265 2013-09-17 Richard Biener <rguenther@suse.de>
1267 PR tree-optimization/58432
1268 * tree-loop-distribution.c (tree_loop_distribution): Also
1269 scan PHIs for outside loop uses and seed a partition from them.
1271 2013-09-17 Bin Cheng <bin.cheng@arm.com>
1273 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
1274 (restructure_reference): Call backtrace_base_for_ref.
1276 2013-09-17 Alan Modra <amodra@gmail.com>
1279 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11 patch.
1281 2013-09-16 DJ Delorie <dj@redhat.com>
1283 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
1284 vregs location for RL78/G10.
1285 (rl78_expand_prologue): Avoid SEL on G10.
1286 (rl78_expand_epilogue): Likewise.
1287 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
1288 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
1289 __RL78_G10__ when appropriate.
1290 (ASM_SPEC): Pass -mg10 along to the assembler.
1291 * config/rl78/rl78.md (sel_rb): Disable for G10.
1292 * config/rl78/rl78.opt: Add -mg10 option.
1293 * config/rl78/t-rl78: Add -mg10 multilib.
1295 2013-09-16 Xinliang David Li <davidxl@google.com>
1297 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
1298 * omp-low.c (omp_max_vf): Ditto.
1299 (expand_omp_simd): Ditto.
1300 * tree-vectorizer.c (vectorize_loops): Ditto.
1301 (gate_vect_slp): Ditto.
1302 (gate_increase_alignment): Ditto.
1303 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
1304 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
1305 (gate_tree_vectorize): Name change.
1306 (tree_vectorize): Ditto.
1307 (pass_vectorize::gate): Call new function.
1308 (pass_vectorize::execute): Ditto.
1309 opts.c: O3 default setting change.
1310 (finish_options): Check new flag.
1311 * doc/invoke.texi: Document new flags.
1312 * common.opt: New flags.
1314 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
1316 * doc/tm.texi.in (Cond Exec Macros): Remove node.
1317 (Condition Code): Don't reference it.
1318 * doc/tm.texi: Regenerate.
1320 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
1323 * lra-constraints.c (undo_optional_reloads): Consider all optional
1324 reload even if it did not get a hard reg.
1326 2013-09-16 Teresa Johnson <tejohnson@google.com>
1328 * dumpfile.c (dump_loc): Remove newline emission.
1329 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
1330 emission to dump_printf_loc calls where missing.
1331 (vect_mark_for_runtime_alias_test): Ditto.
1332 (vect_analyze_data_ref_dependence): Ditto.
1333 (vect_analyze_data_ref_dependences): Ditto.
1334 (vect_slp_analyze_data_ref_dependence): Ditto.
1335 (vect_slp_analyze_data_ref_dependences): Ditto.
1336 (vect_compute_data_ref_alignment): Ditto.
1337 (vect_update_misalignment_for_peel): Ditto.
1338 (vect_verify_datarefs_alignment): Ditto.
1339 (vector_alignment_reachable_p): Ditto.
1340 (vect_get_data_access_cost): Ditto.
1341 (vect_enhance_data_refs_alignment): Ditto.
1342 (vect_find_same_alignment_drs): Ditto.
1343 (vect_analyze_data_refs_alignment): Ditto.
1344 (vect_analyze_group_access): Ditto.
1345 (vect_analyze_data_ref_access): Ditto.
1346 (vect_analyze_data_ref_accesses): Ditto.
1347 (vect_prune_runtime_alias_test_list): Ditto.
1348 (vect_analyze_data_refs): Ditto.
1349 (vect_create_addr_base_for_vector_ref): Ditto.
1350 (vect_create_data_ref_ptr): Ditto.
1351 (vect_grouped_store_supported): Ditto.
1352 (vect_grouped_load_supported): Ditto.
1353 * value-prof.c (check_counter): Ditto.
1354 (check_ic_target): Ditto.
1355 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
1356 (vect_recog_widen_mult_pattern): Ditto.
1357 (vect_recog_widen_sum_pattern): Ditto.
1358 (vect_recog_over_widening_pattern): Ditto.
1359 (vect_recog_widen_shift_pattern): Ditto.
1360 (vect_recog_rotate_pattern): Ditto.
1361 (vect_recog_vector_vector_shift_pattern): Ditto.
1362 (vect_recog_divmod_pattern): Ditto.
1363 (vect_recog_mixed_size_cond_pattern): Ditto.
1364 (vect_recog_bool_pattern): Ditto.
1365 (vect_pattern_recog_1): Ditto.
1366 (vect_pattern_recog): Ditto.
1367 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
1368 (vect_is_simple_iv_evolution): Ditto.
1369 (vect_analyze_scalar_cycles_1): Ditto.
1370 (vect_get_loop_niters): Ditto.
1371 (vect_analyze_loop_1): Ditto.
1372 (vect_analyze_loop_form): Ditto.
1373 (vect_analyze_loop_operations): Ditto.
1374 (vect_analyze_loop_2): Ditto.
1375 (vect_analyze_loop): Ditto.
1376 (report_vect_op): Ditto.
1377 (vect_is_slp_reduction): Ditto.
1378 (vect_is_simple_reduction_1): Ditto.
1379 (vect_get_known_peeling_cost): Ditto.
1380 (vect_estimate_min_profitable_iters): Ditto.
1381 (vect_model_reduction_cost): Ditto.
1382 (vect_model_induction_cost): Ditto.
1383 (get_initial_def_for_induction): Ditto.
1384 (vect_create_epilog_for_reduction): Ditto.
1385 (vectorizable_reduction): Ditto.
1386 (vectorizable_induction): Ditto.
1387 (vectorizable_live_operation): Ditto.
1388 (vect_loop_kill_debug_uses): Ditto.
1389 (vect_transform_loop): Ditto.
1390 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
1391 (vect_stmt_relevant_p): Ditto.
1392 (process_use): Ditto.
1393 (vect_mark_stmts_to_be_vectorized): Ditto.
1394 (vect_model_simple_cost): Ditto.
1395 (vect_model_promotion_demotion_cost): Ditto.
1396 (vect_model_store_cost): Ditto.
1397 (vect_get_store_cost): Ditto.
1398 (vect_model_load_cost): Ditto.
1399 (vect_get_load_cost): Ditto.
1400 (vect_init_vector_1): Ditto.
1401 (vect_get_vec_def_for_operand): Ditto.
1402 (vect_finish_stmt_generation): Ditto.
1403 (vectorizable_call): Ditto.
1404 (vectorizable_conversion): Ditto.
1405 (vectorizable_assignment): Ditto.
1406 (vectorizable_shift): Ditto.
1407 (vectorizable_operation): Ditto.
1408 (vectorizable_store): Ditto.
1409 (vectorizable_load): Ditto.
1410 (vectorizable_condition): Ditto.
1411 (vect_analyze_stmt): Ditto.
1412 (vect_transform_stmt): Ditto.
1413 (vect_is_simple_use): Ditto.
1414 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
1415 (vect_can_advance_ivs_p): Ditto.
1416 (vect_update_ivs_after_vectorizer): Ditto.
1417 (vect_do_peeling_for_loop_bound): Ditto.
1418 (vect_gen_niters_for_prolog_loop): Ditto.
1419 (vect_update_inits_of_drs): Ditto.
1420 (vect_create_cond_for_alias_checks): Ditto.
1421 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
1422 (vect_build_slp_tree_1): Ditto.
1423 (vect_supported_load_permutation_p): Ditto.
1424 (vect_analyze_slp_instance): Ditto.
1425 (vect_analyze_slp): Ditto.
1426 (vect_make_slp_decision): Ditto.
1427 (vect_detect_hybrid_slp): Ditto.
1428 (vect_bb_vectorization_profitable_p): Ditto.
1429 (vect_slp_analyze_bb_1): Ditto.
1430 (vect_update_slp_costs_according_to_vf): Ditto.
1431 (vect_get_mask_element): Ditto.
1432 (vect_transform_slp_perm_load): Ditto.
1433 (vect_schedule_slp_instance): Ditto.
1434 (vect_schedule_slp): Ditto.
1435 (vect_slp_transform_bb): Ditto.
1436 * profile.c (read_profile_edge_counts): Ditto.
1437 (compute_branch_probabilities): Ditto.
1438 * coverage.c (get_coverage_counts): Ditto.
1440 2013-09-16 Diego Novillo <dnovillo@google.com>
1442 * tree-core.h: Add missing comment lines from refactoring
1445 2013-09-16 Jan Hubicka <jh@suse.cz>
1447 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
1448 abstract functions; for static functions check the presence
1451 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
1453 * config/aarch64/aarch64-simd-builtins.def (fma): New.
1454 * config/aarch64/aarch64-simd.md
1455 (aarch64_mla_elt<mode>): New.
1456 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
1457 (aarch64_mls_elt<mode>): Likewise.
1458 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
1459 (aarch64_fma4_elt<mode>): Likewise.
1460 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
1461 (aarch64_fma4_elt_to_128v2df): Likewise.
1462 (aarch64_fma4_elt_to_64df): Likewise.
1463 (fnma<mode>4): Likewise.
1464 (aarch64_fnma4_elt<mode>): Likewise.
1465 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
1466 (aarch64_fnma4_elt_to_128v2df): Likewise.
1467 (aarch64_fnma4_elt_to_64df): Likewise.
1468 * config/aarch64/iterators.md (VDQSF): New.
1469 * config/aarch64/arm_neon.h
1470 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
1471 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
1473 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
1475 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
1476 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
1477 (aarch64_mul3_elt_to_128df): Likewise.
1478 (aarch64_mul3_elt_to_64v2df): Likewise.
1479 * config/aarch64/iterators.md (VEL): Also handle DFmode.
1481 (VMUL_CHANGE_NLANES) Likewise.
1484 * config/aarch64/arm_neon.h
1485 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
1487 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
1489 * config/aarch64/arm_neon.h
1490 (vcvtx_high_f32_f64): Fix parameters.
1492 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1493 Uros Bizjak <ubizjak@gmail.com>
1495 * config/alpha.c: Include tree-ssanames.h.
1497 2013-09-16 Richard Biener <rguenther@suse.de>
1499 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
1500 (dot_rdg_1): Handle control_dd.
1501 (create_edge_for_control_dependence): New function.
1502 (create_rdg_edges): Add control dependences if asked for.
1503 (build_rdg): Likewise.
1504 (generate_loops_for_partition): If there are not necessary
1505 control stmts remove all their dependencies.
1506 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
1507 (distribute_loop): Pass on control dependences.
1508 (tree_loop_distribution): Compute control dependences and remove
1509 restriction on number of loop nodes.
1511 2013-09-16 Jakub Jelinek <jakub@redhat.com>
1513 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
1516 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
1518 * cse.c (try_const_anchors): Punt on CC modes.
1520 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1522 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
1524 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
1527 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
1530 2013-09-14 Iain Sandoe <iain@codesourcery.com>
1533 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
1535 (darwin_objc1_section): Likewise.
1536 (darwin_file_end): Emit Image Info section when required.
1538 2013-09-14 Jan Hubicka <jh@suse.cz>
1540 * tree-into-ssa.c (gate_into_ssa): New.
1541 (pass_data_build_ssa): Use it.
1542 * cgraph.h (expand_thunk): Update prototype.
1543 * cgraphunit.c (analyze_function): Expand thunks early.
1544 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
1545 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
1546 set lowered flag; do not add new function.
1547 (assemble_thunks_and_aliases): Update.
1548 * tree-ssa.c (gate_init_datastructures): New gate.
1549 (pass_data_init_datastructures): Use it.
1551 2013-09-14 Iain Sandoe <iain@codesourcery.com>
1554 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
1555 xmm register set described in the psABI.
1557 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
1559 * dwarf2out.c (should_emit_struct_debug): Add check
1560 for type_decl variable is not NULL.
1562 2013-09-13 Jacek Caban <jacek@codeweavers.com>
1564 * config.gcc: Use new winnt-c.c target hooks
1565 * config/t-winnt: New file
1566 * config/winnt-c.c: New file
1567 * doc/tm.texi.in: Document new hook
1568 * doc/tm.texi: Regenerated
1570 2013-09-13 Jan Hubicka <jh@suse.cz>
1573 * ipa-inline.c (check_callers): New function.
1574 (check_caller_edge): Remove.
1575 (want_inline_function_to_all_callers_p): Also permit alises that are
1577 (inline_to_all_callers): Terminate the walk when devirtualization
1578 introduce new calls.
1580 2013-09-13 Jan Hubicka <jh@suse.cz>
1582 * ipa-inline-analysis.c (struct growth_data): Add node.
1583 (do_estimate_growth_1): Fix detection of recursion.
1585 2013-09-13 Jakub Jelinek <jakub@redhat.com>
1587 PR tree-optimization/58392
1588 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
1589 to avoid shadowing of outer loop variable. If
1590 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
1591 replace_by_duplicate_decl simduid of loops that have it set and
1592 set dest_cfun->has_simduid_loops and/or
1593 dest_cfun->has_force_vect_loops.
1594 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
1595 instead of maybe_lookup_decl.
1596 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
1597 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
1598 call. Copy over force_vect and copy and remap simduid. Set
1599 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
1600 (copy_cfg_body): Remove blocks_to_copy argument. Use
1601 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
1602 caller. Don't set cfun->has_simduid_loops and/or
1603 cfun->has_force_vect_loops here.
1604 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
1606 (expand_call_inline, tree_function_versioning): Adjust copy_body
1609 2013-09-13 Martin Jambor <mjambor@suse.cz>
1612 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
1613 the assert if the edge was a speculative one.
1615 2013-09-13 Richard Biener <rguenther@suse.de>
1617 * tree-data-ref.h (known_dependences_p): Move here ...
1618 * tree-loop-distribution.c (known_dependences_p): ... from here.
1619 (dump_rdg_component, debug_rdg_component): Remove.
1621 (generate_loops_for_partition): Use gimple_uid instead of
1622 relying on matching stmt visit order.
1623 (rdg_build_partitions): Take starting stmt vector.
1624 (ldist_gen): Merge into ...
1625 (distribute_loop): ... this function. Do not compute starting
1627 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
1629 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1631 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
1632 Set type attribute correctly. Set predicable_short_it attribute.
1633 (cmpsi_shiftsi): Remove %? from output template.
1635 2013-09-13 Richard Biener <rguenther@suse.de>
1637 * tree-loop-distribution.c (struct rdg_component,
1638 rdg_defs_used_in_other_loops_p, free_rdg_components,
1639 rdg_build_components): Remove.
1640 (stmts_from_loop): Do not record virtual PHIs.
1641 (generate_loops_for_partition): Skip virtual PHIs.
1642 (build_rdg_partition_for_component): Rename to ...
1643 (build_rdg_partition_for_vertex): ... this and adjust.
1644 (rdg_build_partitions): Take a vector of starting vertices
1645 instead of components. Remove unnecessary leftover handling.
1646 (ldist_gen): Do not build components or record other stores.
1647 (distribute_loop): Do not distribute loops containing stmts
1650 2013-09-13 Christian Bruel <christian.bruel@st.com>
1653 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
1655 2013-09-13 Kai Tietz <ktietz@redhat.com>
1657 * config.gcc: Separate cases for mingw and cygwin targets,
1658 and add 64-bit cygwin target case.
1660 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
1661 dll-export inline-functions.
1662 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
1665 2013-09-13 Jeff Law <law@redhat.com>
1669 2013-09-06 Jeff Law <law@redhat.com>
1671 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
1672 edge implied equivalences into successor phis.
1674 2013-09-12 DJ Delorie <dj@redhat.com>
1676 * config/rl78/rl78-virt.md: Change from | to \; for asm line
1679 2013-09-12 Brooks Moses <bmoses@google.com>
1682 * Makefile.in: Do not install driver binaries in $(target)/bin.
1684 2013-09-12 DJ Delorie <dj@redhat.com>
1686 * config/rl78/rl78.opt (mrelax): New.
1687 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
1688 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
1690 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
1692 (rl78_expand_epilogue): Likewise.
1694 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
1697 * lra-eliminations.c (remove_reg_equal_offset_note): New.
1698 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
1699 pointer elimination with using remove_reg_equal_offset_note.
1701 2013-09-12 DJ Delorie <dj@redhat.com>
1703 * config/msp430/: New port.
1704 * config.gcc (msp430): Added.
1705 * doc/invoke.texi: Document MSP430 options.
1706 * doc/install.texi: Document msp430-elf
1707 * doc/md.texi: Document msp430-elf
1708 * doc/contrib.texi: Document msp430-elf
1710 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
1713 2013-09-12 Martin Jambor <mjambor@suse.cz>
1716 * ipa-prop.c (remove_described_reference): Give up if the edge in the
1717 reference descriptor is NULL.
1718 (ipa_edge_removal_hook): If owning a reference descriptor, set its
1721 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
1723 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
1724 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
1725 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
1726 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
1727 make_temp_ssa_name): move to tree-ssanames.h
1728 * tree-ssa-alias.h: Move prototype.
1729 * tree-ssa.h: Include tree-ssanames.h.
1730 * tree-ssanames.c (FREE_SSANAMES): Move to here.
1731 * tree-ssanames.h: New. Move items from tree-flow*.h
1732 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
1734 2013-09-12 Richard Biener <rguenther@suse.de>
1736 PR tree-optimization/58404
1737 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
1738 propagate non-invariant addresses into dereferences wrapped
1739 in component references.
1741 2013-09-12 Richard Biener <rguenther@suse.de>
1743 PR tree-optimization/58402
1744 * passes.def: Move pass_late_warn_uninitialized later.
1746 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
1748 * tree-ssa.h: New. Move content from tree-flow.h and
1750 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
1751 Move prototypes belonging to tree-ssa.c.
1752 * tree-flow-inline.h (redirect_edge_var_map_def,
1753 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
1755 * gimple.h: Adjust prototypes.
1756 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
1758 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
1759 * tree.h: Move prototype to tree-ssa.h.
1760 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
1761 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
1762 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
1763 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
1764 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
1765 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
1766 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
1767 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
1768 graphite-blocking.c, graphite-clast-to-gimple.c,
1769 graphite-dependences.c, graphite-interchange.c,
1770 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
1771 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
1772 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
1773 ipa-reference.c, ipa-split.c, ipa-utils.c,
1774 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
1775 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
1776 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
1777 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
1778 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
1779 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
1780 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
1781 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
1782 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
1783 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
1784 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
1785 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
1786 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
1787 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
1788 value-prof.c, var-tracking.c,
1789 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
1791 2013-09-12 Richard Biener <rguenther@suse.de>
1793 PR tree-optimization/58396
1794 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
1795 (build_rdg): Take a loop-nest parameter, fix memleaks.
1796 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
1798 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
1800 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
1803 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
1805 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
1806 The FMA instruction names should have a 'v' prefix.
1808 2013-09-12 Richard Biener <rguenther@suse.de>
1810 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
1811 (dot_rdg): Use popen instead of system in optional code.
1812 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
1813 (already_processed_vertex_p): Adjust.
1814 (has_anti_or_output_dependence, predecessor_has_mem_write,
1815 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
1816 rdg_flag_uses): Remove.
1817 (rdg_flag_vertex): Simplify.
1818 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
1820 (build_rdg_partition_for_component): Process the first vertex
1821 of a component only.
1822 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
1824 2013-09-12 Alan Modra <amodra@gmail.com>
1826 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
1828 2013-09-11 DJ Delorie <dj@redhat.com>
1829 Nick Clifton <nickc@redhat.com>
1831 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
1832 (rl78_stack_based_mem): New.
1833 * config/rl78/constraints.md (Iv08): New.
1842 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
1844 (movsi): Change from expand to insn-and-split.
1845 (ashrsi3): Clobber AX.
1849 * config/rl78/rl78.md (CC_REG): Fix.
1850 (addsi3): Allow memory and immediate operands.
1851 (addsi3_internal): Split into...
1852 (addsi3_internal_virt): ...new, and ...
1853 (addsi3_internal_real): ...new.
1855 (subsi3_internal_virt): New.
1856 (subsi3_internal_real): New.
1857 (mulsi3): Add memory operand.
1858 (mulsi3_rl78): Likewise.
1859 (mulsi3_g13): Likewise.
1860 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
1861 (cbranchqi4_real): Add more constraint options.
1862 (cbranchhi4_real): Expand pattern.
1863 (cbranchhi4_real_signed): New.
1864 (cbranchhi4_real_inverted): New.
1865 (cbranchsi4_real_lt): New.
1866 (cbranchsi4_real_ge): New.
1867 (cbranchsi4_real_signed): New.
1868 (cbranchsi4_real): New.
1870 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
1872 (lshrsi3_virt): Likewise.
1873 (ashlsi3_virt): Likewise.
1874 (cbranchqi4_virt_signed): New.
1875 (cbranchhi4_virt_signed): New.
1876 (cbranchsi4_virt): New.
1877 * config/rl78/rl78.c: Whitespace fixes throughout.
1878 (move_elim_pass): New.
1879 (pass_data_rl78_move_elim): New.
1880 (pass_rl78_move_elim): New.
1881 (make_pass_rl78_move_elim): New.
1882 (rl78_devirt_info): Run devirt earlier.
1883 (rl78_move_elim_info): New.
1884 (rl78_asm_file_start): Register it.
1885 (rl78_split_movsi): New.
1886 (rl78_as_legitimate_address): Allow virtual base registers when
1888 (rl78_addr_space_convert): Remove spurious debug stuff.
1889 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
1890 (rl78_print_operand): More cases for not printing '#'.
1891 (rl78_expand_compare): Remove most of the logic.
1892 (content_memory): New.
1893 (clear_content_memory): New.
1894 (get_content_index): New.
1895 (get_content_name): New.
1896 (display_content_memory): New.
1897 (update_content): New.
1898 (record_content): New.
1899 (already_contains): New.
1900 (insn_ok_now): Re-recog insns with virtual registers.
1901 (add_postponed_content_update): New.
1902 (process_postponed_content_update): New.
1903 (gen_and_emit_move): New.
1904 (transcode_memory_rtx): Record new location content.
1905 Use gen_and_emit_move.
1906 (force_into_acc): New.
1907 (move_to_acc): Use gen_and_emit_move.
1908 (move_from_acc): Likewise.
1909 (move_acc_to_reg): Likewise.
1910 (move_to_x): Likewise.
1911 (move_to_hl): Likewise.
1912 (move_to_de): Likewise.
1913 (rl78_alloc_physical_registers_op1): Record location content.
1914 (has_constraint): New.
1915 (rl78_alloc_physical_registers_op2): Record location content.
1917 (rl78_alloc_physical_registers_ro1): Likewise.
1918 (rl78_alloc_physical_registers_cmp): Likewise.
1919 (rl78_alloc_physical_registers_umul): Likewise.
1920 (rl78_alloc_address_registers_macax): New.
1921 (rl78_alloc_physical_registers): Initialize and set location
1922 content memory as needed.
1923 (rl78_reorg): Make sure split2 is called.
1924 (rl78_rtx_costs): New.
1926 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
1928 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
1929 for (not (neg ...)) and (neg (not ...)) cases.
1931 2013-09-11 Richard Biener <rguenther@suse.de>
1934 * passes.def: Split critical edges before late uninit warning passes.
1935 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
1937 2013-09-11 Jakub Jelinek <jakub@redhat.com>
1939 PR tree-optimization/58385
1940 * fold-const.c (build_range_check): If both low and high are NULL,
1941 use omit_one_operand_loc to preserve exp side-effects.
1943 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1945 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
1947 2013-09-11 Richard Biener <rguenther@suse.de>
1949 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
1950 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
1951 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
1952 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
1953 stmts_from_loop, known_dependences_p, build_empty_rdg,
1954 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
1955 * tree-loop-distribution.c: ... here.
1956 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
1957 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
1958 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
1959 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
1960 * tree-loop-distribution.c: ... here.
1961 * tree-loop-distribution.c: Include gimple-pretty-print.h.
1962 (struct partition_s): Add loops member.
1963 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
1964 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
1965 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
1967 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1968 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1969 Sergey Lega <sergey.s.lega@intel.com>
1970 Anna Tikhonova <anna.tikhonova@intel.com>
1971 Ilya Tocar <ilya.tocar@intel.com>
1972 Andrey Turetskiy <andrey.turetskiy@intel.com>
1973 Ilya Verbin <ilya.verbin@intel.com>
1974 Kirill Yukhin <kirill.yukhin@intel.com>
1975 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1977 * config/i386/constraints.md (k): New.
1979 * config/i386/i386.c (const regclass_map): Add new mask registers.
1980 (dbx_register_map): Ditto.
1981 (dbx64_register_map): Ditto.
1982 (svr4_dbx_register_map): Ditto.
1983 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
1985 (ix86_preferred_reload_class): Disable constants for mask registers.
1986 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
1987 (ix86_hard_regno_mode_ok): Support new mask registers.
1988 (x86_order_regs_for_local_alloc): Ditto.
1989 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
1990 (FIXED_REGISTERS): Add new mask registers.
1991 (CALL_USED_REGISTERS): Ditto.
1992 (REG_ALLOC_ORDER): Ditto.
1993 (VALID_MASK_REG_MODE): New.
1994 (FIRST_MASK_REG): Ditto.
1995 (LAST_MASK_REG): Ditto.
1996 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
1997 (MAYBE_MASK_CLASS_P): New.
1998 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
1999 (REG_CLASS_CONTENTS): Ditto.
2000 (MASK_REGNO_P): New.
2001 (ANY_MASK_REG_P): Ditto.
2002 (HI_REGISTER_NAMES): Add new mask registers.
2003 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
2004 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
2006 (attribute "type"): Add mskmov, msklog.
2007 (attribute "length_immediate"): Support them.
2008 (attribute "memory"): Ditto.
2009 (attribute "prefix_0f"): Ditto.
2010 (*movhi_internal): Support new mask registers.
2011 (*movqi_internal): Ditto.
2012 (define_split): Split out clobber pattern is a logic
2013 insn on mask registers.
2014 (*k<logic><mode>): New.
2015 (*andhi_1): Extend to support mask regs.
2016 (*andqi_1): Extend to support mask regs.
2018 (define_split): Split and-not to and and not if operands
2020 (*<code><mode>_1): Separate HI mode to new pattern...
2021 (*<code>hi_1): This.
2022 (*<code>qi_1): Extend to support mask regs.
2024 (kortestzhi): Ditto.
2025 (kortestchi): Ditto.
2027 (*one_cmpl<mode>2_1): Remove HImode and handle it...
2028 (*one_cmplhi2_1): ...Here, now with mask registers support.
2029 (*one_cmplqi2_1): Support new mask registers.
2030 (HI/QImode arithmetics splitter): Don't split if mask registers
2032 (HI/QImode not splitter): Ditto.
2033 * config/i386/predicated.md (mask_reg_operand): New.
2034 (general_reg_operand): Ditto.
2036 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2038 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
2039 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
2041 2013-09-10 Jeff Law <law@redhat.com>
2043 PR tree-optimization/58380
2044 * tree-ssa-threadupdate.c (thread_block): Recognize another case
2045 of threading through a buried loop header.
2047 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
2048 return value for single successor case.
2050 2013-09-10 Jan Hubicka <jh@suse.cz>
2052 * ipa-devirt.c (ipa_devirt): Enable with LTO.
2054 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
2057 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
2058 support conditional execution.
2059 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
2061 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
2063 * lra.c (lra): Clear lra_optional_reload_pseudos before every
2065 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
2066 Check destination too to check move insn.
2067 (undo_optional_reloads): Add check that the original peudo did not
2068 changed its allocation and the optional reload was inherited on last
2069 inheritance pass. Break loop after deciding to keep optional reload.
2070 (lra_undo_inheritance): Add check that inherited pseudo still in
2073 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
2075 * config/aarch64/aarch64.md (generic_sched): New.
2076 * config/aarch64/aarch64-generic.md (load): Make conditional
2077 on generic_sched attribute.
2078 (nonload): Likewise.
2080 2013-09-10 Jan Hubicka <jh@suse.cz>
2082 * lto-cgraph.c: Include ipa-utils.h.
2083 (compute_ltrans_boundary): Also add possible targets into the boundary.
2085 2013-09-10 Jan Hubicka <jh@suse.cz>
2087 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
2088 VAR_DECL of vtable rather than full expression.
2090 2013-09-10 Jan Hubicka <jh@suse.cz>
2091 Paolo Carlini <paolo.carlini@oracle.com>
2093 * cgraphunit.c (analyze_functions): Save input_location, set it
2094 to UNKNOWN_LOCATION and restore it at the end.
2096 2013-09-10 Martin Jambor <mjambor@suse.cz>
2098 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
2099 represented by a thunk.
2101 2013-09-10 Jeff Law <law@redhat.com>
2103 PR tree-optimization/58343
2104 * tree-ssa-threadupdate.c (thread_block): Identify and disable
2105 jump threading requests through loop headers buried in the middle
2106 of a jump threading path.
2108 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
2109 in return value/type.
2111 2013-09-10 Jakub Jelinek <jakub@redhat.com>
2113 PR rtl-optimization/58365
2114 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
2115 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
2118 2013-09-10 Richard Biener <rguenther@suse.de>
2120 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
2121 * tree-data-ref.c (create_rdg_vertices): Collect all data
2122 references, signal failure to the caller, use data-ref API.
2123 (build_rdg): Compute data references only once. Maintain lifetime
2124 of data references and data dependences from within RDG.
2125 (free_rdg): Free dependence relations.
2126 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
2127 inventing extra dependences.
2128 (distribute_loop): Update for RDG API changes.
2130 2013-09-10 Kai Tietz <ktietz@redhat.com>
2132 * doc/invoke.texi (fms-extensions): Document changed
2133 behavior for ms-abi targets.
2134 * config/i386/i386.c (ix86_option_override_internal):
2135 Set default value of option -fms-extension for ms-abi targets.
2137 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
2139 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
2141 2013-09-10 Alan Modra <amodra@gmail.com>
2144 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
2146 2013-09-10 Alan Modra <amodra@gmail.com>
2148 * config/rs6000/predicates.md (add_cint_operand): New.
2149 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
2150 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
2151 using add_cint_operand.
2152 (largetoc_high_plus_aix): Likewise.
2154 2013-09-09 Jakub Jelinek <jakub@redhat.com>
2156 PR tree-optimization/58364
2157 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
2158 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
2159 the current range can't be an unconditional true or false.
2161 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
2163 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
2165 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
2167 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
2169 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
2172 * doc/invoke.texi (-Wdelete-incomplete): Document it.
2174 2013-09-09 Ian Bolton <ian.bolton@arm.com>
2176 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
2177 NO_REGS for immediate that can't be moved directly into FP_REGS.
2179 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2181 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
2182 comparison with negated operand.
2183 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
2186 2013-09-09 Richard Biener <rguenther@suse.de>
2189 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
2190 of a subloop record all its block as affecting loop-closed SSA form.
2192 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
2194 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
2195 of an rtx/bitpos pair.
2196 (store_fixed_bit_field): Update accordingly.
2198 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
2200 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
2202 * builtins.c (expand_errno_check): Likewise.
2203 * dwarf2cfi.c (init_return_column_size): Likewise.
2204 * except.c (sjlj_mark_call_sites): Likewise.
2205 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
2206 * lra-constraints.c (emit_inc): Likewise.
2207 * ree.c (combine_set_extension): Likewise.
2208 * regmove.c (fixup_match_2): Likewise.
2209 * reload1.c (inc_for_reload): Likewise.
2211 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
2213 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
2214 (force_to_mode, simplify_shift_const_1, simplify_comparison):
2215 Use gen_int_mode with the mode of the associated simplify_* call.
2216 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
2217 * expmed.c (expand_shift_1): Likewise.
2218 * function.c (instantiate_virtual_regs_in_insn): Likewise.
2219 * loop-iv.c (iv_number_of_iterations): Likewise.
2220 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
2221 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
2223 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
2225 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
2226 of the associated expand_* call.
2227 (asan_emit_stack_protection): Likewise.
2228 * builtins.c (round_trampoline_addr): Likewise.
2229 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
2230 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
2231 (emit_store_flag): Likewise.
2232 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
2233 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
2235 * function.c (instantiate_virtual_regs_in_insn): Likewise.
2236 * ifcvt.c (noce_try_store_flag_constants): Likewise.
2237 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
2238 * modulo-sched.c (generate_prolog_epilog): Likewise.
2239 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
2240 (expand_ctz, expand_ffs, expand_unop): Likewise.
2242 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
2244 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
2245 of the associated gen_rtx_* call.
2246 * caller-save.c (init_caller_save): Likewise.
2247 * combine.c (find_split_point, make_extraction): Likewise.
2248 (make_compound_operation): Likewise.
2249 * dwarf2out.c (mem_loc_descriptor): Likewise.
2250 * explow.c (plus_constant, probe_stack_range): Likewise.
2251 * expmed.c (expand_mult_const): Likewise.
2252 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
2253 * reload1.c (init_reload): Likewise.
2254 * valtrack.c (cleanup_auto_inc_dec): Likewise.
2255 * var-tracking.c (adjust_mems): Likewise.
2256 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
2257 rather than gen_rtx_fmt_ee.
2259 2013-09-09 Jan Hubicka <jh@suse.cz>
2262 * value-prof.c (gimple_ic): Copy also abnormal edges.
2264 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
2266 * asan.c (asan_shadow_cst): Use gen_int_mode.
2268 2013-09-08 Jan Hubicka <jh@suse.cz>
2270 * ipa-profile.c: Add toplevel comment.
2271 (ipa_propagate_frequency_1): Be more conservative when profile is read.
2272 (contains_hot_call_p): New function.
2273 (ipa_propagate_frequency): Set frequencies based on counts when
2275 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
2276 profile; do not tamper with profile after inlining if it is read.
2278 2013-09-08 Jan Hubicka <jh@suse.cz>
2280 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
2283 2013-09-08 Jan Hubicka <jh@suse.cz>
2285 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
2288 2013-09-08 Jeff Law <law@redhat.com>
2291 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
2294 2013-09-08 Andi Kleen <ak@linux.intel.com>
2296 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
2298 2013-09-08 Jan Hubicka <jh@suse.cz>
2300 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
2302 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
2304 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
2305 for non-debug insns.
2306 * lra.c (new_insn_reg): Take the containing insn as a parameter.
2307 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
2308 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
2311 2013-09-08 Jan Hubicka <jh@suse.cz>
2313 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
2314 targets and devirtualize to BUILT_IN_UNREACHABLE.
2315 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
2316 * ipa.c (walk_polymorphic_call_targets): New function.
2317 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
2318 functions; use the new timevar.
2319 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
2320 was removed from the program.
2321 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
2322 not consider it in the walk when its vtable is dead.
2323 (possible_polymorphic_call_targets_1): Pass anonymous flag to
2325 (devirt_variable_node_removal_hook): New function.
2326 (possible_polymorphic_call_targets): Also register
2327 devirt_variable_node_removal_hook.
2328 (ipa_devirt): Do not do non-speculative devirtualization.
2329 (gate_ipa_devirt): One execute if devirtualizing speculatively.
2331 2013-09-08 Jan Hubicka <jh@suse.cz>
2333 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
2334 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
2335 varpool_remove_variable_insertion_hook): Declare.
2336 * varpool.c (varpool_node_hook_list): New structure.
2337 (first_varpool_node_removal_hook,
2338 first_varpool_variable_insertion_hook): New variables.
2339 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
2340 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
2341 varpool_remove_variable_insertion_hook,
2342 varpool_call_variable_insertion_hooks): New functions.
2343 (varpool_remove_node): Use it.
2345 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
2348 * diagnostic.c (diagnostic_build_prefix): When s.file is
2349 "<built-in>" don't output line and column numbers.
2351 2013-09-06 Jan Hubicka <jh@suse.cz>
2353 * cgraphunit.c (expand_thunk): Get body before touching arguments.
2354 * lto-streamer-out.c: Stream thunks, too.
2355 * lto-streamer-in.c (input_function): Pop cfun here
2356 (lto_read_body): Instead of here.
2358 2013-09-06 Caroline Tice <cmtice@google.com>
2360 * doc/install.texi: Add documentation for the --enable-vtable-verify
2361 and the --disable-libvtv configure options.
2363 2013-09-06 Jeff Law <law@redhat.com>
2365 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
2366 edge implied equivalences into successor phis.
2368 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
2370 * resource.c (mark_referenced_resources): Handle COND_EXEC.
2372 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
2374 * resource.c (mark_target_live_regs): Compute resources taking
2375 into account if a call is predicated or not.
2377 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
2379 * toplev.c (output_stack_usage): Be prepared for suffixes created by
2380 the compiler in the function names.
2382 2013-09-06 Jan Hubicka <jh@suse.cz>
2385 * ipa-inline.c (has_caller_p): New function.
2386 (want_inline_function_to_all_callers_p): Use it.
2387 (sum_callers, inline_to_all_callers): Break out from ...
2388 (ipa_inline): ... here.
2390 2013-09-06 Jan Hubicka <jh@suse.cz>
2392 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
2393 only when AVX is enabled.
2395 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2397 * config/aarch64/aarch64.md
2398 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
2399 is fpsimd_<load/store>2.
2400 (load_pair<mode>): Likewise.
2401 (store_pair<mode>): Likewise.
2403 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2405 * config/arm/types.md (type): Add "mrs" type.
2406 * config/aarch64/aarch64.md
2407 (aarch64_load_tp_hard): Make type "mrs".
2409 (load_tp_hard): Make type "mrs".
2410 * config/arm/cortex-a15.md: Update with new attributes.
2411 * config/arm/cortex-a5.md: Update with new attributes.
2412 * config/arm/cortex-a53.md: Update with new attributes.
2413 * config/arm/cortex-a7.md: Update with new attributes.
2414 * config/arm/cortex-a8.md: Update with new attributes.
2415 * config/arm/cortex-a9.md: Update with new attributes.
2416 * config/arm/cortex-m4.md: Update with new attributes.
2417 * config/arm/cortex-r4.md: Update with new attributes.
2418 * config/arm/fa526.md: Update with new attributes.
2419 * config/arm/fa606te.md: Update with new attributes.
2420 * config/arm/fa626te.md: Update with new attributes.
2421 * config/arm/fa726te.md: Update with new attributes.
2423 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2425 * config/aarch64/aarch64.md
2426 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
2427 (*movtf_aarch64): Likewise.
2429 (thumb1_movdi_insn): Use "multiple" for type where more than one
2430 instruction is used for a move.
2431 (*arm32_movhf): Likewise.
2432 (*thumb_movdf_insn): Likewise.
2434 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2436 * config/arm/types.md (type): Rename fcpys to fmov.
2438 (*arm_movsi_vfp): Rename type fcpys as fmov.
2439 (*thumb2_movsi_vfp): Likewise
2440 (*movhf_vfp_neon): Likewise
2441 (*movhf_vfp): Likewise
2442 (*movsf_vfp): Likewise
2443 (*thumb2_movsf_vfp): Likewise
2444 (*movsfcc_vfp): Likewise
2445 (*thumb2_movsfcc_vfp): Likewise
2446 * config/aarch64/aarch64-simd.md
2447 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
2448 * config/aarch64/aarch64.md
2449 (*movsi_aarch64): Replace type mov_reg with fmovs.
2450 (*movdi_aarch64): Likewise
2451 (*movsf_aarch64): Likewise
2452 (*movdf_aarch64): Likewise
2454 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
2455 * config/arm/iwmmxt.md
2456 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
2457 * config/arm/arm1020e.md: Update with new attributes.
2458 * config/arm/cortex-a15-neon.md: Update with new attributes.
2459 * config/arm/cortex-a5.md: Update with new attributes.
2460 * config/arm/cortex-a53.md: Update with new attributes.
2461 * config/arm/cortex-a7.md: Update with new attributes.
2462 * config/arm/cortex-a8-neon.md: Update with new attributes.
2463 * config/arm/cortex-a9.md: Update with new attributes.
2464 * config/arm/cortex-m4-fpu.md: Update with new attributes.
2465 * config/arm/cortex-r4f.md: Update with new attributes.
2466 * config/arm/marvell-pj4.md: Update with new attributes.
2467 * config/arm/vfp11.md: Update with new attributes.
2469 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2471 * config/aarch64/aarch64.md
2472 (*madd<mode>): Fix type attribute.
2473 (*maddsi_uxtw): Likewise.
2474 (*msub<mode>): Likewise.
2475 (*msubsi_uxtw): Likewise.
2476 (<su_optab>maddsidi4): Likewise.
2477 (<su_optab>msubsidi4): Likewise.
2479 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2481 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
2482 * config/arm/arm.md (core_cycles): Remove fdiv.
2483 * config/arm/vfp.md:
2484 (*sqrtsf2_vfp): Update for attribute changes.
2485 (*sqrtdf2_vfp): Likewise.
2486 * config/aarch64/aarch64.md:
2487 (sqrt<mode>2): Update for attribute changes.
2488 * config/arm/arm1020e.md: Update with new attributes.
2489 * config/arm/cortex-a15-neon.md: Update with new attributes.
2490 * config/arm/cortex-a5.md: Update with new attributes.
2491 * config/arm/cortex-a53.md: Update with new attributes.
2492 * config/arm/cortex-a7.md: Update with new attributes.
2493 * config/arm/cortex-a8-neon.md: Update with new attributes.
2494 * config/arm/cortex-a9.md: Update with new attributes.
2495 * config/arm/cortex-m4-fpu.md: Update with new attributes.
2496 * config/arm/cortex-r4f.md: Update with new attributes.
2497 * config/arm/marvell-pj4.md: Update with new attributes.
2498 * config/arm/vfp11.md: Update with new attributes.
2500 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2502 * config/arm/types.md
2503 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
2504 * config/aarch64/aarch64.md
2505 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
2507 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
2508 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
2509 (float<GPI:mode><GPF:mode>2): Likewise.
2511 (*truncsisf2_vfp): Update with new attributes.
2512 (*truncsidf2_vfp): Likewise.
2513 (fixuns_truncsfsi2): Likewise.
2514 (fixuns_truncdfsi2): Likewise.
2515 (*floatsisf2_vfp): Likewise.
2516 (*floatsidf2_vfp): Likewise.
2517 (floatunssisf2): Likewise.
2518 (floatunssidf2): Likewise.
2519 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
2520 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
2521 * config/arm/arm1020e.md: Update with new attributes.
2522 * config/arm/cortex-a15-neon.md: Update with new attributes.
2523 * config/arm/cortex-a5.md: Update with new attributes.
2524 * config/arm/cortex-a53.md: Update with new attributes.
2525 * config/arm/cortex-a7.md: Update with new attributes.
2526 * config/arm/cortex-a8-neon.md: Update with new attributes.
2527 * config/arm/cortex-a9.md: Update with new attributes.
2528 * config/arm/cortex-m4-fpu.md: Update with new attributes.
2529 * config/arm/cortex-r4f.md: Update with new attributes.
2530 * config/arm/marvell-pj4.md: Update with new attributes.
2531 * config/arm/vfp11.md: Update with new attributes.
2533 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2535 * config/aarch64/arm_neon.h
2536 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
2537 (vqtbx<1,2,3,4><q>_s8): Likewise.
2539 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2541 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
2542 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
2543 (add<mode>3): Add type attribute.
2544 (add<mode>3): Likewise.
2545 (usadd<mode>3): Likewise.
2546 (ssadd<mode>3): Likewise.
2547 (sub<mode>3): Likewise.
2548 (sub<mode>3): Likewise.
2549 (ussub<mode>3): Likewise.
2550 (sssub<mode>3): Likewise.
2551 (ssmulsa3): Likewise.
2552 (usmulusa3): Likewise.
2553 (arm_usatsihi): Likewise.
2555 (*movdi_vfp): Add types for all instructions.
2556 (*movdi_vfp_cortexa8): Likewise.
2557 (*movhf_vfp_neon): Likewise.
2558 (*movhf_vfp): Likewise.
2559 (*movdf_vfp): Likewise.
2560 (*thumb2_movdf_vfp): Likewise.
2561 (*thumb2_movdfcc_vfp): Likewise.
2562 * config/arm/arm.md: Add type attribute to all insn patterns.
2563 (*thumb1_adddi3): Add type attribute.
2564 (*arm_adddi3): Likewise.
2565 (*adddi_sesidi_di): Likewise.
2566 (*adddi_zesidi_di): Likewise.
2567 (*thumb1_addsi3): Likewise.
2568 (addsi3_compare0): Likewise.
2569 (*addsi3_compare0_scratch): Likewise.
2570 (*compare_negsi_si): Likewise.
2571 (cmpsi2_addneg): Likewise.
2572 (*addsi3_carryin_<optab>): Likewise.
2573 (*addsi3_carryin_alt2_<optab>): Likewise.
2574 (*addsi3_carryin_clobercc_<optab>): Likewise.
2575 (*subsi3_carryin): Likewise.
2576 (*subsi3_carryin_const): Likewise.
2577 (*subsi3_carryin_compare): Likewise.
2578 (*subsi3_carryin_compare_const): Likewise.
2579 (*arm_subdi3): Likewise.
2580 (*thumb_subdi3): Likewise.
2581 (*subdi_di_zesidi): Likewise.
2582 (*subdi_di_sesidi): Likewise.
2583 (*subdi_zesidi_di): Likewise.
2584 (*subdi_sesidi_di): Likewise.
2585 (*subdi_zesidi_ze): Likewise.
2586 (thumb1_subsi3_insn): Likewise.
2587 (*arm_subsi3_insn): Likewise.
2588 (*anddi3_insn): Likewise.
2589 (*anddi_zesidi_di): Likewise.
2590 (*anddi_sesdi_di): Likewise.
2591 (*ne_zeroextracts): Likewise.
2592 (*ne_zeroextracts): Likewise.
2593 (*ite_ne_zeroextr): Likewise.
2594 (*ite_ne_zeroextr): Likewise.
2595 (*anddi_notdi_di): Likewise.
2596 (*anddi_notzesidi): Likewise.
2597 (*anddi_notsesidi): Likewise.
2598 (andsi_notsi_si): Likewise.
2599 (thumb1_bicsi3): Likewise.
2600 (*iordi3_insn): Likewise.
2601 (*iordi_zesidi_di): Likewise.
2602 (*iordi_sesidi_di): Likewise.
2603 (*thumb1_iorsi3_insn): Likewise.
2604 (*xordi3_insn): Likewise.
2605 (*xordi_zesidi_di): Likewise.
2606 (*xordi_sesidi_di): Likewise.
2607 (*arm_xorsi3): Likewise.
2608 (*andsi_iorsi3_no): Likewise.
2609 (*smax_0): Likewise.
2610 (*smax_m1): Likewise.
2611 (*arm_smax_insn): Likewise.
2612 (*smin_0): Likewise.
2613 (*arm_smin_insn): Likewise.
2614 (*arm_umaxsi3): Likewise.
2615 (*arm_uminsi3): Likewise.
2616 (*minmax_arithsi): Likewise.
2617 (*minmax_arithsi_): Likewise.
2618 (*satsi_<SAT:code>): Likewise.
2619 (arm_ashldi3_1bit): Likewise.
2620 (arm_ashrdi3_1bit): Likewise.
2621 (arm_lshrdi3_1bit): Likewise.
2622 (*arm_negdi2): Likewise.
2623 (*thumb1_negdi2): Likewise.
2624 (*arm_negsi2): Likewise.
2625 (*thumb1_negsi2): Likewise.
2626 (*negdi_extendsid): Likewise.
2627 (*negdi_zero_extend): Likewise.
2628 (*arm_abssi2): Likewise.
2629 (*thumb1_abssi2): Likewise.
2630 (*arm_neg_abssi2): Likewise.
2631 (*thumb1_neg_abss): Likewise.
2632 (one_cmpldi2): Likewise.
2633 (extend<mode>di2): Likewise.
2634 (*compareqi_eq0): Likewise.
2635 (*arm_extendhisi2addsi): Likewise.
2636 (*arm_movdi): Likewise.
2637 (*thumb1_movdi_insn): Likewise.
2638 (*arm_movt): Likewise.
2639 (*thumb1_movsi_insn): Likewise.
2640 (pic_add_dot_plus_four): Likewise.
2641 (pic_add_dot_plus_eight): Likewise.
2642 (tls_load_dot_plus_eight): Likewise.
2643 (*thumb1_movhi_insn): Likewise.
2644 (*thumb1_movsf_insn): Likewise.
2645 (*movdf_soft_insn): Likewise.
2646 (*thumb_movdf_insn): Likewise.
2647 (cbranchsi4_insn): Likewise.
2648 (cbranchsi4_scratch): Likewise.
2649 (*negated_cbranchsi4): Likewise.
2650 (*tbit_cbranch): Likewise.
2651 (*tlobits_cbranch): Likewise.
2652 (*tstsi3_cbranch): Likewise.
2653 (*cbranchne_decr1): Likewise.
2654 (*addsi3_cbranch): Likewise.
2655 (*addsi3_cbranch_scratch): Likewise.
2656 (*arm_cmpdi_insn): Likewise.
2657 (*arm_cmpdi_unsig): Likewise.
2658 (*arm_cmpdi_zero): Likewise.
2659 (*thumb_cmpdi_zero): Likewise.
2660 (*deleted_compare): Likewise.
2661 (*mov_scc): Likewise.
2662 (*mov_negscc): Likewise.
2663 (*mov_notscc): Likewise.
2664 (*cstoresi_eq0_thumb1_insn): Likewise.
2665 (cstoresi_nltu_thumb1): Likewise.
2666 (cstoresi_ltu_thu): Likewise.
2667 (thumb1_addsi3_addgeu): Likewise.
2668 (*arm_jump): Likewise.
2669 (*thumb_jump): Likewise.
2670 (*check_arch2): Likewise.
2671 (arm_casesi_internal): Likewise.
2672 (thumb1_casesi_dispatch): Likewise.
2673 (*arm_indirect_jump): Likewise.
2674 (*thumb1_indirect_jump): Likewise.
2676 (*and_scc): Likewise.
2677 (*ior_scc): Likewise.
2678 (*compare_scc): Likewise.
2679 (*cond_move): Likewise.
2680 (*cond_arith): Likewise.
2681 (*cond_sub): Likewise.
2682 (*cmp_ite0): Likewise.
2683 (*cmp_ite1): Likewise.
2684 (*cmp_and): Likewise.
2685 (*cmp_ior): Likewise.
2686 (*ior_scc_scc): Likewise.
2687 (*ior_scc_scc_cmp): Likewise.
2688 (*and_scc_scc): Likewise.
2689 (*and_scc_scc_cmp): Likewise.
2690 (*and_scc_scc_nod): Likewise.
2691 (*negscc): Likewise.
2692 (movcond_addsi): Likewise.
2693 (movcond): Likewise.
2694 (*ifcompare_plus_move): Likewise.
2695 (*if_plus_move): Likewise.
2696 (*ifcompare_move_plus): Likewise.
2697 (*if_move_plus): Likewise.
2698 (*ifcompare_arith_arith): Likewise.
2699 (*if_arith_arith): Likewise.
2700 (*ifcompare_arith_move): Likewise.
2701 (*if_arith_move): Likewise.
2702 (*ifcompare_move_arith): Likewise.
2703 (*if_move_arith): Likewise.
2704 (*ifcompare_move_not): Likewise.
2705 (*if_move_not): Likewise.
2706 (*ifcompare_not_move): Likewise.
2707 (*if_not_move): Likewise.
2708 (*ifcompare_shift_move): Likewise.
2709 (*if_shift_move): Likewise.
2710 (*ifcompare_move_shift): Likewise.
2711 (*if_move_shift): Likewise.
2712 (*ifcompare_shift_shift): Likewise.
2713 (*ifcompare_not_arith): Likewise.
2714 (*ifcompare_arith_not): Likewise.
2715 (*if_arith_not): Likewise.
2716 (*ifcompare_neg_move): Likewise.
2717 (*if_neg_move): Likewise.
2718 (*ifcompare_move_neg): Likewise.
2719 (*if_move_neg): Likewise.
2720 (prologue_thumb1_interwork): Likewise.
2721 (*cond_move_not): Likewise.
2722 (*sign_extract_onebit): Likewise.
2723 (*not_signextract_onebit): Likewise.
2724 (stack_tie): Likewise.
2725 (align_4): Likewise.
2726 (align_8): Likewise.
2727 (consttable_end): Likewise.
2728 (consttable_1): Likewise.
2729 (consttable_2): Likewise.
2730 (consttable_4): Likewise.
2731 (consttable_8): Likewise.
2732 (consttable_16): Likewise.
2733 (*thumb1_tablejump): Likewise.
2734 (prefetch): Likewise.
2735 (force_register_use): Likewise.
2736 (thumb_eh_return): Likewise.
2737 (load_tp_hard): Likewise.
2738 (load_tp_soft): Likewise.
2739 (tlscall): Likewise.
2740 (*arm_movtas_ze): Likewise.
2741 (*arm_rev): Likewise.
2742 (*arm_revsh): Likewise.
2743 (*arm_rev16): Likewise.
2744 * config/arm/thumb2.md
2745 (*thumb2_smaxsi3): Likewise.
2746 (*thumb2_sminsi3): Likewise.
2747 (*thumb32_umaxsi3): Likewise.
2748 (*thumb2_uminsi3): Likewise.
2749 (*thumb2_negdi2): Likewise.
2750 (*thumb2_abssi2): Likewise.
2751 (*thumb2_neg_abss): Likewise.
2752 (*thumb2_movsi_insn): Likewise.
2753 (tls_load_dot_plus_four): Likewise.
2754 (*thumb2_movhi_insn): Likewise.
2755 (*thumb2_mov_scc): Likewise.
2756 (*thumb2_mov_negs): Likewise.
2757 (*thumb2_mov_negs): Likewise.
2758 (*thumb2_mov_nots): Likewise.
2759 (*thumb2_mov_nots): Likewise.
2760 (*thumb2_movsicc_): Likewise.
2761 (*thumb2_movsfcc_soft_insn): Likewise.
2762 (*thumb2_indirect_jump): Likewise.
2763 (*thumb2_and_scc): Likewise.
2764 (*thumb2_ior_scc): Likewise.
2765 (*thumb2_ior_scc_strict_it): Likewise.
2766 (*thumb2_cond_move): Likewise.
2767 (*thumb2_cond_arith): Likewise.
2768 (*thumb2_cond_ari): Likewise.
2769 (*thumb2_cond_sub): Likewise.
2770 (*thumb2_negscc): Likewise.
2771 (*thumb2_movcond): Likewise.
2772 (thumb2_casesi_internal): Likewise.
2773 (thumb2_casesi_internal_pic): Likewise.
2774 (*thumb2_alusi3_short): Likewise.
2775 (*thumb2_mov<mode>_shortim): Likewise.
2776 (*thumb2_addsi_short): Likewise.
2777 (*thumb2_subsi_short): Likewise.
2778 (thumb2_addsi3_compare0): Likewise.
2779 (*thumb2_cbz): Likewise.
2780 (*thumb2_cbnz): Likewise.
2781 (*thumb2_one_cmplsi2_short): Likewise.
2782 (*thumb2_negsi2_short): Likewise.
2783 (*orsi_notsi_si): Likewise.
2784 * config/arm/arm1020e.md: Update with new attributes.
2785 * config/arm/arm1026ejs.md: Update with new attributes.
2786 * config/arm/arm1136jfs.md: Update with new attributes.
2787 * config/arm/arm926ejs.md: Update with new attributes.
2788 * config/arm/cortex-a15.md: Update with new attributes.
2789 * config/arm/cortex-a5.md: Update with new attributes.
2790 * config/arm/cortex-a53.md: Update with new attributes.
2791 * config/arm/cortex-a7.md: Update with new attributes.
2792 * config/arm/cortex-a8.md: Update with new attributes.
2793 * config/arm/cortex-a9.md: Update with new attributes.
2794 * config/arm/cortex-m4.md: Update with new attributes.
2795 * config/arm/cortex-r4.md: Update with new attributes.
2796 * config/arm/fa526.md: Update with new attributes.
2797 * config/arm/fa606te.md: Update with new attributes.
2798 * config/arm/fa626te.md: Update with new attributes.
2799 * config/arm/fa726te.md: Update with new attributes.
2801 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2803 * config/aarch64/aarch64-simd.md
2804 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
2805 <vwx> iterator to ensure correct register choice.
2806 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
2807 (aarch64_sqdmull_n<mode>): Likewise.
2808 (aarch64_sqdmull2_n<mode>_internal): Likewise.
2809 * config/aarch64/arm_neon.h
2810 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
2811 (vml<as><q>_n_<su>16): Likewise.
2812 (vml<as>l_high_lane<q>_<su>16): Likewise.
2813 (vml<as>l_high_n_<su>16): Likewise.
2814 (vml<as>l_lane<q>_<su>16): Likewise.
2815 (vml<as>l_n_<su>16): Likewise.
2816 (vmul<q>_lane<q>_<su>16): Likewise.
2817 (vmul<q>_n_<su>16): Likewise.
2818 (vmull_lane<q>_<su>16): Likewise.
2819 (vmull_n_<su>16): Likewise.
2820 (vmull_high_lane<q>_<su>16): Likewise.
2821 (vmull_high_n_<su>16): Likewise.
2822 (vqrdmulh<q>_n_s16): Likewise.
2824 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
2826 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
2827 have the correct lane parameter.
2829 2013-09-06 Richard Biener <rguenther@suse.de>
2831 * cfganal.c (control_dependences::~control_dependences):
2832 Properly free all of the vector.
2834 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
2837 * config/i386/i386.c (ix86_conditional_register_usage):
2838 Proper initialize extended SSE registers.
2840 2013-09-06 Jan Hubicka <jh@suse.cz>
2842 PR tree-optimization/58311
2843 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
2845 2013-09-06 Jan Hubicka <jh@suse.cz>
2847 * Makefile.in (tree-sra.o): Update dependencies.
2848 * tree-sra.c: Include ipa-utils.h
2849 (scan_function): Use recursive_call_p.
2850 (has_caller_p): New function.
2851 (cgraph_for_node_and_aliases): Count also callers of aliases.
2853 2013-09-06 Jan Hubicka <jh@suse.cz>
2856 * cgraph.h (symtab_semantically_equivalent_p): Declare.
2857 * tree-tailcall.c: Include ipa-utils.h.
2858 (find_tail_calls): Use it.
2859 * ipa-pure-const.c (check_call): Likewise.
2860 * ipa-utils.c (recursive_call_p): New function.
2861 * ipa-utils.h (recursive_call_p): Dclare.
2862 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
2863 (symtab_semantically_equivalent_p): New function.
2864 * Makefile.in (tree-tailcall.o): Update dependencies.
2866 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
2868 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
2871 2013-09-06 Richard Biener <rguenther@suse.de>
2873 * lto-streamer.h (lto_global_var_decls): Remove.
2874 * Makefile.in (OBJS): Remove lto-symtab.o.
2875 (lto-symtab.o): Remove.
2876 (GTFILES): Remove lto-symtab.c
2877 * lto-symtab.c: Move to lto/
2879 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2881 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
2882 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
2883 (UNSPEC_FPINT_RINT): New constant definitions.
2884 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
2885 definition with 2 attributes.
2886 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
2887 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
2890 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2892 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
2893 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
2894 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
2895 attribute to "z196_alone".
2896 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
2899 2013-09-06 Richard Biener <rguenther@suse.de>
2901 * basic-block.h (class control_dependences): New.
2902 * tree-ssa-dce.c (control_dependence_map): Remove.
2904 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
2905 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
2906 find_pdom, find_control_dependence, find_all_control_dependences):
2908 (mark_control_dependent_edges_necessary,
2909 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
2910 tree_dce_done, perform_tree_ssa_dce): Adjust.
2911 * cfganal.c (set_control_dependence_map_bit,
2912 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
2913 find_all_control_dependences): Move from tree-ssa-dce.c and
2914 implement as methods of control_dependences class.
2915 (control_dependences::control_dependences): New.
2916 (control_dependences::~control_dependences): Likewise.
2917 (control_dependences::get_edges_dependent_on): Likewise.
2918 (control_dependences::get_edge): Likewise.
2920 2013-09-04 Jan Hubicka <jh@suse.cz>
2922 * tree.c (types_same_for_odr): Drop overactive check.
2923 * ipa-devirt.c (hash_type_name): Likewise.
2925 2013-09-04 Jan Hubicka <jh@suse.cz>
2927 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
2928 (analyze_functions): ... here.
2930 2013-09-04 Jan Hubicka <jh@suse.cz>
2933 * cgraphunit.c (analyze_functions): Clear AUX fields
2934 after processing; initialize assembler name has.
2936 2013-09-05 Jeff Law <law@redhat.com>
2938 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
2939 from thread_around_empty_block. Record threading path into PATH.
2940 Recurse if threading through the initial block is successful.
2941 (thread_across_edge): Corresponding changes to slightly simplify.
2943 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
2945 * config/aarch64/aarch64.md
2946 (type): Remove frecpe, frecps, frecpx.
2947 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
2948 fix to be a TARGET_SIMD instruction.
2949 (aarch64_frecps): Remove.
2950 * config/aarch64/aarch64-simd.md
2951 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
2952 (aarch64_frecps<mode>): Handle all float/vector of float modes.
2954 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
2955 Sofiane Naci <sofiane.naci@arm.com>
2957 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
2958 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
2959 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
2960 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
2961 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
2962 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
2963 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
2964 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
2965 into "clz, "rbit". Rename "shift" to "shift_imm".
2966 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
2967 changes. Update for attribute changes all occurrences of arlo_* and
2969 * config/arm/arm-fixed.md: Update for attribute changes
2970 all occurrences of arlo_* types.
2971 * config/arm/thumb2.md: Update for attribute changes all occurrences
2973 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
2974 (cortexa7_older_only): Likewise.
2975 (cortexa7_younger): Likewise.
2976 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
2977 (1020alu_shift_op): Likewise.
2978 (1020alu_shift_reg_op): Likewise.
2979 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
2980 (alu_shift_op): Likewise.
2981 (alu_shift_reg_op): Likewise.
2982 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
2983 (11_alu_shift_op): Likewise.
2984 (11_alu_shift_reg_op): Likewise.
2985 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
2986 (9_alu_shift_reg_op): Likewise.
2987 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
2989 (cortex_a15_alu_shift): Likewise.
2990 (cortex_a15_alu_shift_reg): Likewise.
2991 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
2993 (cortex_a5_alu_shift): Likewise.
2994 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
2996 (cortex_a53_alu_shift): Likewise.
2997 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
2999 (cortex_a7_alu_reg): Likewise.
3000 (cortex_a7_alu_shift): Likewise.
3001 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
3003 (cortex_a8_alu_shift): Likewise.
3004 (cortex_a8_alu_shift_reg): Likewise.
3005 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
3006 (cortex_a9_dp_shift): Likewise.
3007 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
3009 * config/arm/cortex-r4.md
3010 (cortex_r4_alu): Update for attribute changes.
3011 (cortex_r4_mov): Likewise.
3012 (cortex_r4_alu_shift_reg): Likewise.
3013 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
3014 (526_alu_shift_op): Likewise.
3015 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
3016 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
3017 (626te_alu_shift_op): Likewise.
3018 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
3019 (726te_alu_shift_op): Likewise.
3020 (726te_alu_shift_reg_op): Likewise.
3021 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
3022 (mp626_alu_shift_op): Likewise.
3023 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
3024 (pj4_alu_conds): Likewise.
3025 (pj4_shift): Likewise.
3026 (pj4_shift_conds): Likewise.
3027 (pj4_alu_shift): Likewise.
3028 (pj4_alu_shift_conds): Likewise.
3029 * config/aarch64/aarch64.md: Update for attribute change
3030 all occurrences of arlo_* and shift* types.
3032 2013-09-05 Mike Stump <mikestump@comcast.net>
3034 * tree.h: Move documentation for tree_function_decl to tree-core.h
3035 with the declaration.
3037 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
3040 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
3041 looking for widest mode.
3043 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
3045 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
3047 2013-09-05 Richard Biener <rguenther@suse.de>
3049 PR tree-optimization/58137
3050 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
3051 Do not create vectors of pointers.
3052 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
3053 types for the components of the vector initializer.
3054 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
3055 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
3057 2013-09-05 Martin Jambor <mjambor@suse.cz>
3059 * ipa-prop.c (remove_described_reference): Accept missing references,
3060 return false if that hppens, otherwise return true.
3061 (cgraph_node_for_jfunc): New function.
3062 (try_decrement_rdesc_refcount): Likewise.
3063 (try_make_edge_direct_simple_call): Use them.
3064 (ipa_edge_removal_hook): Remove references from rdescs.
3065 (ipa_edge_duplication_hook): Clone rdescs and their references
3066 when the new edge has the same caller as the old one.
3067 * cgraph.c (cgraph_resolve_speculation): Remove speculative
3068 reference before removing any edges.
3070 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
3072 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
3074 * thumb2.md (thumb2_storewb_parisi): New pattern.
3076 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
3078 * config/aarch64/aarch64-option-extensions.def: Add
3079 AARCH64_OPT_EXTENSION of 'crc'.
3080 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
3081 (AARCH64_ISA_CRC): Ditto.
3082 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
3083 description of the CRC extension.
3085 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
3087 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
3088 * config/rs6000/linux.h: Ditto.
3089 * alpha/linux.h: Ditto.
3090 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
3091 no_c99_libc_has_function.
3092 * config/c6x/uclinux-elf.h: Ditto.
3093 * config/lm32/uclinux-elf.h: Ditto.
3094 * config/m68k/uclinux.h: Ditto.
3095 * config/moxie/uclinux.h: Ditto.
3096 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
3097 (crisv32-*-linux*, cris-*-linux*): Ditto.
3098 * config/bfin/bfin.c: Include "tm_p.h".
3100 2013-09-05 Richard Biener <rguenther@suse.de>
3102 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
3103 check for a definition without a basic-block.
3105 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
3106 Sofiane Naci <sofiane.naci@arm.com>
3108 * config/aarch64/aarch64.md
3109 (*movti_aarch64): Rename r_2_f and f_2_r.
3110 (*movsf_aarch64): Likewise.
3111 (*movdf_aarch64): Likewise.
3112 (*movtf_aarch64): Likewise.
3113 (aarch64_movdi_<mode>low): Likewise.
3114 (aarch64_movdi_<mode>high): Likewise.
3115 (aarch64_mov<mode>high_di): Likewise.
3116 (aarch64_mov<mode>low_di): Likewise.
3117 (aarch64_movtilow_tilow): Likewise.
3118 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
3119 values to config/arm/types.md
3120 (attribute "conds"): Update for attribute change.
3121 (anddi3_insn): Likewise.
3122 (iordi3_insn): Likewise.
3123 (xordi3_insn): Likewise.
3124 (one_cmpldi2): Likewise.
3125 * config/arm/types.md (type): Add Neon types.
3126 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
3127 use "type" attribute.
3128 (movmisalign<mode>_neon_store): Likewise.
3129 (movmisalign<mode>_neon_load): Likewise.
3130 (vec_set<mode>_internal): Likewise.
3131 (vec_setv2di_internal): Likewise.
3132 (vec_extract<mode>): Likewise.
3133 (vec_extractv2di): Likewise.
3134 (add<mode>3_neon): Likewise.
3135 (adddi3_neon): Likewise.
3136 (sub<mode>3_neon): Likewise.
3137 (subdi3_neon): Likewise.
3138 (mul<mode>3_neon): Likewise.
3139 (mul<mode>3add<mode>_neon): Likewise.
3140 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
3141 (fma<VCVTF:mode>4)): Likewise.
3142 (fma<VCVTF:mode>4_intrinsic): Likewise.
3143 (fmsub<VCVTF:mode>4)): Likewise.
3144 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
3145 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
3146 (ior<mode>3): Likewise.
3147 (and<mode>3): Likewise.
3148 (anddi3_neon): Likewise.
3149 (orn<mode>3_neon): Likewise.
3150 (orndi3_neon): Likewise.
3151 (bic<mode>3_neon): Likewise.
3152 (bicdi3_neon): Likewise.
3153 (xor<mode>3): Likewise.
3154 (one_cmpl<mode>2): Likewise.
3155 (abs<mode>2): Likewise.
3156 (neg<mode>2): Likewise.
3157 (umin<mode>3_neon): Likewise.
3158 (umax<mode>3_neon): Likewise.
3159 (smin<mode>3_neon): Likewise.
3160 (smax<mode>3_neon): Likewise.
3161 (vashl<mode>3): Likewise.
3162 (vashr<mode>3_imm): Likewise.
3163 (vlshr<mode>3_imm): Likewise.
3164 (ashl<mode>3_signed): Likewise.
3165 (ashl<mode>3_unsigned): Likewise.
3166 (neon_load_count): Likewise.
3167 (ashldi3_neon_noclobber): Likewise.
3168 (signed_shift_di3_neon): Likewise.
3169 (unsigned_shift_di3_neon): Likewise.
3170 (ashrdi3_neon_imm_noclobber): Likewise.
3171 (lshrdi3_neon_imm_noclobber): Likewise.
3172 (widen_ssum<mode>3): Likewise.
3173 (widen_usum<mode>3): Likewise.
3174 (quad_halves_<code>v4si): Likewise.
3175 (quad_halves_<code>v4sf): Likewise.
3176 (quad_halves_<code>v8hi): Likewise.
3177 (quad_halves_<code>v16qi): Likewise.
3178 (reduc_splus_v2di): Likewise.
3179 (neon_vpadd_internal<mode>): Likewise.
3180 (neon_vpsmin<mode>): Likewise.
3181 (neon_vpsmax<mode>): Likewise.
3182 (neon_vpumin<mode>): Likewise.
3183 (neon_vpumax<mode>): Likewise.
3184 (ss_add<mode>_neon): Likewise.
3185 (us_add<mode>_neon): Likewise.
3186 (ss_sub<mode>_neon): Likewise.
3187 (us_sub<mode>_neon): Likewise.
3188 (neon_vadd<mode>_unspec): Likewise.
3189 (neon_vaddl<mode>): Likewise.
3190 (neon_vaddw<mode>): Likewise.
3191 (neon_vhadd<mode>): Likewise.
3192 (neon_vqadd<mode>): Likewise.
3193 (neon_vaddhn<mode>): Likewise.
3194 (neon_vmul<mode>): Likewise.
3195 (neon_vmla<mode>): Likewise.
3196 (neon_vmlal<mode>): Likewise.
3197 (neon_vmls<mode>): Likewise.
3198 (neon_vmlsl<mode>): Likewise.
3199 (neon_vqdmulh<mode>): Likewise.
3200 (neon_vqdmlal<mode>): Likewise.
3201 (neon_vqdmlsl<mode>): Likewise.
3202 (neon_vmull<mode>): Likewise.
3203 (neon_vqdmull<mode>): Likewise.
3204 (neon_vsub<mode>_unspec): Likewise.
3205 (neon_vsubl<mode>): Likewise.
3206 (neon_vsubw<mode>): Likewise.
3207 (neon_vqsub<mode>): Likewise.
3208 (neon_vhsub<mode>): Likewise.
3209 (neon_vsubhn<mode>): Likewise.
3210 (neon_vceq<mode>): Likewise.
3211 (neon_vcge<mode>): Likewise.
3212 (neon_vcgeu<mode>): Likewise.
3213 (neon_vcgt<mode>): Likewise.
3214 (neon_vcgtu<mode>): Likewise.
3215 (neon_vcle<mode>): Likewise.
3216 (neon_vclt<mode>): Likewise.
3217 (neon_vcage<mode>): Likewise.
3218 (neon_vcagt<mode>): Likewise.
3219 (neon_vtst<mode>): Likewise.
3220 (neon_vabd<mode>): Likewise.
3221 (neon_vabdl<mode>): Likewise.
3222 (neon_vaba<mode>): Likewise.
3223 (neon_vabal<mode>): Likewise.
3224 (neon_vmax<mode>): Likewise.
3225 (neon_vmin<mode>): Likewise.
3226 (neon_vpaddl<mode>): Likewise.
3227 (neon_vpadal<mode>): Likewise.
3228 (neon_vpmax<mode>): Likewise.
3229 (neon_vpmin<mode>): Likewise.
3230 (neon_vrecps<mode>): Likewise.
3231 (neon_vrsqrts<mode>): Likewise.
3232 (neon_vqabs<mode>): Likewise.
3233 (neon_vqneg<mode>): Likewise.
3234 (neon_vcls<mode>): Likewise.
3235 (clz<mode>2): Likewise.
3236 (popcount<mode>2): Likewise.
3237 (neon_vrecpe): Likewise.
3238 (neon_vrsqrte): Likewise.
3239 (neon_vget_lane<mode>_sext_internal): Likewise.
3240 (neon_vget_lane<mode>_zext_internal): Likewise.
3241 (neon_vdup_n<mode>): Likewise.
3242 (neon_vdup_nv2di): Likewise.
3243 (neon_vdpu_lane<mode>_internal): Likewise.
3244 (neon_vswp<mode>): Likewise.
3245 (float<mode><V_cvtto>2): Likewise.
3246 (floatuns<mode><V_cvtto>2): Likewise.
3247 (fix_trunc<mode><V_cvtto>)2): Likewise
3248 (fixuns_trunc<mode><V_cvtto)2): Likewise.
3249 (neon_vcvt<mode>): Likewise.
3250 (neon_vcvtv4sfv4hf): Likewise.
3251 (neon_vcvtv4hfv4sf): Likewise.
3252 (neon_vcvt_n<mode>): Likewise.
3253 (neon_vmovn<mode>): Likewise.
3254 (neon_vqmovn<mode>): Likewise.
3255 (neon_vqmovun<mode>): Likewise.
3256 (neon_vmovl<mode>): Likewise.
3257 (neon_vmul_lane<mode>): Likewise.
3258 (neon_vmull_lane<mode>): Likewise.
3259 (neon_vqdmull_lane<mode>): Likewise.
3260 (neon_vqdmulh_lane<mode>): Likewise.
3261 (neon_vmla_lane<mode>): Likewise.
3262 (neon_vmlal_lane<mode>): Likewise.
3263 (neon_vqdmlal_lane<mode>): Likewise.
3264 (neon_vmls_lane<mode>): Likewise.
3265 (neon_vmlsl_lane<mode>): Likewise.
3266 (neon_vqdmlsl_lane<mode>): Likewise.
3267 (neon_vext<mode>): Likewise.
3268 (neon_vrev64<mode>): Likewise.
3269 (neon_vrev32<mode>): Likewise.
3270 (neon_vrev16<mode>): Likewise.
3271 (neon_vbsl<mode>_internal): Likewise.
3272 (neon_vshl<mode>): Likewise.
3273 (neon_vqshl<mode>): Likewise.
3274 (neon_vshr_n<mode>): Likewise.
3275 (neon_vshrn_n<mode>): Likewise.
3276 (neon_vqshrn_n<mode>): Likewise.
3277 (neon_vqshrun_n<mode>): Likewise.
3278 (neon_vshl_n<mode>): Likewise.
3279 (neon_vqshl_n<mode>): Likewise.
3280 (neon_vqshlu_n<mode>): Likewise.
3281 (neon_vshll_n<mode>): Likewise.
3282 (neon_vsra_n<mode>): Likewise.
3283 (neon_vsri_n<mode>): Likewise.
3284 (neon_vsli_n<mode>): Likewise.
3285 (neon_vtbl1v8qi): Likewise.
3286 (neon_vtbl2v8qi): Likewise.
3287 (neon_vtbl3v8qi): Likewise.
3288 (neon_vtbl4v8qi): Likewise.
3289 (neon_vtbx1v8qi): Likewise.
3290 (neon_vtbx2v8qi): Likewise.
3291 (neon_vtbx3v8qi): Likewise.
3292 (neon_vtbx4v8qi): Likewise.
3293 (neon_vtrn<mode>_internal): Likewise.
3294 (neon_vzip<mode>_internal): Likewise.
3295 (neon_vuzp<mode>_internal): Likewise.
3296 (neon_vld1<mode>): Likewise.
3297 (neon_vld1_lane<mode>): Likewise.
3298 (neon_vld1_dup<mode>): Likewise.
3299 (neon_vld1_dupv2di): Likewise.
3300 (neon_vst1<mode>): Likewise.
3301 (neon_vst1_lane<mode>): Likewise.
3302 (neon_vld2<mode>): Likewise.
3303 (neon_vld2_lane<mode>): Likewise.
3304 (neon_vld2_dup<mode>): Likewise.
3305 (neon_vst2<mode>): Likewise.
3306 (neon_vst2_lane<mode>): Likewise.
3307 (neon_vld3<mode>): Likewise.
3308 (neon_vld3qa<mode>): Likewise.
3309 (neon_vld3qb<mode>): Likewise.
3310 (neon_vld3_lane<mode>): Likewise.
3311 (neon_vld3_dup<mode>): Likewise.
3312 (neon_vst3<mode>): Likewise.
3313 (neon_vst3qa<mode>): Likewise.
3314 (neon_vst3qb<mode>): Likewise.
3315 (neon_vst3_lane<mode>): Likewise.
3316 (neon_vld4<mode>): Likewise.
3317 (neon_vld4qa<mode>): Likewise.
3318 (neon_vld4qb<mode>): Likewise.
3319 (neon_vld4_lane<mode>): Likewise.
3320 (neon_vld4_dup<mode>): Likewise.
3321 (neon_vst4<mode>): Likewise.
3322 (neon_vst4qa<mode>): Likewise.
3323 (neon_vst4qb<mode>): Likewise.
3324 (neon_vst4_lane<mode>): Likewise.
3325 (neon_vec_unpack<US>_lo_<mode>): Likewise.
3326 (neon_vec_unpack<US>_hi_<mode>): Likewise.
3327 (neon_vec_<US>mult_lo_<mode>): Likewise.
3328 (neon_vec_<US>mult_hi_<mode>): Likewise.
3329 (neon_vec_<US>shiftl_<mode>): Likewise.
3330 (neon_unpack<US>_<mode>): Likewise.
3331 (neon_vec_<US>mult_<mode>): Likewise.
3332 (vec_pack_trunc_<mode>): Likewise.
3333 (neon_vec_pack_trunk_<mode>): Likewise.
3334 (neon_vabd<mode>_2): Likewise.
3335 (neon_vabd<mode>_3): Likewise.
3336 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
3337 (thumb2_movsi_vfp): Likewise.
3338 (movdi_vfp): Likewise.
3339 (movdi_vfp_cortexa8): Likewise.
3340 (movhf_vfp_neon): Likewise.
3341 (movhf_vfp): Likewiwse.
3342 (movsf_vfp): Likewiwse.
3343 (thumb2_movsf_vfp): Likewiwse.
3344 (movdf_vfp): Likewise.
3345 (thumb2_movdf_vfp): Likewise.
3346 (movsfcc_vfp): Likewise.
3347 (thumb2_movsfcc_vfp): Likewise.
3348 (movdfcc_vfp): Likewise.
3349 (thumb2_movdfcc_vfp): Likewise.
3350 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
3351 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
3352 (v10_v2c): Likewise.
3353 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
3355 (cortex_a15_neon_int_2): Likewise.
3356 (cortex_a15_neon_int_3): Likewise.
3357 (cortex_a15_neon_int_4): Likewise.
3358 (cortex_a15_neon_int_5): Likewise.
3359 (cortex_a15_neon_vqneg_vqabs): Likewise.
3360 (cortex_a15_neon_vmov): Likewise.
3361 (cortex_a15_neon_vaba): Likewise.
3362 (cortex_a15_neon_vaba_qqq): Likewise.
3363 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3364 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
3365 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
3366 scalar_64_32_long_scalar): Likewise.
3367 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3368 (cortex_a15_neon_mla_qqq_8_16): Likewise.
3369 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
3370 lotype_qdd_64_32_long): Likewise.
3371 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
3372 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
3373 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
3374 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
3375 (cortex_a15_neon_shift_1): Likewise.
3376 (cortex_a15_neon_shift_2): Likewise.
3377 (cortex_a15_neon_shift_3): Likewise.
3378 (cortex_a15_neon_vshl_ddd): Likewise.
3379 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
3380 (cortex_a15_neon_vsra_vrsra): Likewise.
3381 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
3382 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
3383 (cortex_a15_neon_fp_vmul_ddd): Likewise.
3384 (cortex_a15_neon_fp_vmul_qqd): Likewise.
3385 (cortex_a15_neon_fp_vmla_ddd): Likewise.
3386 (cortex_a15_neon_fp_vmla_qqq): Likewise.
3387 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
3388 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
3389 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
3390 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
3391 (cortex_a15_neon_bp_simple): Likewise.
3392 (cortex_a15_neon_bp_2cycle): Likewise.
3393 (cortex_a15_neon_bp_3cycle): Likewise.
3394 (cortex_a15_neon_vld1_1_2_regs): Likewise.
3395 (cortex_a15_neon_vld1_3_4_regs): Likewise.
3396 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
3397 (cortex_a15_neon_vld2_4_regs): Likewise.
3398 (cortex_a15_neon_vld3_vld4): Likewise.
3399 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
3400 (cortex_a15_neon_vst1_3_4_regs): Likewise.
3401 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
3402 (cortex_a15_neon_vst3_vst4): Likewise.
3403 (cortex_a15_neon_vld1_vld2_lane): Likewise.
3404 (cortex_a15_neon_vld3_vld4_lane" 10
3405 (cortex_a15_neon_vst1_vst2_lane): Likewise.
3406 (cortex_a15_neon_vst3_vst4_lane): Likewise.
3407 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
3408 (cortex_a15_neon_ldm_2): Likewise.0
3409 (cortex_a15_neon_stm_2): Likewise.
3410 (cortex_a15_neon_mcr): Likewise.
3411 (cortex_a15_neon_mcr_2_mcrr): Likewise.
3412 (cortex_a15_neon_mrc): Likewise.
3413 (cortex_a15_neon_mrrc): Likewise.
3414 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
3416 (cortex_a15_alu_shift): Likewise.
3417 (cortex_a15_alu_shift_reg): Likewise.
3418 (cortex_a15_mult32): Likewise.
3419 (cortex_a15_mult64): Likewise.
3420 (cortex_a15_block): Likewise.
3421 (cortex_a15_branch): Likewise.
3422 (cortex_a15_load1): Likewise.
3423 (cortex_a15_load3): Likewise.
3424 (cortex_a15_store1): Likewise.
3425 (cortex_a15_store3): Likewise.
3426 (cortex_a15_call): Likewise.
3427 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
3428 (cortex_a5_f2r): Likewise.
3429 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
3431 (cortex_a53_f2r): Likewise.
3432 * config/arm/cortex-a7.md
3433 (cortex_a7_branch): Update for attribute change.
3434 (cortex_a7_call): Likewise.
3435 (cortex_a7_alu_imm): Likewise.
3436 (cortex_a7_alu_reg): Likewise.
3437 (cortex_a7_alu_shift): Likewise.
3438 (cortex_a7_mul): Likewise.
3439 (cortex_a7_load1): Likewise.
3440 (cortex_a7_store1): Likewise.
3441 (cortex_a7_load2): Likewise.
3442 (cortex_a7_store2): Likewise.
3443 (cortex_a7_load3): Likewise.
3444 (cortex_a7_store3): Likewise.
3445 (cortex_a7_load4): Likewise.
3446 (cortex_a7_store4): Likewise.
3447 (cortex_a7_fpalu): Likewise.
3448 (cortex_a7_fconst): Likewise.
3449 (cortex_a7_fpmuls): Likewise.
3450 (cortex_a7_neon_mul): Likewise.
3451 (cortex_a7_fpmacs): Likewise.
3452 (cortex_a7_neon_mla: Likewise.
3453 (cortex_a7_fpmuld: Likewise.
3454 (cortex_a7_fpmacd: Likewise.
3455 (cortex_a7_fpfmad: Likewise.
3456 (cortex_a7_fdivs: Likewise.
3457 (cortex_a7_fdivd: Likewise.
3458 (cortex_a7_r2f: Likewise.
3459 (cortex_a7_f2r: Likewise.
3460 (cortex_a7_f_flags: Likewise.
3461 (cortex_a7_f_loads: Likewise.
3462 (cortex_a7_f_loadd: Likewise.
3463 (cortex_a7_f_stores: Likewise.
3464 (cortex_a7_f_stored: Likewise.
3465 (cortex_a7_neon): Likewise.
3466 * config/arm/cortex-a8-neon.md
3467 (cortex_a8_neon_mrc): Update for attribute change.
3468 (cortex_a8_neon_mrrc): Likewise.
3469 (cortex_a8_neon_int_1): Likewise.
3470 (cortex_a8_neon_int_2): Likewise.
3471 (cortex_a8_neon_int_3): Likewise.
3472 (cortex_a8_neon_int_4): Likewise.
3473 (cortex_a8_neon_int_5): Likewise.
3474 (cortex_a8_neon_vqneg_vqabs): Likewise.
3475 (cortex_a8_neon_vmov): Likewise.
3476 (cortex_a8_neon_vaba): Likewise.
3477 (cortex_a8_neon_vaba_qqq): Likewise.
3478 (cortex_a8_neon_vsma): Likewise.
3479 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3480 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
3481 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
3482 long_scalar): Likewise.
3483 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3484 (cortex_a8_neon_mla_qqq_8_16): Likewise.
3485 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
3486 long_scalar_qdd_64_32_long): Likewise.
3487 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
3488 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
3489 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
3490 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
3491 (cortex_a8_neon_shift_1): Likewise.
3492 (cortex_a8_neon_shift_2): Likewise.
3493 (cortex_a8_neon_shift_3): Likewise.
3494 (cortex_a8_neon_vshl_ddd): Likewise.
3495 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
3496 (cortex_a8_neon_vsra_vrsra): Likewise.
3497 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
3498 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
3499 (cortex_a8_neon_fp_vsum): Likewise.
3500 (cortex_a8_neon_fp_vmul_ddd): Likewise.
3501 (cortex_a8_neon_fp_vmul_qqd): Likewise.
3502 (cortex_a8_neon_fp_vmla_ddd): Likewise.
3503 (cortex_a8_neon_fp_vmla_qqq): Likewise.
3504 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
3505 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
3506 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
3507 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
3508 (cortex_a8_neon_bp_simple): Likewise.
3509 (cortex_a8_neon_bp_2cycle): Likewise.
3510 (cortex_a8_neon_bp_3cycle): Likewise.
3511 (cortex_a8_neon_ldr): Likewise.
3512 (cortex_a8_neon_str): Likewise.
3513 (cortex_a8_neon_vld1_1_2_regs): Likewise.
3514 (cortex_a8_neon_vld1_3_4_regs): Likewise.
3515 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
3516 (cortex_a8_neon_vld2_4_regs): Likewise.
3517 (cortex_a8_neon_vld3_vld4): Likewise.
3518 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
3519 (cortex_a8_neon_vst1_3_4_regs): Likewise.
3520 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
3521 (cortex_a8_neon_vst3_vst4): Likewise.
3522 (cortex_a8_neon_vld1_vld2_lane): Likewise.
3523 (cortex_a8_neon_vld3_vld4_lane): Likewise.
3524 (cortex_a8_neon_vst1_vst2_lane): Likewise.
3525 (cortex_a8_neon_vst3_vst4_lane): Likewise.
3526 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
3527 (cortex_a8_neon_mcr): Likewise.
3528 (cortex_a8_neon_mcr_2_mcrr): Likewise.
3529 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
3530 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
3532 (ca9_neon_mrrc): Likewise.
3533 (cortex_a9_neon_int_1): Likewise.
3534 (cortex_a9_neon_int_2): Likewise.
3535 (cortex_a9_neon_int_3): Likewise.
3536 (cortex_a9_neon_int_4): Likewise.
3537 (cortex_a9_neon_int_5): Likewise.
3538 (cortex_a9_neon_vqneg_vqabs): Likewise.
3539 (cortex_a9_neon_vmov): Likewise.
3540 (cortex_a9_neon_vaba): Likewise.
3541 (cortex_a9_neon_vaba_qqq): Likewise.
3542 (cortex_a9_neon_vsma): Likewise.
3543 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3544 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
3545 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
3546 long_scalar): Likewise.
3547 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3548 (cortex_a9_neon_mla_qqq_8_16): Likewise.
3549 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
3550 long_scalar_qdd_64_32_long): Likewise.
3551 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
3552 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
3553 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
3554 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
3555 (cortex_a9_neon_shift_1): Likewise.
3556 (cortex_a9_neon_shift_2): Likewise.
3557 (cortex_a9_neon_shift_3): Likewise.
3558 (cortex_a9_neon_vshl_ddd): Likewise.
3559 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
3560 (cortex_a9_neon_vsra_vrsra): Likewise.
3561 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
3562 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
3563 (cortex_a9_neon_fp_vsum): Likewise.
3564 (cortex_a9_neon_fp_vmul_ddd): Likewise.
3565 (cortex_a9_neon_fp_vmul_qqd): Likewise.
3566 (cortex_a9_neon_fp_vmla_ddd): Likewise.
3567 (cortex_a9_neon_fp_vmla_qqq): Likewise.
3568 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
3569 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
3570 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
3571 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
3572 (cortex_a9_neon_bp_simple): Likewise.
3573 (cortex_a9_neon_bp_2cycle): Likewise.
3574 (cortex_a9_neon_bp_3cycle): Likewise.
3575 (cortex_a9_neon_ldr): Likewise.
3576 (cortex_a9_neon_str): Likewise.
3577 (cortex_a9_neon_vld1_1_2_regs): Likewise.
3578 (cortex_a9_neon_vld1_3_4_regs): Likewise.
3579 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
3580 (cortex_a9_neon_vld2_4_regs): Likewise.
3581 (cortex_a9_neon_vld3_vld4): Likewise.
3582 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
3583 (cortex_a9_neon_vst1_3_4_regs): Likewise.
3584 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
3585 (cortex_a9_neon_vst3_vst4): Likewise.
3586 (cortex_a9_neon_vld1_vld2_lane): Likewise.
3587 (cortex_a9_neon_vld3_vld4_lane): Likewise.
3588 (cortex_a9_neon_vst1_vst2_lane): Likewise.
3589 (cortex_a9_neon_vst3_vst4_lane): Likewise.
3590 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
3591 (cortex_a9_neon_mcr): Likewise.
3592 (cortex_a9_neon_mcr_2_mcrr): Likewise.
3593 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
3594 (cortex_a9_fps): Likewise.
3595 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
3597 (cortex_m4_fmuls): Likewise.
3598 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
3600 (cortex_r4_mrc): Likewise.
3601 * config/arm/iterators.md: Update comment referring to neon_type.
3602 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
3603 (iwmmxt_movsi_insn): Likewise.
3604 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
3606 (pj4_core_to_vfp): Likewise.
3607 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
3609 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
3610 (vfp_fstore): Likewise.
3611 * doc/md.texi: Change references to neon_type to refer to type.
3613 2013-09-04 Dodji Seketeli <dodji@redhat.com>
3615 * tree.h (DECL_BUILT_IN): Fix typo in comment.
3617 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
3619 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
3622 2013-09-04 Easwaran Raman <eraman@google.com>
3625 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
3626 (build_and_add_sum): Use it.
3627 (appears_later_in_bb): Simplify code.
3629 2013-09-04 Teresa Johnson <tejohnson@google.com>
3631 * dumpfile.c (dump_finish): Don't close stderr/stdout.
3633 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
3635 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
3637 2013-09-04 Jan Hubicka <jh@suse.cz>
3639 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
3640 * ipa-devirt.c: Include diganostic.h
3641 (odr_type_d): Add types and types_set.
3642 (hash_type_name): Work for types with vtables during LTO.
3643 (odr_hasher::remove): Fix comment; destroy types_set.
3644 (add_type_duplicate): New function,
3645 (get_odr_type): Use it.
3646 (dump_type_inheritance_graph): Dump type duplicates.
3647 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
3649 * tree.c (types_same_for_odr): Give exact answers on types with
3652 2013-09-04 Dodji Seketeli <dodji@redhat.com>
3654 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
3655 explaining their differences.
3657 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
3659 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
3661 2013-09-03 Jeff Law <law@redhat.com>
3663 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
3664 when not threading through a joiner block. Pass joiner/no joiner
3665 state to register_jump_thread.
3666 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
3667 state from argument rather than implying on path length.
3668 Dump the entire jump thread path into debugging dump.
3669 * tree-flow.h (register_jump_thread): Update prototype.
3671 2013-08-29 Xinliang David Li <davidxl@google.com>
3673 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
3674 Remove a trivial gcc_assert.
3676 2013-08-29 Xinliang David Li <davidxl@google.com>
3678 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
3679 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
3680 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
3681 Delay base decl alignment adjustment.
3682 * tree-vectorizer.c (vect_destroy_datarefs): New function.
3683 * tree-vectorizer.h: New data structure.
3684 (set_dr_misalignment): New function.
3685 (dr_misalignment): Ditto.
3686 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
3687 (vectorizable_load): Ditto.
3688 (ensure_base_align): New function.
3689 (vectorize_loops): Add dbg_cnt support.
3690 (execute_vect_slp): Ditto.
3691 * dbgcnt.def: New debug counter.
3692 * Makefile: New dependency.
3694 2013-09-03 Meador Inge <meadori@codesourcery.com>
3698 2013-08-30 Meador Inge <meadori@codesourcery.com>
3700 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
3702 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
3704 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
3705 function descriptor.
3707 2013-09-03 Richard Biener <rguenther@suse.de>
3709 * tree-affine.c (add_elt_to_tree): Fix association issue,
3710 avoid useless converts and make sure to always return a
3711 properly typed result.
3713 2013-09-03 Richard Biener <rguenther@suse.de>
3716 * fold-const.c (negate_expr_p): Fix division case.
3717 (negate_expr): Likewise.
3719 2013-09-03 Richard Biener <rguenther@suse.de>
3722 * tree-streamer-out.c: Include tm.h.
3723 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
3725 2013-09-03 Jan Hubicka <jh@suse.cz>
3727 * tree-profile.c (tree_profiling): Cleanup CFG when done.
3729 2013-09-03 Alan Modra <amodra@gmail.com>
3731 * config.gcc (powerpc*-*-linux*): Add support for little-endian
3732 multilibs to big-endian target and vice versa.
3733 * config/rs6000/t-linux64: Use := assignment on all vars.
3734 (MULTILIB_EXTRA_OPTS): Remove fPIC.
3735 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
3736 * config/rs6000/t-linux64le: New file.
3737 * config/rs6000/t-linux64bele: New file.
3738 * config/rs6000/t-linux64lebe: New file.
3740 2013-09-02 Jan Hubicka <jh@suse.cz>
3742 * ipa-inline-transform.c (inline_transform): Do not
3743 optimize_inline_calls when not optimizing.
3745 2013-09-02 Jan Hubicka <jh@suse.cz>
3747 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
3748 duplicated nodes for assembler names.
3749 * symtab.c (symtab_unregister_node): Do not attempt to unlink
3750 hard registers from assembler name hash.
3752 2013-09-02 Jan Hubicka <jh@suse.cz>
3754 * ipa-split.c (execute_split_functions): Split externally visible
3755 functions called once.
3757 2013-09-02 Martin Jambor <mjambor@suse.cz>
3760 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
3761 linked list. When finding the correct duplicate, also consider also
3762 the caller in additon to its inlined_to node.
3764 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
3766 * config/aarch64/aarch64-simd-builtins.def
3767 (dup_lane_scalar): Remove.
3768 * config/aarch64/aarch64-simd.md
3769 (aarch64_simd_dup): Add 'w->w' alternative.
3770 (aarch64_dup_lane<mode>): Allow for VALL.
3771 (aarch64_dup_lane_scalar<mode>): Remove.
3772 (aarch64_dup_lane_<vswap_width_name><mode>): New.
3773 (aarch64_get_lane_signed<mode>): Add w->w altenative.
3774 (aarch64_get_lane_unsigned<mode>): Likewise.
3775 (aarch64_get_lane<mode>): Likewise.
3776 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
3777 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
3778 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
3779 (VCON): Change container of V2SF.
3780 (vswap_width_name): Likewise.
3781 * config/aarch64/arm_neon.h
3782 (__aarch64_vdup_lane_any): New.
3783 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
3784 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
3785 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
3787 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
3790 * expr.c (emit_move_complex): Do not move complex FP values as parts if
3791 the source or the destination is a single hard register.
3793 2013-09-02 Richard Biener <rguenther@suse.de>
3796 * tree-scalar-evolution.c (instantiate_scev_name): Allow
3799 2013-09-02 Richard Biener <rguenther@suse.de>
3801 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
3802 arithmetic to sizetype.
3804 2013-09-02 Bin Cheng <bin.cheng@arm.com>
3806 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
3807 Find auto-increment use both before and after candidate.
3809 2013-09-02 Marek Polacek <polacek@redhat.com>
3811 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
3813 2013-09-01 Jan Hubicka <jh@suse.cz>
3815 * Makefile.in: Add ipa-profile.o
3816 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
3817 * cgraph.c (struct cgraph_propagate_frequency_data,
3818 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
3819 ipa-profile.c; replace cgraph_ by ipa_ prefix.
3820 * cgraph.h (cgraph_propagate_frequency): Remove.
3821 * ipa-inline-analysis.c: Include ipa-utils.h;
3822 drop duplicated cfgloop.h.
3823 (inline_update_callee_summaries): Update.
3824 * ipa-profile.c: New file.
3825 * ipa-utils.h (ipa_propagate_frequency): Declare.
3826 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
3827 data-streamer.h, value-prof.h.
3828 (symtab_remove_unreachable_nodes): Update profile.
3829 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
3830 account_time_size, cmp_counts, dump_histogram,
3831 ipa_profile_generate_summary, ipa_profile_write_summary,
3832 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
3833 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
3834 Move to ipa-profile.c.
3836 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
3838 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
3840 2013-09-01 Jan Hubicka <jh@suse.cz>
3842 * common.opt (fdevirtualize-speculatively): New function.
3843 * invoke.texi (fdevirtualize-speculatively): Document.
3844 * ipa-devirt.c: Include ipa-inline.h
3845 (likely_target_p): New function.
3846 (ipa_devirt): New function.
3847 (gate_ipa_devirt): New function.
3848 (pass_data_ipa_devirt): New static var.
3849 (pass_ipa_devirt): Likewise.
3850 (make_pass_ipa_devirt): New function.
3851 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
3852 (common_handle_option): Disable devirtualization when
3853 value range profiling is available.
3854 * passes.def (pass_ipa_devirt): Add.
3855 * timever.def (TV_IPA_DEVIRT): New timevar.
3856 * tree-pass.h (make_pass_ipa_devirt):
3858 2013-09-01 Iain Sandoe <iain@codesourcery.com>
3860 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
3861 include sanitize(undefined).
3863 2013-08-31 Diego Novillo <dnovillo@google.com>
3865 * Makefile.in (TREE_CORE_H): Define.
3867 (GTFILES): Add tree-core.h.
3868 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
3870 * tree-core.h: New file.
3871 Move all data structures, enum, typedefs, global
3872 declarations and constants from ...
3875 2013-08-31 Jan Hubicka <jh@suse.cz>
3877 * bulitins.c (expand_builtin): Do not early exit for gcov
3878 instrumented functions.
3880 2013-08-31 Marek Polacek <polacek@redhat.com>
3882 * ubsan.c: Include tm_p.h.
3884 2013-08-31 Jan Hubicka <jh@suse.cz>
3886 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
3889 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
3890 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
3891 * tree-cfg.c (verify_gimple_label): ... here.
3892 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
3893 (ipa_merge_profiles): New function.
3894 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
3895 (lto_input_function_body): Likewise.
3896 * ipa-utils.h (ipa_merge_profiles): Declare.
3897 * lto-streamer.h (lto_input_function_body): Update prototype.
3898 (emit_label_in_global_context_p): Remove.
3899 * lto-symtab.c: Include ipa-utils.h
3900 (lto_cgraph_replace_node): Use ipa_merge_profiles.
3902 2013-08-31 Jan Hubicka <jh@suse.cz>
3904 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
3906 2013-08-31 Jan Hubicka <jh@suse.cz>
3908 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
3910 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
3912 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
3913 "cmp" RTX before signed_comparison_operator check to account
3916 2013-08-30 Jan Hubicka <jh@suse.cz>
3918 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
3919 (detect_type_change_1): Rename to ...
3920 (detect_type_change): ... this one; early return on non-polymorphic
3922 (detect_type_change_ssa): Add comp_type parameter; update
3923 use of detect_type_change.
3924 (compute_complex_assign_jump_func): Add param_type parameter;
3925 update use of detect_type_change_ssa.
3926 (compute_complex_ancestor_jump_func): Likewise.
3927 (ipa_get_callee_param_type): New function.
3928 (ipa_compute_jump_functions_for_edge): Compute parameter type;
3929 update calls to the jump function computation functions.
3931 2013-08-30 Teresa Johnson <tejohnson@google.com>
3932 Steven Bosscher <steven@gcc.gnu.org>
3934 * cfgrtl.c (fixup_new_cold_bb): New routine.
3935 (commit_edge_insertions): Invoke fixup_partitions.
3936 (find_partition_fixes): New routine.
3937 (fixup_partitions): Ditto.
3938 (verify_hot_cold_block_grouping): Update comments.
3939 (rtl_verify_edges): Invoke find_partition_fixes.
3940 (rtl_verify_bb_pointers): Update comments.
3941 (rtl_verify_bb_layout): Ditto.
3942 * basic-block.h (probably_never_executed_edge_p): Declare.
3943 (fixup_partitions): Ditto.
3944 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
3945 * bb-reorder.c (sanitize_hot_paths): New function.
3946 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
3948 * predict.c (probably_never_executed_edge_p): New routine.
3949 * cfg.c (check_bb_profile): Add partition insanity warnings.
3951 2013-08-30 Meador Inge <meadori@codesourcery.com>
3953 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
3955 2013-08-30 Marek Polacek <polacek@redhat.com>
3957 * Makefile.in (ubsan.o): Add.
3958 (c-family/c-ubsan.o): Add.
3959 (builtins.o): Add ubsan.h dependency.
3960 * ubsan.h: New file.
3961 * ubsan.c: New file.
3962 * common.opt: Add -fsanitize=undefined option.
3963 (flag_sanitize): Add variable.
3964 (fsanitize=): Add option. Add Driver.
3965 (fsanitize=thread): Remove option.
3966 (fsanitize=address): Likewise.
3967 (static-libubsan): New option.
3968 * doc/invoke.texi: Document the new flag and -static-libubsan.
3969 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
3970 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
3971 * builtin-attrs.def (ATTR_COLD): Define.
3972 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
3973 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
3974 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
3975 * flag-types.h (sanitize_code): New enum.
3976 * opts.c (common_handle_option): Parse command line arguments
3977 of -fsanitize=. Add -fsanitize=unreachable option.
3978 * varasm.c (get_variable_section): Adjust.
3979 (assemble_noswitch_variable): Likewise.
3980 (assemble_variable): Likewise.
3981 (output_constant_def_contents): Likewise.
3982 (categorize_decl_for_section): Likewise.
3983 (place_block_symbol): Likewise.
3984 (output_object_block): Likewise.
3985 * builtins.def: Likewise.
3986 * toplev.c (compile_file): Likewise.
3987 (process_options): Likewise.
3988 * cppbuiltin.c: Likewise.
3989 * tsan.c (tsan_pass): Likewise.
3990 (tsan_gate): Likewise.
3991 (tsan_gate_O0): Likewise.
3992 * cfgexpand.c (partition_stack_vars): Likewise.
3993 (expand_stack_vars): Likewise.
3994 (defer_stack_allocation): Likewise.
3995 (expand_used_vars): Likewise.
3996 * cfgcleanup.c (old_insns_match_p): Likewise.
3997 * asan.c (asan_finish_file): Likewise.
3998 (asan_instrument): Likewise.
3999 (gate_asan): Likewise.
4000 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
4001 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
4002 (asan_global_struct): Use pointer_sized_int_node instead
4003 calling build_nonstandard_integer_type.
4004 (initialize_sanitizer_builtins): Likewise.
4005 (asan_finish_file): Likewise.
4006 * gcc.c: Document %{%:function(args):X}.
4007 (static_spec_functions): Add sanitize.
4008 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
4009 store funcval != NULL there.
4010 (do_spec_1): Adjust handle_spec_function caller.
4011 (handle_braces): Allow %:function(args) as condition.
4012 (sanitize_spec_function): New function.
4013 (ADD_STATIC_LIBUBSAN_LIBS): Define.
4014 (LIBUBSAN_SPEC): Likewise.
4015 (LIBUBSAN_EARLY_SPEC): Likewise.
4016 (SANITIZER_SPEC): Handle libubsan.
4017 (SANITIZER_EARLY_SPEC): Likewise.
4018 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
4019 instead of fsanitize=address.
4020 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
4021 instead of fsanitize=address*.
4022 * builtins.c: Include ubsan.h.
4023 (fold_builtin_0): Instrument __builtin_unreachable.
4024 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
4025 instead of flag_asan.
4026 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
4027 (pointer_sized_int_node): Define.
4028 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
4030 2013-08-30 Mike Stump <mikestump@comcast.net>
4032 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
4035 2013-08-29 Jan Hubicka <jh@suse.cz>
4037 * cgraph.c (cgraph_function_body_availability): Handle weakref
4039 * passes.def: Remove pass_fixup_cfg.
4040 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
4041 track when we need to remove functions.
4042 (gate_ipa_inline): Execute inlining always; add comment why.
4043 (pass_data_ipa_inline): Remove TODO_remove_functions.
4044 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
4045 do not produce summaries.
4046 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
4047 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
4048 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
4049 constant pool and vtable.
4051 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
4053 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
4054 New arm_neon.h's internal macros to specify 64-bit constants.
4055 Avoid using stdint.h's macros.
4057 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
4059 * recog.c (verify_changes): Verify that changes[i].old is non-zero
4060 before applying REG_P.
4062 2013-08-30 Jakub Jelinek <jakub@redhat.com>
4064 PR tree-optimization/58277
4065 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
4066 after seeing too many stmts with vdef in between dombb and current
4067 bb, invalidate everything.
4069 2013-08-30 Richard Biener <rguenther@suse.de>
4071 * fold-const.c (fold_single_bit_test): Fix overflow test.
4073 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
4075 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
4076 and which can live in a register, always retrieve the value on entry.
4077 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
4078 passed by invisible reference specially.
4079 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
4080 (vt_add_function_parameter): Correctly deal with a parameter passed by
4081 invisible reference.
4083 2013-08-30 Jan Hubicka <jh@suse.cz>
4085 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
4087 2013-08-30 Richard Biener <rguenther@suse.de>
4089 PR tree-optimization/58228
4090 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
4091 allow invariant loads in nested loop vectorization.
4093 2013-08-30 Richard Biener <rguenther@suse.de>
4095 PR tree-optimization/58223
4096 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
4097 (has_anti_or_output_dependence): ... this and adjust to also
4098 look for output dependences.
4099 (mark_nodes_having_upstream_mem_writes): Adjust.
4100 (rdg_flag_uses): Likewise.
4102 2013-08-30 Richard Biener <rguenther@suse.de>
4104 PR tree-optimization/58010
4105 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
4106 assert that we have a loop-closed PHI.
4108 2013-08-29 Jan Hubicka <jh@suse.cz>
4110 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
4111 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
4112 * lto-section-in.c (lto_free_function_in_decl_state): New function.
4113 (lto_free_function_in_decl_state_for_node): New function.
4115 2013-08-29 Xinliang David Li <davidxl@google.com>
4117 * loop-unroll.c (report_unroll_peel): Minor message change.
4118 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
4119 Emit alignment peeling message with default -fopt-info.
4120 (vect_loop_versioning): Emit loop version info message.
4121 * tree-vectorizer.c (vectorize_loops): Minor message change.
4122 (execute_vect_slp): Ditto.
4124 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
4126 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
4127 of the clone from the DECL_NAME of the original function.
4129 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
4131 * passes.c (register_pass): Add overload.
4132 * tree-pass.h (register_pass): Forward declare it. Add comment.
4134 2013-08-29 Jan Hubicka <jh@suse.cz>
4136 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
4137 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
4138 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
4139 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
4140 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
4141 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
4142 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
4143 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
4145 2013-08-29 Teresa Johnson <tejohnson@google.com>
4147 * dumpfile.c (dump_loc): Output column number.
4148 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
4149 * doc/invoke.texi: Document optall -fopt-info flag.
4150 * profile.c (read_profile_edge_counts): Use new dump framework.
4151 (compute_branch_probabilities): Ditto.
4152 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
4153 when pass not in any opt group.
4154 * pass_manager.h (pass_manager::get_pass_profile): New method.
4155 * value-prof.c (check_counter): Use new dump framework.
4156 (check_ic_target): Ditto.
4157 * coverage.c (get_coverage_counts): Ditto.
4158 (coverage_init): Setup new dump framework.
4160 2013-08-29 Richard Biener <rguenther@suse.de>
4162 PR tree-optimization/58246
4163 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
4164 handle the dominance check inside a basic-block.
4166 2013-08-29 Richard Biener <rguenther@suse.de>
4169 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
4170 of default defs that appear in abnormal PHI nodes.
4172 2013-08-29 Richard Biener <rguenther@suse.de>
4174 PR tree-optimization/57685
4175 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
4176 single-use operands to avoid exponential complexity.
4178 2013-08-28 Dehao Chen <dehao@google.com>
4180 * ipa-inline.c (edge_badness): Fix integer underflow.
4182 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
4184 * gtm-builtins.def (_ITM_free): Declare leaf.
4186 2013-08-28 Jakub Jelinek <jakub@redhat.com>
4189 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
4190 (*tls_local_dynamic_base_64_largepic): Likewise.
4191 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
4192 Remove predicate from call operand.
4193 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
4194 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
4196 2013-08-28 Jeff Law <law@redhat.com>
4198 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
4199 checks for the number of predecessors and successors allowed.
4200 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
4201 which require copying a joiner block if there is a request which
4202 is a subpath that requires no joiner block copying.
4204 2013-08-28 Jan Hubicka <jh@suse.cz>
4206 * lto-streamer-out.c (DFS_write_tree_body): Drop
4207 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
4208 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
4209 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
4210 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
4211 Do not read DECL_ERROR_ISSUED.
4212 (unpack_ts_decl_with_vis_value_fields): Do not read
4214 (lto_input_ts_binfo_tree_pointers): Do not read
4215 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
4216 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
4217 write DECL_ERROR_ISSUED..
4218 (pack_ts_decl_with_vis_value_fields): Do not write
4220 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
4221 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
4222 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
4223 * tree.h (tree_decl_common): Update comment.
4224 (DECL_ERROR_ISSUED): Remove.
4226 2013-08-28 Jakub Jelinek <jakub@redhat.com>
4229 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
4231 2013-08-28 Jan Hubicka <jh@suse.cz>
4233 * builtins.def (free): Declare leaf.
4235 2013-08-27 David Malcolm <dmalcolm@redhat.com>
4238 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
4239 * configure: Regenerate.
4241 2013-08-27 Martin Jambor <mjambor@suse.cz>
4243 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
4244 (ipa_ancestor_jf_data): Likewise.
4245 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
4246 (ipa_get_jf_pass_through_type_preserved): New function.
4247 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
4248 (ipa_get_jf_ancestor_type_preserved): New function.
4249 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
4250 (ipa_get_jf_ancestor_result): Likewise.
4251 (propagate_vals_accross_pass_through): Use
4252 ipa_get_jf_pass_through_result to do all the value mappings.
4253 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
4254 type_preserved flag.
4255 (ipa_set_jf_cst_copy): New function.
4256 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
4257 (ipa_set_jf_arith_pass_through): Likewise.
4258 (ipa_set_ancestor_jf): Likewise.
4259 (compute_complex_assign_jump_func): Set type_preserved instead of
4261 (ipa_compute_jump_functions_for_edge): Likewise.
4262 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
4263 (update_jump_functions_after_inlining): Update type_preserved.
4264 Explicitely create jump functions when combining one with pass_through.
4265 (ipa_write_jump_function): Stream the type_preserved flags.
4266 (ipa_read_jump_function): Likewise.
4268 2013-08-27 Jakub Jelinek <jakub@redhat.com>
4269 Aldy Hernandez <aldyh@redhat.com>
4271 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
4272 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
4273 * function.h (struct function): Add has_force_vect_loops and
4275 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
4276 * gimple.c (gimple_build_omp_critical): Add KIND argument and
4278 * gimple.def: Update CLAUSES comments.
4279 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
4280 (gimple_build_omp_for): Add argument to prototype.
4281 (gimple_omp_for_kind): New.
4282 (gimple_omp_for_set_kind): New.
4283 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
4284 GOVD_DATA_SHARE_CLASS.
4285 (enum omp_region_type): Add ORT_SIMD.
4286 (gimple_add_tmp_var): Handle ORT_SIMD.
4287 (gimplify_var_or_parm_decl): Same.
4288 (is_gimple_stmt): Same.
4289 (omp_firstprivatize_variable): Same.
4290 (omp_add_variable): Only use splay_tree_insert if lookup failed.
4291 (omp_notice_variable): Handle ORT_SIMD.
4292 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
4293 (omp_check_private): Handle ORT_SIMD.
4294 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
4296 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
4297 Handle OMP_CLAUSE_LASTPRIVATE.
4298 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
4300 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
4301 (gimplify_expr): Handle OMP_SIMD.
4302 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
4303 (expand_GOMP_SIMD_VF): New.
4304 (expand_GOMP_SIMD_LAST_LANE): New.
4305 * internal-fn.def (GOMP_SIMD_LANE): New.
4306 (GOMP_SIMD_VF): New.
4307 (GOMP_SIMD_LAST_LANE): New.
4308 * omp-low.c: Include target.h.
4309 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
4311 (check_omp_nesting_restrictions): Same.
4313 (lower_rec_simd_input_clauses): New.
4314 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
4316 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
4317 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
4318 (expand_omp_build_assign): New.
4319 (expand_omp_for_init_counts): New.
4320 (expand_omp_for_init_vars): New.
4321 (extract_omp_for_update_vars): New.
4322 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
4323 and rewrite accordingly.
4324 (expand_omp_simd): New.
4325 (expand_omp_for): Use expand_omp_simd.
4326 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
4327 (lower_omp_for): Do not lower the body.
4328 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
4330 * tree-flow.h (find_omp_clause): Remove prototype.
4331 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
4332 forcing vectorization of the loop, or if flag_tree_vectorize.
4333 (gate_tree_if_conversion): Similarly.
4334 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
4335 gimple_build_omp_for.
4336 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
4337 * tree-parloops (create_parallel_loop): Pass kind argument to
4338 gimple_build_omp_for.
4339 * tree-pretty-print.c (dump_omp_clause): Add cases for
4340 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
4341 OMP_CLAUSE__SIMDUID_.
4342 (dump_generic_node): Handle OMP_SIMD.
4343 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
4344 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
4345 unroll OMP_SIMD loops here.
4346 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
4347 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
4349 (vect_analyze_data_refs): Handle simd loops.
4350 * tree-vect-loop.c (vectorizable_live_operation): Handle
4352 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
4353 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
4354 (vectorizable_load): Same.
4355 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
4356 (struct simduid_to_vf): New.
4357 (simduid_to_vf::hash): New.
4358 (simduid_to-vf::equal): New.
4359 (struct simd_array_to_simduid): New.
4360 (simd_array_to_simduid::hash): New.
4361 (simd_array_to_simduid::equal): New.
4362 (adjust_simduid_builtins): New.
4363 (struct note_simd_array_uses_struct): New.
4364 (note_simd_array_uses_cb): New.
4365 (note_simd_array_uses): New.
4366 (vectorize_loops): Handle simd hints and adjust simd builtins
4368 * tree-vectorizer.h (struct _stmt_vec_info): Add
4369 simd_lane_access_p field.
4370 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
4371 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
4372 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
4373 (omp_clause_code_name): Same.
4374 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
4375 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
4376 * tree.def (OMP_SIMD): New entry.
4377 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
4378 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
4379 (OMP_CLAUSE_DECL): Adjust range for new clauses.
4380 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
4381 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
4382 (OMP_CLAUSE_LINEAR_STEP): New.
4383 (OMP_CLAUSE_SAFELEN_EXPR): New.
4384 (OMP_CLAUSE__SIMDUID__DECL): New.
4385 (find_omp_clause): New prototype.
4387 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
4389 * config/i386/driver-i386.c (host_detect_local_cpu): Update
4390 Haswell processor detection.
4392 2013-08-27 Christian Widmer <shadow@umbrox.de>
4395 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
4396 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
4397 AVX2 capable processors.
4399 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
4401 * config/aarch64/arm_neon.h: Replace all inline asm implementations
4402 of vget_low_* with implementations in terms of other intrinsics.
4404 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
4407 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
4408 values to -1, 0 and 1.
4410 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
4412 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
4413 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
4414 (<optab><mode>3_insn): Remove.
4415 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
4416 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
4417 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
4418 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
4419 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
4420 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
4421 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
4422 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
4423 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
4424 (ror<mode>3_insn): Likewise.
4425 * config/aarch64/predicates.md (aarch64_simd_register): New.
4427 2013-08-27 Richard Biener <rguenther@suse.de>
4429 PR tree-optimization/57521
4430 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
4431 one edge is non-critical.
4432 (find_phi_replacement_condition): Make sure to use a non-critical
4433 edge. Cleanup and remove old bug workarounds.
4434 (bb_postdominates_preds): Remove.
4435 (if_convertible_loop_p_1): Do not compute post-dominators.
4436 (combine_blocks): Do not free post-dominators.
4437 (main_tree_if_conversion): Likewise.
4438 (pass_data_if_conversion): Add TODO_verify_ssa.
4440 2013-08-27 DJ Delorie <dj@redhat.com>
4442 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
4444 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
4446 * function.c (assign_parm_find_data_types): Set passed_mode and
4447 nominal_mode to the TYPE_MODE of nominal_type for the built
4448 pointer type in case of the struct-pass-by-reference.
4450 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
4452 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
4453 (UINT16_TYPE): Change default to "unsigned int".
4455 * config/avr/avr.opt (mfract-convert-truncate): New option.
4456 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
4457 is set, round negative fractional integers according to n1169
4458 when converting to integer types.
4460 2013-08-26 Jan Hubicka <jh@suse.cz>
4462 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
4463 methods can not be called indirectly when their address is not taken.
4465 2013-08-26 Jan Hubicka <jh@suse.cz>
4467 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
4470 2013-08-26 Jan Hubicka <jh@suse.cz>
4472 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
4475 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
4477 * reload.c (find_valid_class): Allow classes that do not include
4478 FIRST_PSEUDO_REGISTER - 1.
4480 2013-08-26 Jan Hubicka <jh@suse.cz>
4482 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
4483 fix edge count/frequency when speculation failed; fix type check
4484 for the direct call.
4486 2013-08-26 Jan Hubicka <jh@suse.cz>
4488 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
4490 2013-08-26 Jan Hubicka <jh@suse.cz>
4492 * ipa-inline-transform.c (inline_transform): Be ready for basic block
4493 to be changed by edge redirection.
4495 2013-08-26 Jan Hubicka <jh@suse.cz>
4497 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
4498 formating; add sanity check.
4499 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
4500 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
4501 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
4502 (resolve_noninline_speculation): Update callee keys, too.
4504 2013-08-26 Jan Hubicka <jh@suse.cz>
4506 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
4507 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
4509 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
4511 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
4514 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
4516 * config/i386/i386.c (ix86_debug_options): Remove prototype.
4517 (x86_64_elf_select_section): Ditto.
4518 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
4520 (ix86_pass_by_reference): Ditto.
4521 (output_set_got): Ditto.
4522 (ix86_unary_operator_ok): Ditto.
4523 (ix86_expand_builtin): Ditto.
4525 2013-08-23 Jan Hubicka <jh@suse.cz>
4527 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
4529 2013-08-23 Jan Hubicka <jh@suse.cz>
4531 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
4532 (tree_decl_with_vis): Add FINAL field.
4534 2013-08-23 Jeff Law <law@redhat.com>
4536 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
4537 the debugging dump when the expression is fully redundant.
4539 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
4541 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
4542 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
4543 * pretty-print.c (pp_formatted_text_data): Likewise.
4544 (pp_write_text_to_stream): Likewise.
4545 (pp_write_text_as_dot_label_to_stream): Likewise.
4546 (pp_append_r): Likewise.
4547 (pp_format): Likewise.
4548 (pp_flush): Likewise.
4549 (pp_clear_output_area): Likewise.
4550 (pp_append_text): Likewise.
4551 (pp_formatted_text): Likewise.
4552 (pp_remaining_character_count_for_line): Likewise.
4553 (pp_newline): Likewise.
4554 (pp_character): Likewise.
4555 (output_buffer::~output_buffer): Define.
4556 (pretty_printer::~pretty_printer): Destruct output buffer.
4557 * pretty-print.h (output_buffer::~output_buffer): Declare.
4558 (pretty_printer::~pretty_printer): Declare virtual.
4560 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
4563 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
4564 HOST_WIDE_INT_M1U): New macros.
4565 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
4566 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
4567 unsigned -1 for lshift.
4568 * cse.c (cse_insn): Likewise.
4569 * double-int.c (rshift_double, lshift_double): Likewise.
4570 * builtins.c (fold_builtin_bitop): Likewise.
4571 * combine.c (force_to_mode): Likewise.
4572 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
4573 * simplify-rtx.c (simplify_const_unary_operation,
4574 simplify_const_binary_operation): Likewise.
4575 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
4576 check_va_list_escapes): Likewise.
4577 * rtlanal.c (nonzero_bits1): Likewise.
4578 * expmed.c (expand_smod_pow2): Likewise.
4579 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
4581 2013-08-23 Jan Hubicka <jh@suse.cz>
4583 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
4584 as having address taken.
4586 2013-08-23 Jan Hubicka <jh@suse.cz>
4588 * ipa-utils.h (method_class_type): Declare.
4589 * ipa-devirt.c (method_class_type): Export.
4591 * cgraphunit.c (analyze_functions): Do basic devirtualization;
4592 do not walk base classes of anonymous types.
4594 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
4596 PR rtl-optimization/58220
4598 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
4599 handle SEQUENCE insns properly.
4601 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
4603 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
4605 (pp_newline_and_indent): Likewise.
4606 (pp_separate_with): Likewise.
4607 * pretty-print.c (pp_newline_and_flush): Define.
4608 (pp_newline_and_indent): Likewise.
4609 (pp_separate_with): Likewise.
4611 2013-08-23 Jakub Jelinek <jakub@redhat.com>
4614 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
4615 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
4617 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
4619 * config/i386/predicates.md (ext_sse_reg_operand): New.
4620 * config/i386/i386.md (*movti_internal): Use
4621 predicate to determine if EVEX is needed.
4622 (*movsi_internal): Ditto.
4623 (*movdf_internal): Ditto.
4624 (*movsf_internal): Ditto.
4625 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
4627 2013-08-23 Jakub Jelinek <jakub@redhat.com>
4629 PR tree-optimization/58209
4630 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
4631 (find_tail_calls): Give up for pointer result types if m is non-NULL.
4632 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
4633 emit POINTER_PLUS_EXPR.
4634 (create_tailcall_accumulator): For pointer result type accumulate in
4637 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
4639 * configure.ac: Add backslashes missing from the last change.
4640 * configure: Regenerate.
4642 2013-08-22 Jan Hubicka <jh@susue.cz>
4644 * ipa.c (function_and_variable_visibility): First remember function
4645 was global and then make it local.
4647 2013-08-22 Julian Brown <julian@codesourcery.com>
4649 * configure.ac: Add aarch64 to list of arches which use "nop" in
4651 * configure: Regenerate.
4653 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4655 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
4656 gnu_libc_has_function.
4657 * config/s390/tpf.h: Likewise.
4659 2013-08-22 Jan Hubicka <jh@susue.cz>
4661 * timevar.c (validate_phases): Add cast.
4663 2013-08-22 Jan Hubicka <jh@susue.cz>
4665 * timevar.c (validate_phases): Use size_t for memory.
4666 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
4668 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
4670 * pretty-print.h (output_buffer::output_buffer): Declare.
4671 (pretty_printer::pretty_printer): Likewise.
4672 (pp_construct): Remove.
4673 * pretty-print.c (output_buffer::output_buffer): Define.
4674 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
4675 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
4676 (print_gimple_expr): Likewise.
4677 (print_gimple_seq): Likewise.
4678 (gimple_dump_bb): Likewise.
4679 * sched-vis.c (dump_value_slim): Likewise.
4680 (dump_insn_slim): Likewise.
4681 (dump_rtl_slim): Likewise.
4682 (str_pattern_slim): Likewise.
4683 * tree-mudflap.c (mf_varname_tree): Likewise.
4684 * graph.c (print_graph_cfg): Likewise.
4685 (start_graph_dump): Likewise.
4686 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
4688 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
4689 pretty printer initialization.
4690 * coretypes.h (diagnostic_context): Remove superflous type alias
4692 (pretty_printer): Likewise. Declare directly as a class.
4693 (pretty_print_info): Remove declaration as class.
4694 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
4695 and pp_clear_output_area.
4696 (asan_add_global): Likewise.
4698 2013-08-22 Jan Hubicka <jh@suse.cz>
4700 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
4701 * ipa-utils.h (update_type_inheritance_graph): Declare.
4702 (possible_polymorphic_call_target_p): Declare.
4703 (possible_polymorphic_call_target_p): New.
4704 * ipa-devirt.c: Update toplevel comments.
4705 (cached_polymorphic_call_targets): Move up.
4706 (odr_type_d): Move ID down.
4707 (polymorphic_type_binfo_p): Update comment.
4708 (odr_hasher::remove): Likewise;
4709 (get_odr_type): Set anonymous_namespace.
4710 (dump_odr_type): Dump it.
4711 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
4712 (maybe_record_node): Record node in cached_polymorphic_call_targets.
4713 (record_binfo): Add comment.
4714 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
4716 (devirt_node_removal_hook): Do not iCE when cache is freed.
4717 (possible_polymorphic_call_target_p): New predicate.
4718 (update_type_inheritance_graph): New function.
4720 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
4721 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4722 Sergey Lega <sergey.s.lega@intel.com>
4723 Anna Tikhonova <anna.tikhonova@intel.com>
4724 Ilya Tocar <ilya.tocar@intel.com>
4725 Andrey Turetskiy <andrey.turetskiy@intel.com>
4726 Ilya Verbin <ilya.verbin@intel.com>
4727 Kirill Yukhin <kirill.yukhin@intel.com>
4728 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4730 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
4731 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
4732 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
4733 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
4734 (OPTION_MASK_ISA_AVX2_UNSET): Update.
4735 (OPTION_MASK_ISA_AVX512F_UNSET): New.
4736 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
4737 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
4738 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
4739 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
4740 OPT_mavx512pf, OPT_mavx512er cases.
4741 * config/i386/constraints.md (v): New constraint.
4742 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
4743 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
4744 (bit_AVX512CD): New.
4745 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
4746 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
4747 * config/i386/i386-c.c (ix86_target_macros_internal):
4748 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
4750 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
4751 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
4752 * config/i386/i386.c (regclass_map, dbx_register_map)
4753 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
4754 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
4755 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
4757 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
4758 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
4759 -mavx512pf options. Fix formatting.
4760 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
4761 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
4762 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
4763 -mavx512cd, -mavx512pf options.
4764 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
4765 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
4766 (ix86_preferred_output_reload_class): Replace SSE_REGS with
4768 (ix86_hard_regno_mode_ok): Support 512-bit registers.
4769 (ix86_set_reg_reg_cost): Ditto.
4770 (x86_order_regs_for_local_alloc): Ditto.
4771 (MAX_VECT_LEN): Extend to 64-byte.
4772 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
4773 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
4774 (TARGET_AVX512ER, TARGET_AVX512CD): New.
4775 (BIGGEST_ALIGNMENT): Extend to 512-bits.
4776 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
4777 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
4778 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
4779 (SSE_REG_MODE_P): Support new modes.
4780 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
4781 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
4782 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
4783 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
4784 (REG_CLASS_CONTENTS): Add new registers.
4785 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
4786 (EXT_REX_SSE_REGNO_P): New.
4787 (HI_REGISTER_NAMES): Add new registers.
4788 * config/i386/i386.md: Define constants for new registers.
4789 (mode): Add new 512-bit modes.
4790 (prefix): Support evex prefix.
4791 (isa): Support avx512f, noavx512f, fma_avx512f.
4792 (ssemodesuffix): Add new 512-bit modes.
4794 (*movxi_internal_avx512f): Ditto.
4795 (*movdi_internal): Replace constraint "x" with the new constraint "v".
4797 (*movsi_internal): Likewise.
4798 (*movdf_internal): Likewise.
4799 (*movsf_internal): Likewise.
4800 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
4801 (<code><mode>3): Likewise.
4802 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
4804 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
4805 with the new constraint "v".
4806 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
4808 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
4809 with the new constraint "v".
4810 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
4811 (<sse2>_storedqu<avxsizesuffix>): Likewise.
4812 (*<plusminus_insn><mode>3): Likewise.
4813 (<sse>_vm<plusminus_insn><mode>3): Likewise.
4814 (*mul<mode>3): Likewise.
4815 (<sse>_vmmul<mode>3): Likewise.
4816 (<sse>_div<mode>3): Likewise.
4817 (<sse>_vmdiv<mode>3): Likewise.
4818 (<sse>_sqrt<mode>2): Likewise.
4819 (<sse>_vmsqrt<mode>2): Likewise.
4820 (*<code><mode>3_finite): Likewise.
4821 (*<code><mode>3) <smaxmin>: Likewise.
4822 (<sse>_vm<code><mode>3): Likewise.
4823 (*<code><mode>3) <any_logic>: Likewise.
4824 (*fma_fmadd_<mode>): Likewise.
4825 (*fma_fmsub_<mode>): Likewise.
4826 (*fma_fnmadd_<mode>): Likewise.
4827 (*fma_fnmsub_<mode>): Likewise.
4828 (*fma_fmaddsub_<mode>): Likewise.
4829 (*fma_fmsubadd_<mode>): Likewise.
4830 (*fmai_fmadd_<mode>): Likewise.
4831 (*fmai_fmsub_<mode>): Likewise.
4832 (*fmai_fnmadd_<mode>): Likewise.
4833 (*fmai_fnmsub_<mode>): Likewise.
4834 (sse_cvtsi2ss): Likewise.
4835 (sse_cvtsi2ssq): Likewise.
4836 (sse_cvtss2si): Likewise.
4837 (sse_cvtss2si_2): Likewise.
4838 (sse_cvtss2siq): Likewise.
4839 (sse_cvtss2siq_2): Likewise.
4840 (sse_cvttss2si): Likewise.
4841 (sse_cvtss2siq_2): Likewise.
4842 (float<sseintvecmodelower><mode>2): Likewise.
4843 (sse2_cvtsd2si_2): Likewise.
4844 (sse2_cvtsd2siq_2): Likewise.
4845 (*<plusminus_insn><mode>3): Likewise.
4846 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
4847 (*<sse4_1_avx2>_mul<mode>3): Likewise.
4848 (ashr<mode>3): Likewise.
4849 (<shift_insn><mode>3): Likewise.
4850 (avx2_<code><mode>3): Likewise.
4851 (*avx2_<code><mode>3): Likewise.
4852 (*andnot<mode>3): Likewise.
4853 (*<code><mode>3) <any_logic>: Likewise.
4854 (abs<mode>2): Likewise.
4855 (avx2_permvar<mode>): Likewise.
4856 (avx2_perm<mode>_1): Likewise.
4857 (*avx_vpermilp<mode>): Likewise.
4858 (avx_vpermilvar<mode>3): Likewise.
4859 (avx2_ashrv<mode>): Likewise.
4860 (avx2_<shift_insn>v<mode>): Likewise.
4861 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
4863 * doc/rtl.texi: Document XImode.
4865 2013-08-21 Jeff Law <law@redhat.com>
4867 * tree-flow.h (register_jump_thread): Pass vector of edges
4868 instead of each important edge.
4869 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
4870 thread path into a vector and pass that to register_jump_thread.
4871 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
4872 passed in edge vector to the current 3-edge form.
4875 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
4877 * dce.c (fini_dce): Call df_analyze again just in case
4878 delete_unmarked_insns removed anything.
4880 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
4882 * reload.h (struct reg_equivs): Rename to ..
4883 (struct reg_equivs_s): .. this.
4885 2013-08-20 Martin Liska <marxin.liska@gmail.com>
4887 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
4889 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4891 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
4893 2013-08-21 Jeff Law <law@redhat.com>
4895 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
4896 simplify assignments too. If the RHS collapses to a singleton
4897 range, then return the value for the range.
4899 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
4901 * config/i386/sse.md (V16): Rename to...
4903 (mov<mode>): Update iterator name.
4904 (*mov<mode>_internal): Ditto.
4905 (push<mode>1): Ditto.
4906 (movmisalign<mode>): Ditto.
4908 2013-08-20 Jan Hubicka <jh@suse.cz>
4911 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
4912 entry for direct edges.
4913 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
4915 2013-08-20 David Malcolm <dmalcolm@redhat.com>
4917 Revert my last two changes, r201865 and r201864:
4920 2013-08-20 David Malcolm <dmalcolm@redhat.com>
4922 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
4923 instances can own GC refs.
4925 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
4926 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
4927 (gcc::context::gt_pch_nx): Likewise.
4928 (gcc::context::gt_pch_nx): Likewise.
4929 * ggc.h (gt_ggc_mx <T>): New.
4930 (gt_pch_nx_with_op <T>): New.
4931 (gt_pch_nx <T>): New.
4932 * passes.c (opt_pass::gt_ggc_mx): New.
4933 (opt_pass::gt_pch_nx): New.
4934 (opt_pass::gt_pch_nx_with_op): New.
4935 (pass_manager::gt_ggc_mx): New.
4936 (pass_manager::gt_pch_nx): New.
4937 (pass_manager::gt_pch_nx_with_op): New.
4938 (pass_manager::operator new): Use
4939 ggc_internal_cleared_alloc_stat rather than xcalloc.
4940 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
4941 (pass_manager::gt_ggc_mx): New.
4942 (pass_manager::gt_pch_nx): New.
4943 (pass_manager::gt_pch_nx_with_op): New.
4944 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
4945 (opt_pass::operator new): New.
4946 (opt_pass::gt_ggc_mx): New.
4947 (opt_pass::gt_pch_nx): New.
4948 (opt_pass::gt_pch_nx_with_op): New.
4951 2013-08-20 David Malcolm <dmalcolm@redhat.com>
4953 * Makefile.in (GTFILES): Add context.h.
4954 * context.c (gcc::context::operator new): New.
4955 (gcc::context::gt_ggc_mx): New.
4956 (gcc::context::gt_pch_nx): New.
4957 (gcc::context::gt_pch_nx): New.
4958 * context.h (gcc::context): Add GTY((user)) marking.
4959 (gcc::context::operator new): New.
4960 (gcc::context::gt_ggc_mx): New.
4961 (gcc::context::gt_pch_nx): New.
4962 (gcc::context::gt_pch_nx): New.
4963 (g): Add GTY marking.
4964 (gt_ggc_mx (gcc::context *)): New.
4965 (gt_pch_nx (gcc::context *)): New.
4966 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
4967 void *cookie)): New.
4968 * gengtype.c (open_base_files) <ifiles>: Add context.h.
4970 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
4972 * dce.c (fini_dce): Call df_analyze again just in case
4973 delete_unmarked_insns removed anything.
4975 2013-08-20 Teresa Johnson <tejohnson@google.com>
4977 PR rtl-optimizations/57451
4978 * final.c (reemit_insn_block_notes): Prevent lexical blocks
4979 from crossing split section boundaries.
4981 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
4983 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
4984 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
4985 with MULTLIB_DEFAULTS.
4987 2013-08-20 Nick Clifton <nickc@redhat.com>
4989 * target.def (narrow_volatile_bitfield): Note that the default
4990 value is false, not !TARGET_STRICT_ALIGN.
4991 * doc/tm.texi: Regenerate.
4993 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
4995 Fix LIB_SPEC for systems without libpthread.
4997 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
4998 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
5000 * config/i386/linux-common.h: Likewise.
5001 * config/mips/linux-common.h: Likewise.
5003 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
5005 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
5008 2013-08-20 David Malcolm <dmalcolm@redhat.com>
5010 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
5011 instances can own GC refs.
5013 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
5014 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
5015 (gcc::context::gt_pch_nx): Likewise.
5016 (gcc::context::gt_pch_nx): Likewise.
5017 * ggc.h (gt_ggc_mx <T>): New.
5018 (gt_pch_nx_with_op <T>): New.
5019 (gt_pch_nx <T>): New.
5020 * passes.c (opt_pass::gt_ggc_mx): New.
5021 (opt_pass::gt_pch_nx): New.
5022 (opt_pass::gt_pch_nx_with_op): New.
5023 (pass_manager::gt_ggc_mx): New.
5024 (pass_manager::gt_pch_nx): New.
5025 (pass_manager::gt_pch_nx_with_op): New.
5026 (pass_manager::operator new): Use
5027 ggc_internal_cleared_alloc_stat rather than xcalloc.
5028 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
5029 (pass_manager::gt_ggc_mx): New.
5030 (pass_manager::gt_pch_nx): New.
5031 (pass_manager::gt_pch_nx_with_op): New.
5032 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
5033 (opt_pass::operator new): New.
5034 (opt_pass::gt_ggc_mx): New.
5035 (opt_pass::gt_pch_nx): New.
5036 (opt_pass::gt_pch_nx_with_op): New.
5038 2013-08-20 David Malcolm <dmalcolm@redhat.com>
5040 * Makefile.in (GTFILES): Add context.h.
5041 * context.c (gcc::context::operator new): New.
5042 (gcc::context::gt_ggc_mx): New.
5043 (gcc::context::gt_pch_nx): New.
5044 (gcc::context::gt_pch_nx): New.
5045 * context.h (gcc::context): Add GTY((user)) marking.
5046 (gcc::context::operator new): New.
5047 (gcc::context::gt_ggc_mx): New.
5048 (gcc::context::gt_pch_nx): New.
5049 (gcc::context::gt_pch_nx): New.
5050 (g): Add GTY marking.
5051 (gt_ggc_mx (gcc::context *)): New.
5052 (gt_pch_nx (gcc::context *)): New.
5053 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
5054 void *cookie)): New.
5055 * gengtype.c (open_base_files) <ifiles>: Add context.h.
5057 2013-08-20 Alan Modra <amodra@gmail.com>
5060 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
5061 (rs6000_emit_epilogue): Likewise.
5063 2013-08-19 Dehao Chen <dehao@google.com>
5065 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
5067 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
5068 Jakub Jelinek <jakub@redhat.com>
5070 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
5071 (BUILT_IN_FABSD64): Likewise.
5072 (BUILT_IN_FABSD128): Likewise.
5073 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
5074 (fold_builtin_1): Likewise.
5075 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
5076 destination and source operands.
5077 (*abstd2_fpr): Likewise.
5078 (*nabstd2_fpr): Likewise.
5080 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
5082 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
5085 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
5087 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
5090 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
5092 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
5093 * builtins.c (default_libc_has_function): New.
5094 (gnu_libc_has_function): Ditto.
5095 (no_c99_libc_has_function): Ditto.
5096 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
5097 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
5098 (fold_builtin_sincos): Likewise.
5099 (fold_builtin_cexp): Likewise.
5100 * builtins.def (DEF_C94_BUILTIN): Likewise.
5101 (DEF_C99_BUILTIN): Likewise.
5102 (DEF_C99_C90RES_BUILTIN): Likewise.
5103 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
5104 definitions to using this define.
5105 * config/darwin-protos.h (darwin_libc_has_function): New.
5106 * config/darwin.c (darwin_libc_has_function): Ditto.
5107 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
5108 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
5109 * config/darwin.h: Ditto.
5110 * config/elfos.h: Ditto.
5111 * config/freebsd.h: Ditto.
5112 * config/i386/cygming.h: Ditto.
5113 * config/i386/djgpp.h: Ditto.
5114 * config/i386/i386-interix.h: Ditto.
5115 * config/microblaze/microblaze.h: Ditto.
5116 * config/mmix/mmix.h: Ditto.
5117 * config/gnu-user.h: Ditto.
5118 * config/ia64/hpux.h: Ditto.
5119 * config/pa/pa-hpux.h: Ditto.
5120 * config/pdp11/pdp11.h: Ditto.
5121 * config/picochip/picochip.h: Ditto.
5122 * config/linux.h: Ditto.
5123 * config/netbsd.h: Ditto.
5124 * config/openbsd.h: Ditto.
5125 * config/rs6000/aix43.h: Ditto.
5126 * config/rs6000/aix51.h: Ditto.
5127 * config/rs6000/aix52.h: Ditto.
5128 * config/rs6000/aix53.h: Ditto.
5129 * config/rs6000/aix61.h: Ditto.
5130 * config/rs6000/darwin.h: Ditto.
5131 * config/rs6000/linux.h: Ditto.
5132 * config/rs6000/linux64.h: Ditto.
5133 * config/s390/tpf.h: Ditto.
5134 * config/sol2-10.h: Ditto.
5135 * config/sol2.h: Ditto.
5136 * config/vms/vms.h: Ditto.
5137 * config/vxworks.h: Ditto.
5138 * config/linux-android.c (linux_android_libc_has_function):
5139 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
5140 * config/linux-protos.h (linux_android_libc_has_function):
5142 * config/i386/i386.c (ix86_libc_has_function): New.
5143 * config/i386/i386-protos.h
5144 (ix86_libc_has_function): New declaration.
5145 * config/i386/i386.md
5146 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
5147 ("isinf<mode>2): Likewise.
5148 * convert.c (convert_to_integer): Using new target hook
5149 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
5150 TARGET_C99_FUNCTIONS.
5151 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
5152 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
5153 * coretypes.h (function_class): New enum for different
5154 classes of functions.
5155 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
5156 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
5157 (TARGET_HAS_SINCOS): Likewise.
5158 (TARGET_LIBC_HAS_FUNCTION): New.
5159 * doc/tm.texi: Regenerated.
5160 * targhooks.h (default_libc_has_function): New declaration.
5161 (no_c99_libc_has_function): Ditto.
5162 (gnu_libc_has_function): Ditto.
5163 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
5164 and TARGET_HAS_SINCOS.
5166 2013-08-18 Jan Hubicka <jh@suse.cz>
5168 * Makeifle-in (ipa-devirt.o): New.
5169 (GTFILES): Add ipa-utils.h and ipa-devirt.c
5170 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
5171 (analyze_functions): Look into possible targets of polymorphic call.
5172 * dumpfile.c (dump_files): Add type-inheritance dump.
5173 * dumpfile.h (TDI_inheritance): New.
5174 * ipa-devirt.c: New file.
5175 * ipa-utils.h (odr_type_d): Forward declare.
5176 (odr_type): New type.
5177 (build_type_inheritance_graph): Declare.
5178 (possible_polymorphic_call_targets): Declare and introduce inline
5179 variant when only edge is pased.
5180 (dump_possible_polymorphic_call_targets): Likewise.
5181 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
5182 * tree.c (type_in_anonymous_namespace_p): Break out from ...
5183 (types_same_for_odr): ... here.
5184 * tree.h (type_in_anonymous_namespace_p): Declare.
5186 2013-08-18 Jakub Jelinek <jakub@redhat.com>
5188 PR tree-optimization/58006
5189 * tree-parloops.c (take_address_of): Don't ICE if get_name
5191 (eliminate_local_variables_stmt): Remove clobber stmts.
5193 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
5195 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
5196 error message is issued for an alias to undefined symbol.
5198 2013-08-18 Jan Hubicka <jh@suse.cz>
5200 * cgraph.c (cgraph_create_indirect_edge): Discover
5201 polymorphic calls and record basic info into indirect_info.
5202 * gimple-fold.c (gimple_fold_call): When doing BINFO based
5203 devirtualization, ignore objc function calls.
5204 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
5205 call with no parm index info.
5206 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
5207 * tree.c (virtual_method_call_p): New function.
5208 * tree.h (virtual_method_call_p): Declare.
5210 2013-08-16 Jan Hubicka <jh@suse.cz>
5213 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
5215 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
5217 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
5220 2013-08-16 David Malcolm <dmalcolm@redhat.com>
5222 * gengtype.c (type_for_name): Add special-case support for
5223 locating types within the "gcc::" namespace.
5224 (open_base_files): Emit a "using namespace gcc" directive.
5226 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
5229 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
5230 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
5232 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
5233 array instead of each individual operand as a separate argument.
5234 (emit_fusion_gpr_load): Likewise.
5235 (expand_fusion_gpr_load): Add new function declaration.
5237 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
5238 signature to have the operands passed as an array, instead of as
5239 separate arguments. Allow ZERO_EXTEND to be in the memory
5240 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
5241 depend on the register live/dead flags when peepholes are run.
5242 (expand_fusion_gpr_load): New function to be called from the
5243 peephole2 pass, to change the register that addis sets to be the
5245 (emit_fusion_gpr_load): Change the calling signature to have the
5246 operands passed as an array, instead of as separate arguments.
5247 Allow ZERO_EXTEND to be in the memory address, and also
5248 SIGN_EXTEND if -mpower8-fusion-sign.
5250 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
5252 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
5253 adjust the register addis loads up in the peephole2 pass. Do not
5254 depend on the register live/dead state when the peephole pass is done.
5256 2013-08-16 David Malcolm <dmalcolm@redhat.com>
5258 * gengtype.c (create_user_defined_type): Ensure that the kind
5259 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
5260 declaration is seen before the GTY((user)) marking.
5262 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
5265 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
5267 2013-08-16 Jan Hubicka <jh@suse.cz>
5269 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
5270 arugment expected_type.
5271 (gimple_fold_call): Use it.
5272 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
5273 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
5274 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
5275 (try_make_edge_direct_virtual_call): Likewise.
5276 * tree.c (obj_type_ref_class): New.
5277 * tree.h (obj_type_ref_class): Use it.
5279 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
5281 * sched-vis.c (rtl_slim_pp_initialized): Remove.
5282 (rtl_slim_pp): Likewise.
5283 (init_rtl_slim_pretty_print): Likewise.
5284 (dump_value_slim): Don't call it. Use local pretty printer.
5285 (dump_insn_slim): Likewise.
5286 (dump_rtl_slim): Likewise.
5287 (str_pattern_slim): Likewise.
5288 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
5291 2013-08-16 Jakub Jelinek <jakub@redhat.com>
5293 PR tree-optimization/58164
5294 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
5295 walk gimple_goto_dest of GIMPLE_GOTO.
5297 PR tree-optimization/58165
5298 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
5299 bi_call must be the last stmt in a bb, don't split_block, instead
5300 use fallthru edge from it and give up if there is none.
5301 Release conds vector when returning early.
5303 2013-08-14 Xinliang David Li <davidxl@google.com>
5305 * config/i386/i386.c (ix86_option_override_internal):
5306 Remove unused variable and field.
5308 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5311 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
5312 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
5313 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
5314 and Linux, correct BLKmode alignment when 128-bit alignment is
5315 required and compatibility flag is not set.
5316 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
5317 for zero-size arguments when compatibility flag is not set.
5319 2013-08-14 Jakub Jelinek <jakub@redhat.com>
5321 PR tree-optimization/58145
5322 * tree-sra.c (build_ref_for_offset): If prev_base has
5323 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
5325 2013-08-14 Xinliang David Li <davidxl@google.com>
5327 * config/i386/i386.c (ix86_option_override_internal):
5328 Fix uninitialized variable error.
5330 2013-08-14 Xinliang David Li <davidxl@google.com>
5332 * config/i386/i386.opt: Define two new options.
5333 * config/i386/x86-tune.def: Add arch selector field in macros.
5334 * config/i386/i386.h: Adjust macro definition.
5335 * config/i386/i386.c (ix86_option_override_internal):
5337 (parse_mtune_ctrl_str): New function.
5338 (set_ix86_tune_features): New function.
5339 (ix86_function_specific_restore): Call the new helper function.
5341 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
5343 PR rtl-optimization/57662
5344 * sel-sched.c (code_motion_process_successors): When the current insn
5345 is removed after the recursive traversal, break from the loop.
5346 Add comments and debug printouts.
5348 2013-08-14 Jakub Jelinek <jakub@redhat.com>
5349 Alexandre Oliva <aoliva@redhat.com>
5352 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
5353 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
5354 there also UNSPEC_PLTOFF.
5356 2013-08-14 Marek Polacek <polacek@redhat.com>
5358 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
5359 NUM_CONDITIONS bit positions.
5361 2013-08-13 Cary Coutant <ccoutant@google.com>
5363 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
5364 (attr_checksum): Hash vector contents instead of pointer.
5365 (attr_checksum_ordered): Likewise.
5367 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
5369 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
5370 when Pmode != word_mode. Add length_address attribute.
5371 (sse3_monitor_<mode>): Merge from sse3_monitor and
5372 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
5373 Pmode != word_mode. Update insn length attribute.
5374 * config/i386/i386.c (ix86_option_override_internal): Update
5375 ix86_gen_monitor selection for merged sse3_monitor insn.
5377 2013-08-13 Julian Brown <julian@codesourcery.com>
5379 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
5380 perform invalid legitimization on greater-than-word-size modes for
5383 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
5385 * ira.c (setup_class_translate_array): Use aclass instead of cl
5386 for classes not fully covered by allocno classes.
5388 2013-08-13 Jakub Jelinek <jakub@redhat.com>
5390 PR tree-optimization/57661
5391 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
5392 * tree-inline.c (tree_function_versioning): Initialize it.
5393 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
5394 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
5395 that is not being copied.
5398 * asan.c (instrument_strlen_call): Fix typo in comment.
5399 Use char * type even for the lhs of POINTER_PLUS_EXPR.
5401 2013-08-13 Steve Ellcey <sellcey@mips.com>
5403 * config/mips/mips.md (prefetch): Use lw instead of ld on
5404 loongson in 32bit mode.
5406 2013-08-13 Nick Clifton <nickc@redhat.com>
5408 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
5410 2013-08-13 Jan Hubicka <jh@suse.cz>
5412 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
5413 introduced edge; fix typo in sanity check.
5414 (cgraph_resolve_speculation): Export; improve diagnostic.
5415 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
5416 speculation at type mismatch.
5417 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
5418 (cgraph_resolve_speculation): Declare.
5419 (symtab_can_be_discarded): New function.
5420 * value-prof.c (gimple_ic_transform): Remove actual transform code.
5421 * ipa-inline-transform.c (speculation_removed): New global var.
5422 (clone_inlined_nodes): See if speculation can be removed.
5423 (inline_call): If speculations was removed, we growths may not match.
5424 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
5425 (speculation_useful_p): New function.
5426 (resolve_noninline_speculation): New function.
5427 (inline_small_functions): Resolve useless speculations.
5428 * ipa-inline.h (speculation_useful_p): Declare
5429 * ipa.c (can_replace_by_local_alias): Simplify.
5430 (ipa_profile): Produce speculative calls in non-lto, too;
5431 add simple cost model; produce local aliases.
5433 2013-08-13 David Malcolm <dmalcolm@redhat.com>
5435 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
5438 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
5440 * config/i386/i386.c (ix86_function_versions): Use error + inform.
5442 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
5444 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
5445 iterator instead of X87MODEF.
5447 2013-08-12 Perez Read <netfirewall@gmail.com>
5450 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
5451 operand 0 for intel asm alternative.
5452 (*movabs<mode>_2): Ditto for operand 1.
5454 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
5456 * config/aarch64/arm_none.h
5457 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
5459 2013-08-12 Nick Clifton <nickc@redhat.com>
5461 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
5463 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
5465 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
5466 expand for QI/HImode operand to produce more effictive code for
5467 unsigned char(short) --> float(double) conversion.
5469 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
5471 * doc/invoke.texi: Mention that -ftls-model does not force the final
5474 2013-08-12 Marek Polacek <polacek@redhat.com>
5475 Marc Glisse <marc.glisse@inria.fr>
5477 PR tree-optimization/57980
5478 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
5479 when creating -1 constant.
5481 2013-08-10 Jan Hubicka <jh@suse.cz>
5483 Workaround binutils PR14342.
5484 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
5485 (gimple_init_edge_profiler): Likewise.
5486 (gimple_gen_ic_func_profiler): Likewise.
5488 2013-08-09 Jan Hubicka <jh@suse.cz>
5490 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
5492 2013-08-09 Xinliang David Li <davidxl@google.com>
5494 * config/i386/stringop.def: New file.
5495 * config/i386/stringop.opt: New file.
5496 * config/i386/i386-opts.h: Include stringopt.def.
5497 * config/i386/i386.opt: Include stringopt.opt.
5498 * config/i386/i386.c (ix86_option_override_internal):
5499 Override default size based stringop inline strategies with options.
5500 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
5503 2013-08-09 Jan Hubicka <jh@suse.cz>
5505 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
5507 2013-08-09 Jan Hubicka <jh@suse.cz>
5509 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
5511 (dump_cgraph_node): Dump profile-id.
5512 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
5513 and common_target_probability.
5514 * lto-cgraph.c (lto_output_edge): Stream common targets.
5515 (lto_output_node): Stream profile ids.
5516 (input_node): Stream profile ids.
5517 (input_edge): Stream common targets.
5518 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
5519 * ipa.c: Include value-prof.h
5520 (ipa_profile_generate_summary): Turn indirect call statement histograms
5521 into common targets.
5522 (ipa_profile): Turn common targets into speculative edges.
5524 2013-08-09 Jan Hubicka <jh@suse.cz>
5526 * cgraph.h (cgraph_node): Add profile_id.
5527 * value-prof.c (cgraph_node_map): Turn into pointer_map.
5528 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
5529 (del_node_map): Update.
5530 (find_func_by_funcdef_no): Replace by ...
5531 (find_func_by_profile_id): ... this one.
5532 (gimple_ic_transform): Do not remove useful histograms when
5533 speculation is not done; dump info when indirect call removal
5535 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
5536 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
5537 (__gcov_indirect_call_profiler_v2): .. this one.
5538 * profile.h (init_node_map): Update.
5539 * coverage.c (coverage_compute_profile_id): New function.
5540 * coverage.h (coverage_compute_profile_id): Declare.
5541 * tree-profile.c (init_ic_make_global_vars): Make
5542 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
5543 (gimple_init_edge_profiler): Update prototype of
5544 __gcov_indirect_call_profiler.
5545 (gimple_gen_ic_func_profiler): Simplify.
5546 (tree_profiling): Use init_node_map
5548 2013-08-09 Jan Hubicka <jh@suse.cz>
5550 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
5551 non-speculative refs.
5552 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
5553 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
5554 (cgraph_set_call_stmt): Likewise.
5555 (cgraph_create_edge_1): Fix release checking compilatoin;
5557 (cgraph_free_edge): Free indirect info.
5558 (cgraph_turn_edge_to_speculative): New function.
5559 (cgraph_speculative_call_info): New function.
5560 (cgraph_make_edge_direct): Return direct edge; handle speculation.
5561 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
5562 (dump_cgraph_node): Dump speculation.
5563 (verify_edge_count_and_frequency): Accept speculative edges.
5564 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
5565 (verify_cgraph_node): Handle speculation.
5566 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
5567 (cgraph_set_call_stmt): Update prototype.
5568 (cgraph_make_edge_direct): Update prototype.
5569 (cgraph_speculative_call_info): Declare.
5570 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
5571 to change; update call of ipa_find_references.
5572 * ipa-ref.c (ipa_record_reference): Fix return value; clear
5573 lto_stmt_uid and speculative flags.
5574 (ipa_dump_references): Dump speculation.
5575 (ipa_clone_references): Clone speculative flag.
5576 (ipa_clone_referring): Likewise.
5577 (ipa_clone_ref): New function.
5578 (ipa_find_reference): Look into lto_stmt_uids
5579 (ipa_clear_stmts_in_references): Do not clear speculative calls.
5580 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
5581 (ipa_find_reference): Update declaration.
5582 (ipa_clone_ref): Declare.
5583 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
5584 stream speculative flag.
5585 (lto_output_ref): Stream statements uids and speculation.
5586 (input_ref): Likewise.
5587 (input_edge): Stream speuclation.
5588 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
5589 (cgraph_set_call_stmt_including_clones): Handle speculation.
5590 * ipa-inline.c (heap_edge_removal_hook): New function.
5591 (inline_small_functions): Register it.
5592 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
5593 also initialize refs.
5594 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
5596 (try_make_edge_direct_simple_call): Likewise.
5597 (try_make_edge_direct_simple_call): Likewise.
5598 (update_indirect_edges_after_inlining): Likewise.
5599 (remove_described_reference): Look proper lto_stmt_uid.
5600 (propagate_controlled_uses): Likewise.
5601 (propagate_controlled_uses): Liekwise.
5602 * tree-inline.c (copy_bb): Copy speculative edges.
5603 (redirect_all_calls): New function.
5604 (copy_cfg_body): Do redirection after loop info is updated.
5605 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
5607 2013-08-09 Jan Hubicka <jh@suse.cz>
5609 * lto-streamer-out.c (output_function): Renumber PHIs.
5610 * lto-streamer-in.c (input_function): Likewise.
5612 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
5614 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
5615 (get_lane_unsigned): Likewise.
5616 (dup_lane_scalar): Likewise.
5617 (get_lane): enable for VALL.
5618 * config/aarch64/aarch64-simd.md
5619 (aarch64_dup_lane_scalar<mode>): Remove.
5620 (aarch64_get_lane_signed<mode>): Likewise.
5621 (aarch64_get_lane_unsigned<mode>): Likewise.
5622 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
5623 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
5624 (aarch64_get_lane<mode>): Enable for all vector modes.
5625 (aarch64_get_lanedi): Remove misleading constraints.
5626 * config/aarch64/arm_neon.h
5627 (__aarch64_vget_lane_any): Define.
5628 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
5629 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
5630 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
5631 * config/aarch64/iterators.md (VDQQH): New.
5635 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
5637 * configure.ac: Add GAS check for LEON instructions on SPARC.
5638 * configure: Regenerate.
5639 * config.in: Likewise.
5640 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
5642 * config/sparc/sparc.opt (LEON, LEON3): New masks.
5643 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
5645 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
5646 (AS_LEON_FLAG): New macro.
5647 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
5648 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
5649 Deal with LEON and LEON3 for the memory model.
5650 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
5651 (atomic_compare_and_swap<mode>_1): Likewise.
5652 (*atomic_compare_and_swap<mode>_1): Likewise.
5654 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
5656 * config/arm/neon.md (vcond): Fix floating-point vector
5657 comparisons against 0.
5659 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
5661 * lra-constraints.c (emit_spill_move): Remove assert.
5662 (process_alt_operands): Add more debugging
5663 output. Increase reject for spilling into memory. Decrease
5664 reject for reloading scratch.
5665 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
5667 2013-08-08 Steve Ellcey <sellcey@mips.com>
5669 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
5670 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
5671 micromips incompatible. Add nan2008.
5672 (MULTILIB_DIRNAMES): Add nan2008.
5673 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
5674 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
5675 and micromips incompatible. Add nan2008.
5676 (MULTILIB_DIRNAMES): Add nan2008.
5677 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
5679 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
5681 PR rtl-optimization/58079
5682 * combine.c (combine_simplify_rtx): Avoid using SUBST if
5683 simplify_comparison has widened a comparison with an integer.
5685 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5687 * config/arm/neon.md (movmisalign<mode>): Disable when we
5688 don't allow unaligned accesses.
5689 (*movmisalign<mode>_neon_store): Likewise.
5690 (*movmisalign<mode>_neon_load): Likewise.
5691 (*movmisalign<mode>_neon_store): Likewise.
5692 (*movmisalign<mode>_neon_load): Likewise.
5694 2013-08-08 Jan Hubicka <jh@suse.cz>
5696 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
5697 (make_pass_rebuild_cgraph_edges): Also clear references.
5698 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
5699 * ipa-inline-transform.c (inline_transform): Remove all references
5701 * cgraphunit.c (expand_function): Remove all references after
5703 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
5704 (ipa_find_reference): Rewrite to iterator.
5705 (remove_stmt_references): Likewise.
5706 (ipa_clear_stmts_in_references): New function.
5707 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
5708 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
5710 * ipa-split.c (split_function): Remove references in split function.
5712 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
5715 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
5716 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
5718 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
5721 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
5722 suggested mode for the assignment isn't compatible with the
5725 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
5728 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
5730 2013-08-07 Xinliang David Li <davidxl@google.com>
5732 * config/i386/i386.opt: New option -mtune-ctrl=.
5733 * config/i386/x86-tune.def: New file.
5734 * config/i386/i386.h: include x86-tune.def.
5735 * config/i386/i386.c (ix86_option_override_internal):
5736 Parsing -mtune-ctrl= option and set tune features.
5738 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
5741 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
5742 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
5745 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
5747 * rtl.h (update_alignments): Declare.
5748 * final.c (grow_label_align): New function extracted from...
5749 (shorten_branches): ...here. Call it.
5750 (update_alignments): New function.
5751 * reorg.c (sibling_labels): New variable.
5752 (get_label_before): Add SIBLING parameter. If it is non-zero, push
5753 the new label along with it onto the sibling_labels vector.
5754 (fill_simple_delay_slots): Adjust call to get_label_before.
5755 (fill_slots_from_thread): Likewise.
5756 (relax_delay_slots): Likewise.
5757 (make_return_insns): Likewise.
5758 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
5760 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
5762 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
5763 document its semantics.
5764 (diagnostic_report_diagnostic): Adjust accordingly.
5766 2013-08-07 David Malcolm <dmalcolm@redhat.com>
5768 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
5769 (sparc_option_override): ...and port to new C++ pass API.
5770 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
5772 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
5774 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
5776 2013-08-06 Caroline Tice <cmtice@google.com>
5778 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
5779 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
5780 * tree-pass.h: Add pass_vtable_verify.
5781 * varasm.c (assemble_variable): Add code to properly set the comdat
5782 section and name for the .vtable_map_vars section.
5783 (assemble_vtyv_preinit_initializer): New function.
5784 (default_sectin_type_flags): Make sure .vtable_map_vars section has
5786 * output.h: Add function decl for assemble_vtv_preinit_initializer.
5787 * vtable-verify.c: New file.
5788 * vtable-verify.h: New file.
5789 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
5790 initialiation levels.
5791 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
5792 * passes.def: Insert pass_vtable_verify.
5793 * aclocal.m4: Reorder includes.
5794 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
5795 -fvtv-counts options.
5796 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
5797 as appropriate, if -fvtable-verify=... is used.
5798 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
5799 -fvtable-verify=... is used.
5800 * Makefile.in (OBJS): Add vtable-verify.o to list.
5801 (vtable-verify.o): Add new build rule.
5802 (GTFILES): Add vtable-verify.c to list.
5803 * common.opt (fvtable-verify=): New flag.
5804 (vtv_priority): Values for fvtable-verify= flag.
5805 (fvtv-counts): New flag.
5806 (fvtv-debug): New flag.
5807 * tree.h (save_vtable_map_decl): New extern function decl.
5809 2013-08-07 David Malcolm <dmalcolm@redhat.com>
5811 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
5812 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
5813 (pass_data_rl78_devirt): ...new pass_data instance and...
5814 (make_pass_rl78_devirt): ...new function.
5815 (rl78_asm_file_start): Port pass registration to new C++ API.
5817 2013-08-07 David Malcolm <dmalcolm@redhat.com>
5819 * coretypes.h (rtl_opt_pass): Add.
5820 (gcc::context): Add.
5821 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
5822 (epiphany_init): Port to new C++ pass API.
5823 (epiphany_optimize_mode_switching): Likewise.
5824 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
5825 (pass_manager::get_pass_mode_switching): New.
5826 (pass_manager::get_pass_peephole2): New.
5827 * mode-switching.c (pass_mode_switching): Add clone method.
5828 * recog.c (pass_peephole2): Add clone method.
5829 (pass_split_all_insns): Add clone method.
5831 2013-08-06 David Malcolm <dmalcolm@redhat.com>
5833 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
5834 (mips_option_override): ...here, porting to new C++ API for passes.
5836 2013-08-06 Jan Hubicka <jh@suse.cz>
5838 * cgraph.c (cgraph_get_body): New function based on lto.c
5840 * cgraph.h (cgraph_get_body): Declare.
5841 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
5843 * cgraphunit.c (expand_function): Get body prior expanding.
5844 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
5845 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
5847 * passes.c (do_per_function_toporder): Get body.
5848 * tree-inline.c (expand_call_inline): Get body prior inlining it.
5849 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
5851 2013-08-06 Martin Jambor <mjambor@suse.cz>
5854 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
5855 is not re-finalized. Rename second parameter to no_collect.
5857 2013-08-06 Martin Jambor <mjambor@suse.cz>
5860 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
5861 MEM_REF has proper alignment information.
5863 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
5866 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
5868 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
5869 rtx (*) (rtx, ...) with insn_gen_fn.
5870 * genoutput.c (output_insn_data): Cast gen_? function pointers to
5871 insn_gen_fn::stored_funcptr. Add initializer braces.
5873 2013-08-05 David Malcolm <dmalcolm@redhat.com>
5875 Rewrite how instances of passes are cloned to remove assumptions
5876 about their sizes (thus allowing pass subclasses to have
5877 additional data fields, albeit non-GC-managed ones at this point).
5879 * passes.c (make_pass_instance): Now that passes have clone
5880 methods, rewrite this function to eliminate XNEW and memcpy
5881 calls that used hardcoded sizes. Since this function no longer
5882 creates pass instances, rename it to...
5883 (add_pass_instance): ...this. Document the old way that passes were
5884 numbered and flagged, and rework this function to continue using it.
5885 (next_pass_1): Add an initial_pass argument for use by
5887 (position_pass): When adding multiple instances of a pass, use
5888 the pass's clone method, rather than relying on the XNEW/memcpy
5889 within the former make_pass_instance (now add_pass_instance).
5890 (pass_manager::pass_manager): When invoking next_pass_1, also supply
5891 the initial instance of the current pass within the pass manager.
5893 2013-08-05 David Malcolm <dmalcolm@redhat.com>
5895 This is the automated part of the conversion of passes from C
5896 structs to C++ classes.
5898 Patch autogenerated by refactor_passes.py from
5899 https://github.com/davidmalcolm/gcc-refactoring-scripts
5900 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
5902 * asan.c (pass_asan): Convert from a global struct to a subclass of
5903 gimple_opt_pass along with...
5904 (pass_data_asan): ...new pass_data instance and...
5905 (make_pass_asan): ...new function.
5906 (pass_asan_O0): Convert from a global struct to a subclass of
5907 gimple_opt_pass along with...
5908 (pass_data_asan_O0): ...new pass_data instance and...
5909 (make_pass_asan_O0): ...new function.
5910 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
5911 subclass of rtl_opt_pass along with...
5912 (pass_data_inc_dec): ...new pass_data instance and...
5913 (make_pass_inc_dec): ...new function.
5914 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
5915 a subclass of rtl_opt_pass along with...
5916 (pass_data_reorder_blocks): ...new pass_data instance and...
5917 (make_pass_reorder_blocks): ...new function.
5918 (pass_duplicate_computed_gotos): Convert from a global struct to a
5919 subclass of rtl_opt_pass along with...
5920 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
5921 (make_pass_duplicate_computed_gotos): ...new function.
5922 (pass_partition_blocks): Convert from a global struct to a subclass of
5923 rtl_opt_pass along with...
5924 (pass_data_partition_blocks): ...new pass_data instance and...
5925 (make_pass_partition_blocks): ...new function.
5926 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
5927 struct to a subclass of rtl_opt_pass along with...
5928 (pass_data_branch_target_load_optimize1): ...new pass_data instance
5930 (make_pass_branch_target_load_optimize1): ...new function.
5931 (pass_branch_target_load_optimize2): Convert from a global struct to a
5932 subclass of rtl_opt_pass along with...
5933 (pass_data_branch_target_load_optimize2): ...new pass_data instance
5935 (make_pass_branch_target_load_optimize2): ...new function.
5936 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
5937 of rtl_opt_pass along with...
5938 (pass_data_jump): ...new pass_data instance and...
5939 (make_pass_jump): ...new function.
5940 (pass_jump2): Convert from a global struct to a subclass of
5941 rtl_opt_pass along with...
5942 (pass_data_jump2): ...new pass_data instance and...
5943 (make_pass_jump2): ...new function.
5944 * cfgexpand.c (pass_expand): Convert from a global struct to a
5945 subclass of rtl_opt_pass along with...
5946 (pass_data_expand): ...new pass_data instance and...
5947 (make_pass_expand): ...new function.
5948 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
5949 of rtl_opt_pass along with...
5950 (pass_data_free_cfg): ...new pass_data instance and...
5951 (make_pass_free_cfg): ...new function.
5952 (pass_into_cfg_layout_mode): Convert from a global struct to a
5953 subclass of rtl_opt_pass along with...
5954 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
5955 (make_pass_into_cfg_layout_mode): ...new function.
5956 (pass_outof_cfg_layout_mode): Convert from a global struct to a
5957 subclass of rtl_opt_pass along with...
5958 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
5959 (make_pass_outof_cfg_layout_mode): ...new function.
5960 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
5961 struct to a subclass of gimple_opt_pass along with...
5962 (pass_data_build_cgraph_edges): ...new pass_data instance and...
5963 (make_pass_build_cgraph_edges): ...new function.
5964 (pass_rebuild_cgraph_edges): Convert from a global struct to a
5965 subclass of gimple_opt_pass along with...
5966 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
5967 (make_pass_rebuild_cgraph_edges): ...new function.
5968 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
5969 subclass of gimple_opt_pass along with...
5970 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
5972 (make_pass_remove_cgraph_callee_edges): ...new function.
5973 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
5974 struct to a subclass of rtl_opt_pass along with...
5975 (pass_data_stack_adjustments): ...new pass_data instance and...
5976 (make_pass_stack_adjustments): ...new function.
5977 * combine.c (pass_combine): Convert from a global struct to a subclass
5978 of rtl_opt_pass along with...
5979 (pass_data_combine): ...new pass_data instance and...
5980 (make_pass_combine): ...new function.
5981 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
5982 global struct to a subclass of rtl_opt_pass along with...
5983 (pass_data_compare_elim_after_reload): ...new pass_data instance
5985 (make_pass_compare_elim_after_reload): ...new function.
5986 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
5987 of rtl_opt_pass along with...
5988 (pass_data_rtl_cprop): ...new pass_data instance and...
5989 (make_pass_rtl_cprop): ...new function.
5990 * cse.c (pass_cse): Convert from a global struct to a subclass of
5991 rtl_opt_pass along with...
5992 (pass_data_cse): ...new pass_data instance and...
5993 (make_pass_cse): ...new function.
5994 (pass_cse2): Convert from a global struct to a subclass of
5995 rtl_opt_pass along with...
5996 (pass_data_cse2): ...new pass_data instance and...
5997 (make_pass_cse2): ...new function.
5998 (pass_cse_after_global_opts): Convert from a global struct to a
5999 subclass of rtl_opt_pass along with...
6000 (pass_data_cse_after_global_opts): ...new pass_data instance and...
6001 (make_pass_cse_after_global_opts): ...new function.
6002 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
6003 of rtl_opt_pass along with...
6004 (pass_data_ud_rtl_dce): ...new pass_data instance and...
6005 (make_pass_ud_rtl_dce): ...new function.
6006 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
6007 rtl_opt_pass along with...
6008 (pass_data_fast_rtl_dce): ...new pass_data instance and...
6009 (make_pass_fast_rtl_dce): ...new function.
6010 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
6011 a subclass of rtl_opt_pass along with...
6012 (pass_data_df_initialize_opt): ...new pass_data instance and...
6013 (make_pass_df_initialize_opt): ...new function.
6014 (pass_df_initialize_no_opt): Convert from a global struct to a
6015 subclass of rtl_opt_pass along with...
6016 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
6017 (make_pass_df_initialize_no_opt): ...new function.
6018 (pass_df_finish): Convert from a global struct to a subclass of
6019 rtl_opt_pass along with...
6020 (pass_data_df_finish): ...new pass_data instance and...
6021 (make_pass_df_finish): ...new function.
6022 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
6023 rtl_opt_pass along with...
6024 (pass_data_rtl_dse1): ...new pass_data instance and...
6025 (make_pass_rtl_dse1): ...new function.
6026 (pass_rtl_dse2): Convert from a global struct to a subclass of
6027 rtl_opt_pass along with...
6028 (pass_data_rtl_dse2): ...new pass_data instance and...
6029 (make_pass_rtl_dse2): ...new function.
6030 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
6031 subclass of rtl_opt_pass along with...
6032 (pass_data_dwarf2_frame): ...new pass_data instance and...
6033 (make_pass_dwarf2_frame): ...new function.
6034 * except.c (pass_set_nothrow_function_flags): Convert from a global
6035 struct to a subclass of rtl_opt_pass along with...
6036 (pass_data_set_nothrow_function_flags): ...new pass_data instance
6038 (make_pass_set_nothrow_function_flags): ...new function.
6039 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
6040 subclass of rtl_opt_pass along with...
6041 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
6043 (make_pass_convert_to_eh_region_ranges): ...new function.
6044 * final.c (pass_compute_alignments): Convert from a global struct to a
6045 subclass of rtl_opt_pass along with...
6046 (pass_data_compute_alignments): ...new pass_data instance and...
6047 (make_pass_compute_alignments): ...new function.
6048 (pass_final): Convert from a global struct to a subclass of
6049 rtl_opt_pass along with...
6050 (pass_data_final): ...new pass_data instance and...
6051 (make_pass_final): ...new function.
6052 (pass_shorten_branches): Convert from a global struct to a subclass of
6053 rtl_opt_pass along with...
6054 (pass_data_shorten_branches): ...new pass_data instance and...
6055 (make_pass_shorten_branches): ...new function.
6056 (pass_clean_state): Convert from a global struct to a subclass of
6057 rtl_opt_pass along with...
6058 (pass_data_clean_state): ...new pass_data instance and...
6059 (make_pass_clean_state): ...new function.
6060 * function.c (pass_instantiate_virtual_regs): Convert from a global
6061 struct to a subclass of rtl_opt_pass along with...
6062 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
6063 (make_pass_instantiate_virtual_regs): ...new function.
6064 (pass_leaf_regs): Convert from a global struct to a subclass of
6065 rtl_opt_pass along with...
6066 (pass_data_leaf_regs): ...new pass_data instance and...
6067 (make_pass_leaf_regs): ...new function.
6068 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
6069 subclass of rtl_opt_pass along with...
6070 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
6072 (make_pass_thread_prologue_and_epilogue): ...new function.
6073 (pass_match_asm_constraints): Convert from a global struct to a
6074 subclass of rtl_opt_pass along with...
6075 (pass_data_match_asm_constraints): ...new pass_data instance and...
6076 (make_pass_match_asm_constraints): ...new function.
6077 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
6078 subclass of rtl_opt_pass along with...
6079 (pass_data_rtl_fwprop): ...new pass_data instance and...
6080 (make_pass_rtl_fwprop): ...new function.
6081 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
6082 rtl_opt_pass along with...
6083 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
6084 (make_pass_rtl_fwprop_addr): ...new function.
6085 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
6086 rtl_opt_pass along with...
6087 (pass_data_rtl_pre): ...new pass_data instance and...
6088 (make_pass_rtl_pre): ...new function.
6089 (pass_rtl_hoist): Convert from a global struct to a subclass of
6090 rtl_opt_pass along with...
6091 (pass_data_rtl_hoist): ...new pass_data instance and...
6092 (make_pass_rtl_hoist): ...new function.
6093 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
6094 subclass of gimple_opt_pass along with...
6095 (pass_data_lower_cf): ...new pass_data instance and...
6096 (make_pass_lower_cf): ...new function.
6097 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
6098 from a global struct to a subclass of gimple_opt_pass along with...
6099 (pass_data_strength_reduction): ...new pass_data instance and...
6100 (make_pass_strength_reduction): ...new function.
6101 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
6102 of rtl_opt_pass along with...
6103 (pass_data_rtl_ifcvt): ...new pass_data instance and...
6104 (make_pass_rtl_ifcvt): ...new function.
6105 (pass_if_after_combine): Convert from a global struct to a subclass of
6106 rtl_opt_pass along with...
6107 (pass_data_if_after_combine): ...new pass_data instance and...
6108 (make_pass_if_after_combine): ...new function.
6109 (pass_if_after_reload): Convert from a global struct to a subclass of
6110 rtl_opt_pass along with...
6111 (pass_data_if_after_reload): ...new pass_data instance and...
6112 (make_pass_if_after_reload): ...new function.
6113 * init-regs.c (pass_initialize_regs): Convert from a global struct to
6114 a subclass of rtl_opt_pass along with...
6115 (pass_data_initialize_regs): ...new pass_data instance and...
6116 (make_pass_initialize_regs): ...new function.
6117 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
6118 of ipa_opt_pass_d along with...
6119 (pass_data_ipa_cp): ...new pass_data instance and...
6120 (make_pass_ipa_cp): ...new function.
6121 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
6122 global struct to a subclass of gimple_opt_pass along with...
6123 (pass_data_inline_parameters): ...new pass_data instance and...
6124 (make_pass_inline_parameters): ...new function.
6125 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
6126 subclass of gimple_opt_pass along with...
6127 (pass_data_early_inline): ...new pass_data instance and...
6128 (make_pass_early_inline): ...new function.
6129 (pass_ipa_inline): Convert from a global struct to a subclass of
6130 ipa_opt_pass_d along with...
6131 (pass_data_ipa_inline): ...new pass_data instance and...
6132 (make_pass_ipa_inline): ...new function.
6133 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
6134 struct to a subclass of gimple_opt_pass along with...
6135 (pass_data_local_pure_const): ...new pass_data instance and...
6136 (make_pass_local_pure_const): ...new function.
6137 (pass_ipa_pure_const): Convert from a global struct to a subclass of
6138 ipa_opt_pass_d along with...
6139 (pass_data_ipa_pure_const): ...new pass_data instance and...
6140 (make_pass_ipa_pure_const): ...new function.
6141 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
6142 to a subclass of ipa_opt_pass_d along with...
6143 (pass_data_ipa_reference): ...new pass_data instance and...
6144 (make_pass_ipa_reference): ...new function.
6145 * ipa-split.c (pass_split_functions): Convert from a global struct to
6146 a subclass of gimple_opt_pass along with...
6147 (pass_data_split_functions): ...new pass_data instance and...
6148 (make_pass_split_functions): ...new function.
6149 (pass_feedback_split_functions): Convert from a global struct to a
6150 subclass of gimple_opt_pass along with...
6151 (pass_data_feedback_split_functions): ...new pass_data instance and...
6152 (make_pass_feedback_split_functions): ...new function.
6153 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
6154 global struct to a subclass of simple_ipa_opt_pass along with...
6155 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
6157 (make_pass_ipa_function_and_variable_visibility): ...new function.
6158 (pass_ipa_free_inline_summary): Convert from a global struct to a
6159 subclass of simple_ipa_opt_pass along with...
6160 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
6161 (make_pass_ipa_free_inline_summary): ...new function.
6162 (pass_ipa_whole_program_visibility): Convert from a global struct to a
6163 subclass of ipa_opt_pass_d along with...
6164 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
6166 (make_pass_ipa_whole_program_visibility): ...new function.
6167 (pass_ipa_profile): Convert from a global struct to a subclass of
6168 ipa_opt_pass_d along with...
6169 (pass_data_ipa_profile): ...new pass_data instance and...
6170 (make_pass_ipa_profile): ...new function.
6171 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
6172 ipa_opt_pass_d along with...
6173 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
6174 (make_pass_ipa_cdtor_merge): ...new function.
6175 * ira.c (pass_ira): Convert from a global struct to a subclass of
6176 rtl_opt_pass along with...
6177 (pass_data_ira): ...new pass_data instance and...
6178 (make_pass_ira): ...new function.
6179 (pass_reload): Convert from a global struct to a subclass of
6180 rtl_opt_pass along with...
6181 (pass_data_reload): ...new pass_data instance and...
6182 (make_pass_reload): ...new function.
6183 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
6184 subclass of rtl_opt_pass along with...
6185 (pass_data_cleanup_barriers): ...new pass_data instance and...
6186 (make_pass_cleanup_barriers): ...new function.
6187 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
6188 of rtl_opt_pass along with...
6189 (pass_data_loop2): ...new pass_data instance and...
6190 (make_pass_loop2): ...new function.
6191 (pass_rtl_loop_init): Convert from a global struct to a subclass of
6192 rtl_opt_pass along with...
6193 (pass_data_rtl_loop_init): ...new pass_data instance and...
6194 (make_pass_rtl_loop_init): ...new function.
6195 (pass_rtl_loop_done): Convert from a global struct to a subclass of
6196 rtl_opt_pass along with...
6197 (pass_data_rtl_loop_done): ...new pass_data instance and...
6198 (make_pass_rtl_loop_done): ...new function.
6199 (pass_rtl_move_loop_invariants): Convert from a global struct to a
6200 subclass of rtl_opt_pass along with...
6201 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
6202 (make_pass_rtl_move_loop_invariants): ...new function.
6203 (pass_rtl_unswitch): Convert from a global struct to a subclass of
6204 rtl_opt_pass along with...
6205 (pass_data_rtl_unswitch): ...new pass_data instance and...
6206 (make_pass_rtl_unswitch): ...new function.
6207 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
6208 subclass of rtl_opt_pass along with...
6209 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
6211 (make_pass_rtl_unroll_and_peel_loops): ...new function.
6212 (pass_rtl_doloop): Convert from a global struct to a subclass of
6213 rtl_opt_pass along with...
6214 (pass_data_rtl_doloop): ...new pass_data instance and...
6215 (make_pass_rtl_doloop): ...new function.
6216 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
6217 a subclass of rtl_opt_pass along with...
6218 (pass_data_lower_subreg): ...new pass_data instance and...
6219 (make_pass_lower_subreg): ...new function.
6220 (pass_lower_subreg2): Convert from a global struct to a subclass of
6221 rtl_opt_pass along with...
6222 (pass_data_lower_subreg2): ...new pass_data instance and...
6223 (make_pass_lower_subreg2): ...new function.
6224 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
6225 struct to a subclass of ipa_opt_pass_d along with...
6226 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
6227 (make_pass_ipa_lto_gimple_out): ...new function.
6228 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
6229 of ipa_opt_pass_d along with...
6230 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
6231 (make_pass_ipa_lto_finish_out): ...new function.
6232 * mode-switching.c (pass_mode_switching): Convert from a global struct
6233 to a subclass of rtl_opt_pass along with...
6234 (pass_data_mode_switching): ...new pass_data instance and...
6235 (make_pass_mode_switching): ...new function.
6236 * modulo-sched.c (pass_sms): Convert from a global struct to a
6237 subclass of rtl_opt_pass along with...
6238 (pass_data_sms): ...new pass_data instance and...
6239 (make_pass_sms): ...new function.
6240 * omp-low.c (pass_expand_omp): Convert from a global struct to a
6241 subclass of gimple_opt_pass along with...
6242 (pass_data_expand_omp): ...new pass_data instance and...
6243 (make_pass_expand_omp): ...new function.
6244 (pass_lower_omp): Convert from a global struct to a subclass of
6245 gimple_opt_pass along with...
6246 (pass_data_lower_omp): ...new pass_data instance and...
6247 (make_pass_lower_omp): ...new function.
6248 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
6249 of gimple_opt_pass along with...
6250 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
6251 (make_pass_diagnose_omp_blocks): ...new function.
6252 * passes.c (pass_early_local_passes): Convert from a global struct to
6253 a subclass of simple_ipa_opt_pass along with...
6254 (pass_data_early_local_passes): ...new pass_data instance and...
6255 (make_pass_early_local_passes): ...new function.
6256 (pass_all_early_optimizations): Convert from a global struct to a
6257 subclass of gimple_opt_pass along with...
6258 (pass_data_all_early_optimizations): ...new pass_data instance and...
6259 (make_pass_all_early_optimizations): ...new function.
6260 (pass_all_optimizations): Convert from a global struct to a subclass
6261 of gimple_opt_pass along with...
6262 (pass_data_all_optimizations): ...new pass_data instance and...
6263 (make_pass_all_optimizations): ...new function.
6264 (pass_all_optimizations_g): Convert from a global struct to a subclass
6265 of gimple_opt_pass along with...
6266 (pass_data_all_optimizations_g): ...new pass_data instance and...
6267 (make_pass_all_optimizations_g): ...new function.
6268 (pass_rest_of_compilation): Convert from a global struct to a subclass
6269 of rtl_opt_pass along with...
6270 (pass_data_rest_of_compilation): ...new pass_data instance and...
6271 (make_pass_rest_of_compilation): ...new function.
6272 (pass_postreload): Convert from a global struct to a subclass of
6273 rtl_opt_pass along with...
6274 (pass_data_postreload): ...new pass_data instance and...
6275 (make_pass_postreload): ...new function.
6276 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
6277 subclass of rtl_opt_pass along with...
6278 (pass_data_gcse2): ...new pass_data instance and...
6279 (make_pass_gcse2): ...new function.
6280 * postreload.c (pass_postreload_cse): Convert from a global struct to
6281 a subclass of rtl_opt_pass along with...
6282 (pass_data_postreload_cse): ...new pass_data instance and...
6283 (make_pass_postreload_cse): ...new function.
6284 * predict.c (pass_profile): Convert from a global struct to a subclass
6285 of gimple_opt_pass along with...
6286 (pass_data_profile): ...new pass_data instance and...
6287 (make_pass_profile): ...new function.
6288 (pass_strip_predict_hints): Convert from a global struct to a subclass
6289 of gimple_opt_pass along with...
6290 (pass_data_strip_predict_hints): ...new pass_data instance and...
6291 (make_pass_strip_predict_hints): ...new function.
6292 * recog.c (pass_peephole2): Convert from a global struct to a subclass
6293 of rtl_opt_pass along with...
6294 (pass_data_peephole2): ...new pass_data instance and...
6295 (make_pass_peephole2): ...new function.
6296 (pass_split_all_insns): Convert from a global struct to a subclass of
6297 rtl_opt_pass along with...
6298 (pass_data_split_all_insns): ...new pass_data instance and...
6299 (make_pass_split_all_insns): ...new function.
6300 (pass_split_after_reload): Convert from a global struct to a subclass
6301 of rtl_opt_pass along with...
6302 (pass_data_split_after_reload): ...new pass_data instance and...
6303 (make_pass_split_after_reload): ...new function.
6304 (pass_split_before_regstack): Convert from a global struct to a
6305 subclass of rtl_opt_pass along with...
6306 (pass_data_split_before_regstack): ...new pass_data instance and...
6307 (make_pass_split_before_regstack): ...new function.
6308 (pass_split_before_sched2): Convert from a global struct to a subclass
6309 of rtl_opt_pass along with...
6310 (pass_data_split_before_sched2): ...new pass_data instance and...
6311 (make_pass_split_before_sched2): ...new function.
6312 (pass_split_for_shorten_branches): Convert from a global struct to a
6313 subclass of rtl_opt_pass along with...
6314 (pass_data_split_for_shorten_branches): ...new pass_data instance
6316 (make_pass_split_for_shorten_branches): ...new function.
6317 * ree.c (pass_ree): Convert from a global struct to a subclass of
6318 rtl_opt_pass along with...
6319 (pass_data_ree): ...new pass_data instance and...
6320 (make_pass_ree): ...new function.
6321 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
6322 subclass of rtl_opt_pass along with...
6323 (pass_data_stack_regs): ...new pass_data instance and...
6324 (make_pass_stack_regs): ...new function.
6325 (pass_stack_regs_run): Convert from a global struct to a subclass of
6326 rtl_opt_pass along with...
6327 (pass_data_stack_regs_run): ...new pass_data instance and...
6328 (make_pass_stack_regs_run): ...new function.
6329 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
6330 subclass of rtl_opt_pass along with...
6331 (pass_data_cprop_hardreg): ...new pass_data instance and...
6332 (make_pass_cprop_hardreg): ...new function.
6333 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
6334 subclass of rtl_opt_pass along with...
6335 (pass_data_reginfo_init): ...new pass_data instance and...
6336 (make_pass_reginfo_init): ...new function.
6337 * regmove.c (pass_regmove): Convert from a global struct to a subclass
6338 of rtl_opt_pass along with...
6339 (pass_data_regmove): ...new pass_data instance and...
6340 (make_pass_regmove): ...new function.
6341 * regrename.c (pass_regrename): Convert from a global struct to a
6342 subclass of rtl_opt_pass along with...
6343 (pass_data_regrename): ...new pass_data instance and...
6344 (make_pass_regrename): ...new function.
6345 * reorg.c (pass_delay_slots): Convert from a global struct to a
6346 subclass of rtl_opt_pass along with...
6347 (pass_data_delay_slots): ...new pass_data instance and...
6348 (make_pass_delay_slots): ...new function.
6349 (pass_machine_reorg): Convert from a global struct to a subclass of
6350 rtl_opt_pass along with...
6351 (pass_data_machine_reorg): ...new pass_data instance and...
6352 (make_pass_machine_reorg): ...new function.
6353 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
6354 of rtl_opt_pass along with...
6355 (pass_data_sched): ...new pass_data instance and...
6356 (make_pass_sched): ...new function.
6357 (pass_sched2): Convert from a global struct to a subclass of
6358 rtl_opt_pass along with...
6359 (pass_data_sched2): ...new pass_data instance and...
6360 (make_pass_sched2): ...new function.
6361 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
6362 to a subclass of rtl_opt_pass along with...
6363 (pass_data_stack_ptr_mod): ...new pass_data instance and...
6364 (make_pass_stack_ptr_mod): ...new function.
6365 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
6366 to a subclass of rtl_opt_pass along with...
6367 (pass_data_rtl_store_motion): ...new pass_data instance and...
6368 (make_pass_rtl_store_motion): ...new function.
6369 * tracer.c (pass_tracer): Convert from a global struct to a subclass
6370 of gimple_opt_pass along with...
6371 (pass_data_tracer): ...new pass_data instance and...
6372 (make_pass_tracer): ...new function.
6373 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
6374 to a subclass of gimple_opt_pass along with...
6375 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
6376 (make_pass_diagnose_tm_blocks): ...new function.
6377 (pass_lower_tm): Convert from a global struct to a subclass of
6378 gimple_opt_pass along with...
6379 (pass_data_lower_tm): ...new pass_data instance and...
6380 (make_pass_lower_tm): ...new function.
6381 (pass_tm_init): Convert from a global struct to a subclass of
6382 gimple_opt_pass along with...
6383 (pass_data_tm_init): ...new pass_data instance and...
6384 (make_pass_tm_init): ...new function.
6385 (pass_tm_mark): Convert from a global struct to a subclass of
6386 gimple_opt_pass along with...
6387 (pass_data_tm_mark): ...new pass_data instance and...
6388 (make_pass_tm_mark): ...new function.
6389 (pass_tm_edges): Convert from a global struct to a subclass of
6390 gimple_opt_pass along with...
6391 (pass_data_tm_edges): ...new pass_data instance and...
6392 (make_pass_tm_edges): ...new function.
6393 (pass_tm_memopt): Convert from a global struct to a subclass of
6394 gimple_opt_pass along with...
6395 (pass_data_tm_memopt): ...new pass_data instance and...
6396 (make_pass_tm_memopt): ...new function.
6397 (pass_ipa_tm): Convert from a global struct to a subclass of
6398 simple_ipa_opt_pass along with...
6399 (pass_data_ipa_tm): ...new pass_data instance and...
6400 (make_pass_ipa_tm): ...new function.
6401 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
6402 subclass of gimple_opt_pass along with...
6403 (pass_data_call_cdce): ...new pass_data instance and...
6404 (make_pass_call_cdce): ...new function.
6405 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
6406 subclass of gimple_opt_pass along with...
6407 (pass_data_build_cfg): ...new pass_data instance and...
6408 (make_pass_build_cfg): ...new function.
6409 (pass_split_crit_edges): Convert from a global struct to a subclass of
6410 gimple_opt_pass along with...
6411 (pass_data_split_crit_edges): ...new pass_data instance and...
6412 (make_pass_split_crit_edges): ...new function.
6413 (pass_warn_function_return): Convert from a global struct to a
6414 subclass of gimple_opt_pass along with...
6415 (pass_data_warn_function_return): ...new pass_data instance and...
6416 (make_pass_warn_function_return): ...new function.
6417 (pass_warn_function_noreturn): Convert from a global struct to a
6418 subclass of gimple_opt_pass along with...
6419 (pass_data_warn_function_noreturn): ...new pass_data instance and...
6420 (make_pass_warn_function_noreturn): ...new function.
6421 (pass_warn_unused_result): Convert from a global struct to a subclass
6422 of gimple_opt_pass along with...
6423 (pass_data_warn_unused_result): ...new pass_data instance and...
6424 (make_pass_warn_unused_result): ...new function.
6425 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
6426 a subclass of gimple_opt_pass along with...
6427 (pass_data_merge_phi): ...new pass_data instance and...
6428 (make_pass_merge_phi): ...new function.
6429 * tree-complex.c (pass_lower_complex): Convert from a global struct to
6430 a subclass of gimple_opt_pass along with...
6431 (pass_data_lower_complex): ...new pass_data instance and...
6432 (make_pass_lower_complex): ...new function.
6433 (pass_lower_complex_O0): Convert from a global struct to a subclass of
6434 gimple_opt_pass along with...
6435 (pass_data_lower_complex_O0): ...new pass_data instance and...
6436 (make_pass_lower_complex_O0): ...new function.
6437 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
6438 subclass of gimple_opt_pass along with...
6439 (pass_data_lower_eh): ...new pass_data instance and...
6440 (make_pass_lower_eh): ...new function.
6441 (pass_refactor_eh): Convert from a global struct to a subclass of
6442 gimple_opt_pass along with...
6443 (pass_data_refactor_eh): ...new pass_data instance and...
6444 (make_pass_refactor_eh): ...new function.
6445 (pass_lower_resx): Convert from a global struct to a subclass of
6446 gimple_opt_pass along with...
6447 (pass_data_lower_resx): ...new pass_data instance and...
6448 (make_pass_lower_resx): ...new function.
6449 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
6450 of gimple_opt_pass along with...
6451 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
6452 (make_pass_lower_eh_dispatch): ...new function.
6453 (pass_cleanup_eh): Convert from a global struct to a subclass of
6454 gimple_opt_pass along with...
6455 (pass_data_cleanup_eh): ...new pass_data instance and...
6456 (make_pass_cleanup_eh): ...new function.
6457 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
6458 to a subclass of simple_ipa_opt_pass along with...
6459 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
6460 (make_pass_ipa_lower_emutls): ...new function.
6461 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
6462 a subclass of gimple_opt_pass along with...
6463 (pass_data_if_conversion): ...new pass_data instance and...
6464 (make_pass_if_conversion): ...new function.
6465 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
6466 subclass of gimple_opt_pass along with...
6467 (pass_data_build_ssa): ...new pass_data instance and...
6468 (make_pass_build_ssa): ...new function.
6469 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
6470 global struct to a subclass of gimple_opt_pass along with...
6471 (pass_data_loop_distribution): ...new pass_data instance and...
6472 (make_pass_loop_distribution): ...new function.
6473 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
6474 subclass of gimple_opt_pass along with...
6475 (pass_data_mudflap_1): ...new pass_data instance and...
6476 (make_pass_mudflap_1): ...new function.
6477 (pass_mudflap_2): Convert from a global struct to a subclass of
6478 gimple_opt_pass along with...
6479 (pass_data_mudflap_2): ...new pass_data instance and...
6480 (make_pass_mudflap_2): ...new function.
6481 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
6482 subclass of gimple_opt_pass along with...
6483 (pass_data_mudflap_1): ...new pass_data instance and...
6484 (make_pass_mudflap_1): ...new function.
6485 (pass_mudflap_2): Convert from a global struct to a subclass of
6486 gimple_opt_pass along with...
6487 (pass_data_mudflap_2): ...new pass_data instance and...
6488 (make_pass_mudflap_2): ...new function.
6489 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
6490 gimple_opt_pass along with...
6491 (pass_data_nrv): ...new pass_data instance and...
6492 (make_pass_nrv): ...new function.
6493 (pass_return_slot): Convert from a global struct to a subclass of
6494 gimple_opt_pass along with...
6495 (pass_data_return_slot): ...new pass_data instance and...
6496 (make_pass_return_slot): ...new function.
6497 * tree-object-size.c (pass_object_sizes): Convert from a global struct
6498 to a subclass of gimple_opt_pass along with...
6499 (pass_data_object_sizes): ...new pass_data instance and...
6500 (make_pass_object_sizes): ...new function.
6501 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
6502 global struct to a subclass of gimple_opt_pass along with...
6503 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
6505 (make_pass_cleanup_cfg_post_optimizing): ...new function.
6506 (pass_fixup_cfg): Convert from a global struct to a subclass of
6507 gimple_opt_pass along with...
6508 (pass_data_fixup_cfg): ...new pass_data instance and...
6509 (make_pass_fixup_cfg): ...new function.
6510 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
6511 (make_pass_mudflap_1): ...new function.
6512 (pass_mudflap_2): Replace declaration with that of...
6513 (make_pass_mudflap_2): ...new function.
6514 (pass_asan): Replace declaration with that of...
6515 (make_pass_asan): ...new function.
6516 (pass_asan_O0): Replace declaration with that of...
6517 (make_pass_asan_O0): ...new function.
6518 (pass_tsan): Replace declaration with that of...
6519 (make_pass_tsan): ...new function.
6520 (pass_tsan_O0): Replace declaration with that of...
6521 (make_pass_tsan_O0): ...new function.
6522 (pass_lower_cf): Replace declaration with that of...
6523 (make_pass_lower_cf): ...new function.
6524 (pass_refactor_eh): Replace declaration with that of...
6525 (make_pass_refactor_eh): ...new function.
6526 (pass_lower_eh): Replace declaration with that of...
6527 (make_pass_lower_eh): ...new function.
6528 (pass_lower_eh_dispatch): Replace declaration with that of...
6529 (make_pass_lower_eh_dispatch): ...new function.
6530 (pass_lower_resx): Replace declaration with that of...
6531 (make_pass_lower_resx): ...new function.
6532 (pass_build_cfg): Replace declaration with that of...
6533 (make_pass_build_cfg): ...new function.
6534 (pass_early_tree_profile): Replace declaration with that of...
6535 (make_pass_early_tree_profile): ...new function.
6536 (pass_cleanup_eh): Replace declaration with that of...
6537 (make_pass_cleanup_eh): ...new function.
6538 (pass_sra): Replace declaration with that of...
6539 (make_pass_sra): ...new function.
6540 (pass_sra_early): Replace declaration with that of...
6541 (make_pass_sra_early): ...new function.
6542 (pass_early_ipa_sra): Replace declaration with that of...
6543 (make_pass_early_ipa_sra): ...new function.
6544 (pass_tail_recursion): Replace declaration with that of...
6545 (make_pass_tail_recursion): ...new function.
6546 (pass_tail_calls): Replace declaration with that of...
6547 (make_pass_tail_calls): ...new function.
6548 (pass_tree_loop): Replace declaration with that of...
6549 (make_pass_tree_loop): ...new function.
6550 (pass_tree_loop_init): Replace declaration with that of...
6551 (make_pass_tree_loop_init): ...new function.
6552 (pass_lim): Replace declaration with that of...
6553 (make_pass_lim): ...new function.
6554 (pass_tree_unswitch): Replace declaration with that of...
6555 (make_pass_tree_unswitch): ...new function.
6556 (pass_predcom): Replace declaration with that of...
6557 (make_pass_predcom): ...new function.
6558 (pass_iv_canon): Replace declaration with that of...
6559 (make_pass_iv_canon): ...new function.
6560 (pass_scev_cprop): Replace declaration with that of...
6561 (make_pass_scev_cprop): ...new function.
6562 (pass_empty_loop): Replace declaration with that of...
6563 (make_pass_empty_loop): ...new function.
6564 (pass_record_bounds): Replace declaration with that of...
6565 (make_pass_record_bounds): ...new function.
6566 (pass_graphite): Replace declaration with that of...
6567 (make_pass_graphite): ...new function.
6568 (pass_graphite_transforms): Replace declaration with that of...
6569 (make_pass_graphite_transforms): ...new function.
6570 (pass_if_conversion): Replace declaration with that of...
6571 (make_pass_if_conversion): ...new function.
6572 (pass_loop_distribution): Replace declaration with that of...
6573 (make_pass_loop_distribution): ...new function.
6574 (pass_vectorize): Replace declaration with that of...
6575 (make_pass_vectorize): ...new function.
6576 (pass_slp_vectorize): Replace declaration with that of...
6577 (make_pass_slp_vectorize): ...new function.
6578 (pass_complete_unroll): Replace declaration with that of...
6579 (make_pass_complete_unroll): ...new function.
6580 (pass_complete_unrolli): Replace declaration with that of...
6581 (make_pass_complete_unrolli): ...new function.
6582 (pass_parallelize_loops): Replace declaration with that of...
6583 (make_pass_parallelize_loops): ...new function.
6584 (pass_loop_prefetch): Replace declaration with that of...
6585 (make_pass_loop_prefetch): ...new function.
6586 (pass_iv_optimize): Replace declaration with that of...
6587 (make_pass_iv_optimize): ...new function.
6588 (pass_tree_loop_done): Replace declaration with that of...
6589 (make_pass_tree_loop_done): ...new function.
6590 (pass_ch): Replace declaration with that of...
6591 (make_pass_ch): ...new function.
6592 (pass_ccp): Replace declaration with that of...
6593 (make_pass_ccp): ...new function.
6594 (pass_phi_only_cprop): Replace declaration with that of...
6595 (make_pass_phi_only_cprop): ...new function.
6596 (pass_build_ssa): Replace declaration with that of...
6597 (make_pass_build_ssa): ...new function.
6598 (pass_build_alias): Replace declaration with that of...
6599 (make_pass_build_alias): ...new function.
6600 (pass_build_ealias): Replace declaration with that of...
6601 (make_pass_build_ealias): ...new function.
6602 (pass_dominator): Replace declaration with that of...
6603 (make_pass_dominator): ...new function.
6604 (pass_dce): Replace declaration with that of...
6605 (make_pass_dce): ...new function.
6606 (pass_dce_loop): Replace declaration with that of...
6607 (make_pass_dce_loop): ...new function.
6608 (pass_cd_dce): Replace declaration with that of...
6609 (make_pass_cd_dce): ...new function.
6610 (pass_call_cdce): Replace declaration with that of...
6611 (make_pass_call_cdce): ...new function.
6612 (pass_merge_phi): Replace declaration with that of...
6613 (make_pass_merge_phi): ...new function.
6614 (pass_split_crit_edges): Replace declaration with that of...
6615 (make_pass_split_crit_edges): ...new function.
6616 (pass_pre): Replace declaration with that of...
6617 (make_pass_pre): ...new function.
6618 (pass_profile): Replace declaration with that of...
6619 (make_pass_profile): ...new function.
6620 (pass_strip_predict_hints): Replace declaration with that of...
6621 (make_pass_strip_predict_hints): ...new function.
6622 (pass_lower_complex_O0): Replace declaration with that of...
6623 (make_pass_lower_complex_O0): ...new function.
6624 (pass_lower_complex): Replace declaration with that of...
6625 (make_pass_lower_complex): ...new function.
6626 (pass_lower_vector): Replace declaration with that of...
6627 (make_pass_lower_vector): ...new function.
6628 (pass_lower_vector_ssa): Replace declaration with that of...
6629 (make_pass_lower_vector_ssa): ...new function.
6630 (pass_lower_omp): Replace declaration with that of...
6631 (make_pass_lower_omp): ...new function.
6632 (pass_diagnose_omp_blocks): Replace declaration with that of...
6633 (make_pass_diagnose_omp_blocks): ...new function.
6634 (pass_expand_omp): Replace declaration with that of...
6635 (make_pass_expand_omp): ...new function.
6636 (pass_expand_omp_ssa): Replace declaration with that of...
6637 (make_pass_expand_omp_ssa): ...new function.
6638 (pass_object_sizes): Replace declaration with that of...
6639 (make_pass_object_sizes): ...new function.
6640 (pass_strlen): Replace declaration with that of...
6641 (make_pass_strlen): ...new function.
6642 (pass_fold_builtins): Replace declaration with that of...
6643 (make_pass_fold_builtins): ...new function.
6644 (pass_stdarg): Replace declaration with that of...
6645 (make_pass_stdarg): ...new function.
6646 (pass_early_warn_uninitialized): Replace declaration with that of...
6647 (make_pass_early_warn_uninitialized): ...new function.
6648 (pass_late_warn_uninitialized): Replace declaration with that of...
6649 (make_pass_late_warn_uninitialized): ...new function.
6650 (pass_cse_reciprocals): Replace declaration with that of...
6651 (make_pass_cse_reciprocals): ...new function.
6652 (pass_cse_sincos): Replace declaration with that of...
6653 (make_pass_cse_sincos): ...new function.
6654 (pass_optimize_bswap): Replace declaration with that of...
6655 (make_pass_optimize_bswap): ...new function.
6656 (pass_optimize_widening_mul): Replace declaration with that of...
6657 (make_pass_optimize_widening_mul): ...new function.
6658 (pass_warn_function_return): Replace declaration with that of...
6659 (make_pass_warn_function_return): ...new function.
6660 (pass_warn_function_noreturn): Replace declaration with that of...
6661 (make_pass_warn_function_noreturn): ...new function.
6662 (pass_cselim): Replace declaration with that of...
6663 (make_pass_cselim): ...new function.
6664 (pass_phiopt): Replace declaration with that of...
6665 (make_pass_phiopt): ...new function.
6666 (pass_forwprop): Replace declaration with that of...
6667 (make_pass_forwprop): ...new function.
6668 (pass_phiprop): Replace declaration with that of...
6669 (make_pass_phiprop): ...new function.
6670 (pass_tree_ifcombine): Replace declaration with that of...
6671 (make_pass_tree_ifcombine): ...new function.
6672 (pass_dse): Replace declaration with that of...
6673 (make_pass_dse): ...new function.
6674 (pass_nrv): Replace declaration with that of...
6675 (make_pass_nrv): ...new function.
6676 (pass_rename_ssa_copies): Replace declaration with that of...
6677 (make_pass_rename_ssa_copies): ...new function.
6678 (pass_sink_code): Replace declaration with that of...
6679 (make_pass_sink_code): ...new function.
6680 (pass_fre): Replace declaration with that of...
6681 (make_pass_fre): ...new function.
6682 (pass_check_data_deps): Replace declaration with that of...
6683 (make_pass_check_data_deps): ...new function.
6684 (pass_copy_prop): Replace declaration with that of...
6685 (make_pass_copy_prop): ...new function.
6686 (pass_vrp): Replace declaration with that of...
6687 (make_pass_vrp): ...new function.
6688 (pass_uncprop): Replace declaration with that of...
6689 (make_pass_uncprop): ...new function.
6690 (pass_return_slot): Replace declaration with that of...
6691 (make_pass_return_slot): ...new function.
6692 (pass_reassoc): Replace declaration with that of...
6693 (make_pass_reassoc): ...new function.
6694 (pass_rebuild_cgraph_edges): Replace declaration with that of...
6695 (make_pass_rebuild_cgraph_edges): ...new function.
6696 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
6697 (make_pass_remove_cgraph_callee_edges): ...new function.
6698 (pass_build_cgraph_edges): Replace declaration with that of...
6699 (make_pass_build_cgraph_edges): ...new function.
6700 (pass_local_pure_const): Replace declaration with that of...
6701 (make_pass_local_pure_const): ...new function.
6702 (pass_tracer): Replace declaration with that of...
6703 (make_pass_tracer): ...new function.
6704 (pass_warn_unused_result): Replace declaration with that of...
6705 (make_pass_warn_unused_result): ...new function.
6706 (pass_diagnose_tm_blocks): Replace declaration with that of...
6707 (make_pass_diagnose_tm_blocks): ...new function.
6708 (pass_lower_tm): Replace declaration with that of...
6709 (make_pass_lower_tm): ...new function.
6710 (pass_tm_init): Replace declaration with that of...
6711 (make_pass_tm_init): ...new function.
6712 (pass_tm_mark): Replace declaration with that of...
6713 (make_pass_tm_mark): ...new function.
6714 (pass_tm_memopt): Replace declaration with that of...
6715 (make_pass_tm_memopt): ...new function.
6716 (pass_tm_edges): Replace declaration with that of...
6717 (make_pass_tm_edges): ...new function.
6718 (pass_split_functions): Replace declaration with that of...
6719 (make_pass_split_functions): ...new function.
6720 (pass_feedback_split_functions): Replace declaration with that of...
6721 (make_pass_feedback_split_functions): ...new function.
6722 (pass_strength_reduction): Replace declaration with that of...
6723 (make_pass_strength_reduction): ...new function.
6724 (pass_ipa_lower_emutls): Replace declaration with that of...
6725 (make_pass_ipa_lower_emutls): ...new function.
6726 (pass_ipa_function_and_variable_visibility): Replace declaration with
6728 (make_pass_ipa_function_and_variable_visibility): ...new function.
6729 (pass_ipa_tree_profile): Replace declaration with that of...
6730 (make_pass_ipa_tree_profile): ...new function.
6731 (pass_early_local_passes): Replace declaration with that of...
6732 (make_pass_early_local_passes): ...new function.
6733 (pass_ipa_whole_program_visibility): Replace declaration with that
6735 (make_pass_ipa_whole_program_visibility): ...new function.
6736 (pass_ipa_lto_gimple_out): Replace declaration with that of...
6737 (make_pass_ipa_lto_gimple_out): ...new function.
6738 (pass_ipa_increase_alignment): Replace declaration with that of...
6739 (make_pass_ipa_increase_alignment): ...new function.
6740 (pass_ipa_inline): Replace declaration with that of...
6741 (make_pass_ipa_inline): ...new function.
6742 (pass_ipa_free_lang_data): Replace declaration with that of...
6743 (make_pass_ipa_free_lang_data): ...new function.
6744 (pass_ipa_free_inline_summary): Replace declaration with that of...
6745 (make_pass_ipa_free_inline_summary): ...new function.
6746 (pass_ipa_cp): Replace declaration with that of...
6747 (make_pass_ipa_cp): ...new function.
6748 (pass_ipa_reference): Replace declaration with that of...
6749 (make_pass_ipa_reference): ...new function.
6750 (pass_ipa_pure_const): Replace declaration with that of...
6751 (make_pass_ipa_pure_const): ...new function.
6752 (pass_ipa_pta): Replace declaration with that of...
6753 (make_pass_ipa_pta): ...new function.
6754 (pass_ipa_lto_finish_out): Replace declaration with that of...
6755 (make_pass_ipa_lto_finish_out): ...new function.
6756 (pass_ipa_tm): Replace declaration with that of...
6757 (make_pass_ipa_tm): ...new function.
6758 (pass_ipa_profile): Replace declaration with that of...
6759 (make_pass_ipa_profile): ...new function.
6760 (pass_ipa_cdtor_merge): Replace declaration with that of...
6761 (make_pass_ipa_cdtor_merge): ...new function.
6762 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
6764 (make_pass_cleanup_cfg_post_optimizing): ...new function.
6765 (pass_init_datastructures): Replace declaration with that of...
6766 (make_pass_init_datastructures): ...new function.
6767 (pass_fixup_cfg): Replace declaration with that of...
6768 (make_pass_fixup_cfg): ...new function.
6769 (pass_expand): Replace declaration with that of...
6770 (make_pass_expand): ...new function.
6771 (pass_instantiate_virtual_regs): Replace declaration with that of...
6772 (make_pass_instantiate_virtual_regs): ...new function.
6773 (pass_rtl_fwprop): Replace declaration with that of...
6774 (make_pass_rtl_fwprop): ...new function.
6775 (pass_rtl_fwprop_addr): Replace declaration with that of...
6776 (make_pass_rtl_fwprop_addr): ...new function.
6777 (pass_jump): Replace declaration with that of...
6778 (make_pass_jump): ...new function.
6779 (pass_jump2): Replace declaration with that of...
6780 (make_pass_jump2): ...new function.
6781 (pass_lower_subreg): Replace declaration with that of...
6782 (make_pass_lower_subreg): ...new function.
6783 (pass_cse): Replace declaration with that of...
6784 (make_pass_cse): ...new function.
6785 (pass_fast_rtl_dce): Replace declaration with that of...
6786 (make_pass_fast_rtl_dce): ...new function.
6787 (pass_ud_rtl_dce): Replace declaration with that of...
6788 (make_pass_ud_rtl_dce): ...new function.
6789 (pass_rtl_dce): Replace declaration with that of...
6790 (make_pass_rtl_dce): ...new function.
6791 (pass_rtl_dse1): Replace declaration with that of...
6792 (make_pass_rtl_dse1): ...new function.
6793 (pass_rtl_dse2): Replace declaration with that of...
6794 (make_pass_rtl_dse2): ...new function.
6795 (pass_rtl_dse3): Replace declaration with that of...
6796 (make_pass_rtl_dse3): ...new function.
6797 (pass_rtl_cprop): Replace declaration with that of...
6798 (make_pass_rtl_cprop): ...new function.
6799 (pass_rtl_pre): Replace declaration with that of...
6800 (make_pass_rtl_pre): ...new function.
6801 (pass_rtl_hoist): Replace declaration with that of...
6802 (make_pass_rtl_hoist): ...new function.
6803 (pass_rtl_store_motion): Replace declaration with that of...
6804 (make_pass_rtl_store_motion): ...new function.
6805 (pass_cse_after_global_opts): Replace declaration with that of...
6806 (make_pass_cse_after_global_opts): ...new function.
6807 (pass_rtl_ifcvt): Replace declaration with that of...
6808 (make_pass_rtl_ifcvt): ...new function.
6809 (pass_into_cfg_layout_mode): Replace declaration with that of...
6810 (make_pass_into_cfg_layout_mode): ...new function.
6811 (pass_outof_cfg_layout_mode): Replace declaration with that of...
6812 (make_pass_outof_cfg_layout_mode): ...new function.
6813 (pass_loop2): Replace declaration with that of...
6814 (make_pass_loop2): ...new function.
6815 (pass_rtl_loop_init): Replace declaration with that of...
6816 (make_pass_rtl_loop_init): ...new function.
6817 (pass_rtl_move_loop_invariants): Replace declaration with that of...
6818 (make_pass_rtl_move_loop_invariants): ...new function.
6819 (pass_rtl_unswitch): Replace declaration with that of...
6820 (make_pass_rtl_unswitch): ...new function.
6821 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
6822 (make_pass_rtl_unroll_and_peel_loops): ...new function.
6823 (pass_rtl_doloop): Replace declaration with that of...
6824 (make_pass_rtl_doloop): ...new function.
6825 (pass_rtl_loop_done): Replace declaration with that of...
6826 (make_pass_rtl_loop_done): ...new function.
6827 (pass_web): Replace declaration with that of...
6828 (make_pass_web): ...new function.
6829 (pass_cse2): Replace declaration with that of...
6830 (make_pass_cse2): ...new function.
6831 (pass_df_initialize_opt): Replace declaration with that of...
6832 (make_pass_df_initialize_opt): ...new function.
6833 (pass_df_initialize_no_opt): Replace declaration with that of...
6834 (make_pass_df_initialize_no_opt): ...new function.
6835 (pass_reginfo_init): Replace declaration with that of...
6836 (make_pass_reginfo_init): ...new function.
6837 (pass_inc_dec): Replace declaration with that of...
6838 (make_pass_inc_dec): ...new function.
6839 (pass_stack_ptr_mod): Replace declaration with that of...
6840 (make_pass_stack_ptr_mod): ...new function.
6841 (pass_initialize_regs): Replace declaration with that of...
6842 (make_pass_initialize_regs): ...new function.
6843 (pass_combine): Replace declaration with that of...
6844 (make_pass_combine): ...new function.
6845 (pass_if_after_combine): Replace declaration with that of...
6846 (make_pass_if_after_combine): ...new function.
6847 (pass_ree): Replace declaration with that of...
6848 (make_pass_ree): ...new function.
6849 (pass_partition_blocks): Replace declaration with that of...
6850 (make_pass_partition_blocks): ...new function.
6851 (pass_match_asm_constraints): Replace declaration with that of...
6852 (make_pass_match_asm_constraints): ...new function.
6853 (pass_regmove): Replace declaration with that of...
6854 (make_pass_regmove): ...new function.
6855 (pass_split_all_insns): Replace declaration with that of...
6856 (make_pass_split_all_insns): ...new function.
6857 (pass_fast_rtl_byte_dce): Replace declaration with that of...
6858 (make_pass_fast_rtl_byte_dce): ...new function.
6859 (pass_lower_subreg2): Replace declaration with that of...
6860 (make_pass_lower_subreg2): ...new function.
6861 (pass_mode_switching): Replace declaration with that of...
6862 (make_pass_mode_switching): ...new function.
6863 (pass_sms): Replace declaration with that of...
6864 (make_pass_sms): ...new function.
6865 (pass_sched): Replace declaration with that of...
6866 (make_pass_sched): ...new function.
6867 (pass_ira): Replace declaration with that of...
6868 (make_pass_ira): ...new function.
6869 (pass_reload): Replace declaration with that of...
6870 (make_pass_reload): ...new function.
6871 (pass_clean_state): Replace declaration with that of...
6872 (make_pass_clean_state): ...new function.
6873 (pass_branch_prob): Replace declaration with that of...
6874 (make_pass_branch_prob): ...new function.
6875 (pass_value_profile_transformations): Replace declaration with that
6877 (make_pass_value_profile_transformations): ...new function.
6878 (pass_postreload_cse): Replace declaration with that of...
6879 (make_pass_postreload_cse): ...new function.
6880 (pass_gcse2): Replace declaration with that of...
6881 (make_pass_gcse2): ...new function.
6882 (pass_split_after_reload): Replace declaration with that of...
6883 (make_pass_split_after_reload): ...new function.
6884 (pass_branch_target_load_optimize1): Replace declaration with that
6886 (make_pass_branch_target_load_optimize1): ...new function.
6887 (pass_thread_prologue_and_epilogue): Replace declaration with that
6889 (make_pass_thread_prologue_and_epilogue): ...new function.
6890 (pass_stack_adjustments): Replace declaration with that of...
6891 (make_pass_stack_adjustments): ...new function.
6892 (pass_peephole2): Replace declaration with that of...
6893 (make_pass_peephole2): ...new function.
6894 (pass_if_after_reload): Replace declaration with that of...
6895 (make_pass_if_after_reload): ...new function.
6896 (pass_regrename): Replace declaration with that of...
6897 (make_pass_regrename): ...new function.
6898 (pass_cprop_hardreg): Replace declaration with that of...
6899 (make_pass_cprop_hardreg): ...new function.
6900 (pass_reorder_blocks): Replace declaration with that of...
6901 (make_pass_reorder_blocks): ...new function.
6902 (pass_branch_target_load_optimize2): Replace declaration with that
6904 (make_pass_branch_target_load_optimize2): ...new function.
6905 (pass_leaf_regs): Replace declaration with that of...
6906 (make_pass_leaf_regs): ...new function.
6907 (pass_split_before_sched2): Replace declaration with that of...
6908 (make_pass_split_before_sched2): ...new function.
6909 (pass_compare_elim_after_reload): Replace declaration with that of...
6910 (make_pass_compare_elim_after_reload): ...new function.
6911 (pass_sched2): Replace declaration with that of...
6912 (make_pass_sched2): ...new function.
6913 (pass_stack_regs): Replace declaration with that of...
6914 (make_pass_stack_regs): ...new function.
6915 (pass_stack_regs_run): Replace declaration with that of...
6916 (make_pass_stack_regs_run): ...new function.
6917 (pass_df_finish): Replace declaration with that of...
6918 (make_pass_df_finish): ...new function.
6919 (pass_compute_alignments): Replace declaration with that of...
6920 (make_pass_compute_alignments): ...new function.
6921 (pass_duplicate_computed_gotos): Replace declaration with that of...
6922 (make_pass_duplicate_computed_gotos): ...new function.
6923 (pass_variable_tracking): Replace declaration with that of...
6924 (make_pass_variable_tracking): ...new function.
6925 (pass_free_cfg): Replace declaration with that of...
6926 (make_pass_free_cfg): ...new function.
6927 (pass_machine_reorg): Replace declaration with that of...
6928 (make_pass_machine_reorg): ...new function.
6929 (pass_cleanup_barriers): Replace declaration with that of...
6930 (make_pass_cleanup_barriers): ...new function.
6931 (pass_delay_slots): Replace declaration with that of...
6932 (make_pass_delay_slots): ...new function.
6933 (pass_split_for_shorten_branches): Replace declaration with that of...
6934 (make_pass_split_for_shorten_branches): ...new function.
6935 (pass_split_before_regstack): Replace declaration with that of...
6936 (make_pass_split_before_regstack): ...new function.
6937 (pass_convert_to_eh_region_ranges): Replace declaration with that
6939 (make_pass_convert_to_eh_region_ranges): ...new function.
6940 (pass_shorten_branches): Replace declaration with that of...
6941 (make_pass_shorten_branches): ...new function.
6942 (pass_set_nothrow_function_flags): Replace declaration with that of...
6943 (make_pass_set_nothrow_function_flags): ...new function.
6944 (pass_dwarf2_frame): Replace declaration with that of...
6945 (make_pass_dwarf2_frame): ...new function.
6946 (pass_final): Replace declaration with that of...
6947 (make_pass_final): ...new function.
6948 (pass_rtl_seqabstr): Replace declaration with that of...
6949 (make_pass_rtl_seqabstr): ...new function.
6950 (pass_release_ssa_names): Replace declaration with that of...
6951 (make_pass_release_ssa_names): ...new function.
6952 (pass_early_inline): Replace declaration with that of...
6953 (make_pass_early_inline): ...new function.
6954 (pass_inline_parameters): Replace declaration with that of...
6955 (make_pass_inline_parameters): ...new function.
6956 (pass_update_address_taken): Replace declaration with that of...
6957 (make_pass_update_address_taken): ...new function.
6958 (pass_convert_switch): Replace declaration with that of...
6959 (make_pass_convert_switch): ...new function.
6960 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
6961 to a subclass of simple_ipa_opt_pass along with...
6962 (pass_data_ipa_tree_profile): ...new pass_data instance and...
6963 (make_pass_ipa_tree_profile): ...new function.
6964 * tree-sra.c (pass_sra_early): Convert from a global struct to a
6965 subclass of gimple_opt_pass along with...
6966 (pass_data_sra_early): ...new pass_data instance and...
6967 (make_pass_sra_early): ...new function.
6968 (pass_sra): Convert from a global struct to a subclass of
6969 gimple_opt_pass along with...
6970 (pass_data_sra): ...new pass_data instance and...
6971 (make_pass_sra): ...new function.
6972 (pass_early_ipa_sra): Convert from a global struct to a subclass of
6973 gimple_opt_pass along with...
6974 (pass_data_early_ipa_sra): ...new pass_data instance and...
6975 (make_pass_early_ipa_sra): ...new function.
6976 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
6977 subclass of gimple_opt_pass along with...
6978 (pass_data_ccp): ...new pass_data instance and...
6979 (make_pass_ccp): ...new function.
6980 (pass_fold_builtins): Convert from a global struct to a subclass of
6981 gimple_opt_pass along with...
6982 (pass_data_fold_builtins): ...new pass_data instance and...
6983 (make_pass_fold_builtins): ...new function.
6984 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
6985 subclass of gimple_opt_pass along with...
6986 (pass_data_copy_prop): ...new pass_data instance and...
6987 (make_pass_copy_prop): ...new function.
6988 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
6989 global struct to a subclass of gimple_opt_pass along with...
6990 (pass_data_rename_ssa_copies): ...new pass_data instance and...
6991 (make_pass_rename_ssa_copies): ...new function.
6992 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
6993 subclass of gimple_opt_pass along with...
6994 (pass_data_dce): ...new pass_data instance and...
6995 (make_pass_dce): ...new function.
6996 (pass_dce_loop): Convert from a global struct to a subclass of
6997 gimple_opt_pass along with...
6998 (pass_data_dce_loop): ...new pass_data instance and...
6999 (make_pass_dce_loop): ...new function.
7000 (pass_cd_dce): Convert from a global struct to a subclass of
7001 gimple_opt_pass along with...
7002 (pass_data_cd_dce): ...new pass_data instance and...
7003 (make_pass_cd_dce): ...new function.
7004 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
7005 subclass of gimple_opt_pass along with...
7006 (pass_data_dominator): ...new pass_data instance and...
7007 (make_pass_dominator): ...new function.
7008 (pass_phi_only_cprop): Convert from a global struct to a subclass of
7009 gimple_opt_pass along with...
7010 (pass_data_phi_only_cprop): ...new pass_data instance and...
7011 (make_pass_phi_only_cprop): ...new function.
7012 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
7013 subclass of gimple_opt_pass along with...
7014 (pass_data_dse): ...new pass_data instance and...
7015 (make_pass_dse): ...new function.
7016 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
7017 a subclass of gimple_opt_pass along with...
7018 (pass_data_forwprop): ...new pass_data instance and...
7019 (make_pass_forwprop): ...new function.
7020 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
7021 struct to a subclass of gimple_opt_pass along with...
7022 (pass_data_tree_ifcombine): ...new pass_data instance and...
7023 (make_pass_tree_ifcombine): ...new function.
7024 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
7025 subclass of gimple_opt_pass along with...
7026 (pass_data_ch): ...new pass_data instance and...
7027 (make_pass_ch): ...new function.
7028 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
7029 subclass of gimple_opt_pass along with...
7030 (pass_data_tree_loop): ...new pass_data instance and...
7031 (make_pass_tree_loop): ...new function.
7032 (pass_tree_loop_init): Convert from a global struct to a subclass of
7033 gimple_opt_pass along with...
7034 (pass_data_tree_loop_init): ...new pass_data instance and...
7035 (make_pass_tree_loop_init): ...new function.
7036 (pass_lim): Convert from a global struct to a subclass of
7037 gimple_opt_pass along with...
7038 (pass_data_lim): ...new pass_data instance and...
7039 (make_pass_lim): ...new function.
7040 (pass_tree_unswitch): Convert from a global struct to a subclass of
7041 gimple_opt_pass along with...
7042 (pass_data_tree_unswitch): ...new pass_data instance and...
7043 (make_pass_tree_unswitch): ...new function.
7044 (pass_predcom): Convert from a global struct to a subclass of
7045 gimple_opt_pass along with...
7046 (pass_data_predcom): ...new pass_data instance and...
7047 (make_pass_predcom): ...new function.
7048 (pass_vectorize): Convert from a global struct to a subclass of
7049 gimple_opt_pass along with...
7050 (pass_data_vectorize): ...new pass_data instance and...
7051 (make_pass_vectorize): ...new function.
7052 (pass_graphite): Convert from a global struct to a subclass of
7053 gimple_opt_pass along with...
7054 (pass_data_graphite): ...new pass_data instance and...
7055 (make_pass_graphite): ...new function.
7056 (pass_graphite_transforms): Convert from a global struct to a subclass
7057 of gimple_opt_pass along with...
7058 (pass_data_graphite_transforms): ...new pass_data instance and...
7059 (make_pass_graphite_transforms): ...new function.
7060 (pass_check_data_deps): Convert from a global struct to a subclass of
7061 gimple_opt_pass along with...
7062 (pass_data_check_data_deps): ...new pass_data instance and...
7063 (make_pass_check_data_deps): ...new function.
7064 (pass_iv_canon): Convert from a global struct to a subclass of
7065 gimple_opt_pass along with...
7066 (pass_data_iv_canon): ...new pass_data instance and...
7067 (make_pass_iv_canon): ...new function.
7068 (pass_scev_cprop): Convert from a global struct to a subclass of
7069 gimple_opt_pass along with...
7070 (pass_data_scev_cprop): ...new pass_data instance and...
7071 (make_pass_scev_cprop): ...new function.
7072 (pass_record_bounds): Convert from a global struct to a subclass of
7073 gimple_opt_pass along with...
7074 (pass_data_record_bounds): ...new pass_data instance and...
7075 (make_pass_record_bounds): ...new function.
7076 (pass_complete_unroll): Convert from a global struct to a subclass of
7077 gimple_opt_pass along with...
7078 (pass_data_complete_unroll): ...new pass_data instance and...
7079 (make_pass_complete_unroll): ...new function.
7080 (pass_complete_unrolli): Convert from a global struct to a subclass of
7081 gimple_opt_pass along with...
7082 (pass_data_complete_unrolli): ...new pass_data instance and...
7083 (make_pass_complete_unrolli): ...new function.
7084 (pass_parallelize_loops): Convert from a global struct to a subclass
7085 of gimple_opt_pass along with...
7086 (pass_data_parallelize_loops): ...new pass_data instance and...
7087 (make_pass_parallelize_loops): ...new function.
7088 (pass_loop_prefetch): Convert from a global struct to a subclass of
7089 gimple_opt_pass along with...
7090 (pass_data_loop_prefetch): ...new pass_data instance and...
7091 (make_pass_loop_prefetch): ...new function.
7092 (pass_iv_optimize): Convert from a global struct to a subclass of
7093 gimple_opt_pass along with...
7094 (pass_data_iv_optimize): ...new pass_data instance and...
7095 (make_pass_iv_optimize): ...new function.
7096 (pass_tree_loop_done): Convert from a global struct to a subclass of
7097 gimple_opt_pass along with...
7098 (pass_data_tree_loop_done): ...new pass_data instance and...
7099 (make_pass_tree_loop_done): ...new function.
7100 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
7101 struct to a subclass of gimple_opt_pass along with...
7102 (pass_data_cse_reciprocals): ...new pass_data instance and...
7103 (make_pass_cse_reciprocals): ...new function.
7104 (pass_cse_sincos): Convert from a global struct to a subclass of
7105 gimple_opt_pass along with...
7106 (pass_data_cse_sincos): ...new pass_data instance and...
7107 (make_pass_cse_sincos): ...new function.
7108 (pass_optimize_bswap): Convert from a global struct to a subclass of
7109 gimple_opt_pass along with...
7110 (pass_data_optimize_bswap): ...new pass_data instance and...
7111 (make_pass_optimize_bswap): ...new function.
7112 (pass_optimize_widening_mul): Convert from a global struct to a
7113 subclass of gimple_opt_pass along with...
7114 (pass_data_optimize_widening_mul): ...new pass_data instance and...
7115 (make_pass_optimize_widening_mul): ...new function.
7116 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
7117 subclass of gimple_opt_pass along with...
7118 (pass_data_phiopt): ...new pass_data instance and...
7119 (make_pass_phiopt): ...new function.
7120 (pass_cselim): Convert from a global struct to a subclass of
7121 gimple_opt_pass along with...
7122 (pass_data_cselim): ...new pass_data instance and...
7123 (make_pass_cselim): ...new function.
7124 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
7125 subclass of gimple_opt_pass along with...
7126 (pass_data_phiprop): ...new pass_data instance and...
7127 (make_pass_phiprop): ...new function.
7128 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
7129 subclass of gimple_opt_pass along with...
7130 (pass_data_pre): ...new pass_data instance and...
7131 (make_pass_pre): ...new function.
7132 (pass_fre): Convert from a global struct to a subclass of
7133 gimple_opt_pass along with...
7134 (pass_data_fre): ...new pass_data instance and...
7135 (make_pass_fre): ...new function.
7136 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
7137 subclass of gimple_opt_pass along with...
7138 (pass_data_reassoc): ...new pass_data instance and...
7139 (make_pass_reassoc): ...new function.
7140 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
7141 subclass of gimple_opt_pass along with...
7142 (pass_data_sink_code): ...new pass_data instance and...
7143 (make_pass_sink_code): ...new function.
7144 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
7145 subclass of gimple_opt_pass along with...
7146 (pass_data_strlen): ...new pass_data instance and...
7147 (make_pass_strlen): ...new function.
7148 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
7149 struct to a subclass of gimple_opt_pass along with...
7150 (pass_data_build_alias): ...new pass_data instance and...
7151 (make_pass_build_alias): ...new function.
7152 (pass_build_ealias): Convert from a global struct to a subclass of
7153 gimple_opt_pass along with...
7154 (pass_data_build_ealias): ...new pass_data instance and...
7155 (make_pass_build_ealias): ...new function.
7156 (pass_ipa_pta): Convert from a global struct to a subclass of
7157 simple_ipa_opt_pass along with...
7158 (pass_data_ipa_pta): ...new pass_data instance and...
7159 (make_pass_ipa_pta): ...new function.
7160 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
7161 subclass of gimple_opt_pass along with...
7162 (pass_data_uncprop): ...new pass_data instance and...
7163 (make_pass_uncprop): ...new function.
7164 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
7165 global struct to a subclass of gimple_opt_pass along with...
7166 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
7167 (make_pass_late_warn_uninitialized): ...new function.
7168 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
7169 to a subclass of gimple_opt_pass along with...
7170 (pass_data_init_datastructures): ...new pass_data instance and...
7171 (make_pass_init_datastructures): ...new function.
7172 (pass_early_warn_uninitialized): Convert from a global struct to a
7173 subclass of gimple_opt_pass along with...
7174 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
7175 (make_pass_early_warn_uninitialized): ...new function.
7176 (pass_update_address_taken): Convert from a global struct to a
7177 subclass of gimple_opt_pass along with...
7178 (pass_data_update_address_taken): ...new pass_data instance and...
7179 (make_pass_update_address_taken): ...new function.
7180 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
7181 struct to a subclass of gimple_opt_pass along with...
7182 (pass_data_release_ssa_names): ...new pass_data instance and...
7183 (make_pass_release_ssa_names): ...new function.
7184 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
7185 subclass of gimple_opt_pass along with...
7186 (pass_data_stdarg): ...new pass_data instance and...
7187 (make_pass_stdarg): ...new function.
7188 * tree-switch-conversion.c (pass_convert_switch): Convert from a
7189 global struct to a subclass of gimple_opt_pass along with...
7190 (pass_data_convert_switch): ...new pass_data instance and...
7191 (make_pass_convert_switch): ...new function.
7192 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
7193 to a subclass of gimple_opt_pass along with...
7194 (pass_data_tail_recursion): ...new pass_data instance and...
7195 (make_pass_tail_recursion): ...new function.
7196 (pass_tail_calls): Convert from a global struct to a subclass of
7197 gimple_opt_pass along with...
7198 (pass_data_tail_calls): ...new pass_data instance and...
7199 (make_pass_tail_calls): ...new function.
7200 * tree-vect-generic.c (pass_lower_vector): Convert from a global
7201 struct to a subclass of gimple_opt_pass along with...
7202 (pass_data_lower_vector): ...new pass_data instance and...
7203 (make_pass_lower_vector): ...new function.
7204 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
7205 gimple_opt_pass along with...
7206 (pass_data_lower_vector_ssa): ...new pass_data instance and...
7207 (make_pass_lower_vector_ssa): ...new function.
7208 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
7209 to a subclass of gimple_opt_pass along with...
7210 (pass_data_slp_vectorize): ...new pass_data instance and...
7211 (make_pass_slp_vectorize): ...new function.
7212 (pass_ipa_increase_alignment): Convert from a global struct to a
7213 subclass of simple_ipa_opt_pass along with...
7214 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
7215 (make_pass_ipa_increase_alignment): ...new function.
7216 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
7217 gimple_opt_pass along with...
7218 (pass_data_vrp): ...new pass_data instance and...
7219 (make_pass_vrp): ...new function.
7220 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
7221 subclass of simple_ipa_opt_pass along with...
7222 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
7223 (make_pass_ipa_free_lang_data): ...new function.
7224 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
7225 gimple_opt_pass along with...
7226 (pass_data_tsan): ...new pass_data instance and...
7227 (make_pass_tsan): ...new function.
7228 (pass_tsan_O0): Convert from a global struct to a subclass of
7229 gimple_opt_pass along with...
7230 (pass_data_tsan_O0): ...new pass_data instance and...
7231 (make_pass_tsan_O0): ...new function.
7232 * var-tracking.c (pass_variable_tracking): Convert from a global
7233 struct to a subclass of rtl_opt_pass along with...
7234 (pass_data_variable_tracking): ...new pass_data instance and...
7235 (make_pass_variable_tracking): ...new function.
7236 * web.c (pass_web): Convert from a global struct to a subclass of
7237 rtl_opt_pass along with...
7238 (pass_data_web): ...new pass_data instance and...
7239 (make_pass_web): ...new function.
7240 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
7241 declaration with that of...
7242 (make_pass_mode_switch_use): ...new function.
7243 (pass_resolve_sw_modes): Replace declaration with that of...
7244 (make_pass_resolve_sw_modes): ...new function.
7245 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
7246 from a global struct to a subclass of rtl_opt_pass along with...
7247 (pass_data_mode_switch_use): ...new pass_data instance and...
7248 (make_pass_mode_switch_use): ...new function.
7249 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
7250 from a global struct to a subclass of rtl_opt_pass along with...
7251 (pass_data_resolve_sw_modes): ...new pass_data instance and...
7252 (make_pass_resolve_sw_modes): ...new function.
7253 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
7254 struct to a subclass of rtl_opt_pass along with...
7255 (pass_data_insert_vzeroupper): ...new pass_data instance and...
7256 (make_pass_insert_vzeroupper): ...new function.
7257 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
7258 global struct to a subclass of rtl_opt_pass along with...
7259 (pass_data_work_around_errata): ...new pass_data instance and...
7260 (make_pass_work_around_errata): ...new function.
7261 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
7262 struct to a subclass of rtl_opt_pass along with...
7263 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
7264 (make_pass_mips_machine_reorg2): ...new function.
7266 2013-08-05 David Malcolm <dmalcolm@redhat.com>
7268 * passes.c (pass_manager::operator new): New.
7270 2013-08-05 David Malcolm <dmalcolm@redhat.com>
7272 Handwritten part of conversion of passes to C++ classes.
7274 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
7275 (toplev.o): Add dep on PASS_MANAGER_H.
7276 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
7277 of early local pases to reflect this moving from a global to a
7278 member of gcc::pass_manager.
7279 (cgraph_add_new_function): Likewise.
7280 * lto-cgraph.c (lto_output_node): Update for conversion of
7281 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
7282 * passes.c (opt_pass::clone): New.
7283 (opt_pass::gate): New.
7284 (opt_pass::execute): New.
7285 (opt_pass::opt_pass): New.
7286 (pass_manager::execute_early_local_passes): New.
7287 (pass_manager::execute_pass_mode_switching): new.
7288 (finish_optimization_passes): Convert to...
7289 (pass_manager::finish_optimization_passes): ...this.
7290 (finish_optimization_passes): Update for conversion of passes to
7292 (register_dump_files_1): Use has_gate since we cannot portably
7293 check a vtable entry against NULL.
7294 (dump_one_pass): Likewise.
7295 (ipa_write_summaries_2): Likewise.
7296 (ipa_write_optimization_summaries_1): Likewise.
7297 (ipa_read_summaries_1): Likewise.
7298 (ipa_read_optimization_summaries_1): Likewise.
7299 (execute_ipa_stmt_fixups): Likewise.
7300 (pass_manager::pass_manager): Rewrite pass-creation, invoking
7301 pass-creation functions rather than wiring up globals, and
7302 storing the results in fields of pass_manager generated using
7304 (pass_manager::dump_profile_report): Update for conversion of
7305 passes to C++ classes.
7306 (pass_manager::execute_ipa_summary_passes): Likewise.
7307 (execute_one_ipa_transform_pass): Likewise.
7308 (execute_one_pass): Use has_gate and has_execute since we cannot
7309 portably check a vtable entry against NULL.
7310 * pass_manager.h (pass_manager::finish_optimization_passes): New.
7311 (pass_manager): Use pass-instances.def to add fields for the
7312 various pass instances.
7313 * toplev.c (finalize): Update for move of
7314 finish_optimization_passes to a method of gcc::pass_manager.
7315 * toplev.h (finish_optimization_passes): Move to method of class
7317 * tree-pass.h (struct pass_data): New.
7318 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
7319 (opt_pass::gate): Convert to virtual function.
7320 (opt_pass::~opt_pass): New.
7321 (opt_pass::clone): New.
7322 (opt_pass::execute): Convert to virtual function.
7323 (opt_pass::opt_pass): New.
7324 (opt_pass::ctxt_): new.
7325 (gimple_opt_pass): Convert to subclass of opt_pass.
7326 (gimple_opt_pass::gimple_opt_pass): New.
7327 (rtl_opt_pass): Convert to subclass of opt_pass.
7328 (rtl_opt_pass::rtl_opt_pass): New.
7329 (ipa_opt_pass_d): Convert to subclass of opt_pass.
7330 (ipa_opt_pass_d::ipa_opt_pass_d): New.
7331 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
7332 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
7333 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
7334 invocation of pass_mode_switching to reflect this moving from a
7335 global to a member of gcc::pass_manager.
7336 (ix86_option_override): Rework how pass_insert_vzeroupper is
7337 added to the pass_manager to reflect autogenerated changes.
7338 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
7340 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
7342 PR rtl-optimization/57708
7343 * recog.c (peep2_find_free_register): Validate all regs in a
7346 2013-08-05 Jan Hubicka <jh@suse.cz>
7349 * cgraph.c (verify_cgraph_node): Accept local flags from other
7351 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
7352 (function_and_variable_visibility): Likewise.
7353 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
7355 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
7357 * graph.c (init_graph_slim_pretty_print): Remove.
7358 (print_graph_cfg): Do not call it. Use local pretty printer.
7359 (start_graph_dump): Likewise.
7361 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
7363 * gimple-pretty-print.c (buffer): Remove.
7364 (initialized): Likewise.
7365 (maybe_init_pretty_print): Likewise.
7366 (print_gimple_stmt): Do not call it. Use non-static local
7367 pretty_printer variable.
7368 (print_gimple_expr): Likewise.
7369 (print_gimple_seq): Likewise.
7370 (gimple_dump_bb): Likewise.
7372 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
7374 * asan.c (asan_pp): Remove.
7375 (asan_pp_initialized): Likewise.
7376 (asan_pp_initialize): Likewise.
7377 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
7378 (asan_emit_stack_protection): Tidy. Use local pretty printer.
7379 (asan_add_global): Likewise.
7381 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
7383 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
7384 * diagnostic.h (diagnostic_flush_buffer): Adjust.
7385 * pretty-print.c (pp_formatted_text_data): Likewise.
7386 (pp_indent): Rename from pp_base_indent.
7387 (pp_format): Rename from pp_base_format.
7388 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
7389 (pp_format_verbatim): Rename from pp_base_format_verbatim.
7390 (pp_flush): Rename from pp_base_flush.
7391 (pp_set_line_maximum_length): Rename from
7392 pp_base_set_line_maximum_length.
7393 (pp_clear_output_area): Rename from pp_base_clear_output_area.
7394 (pp_set_prefix): Rename from pp_base_set_prefix.
7395 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
7396 (pp_emit_prefix): Rename from pp_base_emit_prefix.
7397 (pp_append_text): Rename from pp_base_append_text.
7398 (pp_formatted_text): Rename from pp_base_formatted_text.
7399 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
7400 (pp_remaining_character_count_for_line): Rename from
7401 pp_base_remaining_character_count_for_line.
7402 (pp_newline): Rename from pp_base_newline.
7403 (pp_character): Rename from pp_base_character.
7404 (pp_string): Rename from pp_base_string.
7405 (pp_maybe_space): Rename from pp_base_maybe_space.
7406 * asan.c (asan_pp_string): Adjust.
7407 (asan_emit_stack_protection): Likewise.
7408 (asan_add_global): Likewise.
7409 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
7410 * tree-mudflap.c (mf_varname_tree): Likewise.
7411 * tree-pretty-print.c (pp_tree_identifier): Rename from
7412 pp_base_tree_identifier.
7413 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
7414 Declare as function.
7416 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
7418 * pretty-print.h (pp_bar_bar): New.
7419 (pp_ampersand_ampersand): Likewise.
7420 (pp_less_equal): Likewise.
7421 (pp_greater_equal): Likewise.
7422 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
7423 printer functions instead of pp_string or operators and punctuators.
7424 (dump_gimple_call): Likewise.
7425 (dump_gimple_omp_for): Likewise.
7426 (dump_gimple_transaction): Likewise.
7427 (dump_gimple_phi): Likewise.
7428 (pp_gimple_stmt_1): Likewise.
7429 * sched-vis.c (print_insn): Likewise.
7430 * tree-mudflap.c (mf_varname_tree): Likewise.
7431 * tree-pretty-print.c (dump_block_node): Likewise.
7432 (dump_generic_node): Likewise.
7434 2013-08-02 Jan Hubicka <jh@suse.cz>
7436 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
7438 * lto-streamer-out.c (tree_is_indexable): Results decls and
7439 parm decls are not indexable.
7440 (DFS_write_tree_body): Do not follow args and results.
7441 (hash_tree): Likewise.
7442 (output_functions): Rearrange so struct function is needed
7443 only when real body is output; be able to also ouptut abstract
7444 functions; output DECL_ARGUMENTS and DECL_RESULT.
7445 (lto_output): When not in WPA, ale store abstract functions.
7446 (write_symbol): Do not care about RESULT_DECL.
7447 (output_symbol_p): Handle correctly sbtract decls.
7448 * lto-streamer-in.c (input_function): Rearrange so struct
7449 function can be NULL at entry; allow streaming of
7450 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
7451 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
7452 sanity check during LTO.
7453 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
7454 RESULT_DECl and DECL_ARGUMENTS.
7455 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
7458 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
7460 * pretty-print.h (pp_underscore): New.
7462 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
7463 printer functions instead of pp_character.
7464 (dump_binary_rhs): Likewise.
7465 (dump_ternary_rhs): Likewise.
7466 (dump_gimple_call_args): Likewise.
7467 (pp_points_to_solution): Likewise.
7468 (dump_gimple_call): Likewise.
7469 (dump_gimple_switch): Likewise.
7470 (dump_gimple_cond): Likewise.
7471 (dump_gimple_bind): Likewise.
7472 (dump_gimple_try): Likewise.
7473 (dump_gimple_omp_for): Likewise.
7474 (dump_gimple_omp_continue): Likewise.
7475 (dump_gimple_omp_single): Likewise.
7476 (dump_gimple_omp_sections): Likewise.
7477 (dump_gimple_omp_block): Likewise.
7478 (dump_gimple_omp_critical): Likewise.
7479 (dump_gimple_transaction): Likewise.
7480 (dump_gimple_asm): Likewise.
7481 (dump_gimple_phi): Likewise.
7482 (dump_gimple_omp_parallel): Likewise.
7483 (dump_gimple_omp_task): Likewise.
7484 (dump_gimple_omp_atomic_load): Likewise.
7485 (dump_gimple_omp_atomic_store): Likewise.
7486 (dump_gimple_mem_ops): Likewise.
7487 (pp_gimple_stmt_1): Likewise.
7488 (pp_cfg_jump): Likewise.
7489 (dump_implicit_edges): Likewise.
7490 (gimple_dump_bb_for_graph): Likewise.
7491 * graph.c (draw_cfg_node): Likewise.
7492 * langhooks.c (lhd_print_error_function): Likewise.
7493 * sched-vis.c (print_exp): Likewise.
7494 (print_value): Likewise.
7495 (print_pattern): Likewise.
7496 (print_insn): Likewise.
7497 (rtl_dump_bb_for_graph): Likewise.
7498 * tree-pretty-print.c (dump_function_declaration): Likewise.
7499 (dump_array_domain): Likewise.
7500 (dump_omp_clause): Likewise.
7501 (dump_location): Likewise.
7502 (dump_generic_node): Likewise.
7503 (print_struct_decl): Likewise.
7504 * diagnostic.c (diagnostic_show_locus): Use pp_space.
7506 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
7508 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
7509 candidate table when replacing a candidate statement.
7510 (replace_rhs_if_not_dup): Likewise.
7511 (replace_one_candidate): Likewise.
7513 2013-08-02 Jan Hubicka <jh@suse.cz>
7514 Martin Liska <marxin.liska@gmail.com>
7516 * cgraphunit.c (add_new_function): Fix logic when adding from
7518 (assemble_thunk): Rename to ...
7519 (expand_thunk); .. this one; export; get it working with
7520 general functions; make produced gimple valid.
7521 * cgraph.h (expand_thunk): Declare.
7523 2013-08-02 Jan Hubicka <jh@suse.cz>
7525 * ipa-cp.c (gather_context_independent_values): Use
7526 ipa_get_param_move_cost.
7527 (get_replacement_map): Remove PARAM; move parameter folding
7529 (create_specialized_node): Update.
7530 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
7531 assert that we have gimple body; update move_cost.
7532 (count_formal_params): Assert that we have gimple body.
7533 (ipa_dump_param): New function.
7534 (ipa_alloc_node_params): Break out from ...
7535 (ipa_initialize_node_params): ... here.
7536 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
7537 (ipa_write_node_info): Stream move costs.
7538 (ipa_read_node_info): Read move costs.
7539 (ipa_update_after_lto_read): Do not recompute node params.
7540 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
7541 (ipa_get_param): Check we are not in WPA.
7542 (ipa_get_param_move_cost): New.
7543 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
7544 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
7545 parm numbers to be present.
7547 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
7549 PR rtl-optimization/58048
7550 * lra-constraints.c (process_alt_operands): Don't check asm
7551 operand on register.
7553 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
7555 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
7556 the implied StoreLoad barrier for atomic operations if before.
7558 2013-08-02 Jan Hubicka <jh@suse.cz>
7559 Martin Liska <marxin.liska@gmail.com>
7561 * cgraph.c (cgraph_function_body_availability): Do not check
7563 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
7564 symtab_node_availability): Declare.
7565 * ipa.c (can_replace_by_local_alias): New.
7566 (function_and_variable_visibility): Use it.
7567 * symtab.c (symtab_for_node_and_aliases,
7568 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
7570 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
7572 PR rtl-optimization/57963
7573 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
7574 (lra_constraints): Use them.
7576 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
7578 * config/arm/types.md (define_attr "type"): Add "load_acq"
7580 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
7582 (cortex_a53_store1): Likewise.
7584 2013-08-01 Jan Hubicka <jh@suse.cz>
7586 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
7587 partitions are not needed.
7589 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
7591 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
7592 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
7593 MAYBE_NON_Q_CLASS_P where appropriate.
7595 2013-08-01 Jan Hubicka <jh@suse.cz>
7597 * cgraph.h (release_function_body): Declare.
7598 * tree.c (free_lang_data_in_decl): Free, parameters and return values
7599 of unused delcarations.
7601 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7603 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
7604 RTL form when subtracting a constant.
7606 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7608 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
7609 Generate canonical plus rtx with negated immediate instead of minus
7611 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
7613 2013-08-01 Jan Hubicka <jh@suse.cz>
7615 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
7616 (cgraph_release_function_body): Likewise.
7617 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
7618 * cgraph.h (cgrpah_node): Rename abstract_and_needed
7619 to used_as_abstract_origin.
7620 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
7621 symbols used as abstract origins.
7622 * cgraphunit.c (analyze_functions): Update.
7623 * ipa.c (symtab_remove_unreachable_nodes): Recompute
7624 used_as_abstract_origin.
7625 * tree-inline.c (tree_function_versioning): Update
7626 used_as_abstract_origin; be ready for DECL_RESULT and
7627 DECL_ARGUMENTS to be NULL.
7629 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
7630 for abstract functions.
7631 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
7634 2013-08-01 Jan Hubicka <jh@suse.cz>
7636 * profile.c (compute_value_histograms): Fix thinko.
7638 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
7640 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
7641 aarch-common-protos.h to extra_headers.
7642 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
7643 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
7644 * config/aarch64/t-aarch64 (aarch-common.o): Define.
7646 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
7648 * config/aarch64/aarch64.md (define_attr "type"): Delete.
7649 Include "../arm/types.md". Define "type" attribute for all patterns.
7650 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
7653 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
7655 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
7656 to support power8 load fusion.
7657 (fusion_gpr_mem_load): Likewise.
7659 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
7661 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
7662 declarations for power8 load fusion.
7663 (emit_fusion_gpr_load): Likewise.
7665 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
7666 tuning for power8, turn on fusion mode by default. Turn on sign
7667 extending fusion mode if normal fusion mode is on, and we are at
7669 (fusion_gpr_load_p): New function, return true if we can fuse an
7670 addis instruction with a dependent load to a GPR.
7671 (emit_fusion_gpr_load): Emit the instructions for power8 load
7674 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
7675 (VSX load fusion peepholes): New peepholes to fuse together an
7676 addi instruction with a VSX load instruction.
7678 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
7679 peepholes to fuse an addis instruction with a load to a GPR base
7680 register. If we are supporting sign extending fusions, convert
7681 sign extending loads to zero extending loads and add an explicit
7684 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
7686 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
7687 aarch-common-protos.h to extra_headers.
7688 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
7689 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
7690 (arm_early_store_addr_dep): Likewise.
7691 (arm_no_early_alu_shift_dep): Likewise.
7692 (arm_no_early_alu_shift_value_dep): Likewise.
7693 (arm_no_early_mul_dep): Likewise.
7694 (arm_no_early_store_addr_dep): Likewise.
7695 (arm_mac_accumulator_is_mul_result): Likewise.
7696 (arm_mac_accumulator_is_result): Likewise.
7697 * config/arm/aarch-common.c: ... here. New file.
7698 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
7700 (arm_early_store_addr_dep): Likewise.
7701 (arm_no_early_alu_shift_dep): Likewise.
7702 (arm_no_early_alu_shift_value_dep): Likewise.
7703 (arm_no_early_mul_dep): Likewise.
7704 (arm_no_early_store_addr_dep): Likewise.
7705 (arm_mac_accumulator_is_mul_result): Likewise.
7706 (arm_mac_accumulator_is_result): Likewise.
7707 * config/arm/aarch-common-protos.h: ... here. New file.
7708 * config/arm/t-arm (aarch-common.o): Define.
7710 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
7712 * config/arm/arm.md: Include new file "types.md".
7713 (define_attr "type"): Move from here to ...
7714 (define_attr "mul32"): Likewise.
7715 (define_attr "mul64"): Likewise.
7716 * config/arm/types.md: ... here. New file.
7718 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
7720 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
7721 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
7723 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7725 * gen-pass-instances.awk: Fix offset of substr().
7727 2013-07-31 David Malcolm <dmalcolm@redhat.com>
7729 * Makefile.in (pass-instances.def): New.
7730 (passes.o): Replace dependency on passes.def with one on
7733 * gen-pass-instances.awk: New.
7735 * passes.c (pass_manager::pass_manager): Use pass-instances.def
7736 rather than passes.def, updating local definition of NEXT_PASS
7737 macro to add an extra NUM parameter (currently unused).
7739 2013-07-30 David Malcolm <dmalcolm@redhat.com>
7741 * Makefile.in (PASS_MANAGER_H): New.
7742 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
7743 (passes.o): Likewise.
7744 (statistics.o): Likewise.
7745 (cgraphunit.o): Likewise.
7746 (context.o): Depend on PASS_MANAGER_H.
7748 * pass_manager.h: New.
7750 * cgraphunit.c (cgraph_add_new_function): Update for moves
7751 of globals to fields of pass_manager.
7752 (analyze_function): Likewise.
7753 (expand_function): Likewise.
7754 (ipa_passes): Likewise.
7755 (compile): Likewise.
7757 * context.c (context::context): New.
7758 * context.h (context::context): New.
7759 (context::get_passes): New.
7760 (context::passes_): New.
7762 * lto-cgraph.c (input_node): Update for moves of globals to
7763 fields of pass_manager.
7765 * passes.c (all_passes): Remove, in favor of a field of the
7766 same name within the new class pass_manager.
7767 (all_small_ipa_passes): Likewise.
7768 (all_lowering_passes): Likewise.
7769 (all_regular_ipa_passes): Likewise.
7770 (all_late_ipa_passes): Likewise.
7771 (all_lto_gen_passes): Likewise.
7772 (passes_by_id): Likewise.
7773 (passes_by_id_size): Likewise.
7774 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
7775 the new class pass_manager.
7776 (set_pass_for_id): Convert to...
7777 (pass_manager::set_pass_for_id): ...method.
7778 (get_pass_for_id): Convert to...
7779 (pass_manager::get_pass_for_id): ...method.
7780 (register_one_dump_file): Move body of implementation into...
7781 (pass_manager::register_one_dump_file): ...here.
7782 (register_dump_files_1): Convert to...
7783 (pass_manager::register_dump_files_1): ...method.
7784 (register_dump_files): Convert to...
7785 (pass_manager::register_dump_files): ...method.
7786 (create_pass_tab): Update for moves of globals to fields of
7788 (dump_passes): Move body of implementation into...
7789 (pass_manager::dump_passes): ...here.
7790 (register_pass): Move body of implementation into...
7791 (pass_manager::register_pass): ...here.
7792 (init_optimization_passes): Convert into...
7793 (pass_manager::pass_manager): ...constructor for new
7794 pass_manager class, and initialize the pass_lists array.
7795 (check_profile_consistency): Update for moves of globals to
7796 fields of pass_manager.
7797 (dump_profile_report): Move body of implementation into...
7798 (pass_manager::dump_profile_report): ...here.
7799 (ipa_write_summaries_1): Update for moves of pass lists from
7800 being globals to fields of pass_manager.
7801 (ipa_write_optimization_summaries): Likewise.
7802 (ipa_read_summaries): Likewise.
7803 (ipa_read_optimization_summaries): Likewise.
7804 (execute_all_ipa_stmt_fixups): Likewise.
7806 * statistics.c (statistics_fini): Update for moves of globals to
7807 fields of pass_manager.
7809 * toplev.c (general_init): Replace call to
7810 init_optimization_passes with construction of the pass_manager
7813 * tree-pass.h (all_passes): Remove, in favor of a field of the
7814 same name within the new class pass_manager.
7815 (all_small_ipa_passes): Likewise.
7816 (all_lowering_passes): Likewise.
7817 (all_regular_ipa_passes): Likewise.
7818 (all_lto_gen_passes): Likewise.
7819 (all_late_ipa_passes): Likewise.
7820 (passes_by_id): Likewise.
7821 (passes_by_id_size): Likewise.
7822 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
7823 the new class pass_manager.
7824 (get_pass_for_id): Remove.
7826 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
7828 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
7831 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
7833 * arm.md (mulhi3): New expand pattern.
7835 2013-07-30 Jan Hubicka <jh@suse.cz>
7836 Martin Liska <marxin.liska@gmail.com>
7838 * profile.c (compute_value_histograms): Do not ICE when
7839 there is mismatch only on some counters.
7841 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
7843 PR rtl-optimization/57637
7844 * function.c (move_insn_for_shrink_wrap): Also check the
7845 GEN set of the LIVE problem for the liveness analysis
7846 if it exists, otherwise give up.
7848 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
7850 PR tree-optimization/57993
7851 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
7852 replaced statement in the candidate table.
7853 (phi_add_costs): Return infinite cost when the hidden basis does
7854 not dominate all phis on which the candidate is dependent.
7855 (replace_one_candidate): Record replaced statement in the
7858 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
7860 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
7861 (ashlv2si3): New expander.
7862 (*ashlv2si3_i): New define_insn_and_split.
7863 * predicates.md (float_operation): Allow patterns with three
7866 PR rtl-optimization/58021
7867 * mode-switching.c (create_pre_exit): Always split off preceding
7868 insns if we are not at the basic block head.
7870 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
7872 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
7873 (UCLIBC_DYNAMIC_LINKER): New macro.
7874 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
7876 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
7877 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
7879 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
7880 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
7881 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
7882 for SF and DF modes. Use ieee_quad_format for TF mode.
7883 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
7884 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
7885 (mips_option_override): Handle `-mnan=legacy'.
7886 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
7887 `-mabs=2008' and `-mnan=2008'.
7888 (OPTION_DEFAULT_SPECS): Add "nan" default.
7889 (ASM_SPEC): Handle `-mnan='.
7890 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
7891 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
7892 comment accordingly.
7893 (neg<mode>2): Likewise.
7894 * config/mips/mips.opt (mabs, mnan): New options.
7895 * doc/install.texi (Configuration): Document `--with-nan=' option.
7896 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
7898 (MIPS Options): Document them.
7899 * config.gcc <mips*-*-*>: Handle `--with-nan='.
7900 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
7901 * configure: Regenerate.
7902 * config.in: Regenerate.
7904 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
7906 * config/i386/i386.md (float post-reload splitters): Do not check
7907 for subregs of SSE registers.
7909 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
7910 H.J. Lu <hongjiu.lu@intel.com>
7914 * config/i386/i386.md (post-reload splitter
7915 to avoid partial SSE reg dependency stalls): New pattern.
7917 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
7919 * config/s390/s390.md ("movcc"): Swap load and store instructions.
7921 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
7923 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
7924 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
7926 2013-07-26 Cary Coutant <ccoutant@google.com>
7928 * dwarf2out.c (die_checksum_ordered): Don't include template
7929 instantiations in signature.
7930 (is_template_parameter): New function.
7931 (is_template_instantiation): New function.
7932 (generate_skeleton_bottom_up): Don't include template instantiations
7934 (generate_skeleton): Likewise.
7935 (break_out_comdat_types): Move recursive call to break out nested
7937 (prune_unused_types_mark_generic_parms_dies): Call
7938 is_template_parameter.
7940 2013-07-26 Ian Bolton <ian.bolton@arm.com>
7942 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
7943 uses vector registers.
7944 * config/aarch64/iterators.md: Add attributes rtn and vas.
7946 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7947 Richard Earnshaw <richard.earnshaw@arm.com>
7949 * combine.c (simplify_comparison): Re-canonicalize operands
7951 * config/arm/arm.md (movcond_addsi): New splitter.
7953 2013-07-25 Sterling Augustine <saugustine@google.com>
7955 * dwarf2out.c (size_of_pubnames): Move code to...
7956 (include_pubname_in_output): ...here. New.
7957 (want_pubnames): Rearrange.
7958 (output_pubnames): Call include_pubname_in_output. Move assertion.
7960 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
7962 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
7964 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
7967 * doc/extend.texi: Remove obsolete builtins. Fix
7968 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
7970 2013-07-25 Jan Hubicka <jh@suse.cz>
7972 * cgraph.c (release_function_body): Break out from ...
7973 (cgraph_release_function_body): ... this one; also release DECL_RESULT
7975 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
7976 old_tree in the map.
7977 (create_specialized_node): Update.
7978 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
7980 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
7981 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
7982 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
7983 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
7986 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7988 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
7989 addsi3_carryin_alt2_<optab>): Correct output template.
7991 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7993 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
7994 Adjust for arm_restrict_it.
7995 Remove trailing whitespace.
7997 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
7999 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
8000 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
8002 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
8004 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
8006 PR rtl-optimization/57960
8007 * lra-constraints.c (process_alt_operands): Use the right mode
8008 when checking strict_low.
8010 2013-07-25 Jan Hubicka <jh@suse.cz>
8012 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
8013 * cgraph.c (cgraph_remove_node): Do not release function body
8014 when in cgraph streaming.
8015 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
8016 in other partitions are not considered reachable; fix handling of
8019 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
8021 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
8023 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
8028 * config/arm/arm.md ("*sibcall_insn): Replace use of
8029 Ss with US. Adjust output for v5 and v4t.
8030 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
8032 * config/arm/constraints.md ("Ss"): Rename to US.
8034 2013-07-25 Terry Guo <terry.guo@arm.com>
8036 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
8037 shift_add/shift_sub0/shift_sub1 RTXs.
8039 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
8040 Anton Blanchard <anton@au1.ibm.com>
8042 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
8043 (altivec_vpks<VI_char>ss): Likewise.
8044 (altivec_vpks<VI_char>us): Likewise.
8045 (altivec_vpku<VI_char>us): Likewise.
8046 (altivec_vpku<VI_char>um): Likewise.
8048 2013-07-24 David Malcolm <dmalcolm@redhat.com>
8050 Introduce context class.
8052 * Makefile.in (CONTEXT_H): New.
8053 (OBJS): Add context.o.
8054 (toplev.o): Add CONTEXT_H to dependencies.
8057 * toplev.c (general_init): Create the singleton gcc::context instance.
8063 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
8065 PR rtl-optimization/57968
8066 * mode-switching.c (create_pre_exit): Allow instructions that
8067 don't set a return register to need a non-exit mode.
8069 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
8070 Anton Blanchard <anton@au1.ibm.com>
8072 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
8073 operands to vperm for little endian.
8074 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
8075 of lvsl to create the control mask for a vperm for little endian.
8077 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8078 Anton Blanchard <anton@au1.ibm.com>
8080 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
8081 two operands for little-endian.
8083 2013-07-23 Steve Ellcey <sellcey@mips.com>
8085 * config/mips/mips.c (mips_case_values_threshold): New.
8086 (TARGET_CASE_VALUES_THRESHOLD): Define.
8088 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8089 Anton Blanchard <anton@au1.ibm.com>
8091 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
8092 selection of field for vector splat in little endian mode.
8094 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
8096 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
8097 expanders to rs6000.md.
8098 (ior<mode>3): Likewise.
8099 (and<mode>3): Likewise.
8100 (one_cmpl<mode>2): Likewise.
8101 (nor<mode>3): Likewise.
8102 (andc<mode>3): Likewise.
8103 (eqv<mode>3): Likewise.
8104 (nand<mode>3): Likewise.
8105 (orc<mode>3): Likewise.
8107 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
8110 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
8111 to split multi-word logical operations.
8112 (rs6000_split_logical_di): Likewise.
8113 (rs6000_split_logical): Likewise.
8115 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
8116 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
8117 and allow TImode operations in 32-bit.
8118 (vsx_and<mode>3_64bit): Likewise.
8119 (vsx_ior<mode>3_32bit): Likewise.
8120 (vsx_ior<mode>3_64bit): Likewise.
8121 (vsx_xor<mode>3_32bit): Likewise.
8122 (vsx_xor<mode>3_64bit): Likewise.
8123 (vsx_one_cmpl<mode>2_32bit): Likewise.
8124 (vsx_one_cmpl<mode>2_64bit): Likewise.
8125 (vsx_nor<mode>3_32bit): Likewise.
8126 (vsx_nor<mode>3_64bit): Likewise.
8127 (vsx_andc<mode>3_32bit): Likewise.
8128 (vsx_andc<mode>3_64bit): Likewise.
8129 (vsx_eqv<mode>3_32bit): Likewise.
8130 (vsx_eqv<mode>3_64bit): Likewise.
8131 (vsx_nand<mode>3_32bit): Likewise.
8132 (vsx_nand<mode>3_64bit): Likewise.
8133 (vsx_orc<mode>3_32bit): Likewise.
8134 (vsx_orc<mode>3_64bit): Likewise.
8136 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
8137 logical types in GPRs.
8139 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
8140 logical insns to rs6000.md, and allow TImode operations in
8142 (altivec_ior<mode>3): Likewise.
8143 (altivec_xor<mode>3): Likewise.
8144 (altivec_one_cmpl<mode>2): Likewise.
8145 (altivec_nor<mode>3): Likewise.
8146 (altivec_andc<mode>3): Likewise.
8148 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
8149 attributes for moving the 128-bit logical operations into
8151 (BOOL_REGS_OUTPUT): Likewise.
8152 (BOOL_REGS_OP1): Likewise.
8153 (BOOL_REGS_OP2): Likewise.
8154 (BOOL_REGS_UNARY): Likewise.
8155 (BOOL_REGS_AND_CR0): Likewise.
8156 (one_cmpl<mode>2): Add support for DI logical operations on
8157 32-bit, splitting the operations to 32-bit.
8161 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
8162 changes to combine the 32/64-bit code, allow logical operations on
8163 TI mode in 32-bit, and to use similar match_operator patterns like
8164 scalar mode uses. Combine the Altivec and VSX code for logical
8165 operations, and move it here.
8166 (ior<mode>3, 128-bit types): Likewise.
8167 (xor<mode>3, 128-bit types): Likewise.
8168 (one_cmpl<mode>3, 128-bit types): Likewise.
8169 (nor<mode>3, 128-bit types): Likewise.
8170 (andc<mode>3, 128-bit types): Likewise.
8171 (eqv<mode>3, 128-bit types): Likewise.
8172 (nand<mode>3, 128-bit types): Likewise.
8173 (orc<mode>3, 128-bit types): Likewise.
8174 (and<mode>3_internal): Likewise.
8175 (bool<mode>3_internal): Likewise.
8176 (boolc<mode>3_internal1): Likewise.
8177 (boolc<mode>3_internal2): Likewise.
8178 (boolcc<mode>3_internal1): Likewise.
8179 (boolcc<mode>3_internal2): Likewise.
8180 (eqv<mode>3_internal1): Likewise.
8181 (eqv<mode>3_internal2): Likewise.
8182 (one_cmpl1<mode>3_internal): Likewise.
8184 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
8186 * config/microblaze/microblaze.c (microblaze_expand_prologue):
8187 Rename flag_stack_usage to flag_stack_usage_info.
8189 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
8191 * config/microblaze/sync.md: New file.
8192 * config/microblaze/microblaze.md: Include sync.md
8193 * config/microblaze/microblaze.c: Add print_operand 'y'.
8194 * config/microblaze/constraints.md: Add memory_contraint
8195 'Q' which is a single register.
8197 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
8199 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
8201 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
8203 * reload.c (find_reloads): Exit loop once we find this operand
8204 cannot be reloaded somehow for this alternative.
8206 * reload.c (find_reloads): Exit loop once we find a hard register.
8208 * rtlanal.c (computed_jump_p): Exit loop once we find label
8211 * i386.c (ix86_pad_returns): Exit loop after setting replace.
8213 * cfgloopmanip.c (remove_path): Exit loop after setting
8216 * gensupport.c (subst_dup): Avoid loop if code is not
8217 MATCH_DUP nor MATCH_OP_DUP.
8219 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
8221 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
8223 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
8225 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
8226 true for SP_REGNUM if mode == ptr_mode.
8227 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
8228 with value R0_REGNUM + 31.
8230 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
8232 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
8233 pad pointer-typed argument downward.
8235 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
8237 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
8238 and __ILP32__ when the ILP32 model is in use.
8240 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
8242 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
8243 (aarch64_load_symref_appropriately): In the case of
8244 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
8245 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
8246 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
8247 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
8248 if 'mode' doesn't equal to 'ptr_mode'.
8249 (aarch64_output_mi_thunk): Add an assertion on the alignment of
8250 'vcall_offset'; change to call aarch64_emit_move differently depending
8251 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
8252 to calculate the upper bound of 'vcall_offset'.
8253 (aarch64_cannot_force_const_mem): Change to also return true if
8255 (aarch64_legitimize_reload_address): In the case of large
8256 displacements, add new local variable 'xmode' and an assertion
8257 based on it; change to use 'xmode' to generate the new rtx and
8259 (aarch64_asm_trampoline_template): Change to generate the template
8260 differently depending on TARGET_ILP32 or not; change to use
8261 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
8262 (aarch64_trampoline_size): Removed.
8263 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
8264 and replace immediate literals with it. Change to use 'ptr_mode'
8265 instead of 'DImode' and call convert_memory_address if the mode
8266 of 'fnaddr' doesn't equal to 'ptr_mode'.
8267 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
8269 (aarch64_elf_asm_destructor): Likewise.
8270 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
8271 on TARGET_ILP32 instead of aarch64_trampoline_size.
8272 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
8273 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
8274 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
8275 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
8276 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
8277 (storewb_pair<GPI:mode>_<P:mode>): ... this.
8278 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
8279 depending on the value of 'mode'.
8280 (add_losym_<mode>): New.
8281 (ldr_got_small_<mode>): New, based on ldr_got_small.
8282 (ldr_got_small): Remove.
8283 (ldr_got_small_sidi): New.
8284 * config/aarch64/iterators.md (P): New.
8285 (PTR): Change to 'ptr_mode' in the condition.
8287 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
8289 * config.gcc (aarch64*-*-*): Support --with-abi.
8290 (aarch64*-*-elf): Support --with-multilib-list.
8291 (aarch64*-*-linux*): Likewise.
8292 (supported_defaults): Add abi to aarch64*-*-*.
8293 * configure.ac: Mention AArch64 for --with-multilib-list.
8294 * configure: Re-generated.
8295 * config/aarch64/biarchilp32.h: New file.
8296 * config/aarch64/biarchlp64.h: New file.
8297 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
8299 (MULTILIB_DEFAULTS): Ditto.
8300 (DRIVER_SELF_SPECS): Ditto.
8301 (ASM_SPEC): Update to also substitute -mabi.
8302 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
8303 file whose name depends on -mabi= and -mbig-endian.
8304 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
8306 (POINTER_SIZE): New define.
8307 (POINTERS_EXTEND_UNSIGNED): Ditto.
8308 (enum aarch64_abi_type): New enumeration tag.
8309 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
8310 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
8311 (TARGET_ILP32): New define.
8312 * config/aarch64/aarch64.opt (mabi): New.
8314 (ilp32, lp64): New values for -mabi.
8315 * config/aarch64/t-aarch64 (comma): New define.
8316 (MULTILIB_OPTIONS): Ditto.
8317 (MULTILIB_DIRNAMES): Ditto.
8318 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
8319 * doc/invoke.texi: Document -mabi for AArch64.
8321 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
8323 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
8325 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
8326 Anton Blanchard <anton@au1.ibm.com>
8328 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
8329 endianness when selecting field to splat.
8331 2013-07-22 Eric Christopher <echristo@gmail.com>
8333 * dwarf2out.c (die_odr_checksum): New function to use
8334 CHECKSUM_ macros and ULEB128 for DIE tag.
8335 (generate_type_signature): Use.
8337 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
8339 * config.gcc (sparc*-*-*): Accept leon3 processor.
8340 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
8341 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
8342 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
8343 * config/sparc/sparc.opt (enum processor_type): Add leon3.
8344 (mfix-ut699): Adjust comment.
8345 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
8346 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
8347 (CPP_CPU_SPEC): Likewise.
8348 (ASM_CPU_SPEC): Likewise.
8349 * config/sparc/sparc.c (leon3_cost): New constant.
8350 (sparc_option_override): Add leon3 support.
8351 (mem_ref): New function.
8352 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
8353 (sparc_do_work_around_errata): Look into the instruction in the delay
8354 slot and adjust accordingly. Add fix for the data cache nullify issues
8355 of the UT699. Change insertion position for the NOP.
8356 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
8357 (leon3_load): New reservation.
8358 (leon_store): Bump latency to 2.
8359 (grfpu): New automaton.
8360 (grfpu_alu): New unit.
8361 (grfpu_ds): Likewise.
8362 (leon_fp_alu): Adjust.
8363 (leon_fp_mult): Delete.
8364 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
8365 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
8366 * config/sparc/sparc.md (cpu): Add leon3.
8367 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
8369 (atomic_test_and_set): Likewise.
8372 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
8374 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
8375 default for R5900 targets.
8376 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
8377 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
8378 * config/mips/mips.c (mips_option_override): Report an error for
8379 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
8380 for -march=r5900 -mhard-float.
8382 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
8384 * df-problems.c (can_move_insns_across): Exit loop once we
8385 find a non-fixed, non-global register.
8387 * ipa-pure-const.c (propagate_nothrow): Exit loop after
8390 * omega.c (omega_eliminate_red): Break after setting red_found.
8391 (omega_problem_has_red_equations): Similarly after setting found.
8392 (omega_query_variable): Similarly after setting coupled.
8394 2013-07-22 Marek Polacek <polacek@redhat.com>
8396 * gimplify.c: Don't include gimple.h twice.
8398 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8400 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
8401 instead of TARGET_THUMB1.
8402 (Pz): New constraint.
8403 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
8405 (compare_negsi_si): Likewise.
8406 (compare_addsi2_op0): Likewise.
8407 (compare_addsi2_op1): Likewise.
8408 (addsi3_carryin_<optab>): Likewise.
8409 (addsi3_carryin_alt2_<optab>): Likewise.
8410 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
8411 for arm_restrict_it.
8412 (subsi3_carryin): Likewise.
8413 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
8414 (minmax_arithsi): Disable for arm_restrict_it.
8415 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
8416 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
8417 (satsi_<SAT:code>_shift): Likewise.
8418 (arm_shiftsi3): Add alternative for 16-bit encoding.
8419 (arm32_movhf): Disable for arm_restrict_it.
8420 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
8421 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
8423 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
8425 * config/arm/arm.md (attribute "insn"): Delete.
8426 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
8427 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
8428 (not_shiftsi): Update for attribute change.
8429 (not_shiftsi_compare0): Likewise.
8430 (not_shiftsi_compare0_scratch): Likewise.
8431 (arm_one_cmplsi2): Likewise.
8432 (thumb1_one_cmplsi2): Likewise.
8433 (notsi_compare0): Likewise.
8434 (notsi_compare0_scratch): Likewise.
8435 (thumb1_movdi_insn): Likewise.
8436 (arm_movsi_insn): Likewise.
8437 (movhi_insn_arch4): Likewise.
8438 (movhi_bytes): Likewise.
8439 (arm_movqi_insn): Likewise.
8440 (thumb1_movqi_insn): Likewise.
8441 (arm32_movhf): Likewise.
8442 (thumb1_movhf): Likewise.
8443 (arm_movsf_soft_insn): Likewise.
8444 (thumb1_movsf_insn): Likewise.
8445 (thumb_movdf_insn): Likewise.
8446 (movsicc_insn): Likewise.
8447 (movsfcc_soft_insn): Likewise.
8448 (and_scc): Likewise.
8449 (cond_move): Likewise.
8450 (if_move_not): Likewise.
8451 (if_not_move): Likewise.
8452 (if_shift_move): Likewise.
8453 (if_move_shift): Likewise.
8454 (if_shift_shift): Likewise.
8455 (if_not_arith): Likewise.
8456 (if_arith_not): Likewise.
8457 (cond_move_not): Likewise.
8458 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
8459 (neon_mov<mode>): Likewise.
8460 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
8461 (thumb2_movsi_vfp): Likewise.
8462 (movsf_vfp): Likewise.
8463 (thumb2_movsf_vfp): Likewise.
8464 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
8466 (cortexa7_older_only): Likewise.
8467 (cortexa7_younger): Likewise.
8468 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
8469 (1020alu_shift_op): Likewise.
8470 (1020alu_shift_reg_op): Likewise.
8471 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
8472 (alu_shift_op): Likewise.
8473 (alu_shift_reg_op): Likewise.
8474 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
8475 (11_alu_shift_op): Likewise.
8476 (11_alu_shift_reg_op): Likewise.
8477 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
8478 (9_alu_shift_reg_op): Likewise.
8479 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
8481 (cortex_a15_alu_shift): Likewise.
8482 (cortex_a15_alu_shift_reg): Likewise.
8483 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
8484 (cortex_a5_alu_shift): Likewise.
8485 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
8487 (cortex_a53_alu_shift): Likewise.
8488 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
8490 (cortex_a7_alu_reg): Likewise.
8491 (cortex_a7_alu_shift): Likewise.
8492 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
8493 (cortex_a8_alu_shift): Likewise.
8494 (cortex_a8_alu_shift_reg): Likewise.
8495 (cortex_a8_mov): Likewise.
8496 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
8497 (cortex_a9_dp_shift): Likewise.
8498 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
8499 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
8500 (cortex_r4_mov): Likewise.
8501 (cortex_r4_alu_shift): Likewise.
8502 (cortex_r4_alu_shift_reg): Likewise.
8503 * config/arm/fa526.md (526_alu_op): Update for attribute change.
8504 (526_alu_shift_op): Likewise.
8505 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
8506 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
8507 (626te_alu_shift_op): Likewise.
8508 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
8509 (726te_alu_op): Likewise.
8510 (726te_alu_shift_op): Likewise.
8511 (726te_alu_shift_reg_op): Likewise.
8512 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
8513 (mp626_alu_shift_op): Likewise.
8514 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
8515 (pj4_alu_e1_conds): Likewise.
8516 (pj4_alu): Likewise.
8517 (pj4_alu_conds): Likewise.
8518 (pj4_shift): Likewise.
8519 (pj4_shift_conds): Likewise.
8520 (pj4_alu_shift): Likewise.
8521 (pj4_alu_shift_conds): Likewise.
8523 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8525 * config/arm/predicates.md (shiftable_operator_strict_it):
8527 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
8528 Disable cond_exec version for arm_restrict_it.
8529 (thumb2_smaxsi3): Convert to generate cond_exec.
8530 (thumb2_sminsi3): Likewise.
8531 (thumb32_umaxsi3): Likewise.
8532 (thumb2_uminsi3): Likewise.
8533 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
8534 (thumb2_neg_abssi2): Likewise.
8535 (thumb2_mov_scc): Add alternative for 16-bit encoding.
8536 (thumb2_movsicc_insn): Adjust alternatives.
8537 (thumb2_mov_negscc): Disable for arm_restrict_it.
8538 (thumb2_mov_negscc_strict_it): New pattern.
8539 (thumb2_mov_notscc_strict_it): New pattern.
8540 (thumb2_mov_notscc): Disable for arm_restrict_it.
8541 (thumb2_ior_scc): Likewise.
8542 (thumb2_ior_scc_strict_it): New pattern.
8543 (thumb2_cond_move): Adjust for arm_restrict_it.
8544 (thumb2_cond_arith): Disable for arm_restrict_it.
8545 (thumb2_cond_arith_strict_it): New pattern.
8546 (thumb2_cond_sub): Adjust for arm_restrict_it.
8547 (thumb2_movcond): Likewise.
8548 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
8549 (thumb2_zero_extendhisi2_v6): Likewise.
8550 (thumb2_zero_extendqisi2_v6): Likewise.
8551 (orsi_notsi_si): Likewise.
8552 (orsi_not_shiftsi_si): Likewise.
8554 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
8556 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
8557 instruction sequence is 1 byte shorter.
8559 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
8561 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
8562 it is not needed after split.
8564 2013-07-20 Iain Sandoe <iain@codesourcery.com>
8567 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
8568 second label for nonlocal goto receivers. Don't output pic base labels
8569 unless we're producing PIC; mark that action unreachable().
8570 (ix86_save_reg): If the function contains a nonlocal label, save the
8572 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
8573 * config/darwin.c (emitted_pic_label_num): New GTY.
8574 (update_pic_label_number_if_needed): New.
8575 (machopic_output_function_base_name): Adjust for nonlocal receiver
8577 (machopic_should_output_picbase_label): New.
8578 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
8579 (nonlocal_goto_receiver): New insn and split.
8581 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
8583 * config/aarch64/aarch64-builtins.c
8584 (aarch64_fold_builtin): Fold abs in all modes.
8585 * config/aarch64/aarch64-simd-builtins.def
8586 (abs): Enable for all modes.
8587 * config/aarch64/arm_neon.h
8588 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
8589 (vabs_f64): Add missing intrinsic.
8591 2013-07-19 Ian Bolton <ian.bolton@arm.com>
8593 * config/aarch64/arm_neon.h (vabs_s64): New function
8595 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
8598 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
8599 * config/avr/avr.md (adjust_len): Add `round'.
8600 * config/avr/avr-protos.h (avr_out_round): New prototype.
8601 (avr_out_plus): Add `out_label' argument.
8602 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
8603 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
8604 Handle the case where `insn' is just a pattern.
8605 (avr_out_bitop): Handle the case where `insn' is just a pattern.
8606 (avr_out_round): New function.
8607 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
8609 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
8611 * config/microblaze/microblaze.c (microblaze_expand_prologue):
8612 Add check for flag_stack_usage to handle -fstack-usage support
8614 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
8616 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
8617 interaction for new Power8 flags and VSX.
8619 2013-07-18 Sriraman Tallam <tmsriram@google.com>
8622 * tree-inline.c (expand_call_inline): Emit errors during
8623 early_inlining only if optimization is not turned on.
8625 2013-07-18 David Malcolm <dmalcolm@redhat.com>
8629 * passes.c (init_optimization_passes): Move the construction of
8630 the pass hierarchy into a new passes.def file.
8632 * Makefile.in (passes.o): Add dependency on passes.def.
8634 2013-07-18 David Malcolm <dmalcolm@redhat.com>
8636 * passes.c (init_optimization_passes): Introduce macros for
8637 constructing the tree of passes (INSERT_PASSES_AFTER,
8638 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
8639 TERMINATE_PASS_LIST).
8641 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
8642 Wei Mi <wmi@google.com>
8644 PR rtl-optimization/57878
8645 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
8647 (reload_pseudo_compare_func): Check nregs first for reload
8650 2013-07-18 David Malcolm <dmalcolm@redhat.com>
8652 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
8654 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
8656 * read-rtl.c (validate_const_int): Once an invalid character is
8657 seen, quit the loop.
8659 * gengtype.c (write_roots): Similarly once we find the "deletable"
8660 or "if_marked" option.
8662 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
8664 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
8665 "xtab" and "sat". Move value "clz" from here to ...
8666 (attriubte "type"): ... here.
8667 (satsi_<SAT:code>): Delete "insn" attribute.
8668 (satsi_<SAT:code>_shift): Likewise.
8669 (arm_zero_extendqisi2addsi): Likewise.
8670 (arm_extendqisi2addsi): Likewise.
8671 (clzsi2): Update for attribute changes.
8672 (rbitsi2): Likewise.
8673 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
8675 (arm_usatsihi): Likewise.
8676 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
8678 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
8680 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
8681 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
8682 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
8683 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
8684 in alphabetical order.
8685 (attribute "core_cycles"): Update for attribute changes.
8686 (arm_addsi3): Likewise.
8687 (addsi3_compare0): Likewise.
8688 (addsi3_compare0_scratch): Likewise.
8689 (addsi3_compare_op1): Likewise.
8690 (addsi3_compare_op2): Likewise.
8691 (compare_addsi2_op0): Likewise.
8692 (compare_addsi2_op1): Likewise.
8693 (addsi3_carryin_shift_<optab>): Likewise.
8694 (subsi3_carryin_shift): Likewise.
8695 (rsbsi3_carryin_shift): Likewise.
8696 (arm_subsi3_insn): Likewise.
8697 (subsi3_compare0): Likewise.
8698 (subsi3_compare): Likewise.
8699 (arm_andsi3_insn): Likewise.
8700 (thumb1_andsi3_insn): Likewise.
8701 (andsi3_compare0): Likewise.
8702 (andsi3_compare0_scratch): Likewise.
8703 (zeroextractsi_compare0_scratch
8704 (andsi_not_shiftsi_si): Likewise.
8705 (iorsi3_insn): Likewise.
8706 (iorsi3_compare0): Likewise.
8707 (iorsi3_compare0_scratch): Likewise.
8708 (arm_xorsi3): Likewise.
8709 (thumb1_xorsi3_insn): Likewise.
8710 (xorsi3_compare0): Likewise.
8711 (xorsi3_compare0_scratch): Likewise.
8712 (satsi_<SAT:code>_shift): Likewise.
8714 (arm_shiftsi3): Likewise.
8715 (shiftsi3_compare0): Likewise.
8716 (not_shiftsi): Likewise.
8717 (not_shiftsi_compare0): Likewise.
8718 (not_shiftsi_compare0_scratch): Likewise.
8719 (arm_one_cmplsi2): Likewise.
8720 (thumb_one_complsi2): Likewise.
8721 (notsi_compare0): Likewise.
8722 (notsi_compare0_scratch): Likewise.
8723 (thumb1_zero_extendhisi2): Likewise.
8724 (arm_zero_extendhisi2): Likewise.
8725 (arm_zero_extendhisi2_v6): Likewise.
8726 (arm_zero_extendhisi2addsi): Likewise.
8727 (thumb1_zero_extendqisi2): Likewise.
8728 (thumb1_zero_extendqisi2_v6): Likewise.
8729 (arm_zero_extendqisi2): Likewise.
8730 (arm_zero_extendqisi2_v6): Likewise.
8731 (arm_zero_extendqisi2addsi): Likewise.
8732 (thumb1_extendhisi2): Likewise.
8733 (arm_extendhisi2): Likewise.
8734 (arm_extendhisi2_v6): Likewise.
8735 (arm_extendqisi): Likewise.
8736 (arm_extendqisi_v6): Likewise.
8737 (arm_extendqisi2addsi): Likewise.
8738 (thumb1_extendqisi2): Likewise.
8739 (thumb1_movdi_insn): Likewise.
8740 (arm_movsi_insn): Likewise.
8741 (movsi_compare0): Likewise.
8742 (movhi_insn_arch4): Likewise.
8743 (movhi_bytes): Likewise.
8744 (arm_movqi_insn): Likewise.
8745 (thumb1_movqi_insn): Likewise.
8746 (arm32_movhf): Likewise.
8747 (thumb1_movhf): Likewise.
8748 (arm_movsf_soft_insn): Likewise.
8749 (thumb1_movsf_insn): Likewise.
8750 (movdf_soft_insn): Likewise.
8751 (thumb_movdf_insn): Likewise.
8752 (arm_cmpsi_insn): Likewise.
8753 (cmpsi_shiftsi): Likewise.
8754 (cmpsi_shiftsi_swp): Likewise.
8755 (arm_cmpsi_negshiftsi_si): Likewise.
8756 (movsicc_insn): Likewise.
8757 (movsfcc_soft_insn): Likewise.
8758 (arith_shiftsi): Likewise.
8759 (arith_shiftsi_compare0
8760 (arith_shiftsi_compare0_scratch
8761 (sub_shiftsi): Likewise.
8762 (sub_shiftsi_compare0
8763 (sub_shiftsi_compare0_scratch
8764 (and_scc): Likewise.
8765 (cond_move): Likewise.
8766 (if_plus_move): Likewise.
8767 (if_move_plus): Likewise.
8768 (if_move_not): Likewise.
8769 (if_not_move): Likewise.
8770 (if_shift_move): Likewise.
8771 (if_move_shift): Likewise.
8772 (if_shift_shift): Likewise.
8773 (if_not_arith): Likewise.
8774 (if_arith_not): Likewise.
8775 (cond_move_not): Likewise.
8776 (thumb1_ashlsi3): Set type attribute.
8777 (thumb1_ashrsi3): Likewise.
8778 (thumb1_lshrsi3): Likewise.
8779 (thumb1_rotrsi3): Likewise.
8780 (shiftsi3_compare0_scratch): Likewise.
8781 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
8782 (neon_mov<mode>): Likewise.
8783 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
8785 (thumb2_movsi_insn): Likewise.
8786 (thumb2_cmpsi_neg_shiftsi): Likewise.
8787 (thumb2_extendqisi_v6): Likewise.
8788 (thumb2_zero_extendhisi2_v6): Likewise.
8789 (thumb2_zero_extendqisi2_v6): Likewise.
8790 (thumb2_shiftsi3_short): Likewise.
8791 (thumb2_addsi3_compare0_scratch): Likewise.
8792 (orsi_not_shiftsi_si): Likewise.
8793 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
8794 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
8796 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
8797 (1020alu_shift_op): Likewise.
8798 (1020alu_shift_reg_op): Likewise.
8799 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
8800 (alu_shift_op): Likewise.
8801 (alu_shift_reg_op): Likewise.
8802 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
8803 (11_alu_shift_op): Likewise.
8804 (11_alu_shift_reg_op): Likewise.
8805 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
8806 (9_alu_shift_reg_op): Likewise.
8807 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
8809 (cortex_a15_alu_shift): Likewise.
8810 (cortex_a15_alu_shift_reg): Likewise.
8811 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
8813 (cortex_a5_alu_shift): Likewise.
8814 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
8816 (cortex_a53_alu_shift): Likewise.
8817 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
8819 (cortex_a7_alu_reg): Likewise.
8820 (cortex_a7_alu_shift): Likewise.
8821 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
8823 (cortex_a8_alu_shift): Likewise.
8824 (cortex_a8_alu_shift_reg): Likewise.
8825 (cortex_a8_mov): Likewise.
8826 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
8827 (cortex_a9_dp_shift): Likewise.
8828 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
8830 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
8832 (cortex_r4_mov): Likewise.
8833 (cortex_r4_alu_shift): Likewise.
8834 (cortex_r4_alu_shift_reg): Likewise.
8835 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
8836 (526_alu_shift_op): Likewise.
8837 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
8838 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
8839 (626te_alu_shift_op): Likewise.
8840 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
8841 (726te_alu_op): Likewise.
8842 (726te_alu_shift_op): Likewise.
8843 (726te_alu_shift_reg_op): Likewise.
8844 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
8845 (mp626_alu_shift_op): Likewise.
8846 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
8847 (pj4_alu_e1_conds): Likewise.
8848 (pj4_alu): Likewise.
8849 (pj4_alu_conds): Likewise.
8850 (pj4_shift): Likewise.
8851 (pj4_shift_conds): Likewise.
8852 (pj4_alu_shift): Likewise.
8853 (pj4_alu_shift_conds): Likewise.
8854 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
8856 (cortexa7_older_only): Likewise.
8857 (cortexa7_younger): Likewise.
8859 2013-07-18 David Malcolm <dmalcolm@redhat.com>
8861 * ipa-pure-const.c (generate_summary): Rename to...
8862 (pure_const_generate_summary): ... this.
8864 2013-07-17 Iain Sandoe <iain@codesourcery.com>
8866 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
8868 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
8871 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
8874 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8876 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
8877 enabled without -march=zEC12.
8878 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
8881 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
8883 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
8884 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
8885 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
8886 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
8887 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
8888 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
8889 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
8891 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
8892 and ISA_HAS_NMADD3_NMSUB3.
8893 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
8894 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
8895 (nmsub4<mode>, nmsub3<mode>): Likewise.
8896 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
8898 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
8900 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
8901 TARGET_MIPS5400 checking.
8903 2013-07-16 Jakub Jelinek <jakub@redhat.com>
8904 Peter Bergner <bergner@vnet.ibm.com>
8906 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
8907 registers in the comment.
8908 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
8909 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
8910 rather than FIRST_PSEUDO_REGISTERS.
8912 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
8914 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
8915 enable extra ISA flags with TARGET_HTM.
8917 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
8919 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
8922 2013-07-15 Cong Hou <congh@google.com>
8924 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
8925 in compare function for sorting.
8927 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
8929 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
8930 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
8931 * config/rs6000/rs6000.opt: Add -mhtm option.
8932 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
8933 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
8934 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
8935 __HTM__ if the HTM instructions are available.
8936 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
8937 htm_spr_reg_operand): New define_predicates.
8938 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
8939 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
8941 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
8942 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
8943 HTM builtin functions.
8944 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
8945 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
8946 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
8947 (rs6000_builtin_mask_calculate): Likewise.
8948 (rs6000_option_override_internal): Likewise.
8949 (bdesc_htm): Add new HTM builtin support.
8950 (htm_spr_num): New function.
8951 (htm_spr_regno): Likewise.
8952 (rs6000_htm_spr_icode): Likewise.
8953 (htm_expand_builtin): Likewise.
8954 (htm_init_builtins): Likewise.
8955 (rs6000_expand_builtin): Add support for HTM builtin functions.
8956 (rs6000_init_builtins): Likewise.
8957 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
8959 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
8960 (TARGET_HTM, MASK_HTM): Define macros.
8961 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
8962 (FIXED_REGISTERS): Likewise.
8963 (CALL_USED_REGISTERS): Likewise.
8964 (CALL_REALLY_USED_REGISTERS): Likewise.
8965 (REG_ALLOC_ORDER): Likewise.
8966 (enum reg_class): Likewise.
8967 (REG_CLASS_NAMES): Likewise.
8968 (REG_CLASS_CONTENTS): Likewise.
8969 (REGISTER_NAMES): Likewise.
8970 (ADDITIONAL_REGISTER_NAMES): Likewise.
8971 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
8972 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
8973 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
8974 * config/rs6000/htm.md: New file.
8975 * config/rs6000/htmintrin.h: New file.
8976 * config/rs6000/htmxlintrin.h: New file.
8978 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
8980 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
8981 Define SYMBOL_TINY_GOT, update comment.
8982 * config/aarch64/aarch64.c
8983 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
8984 (aarch64_expand_mov_immediate): Likewise.
8985 (aarch64_print_operand): Likewise.
8986 (aarch64_classify_symbol): Likewise.
8987 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
8988 (ldr_got_tiny): Define.
8990 2013-07-13 Tobias Grosser <tobias@grosser.es>
8992 PR tree-optimization/54094
8993 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
8994 scheduling dimension for the parallelism check from the polyhedral
8995 information in the AST.
8996 * graphite-dependences.c (carries_deps): Do not assume the schedule is
8999 2013-07-13 Jason Merrill <jason@redhat.com>
9001 * print-tree.c (debug_vec_tree): Use debug_raw.
9002 (debug_raw (vec<tree, va_gc> &)): New.
9003 (debug_raw (vec<tree, va_gc> *)): New.
9004 * tree.h: Declare them.
9006 2013-07-13 Bin Cheng <bin.cheng@arm.com>
9008 * ifcvt.c (ifcvt_after_combine): New static variable.
9009 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
9011 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
9012 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
9013 rest_of_handle_if_after_reload): Pass new argument for if_convert.
9015 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
9017 * config/mips/mips.c (mips_expand_call): Remove empty statement.
9019 2013-07-12 Michael Matz <matz@suse.de>
9022 * convert.c (convert_to_real): Reject non-float inner types.
9024 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
9026 * config/aarch64/aarch64-protos.h
9027 (aarch64_simd_immediate_valid_for_move): Remove.
9028 * config/aarch64/aarch64.c (simd_immediate_info): New member.
9029 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
9031 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
9033 2013-07-11 Steve Ellcey <sellcey@mips.com>
9035 * config/mips/mips.c (mips_conditional_register_usage): Do not
9036 use t[0-7] registers in MIPS16 mode when optimizing for size.
9038 2013-07-11 Sriraman Tallam <tmsriram@google.com>
9040 * config/i386/i386.c (dispatch_function_versions): Fix array
9041 indexing of function_version_info to match actual_versions.
9043 2013-07-11 Teresa Johnson <tejohnson@google.com>
9045 * vec.h (struct va_gc): Move release out-of-line.
9046 (va_gc::release): Call ggc_free on released vec.
9048 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
9050 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
9051 Require GOT register as additional operand in UNSPEC.
9052 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
9053 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
9054 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
9055 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
9056 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
9057 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
9058 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
9060 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
9063 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
9064 name seen by assembler/linker rather if available.
9066 2013-07-11 Andreas Schwab <schwab@suse.de>
9068 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
9070 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
9072 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
9074 2013-07-10 Joseph Myers <joseph@codesourcery.com>
9076 * doc/tm.texi.in: Move hook documentation to ....
9077 * target.def: ... here.
9079 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
9081 * doc/tm.texi: Regenerate.
9083 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
9086 * doc/invoke.texi: Document -Wconditionally-supported.
9088 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
9091 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
9094 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
9097 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
9098 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
9099 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
9100 Remove duplicate devices.
9101 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
9102 * config/avr/t-multilib: Regenerate.
9103 * config/avr/avr-tables.opt: Regenerate.
9104 * doc/avr-mmcu.texi: Regenerate.
9106 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
9109 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
9111 2013-07-10 Graham Stott <graham.stott@btinternet.com>
9113 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
9114 the cost of MULT when optimizing for size.
9116 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
9118 * config/cr16/cr16-protos.h: Don't include target.h.
9120 2013-07-09 Joseph Myers <joseph@codesourcery.com>
9122 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
9123 adjust register size for TDmode and TFmode for VSX registers.
9125 2013-07-08 Kai Tietz <ktietz@redhat.com>
9128 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
9129 hook_bool_const_tree_true.
9131 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9133 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
9134 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
9135 * config/s390/s390.md: Define FPR*_REGNUM constants.
9136 Fix FPR2_REGNUM constant (18 -> 17).
9137 ("*trunc<BFP:mode><DFP_ALL:mode>2")
9138 ("*trunc<DFP_ALL:mode><BFP:mode>2")
9139 ("trunc<BFP:mode><DFP_ALL:mode>2")
9140 ("trunc<DFP_ALL:mode><BFP:mode>2")
9141 ("*extend<BFP:mode><DFP_ALL:mode>2")
9142 ("*extend<DFP_ALL:mode><BFP:mode>2")
9143 ("extend<BFP:mode><DFP_ALL:mode>2")
9144 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
9147 2013-07-08 Graham Stott <graham.stott@btinternet.com>
9149 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
9151 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9153 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
9154 and cfun_fpr_bit_p to cfun_fpr_save_p.
9155 (s390_frame_area, s390_register_info, s390_frame_info)
9156 (s390_emit_prologue, s390_emit_epilogue)
9157 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
9159 * config/s390/s390.h: Define *_REGNUM macros for floating point
9162 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
9164 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
9166 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
9168 PR rtl-optimization/57786
9169 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
9170 and break out of the loop when it is set to false.
9172 2013-07-08 Jakub Jelinek <jakub@redhat.com>
9175 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
9176 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
9177 (const_int 63)) 0)).
9178 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
9179 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
9180 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
9182 PR rtl-optimization/57829
9183 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
9184 mask bits outside of mode are just sign-extension from mode to HWI.
9186 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
9188 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
9189 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
9190 adjust_address instead of change_address to keep info about alignment.
9191 (emit_strmov): Remove.
9192 (emit_memmov): New function.
9193 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
9194 (expand_movmem_epilogue): Likewise and return updated rtx for
9196 (expand_constant_movmem_prologue): Likewise and return updated rtx for
9197 destination and source.
9198 (decide_alignment): Refactor, handle vector_loop.
9199 (ix86_expand_movmem): Likewise.
9200 (ix86_expand_setmem): Likewise.
9201 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
9203 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
9205 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
9206 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
9208 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
9210 * config/i386/sse.md (sse_movlhps): Change alternative 3
9211 of operand 2 to "m".
9213 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
9216 * config/i386/sse.md (iptr): New mode attribute.
9217 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
9218 (<sse>_vm<plusminus_insn><mode>3): Ditto.
9219 (<sse>_vmmul<mode>3): Ditto.
9220 (<sse>_vmdiv<mode>3): Ditto.
9221 (sse_vmrcpv4sf2): Ditto.
9222 (<sse>_vmsqrt<mode>2): Ditto.
9223 (sse_vmrsqrtv4sf2): Ditto.
9224 (<sse>_vm<code><mode>3): Ditto.
9225 (avx_vmcmp<mode>3): Ditto.
9226 (<sse>_vmmaskcmp<mode>3): Ditto.
9227 (<sse>_comi): Ditto.
9228 (<sse>_ucomi): Ditto.
9229 (*xop_vmfrcz_<mode>): Ditto.
9230 (*fmai_fmadd_<mode>): Ditto.
9231 (*fmai_fmsub_<mode>): Ditto.
9232 (*fmai_fnmadd_<mode>): Ditto.
9233 (*fmai_fnmsub_<mode>): Ditto.
9234 (*fma4i_vmfmadd_<mode>): Ditto.
9235 (*fma4i_vmfmsub_<mode>): Ditto.
9236 (*fma4i_vmfnmadd_<mode>): Ditto.
9237 (*fma4i_vmfnmsub_<mode>): Ditto.
9238 (*xop_vmfrcz_<mode>): Ditto.
9239 (sse_cvtps2pi): Ditto.
9240 (sse_cvttps2pi): Ditto.
9241 (sse_cvtss2si): Ditto.
9242 (sse_cvtss2si_2): Ditto.
9243 (sse_cvtss2siq_2): Ditto.
9244 (sse_cvttss2si): Ditto.
9245 (sse_cvttss2siq): Ditto.
9246 (sse_cvtsd2si): Ditto.
9247 (sse_cvtsd2si_2): Ditto.
9248 (sse_cvtsd2siq_2): Ditto.
9249 (sse_cvttsd2si): Ditto.
9250 (sse_cvttsd2siq): Ditto.
9251 (sse_cvtsd2ss): Ditto.
9252 (sse_cvtss2sd): Ditto.
9253 (avx2_pbroadcast<mode>): Ditto.
9254 (avx2_pbroadcast<mode>_1): Ditto.
9255 (*avx_vperm_broadcast_v4sf): Ditto.
9257 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
9258 (sse_movlhps): Ditto.
9259 (sse_storehps): Ditto.
9260 (sse_loadhps): Ditto.
9261 (sse_storelps): Ditto.
9262 (sse_loadlps): Ditto.
9263 (*vec_concatv4sf): Ditto.
9264 (*vec_interleave_highv2df): Ditto.
9265 (*vec_interleave_lowv2df): Ditto.
9266 (*vec_extractv2df_1_sse): Ditto.
9267 (*vec_extractv2df_0_sse): Ditto.
9268 (sse2_storelpd): Ditto.
9269 (sse2_loadlpd): Ditto.
9270 (sse2_movsd): Ditto.
9271 (*vec_concatv4si): Ditto.
9272 (vec_concatv2di): Ditto.
9274 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
9275 for Intel asm dialect.
9276 (mmx_punpcklwd): Ditto.
9277 (mmx_punpckldq): Ditto.
9279 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
9280 for intel assembler dialect.
9282 2013-07-06 Jakub Jelinek <jakub@redhat.com>
9285 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
9286 for BUILT_IN_C{LZ,LRSB}*.
9287 * tree.h (CASE_INT_FN): Add FN##IMAX case.
9288 * tree-vrp.c (extract_range_basic): Handle
9289 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
9290 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
9291 fall thru to code calling set_value*.
9292 * builtins.c (expand_builtin): Remove *IMAX cases.
9293 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
9294 if width is bigger than 2*HWI.
9296 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
9298 PR rtl-optimization/55342
9299 * lra-int.h (lra_subreg_reload_pseudos): New.
9300 * lra.c: Add undoing optional reloads to the block diagram.
9301 (lra_subreg_reload_pseudos): New.
9302 (lra_optional_reload_pseudos): Change comments.
9303 (lra): Init and clear lra_subreg_reload_pseudos. Clear
9304 lra_optional_reload_pseudos after undo transformations.
9305 * lra-assigns.c (pseudo_prefix_title): New.
9306 (lra_setup_reg_renumber): Use it.
9307 (spill_for): Ditto. Check subreg reload pseudos too.
9308 (assign_by_spills): Consider subreg reload pseudos too.
9309 * lra-constraints.c (simplify_operand_subreg): Use
9310 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
9311 (curr_insn_transform): Recognize and do optional reloads.
9312 (undo_optional_reloads): New.
9313 (lra_undo_inheritance): Call undo_optional_reloads.
9315 2013-07-05 Thomas Quinot <quinot@adacore.com>
9317 * tree-complex.c (expand_complex_operations_1): Fix typo.
9319 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
9321 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
9322 (tune_params): New member 'const vec_costs'.
9323 * config/aarch64/aarch64.c (generic_vector_cost): New.
9324 (generic_tunings): New member 'generic_vector_cost'.
9325 (aarch64_builtin_vectorization_cost): New.
9326 (aarch64_add_stmt_cost): New.
9327 (TARGET_VECTORIZE_ADD_STMT_COST): New.
9328 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
9330 2013-07-03 Jakub Jelinek <jakub@redhat.com>
9333 * config/i386/predicates.md (vsib_address_operand): Disallow
9334 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
9336 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
9339 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
9340 expand_builtin_setjmp_receiver.
9341 (expand_label): Adjust, call expand_builtin_setjmp_receiver
9342 with NULL for the label parameter.
9343 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
9344 the frame-pointer. Adjust comments.
9345 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
9346 only if LABEL is non-NULL.
9348 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
9350 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
9351 (ARM_ABI_AAPCS64): Ditto.
9353 (ARM_DEFAULT_ABI): Ditto.
9355 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
9357 * config/aarch64/aarch64-builtins.c
9358 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
9359 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
9361 * config/aarch64/aarch64-simd.md
9362 (aarch64_ld1<VALL:mode>): New.
9363 (aarch64_st1<VALL:mode>): Likewise.
9364 * config/aarch64/arm_neon.h
9365 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
9367 2013-07-02 Sriraman Tallam <tmsriram@google.com>
9369 * config/i386/i386.c (gate_insert_vzeroupper): Check if
9371 (ix86_option_override_internal):Turn on all -mavx target flags by
9372 default as they are dependent on AVX anyway.
9374 2013-07-02 Cary Coutant <ccoutant@google.com>
9376 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
9378 (loc_checksum_ordered): Likewise.
9379 (hash_loc_operands): Remove inline keyword.
9381 2013-07-02 Jakub Jelinek <jakub@redhat.com>
9383 PR tree-optimization/57741
9384 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
9385 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
9386 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
9387 Allow REAL_CST step_exprs if flag_associative_math.
9388 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
9390 2013-07-02 Ian Bolton <ian.bolton@arm.com>
9392 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
9394 2013-07-02 Ian Bolton <ian.bolton@arm.com>
9396 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
9398 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9400 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
9402 (iorsi3_insn): Likewise.
9403 (arm_xorsi3): Likewise.
9405 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
9407 * arm.md (attribute "wtype"): Delete. Move attribute values from here
9409 (attribute "type"): ... here, and prefix with "wmmx_".
9410 (attribute "core_cycles"): Update for attribute changes.
9411 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
9412 (tbcstv4hi): Likewise.
9413 (tbcstv2si): Likewise.
9414 (iwmmxt_iordi3): Likewise.
9415 (iwmmxt_xordi3): Likewise.
9416 (iwmmxt_anddi3): Likewise.
9417 (iwmmxt_nanddi3): Likewise.
9418 (iwmmxt_arm_movdi): Likewise.
9419 (iwmmxt_movsi_insn): Likewise.
9420 (mov<mode>_internal): Likewise.
9421 (and<mode>3_iwmmxt): Likewise.
9422 (ior<mode>3_iwmmxt): Likewise.
9423 (xor<mode>3_iwmmxt): Likewise.
9424 (add<mode>3_iwmmxt): Likewise.
9425 (ssaddv8qi3): Likewise.
9426 (ssaddv4hi3): Likewise.
9427 (ssaddv2si3): Likewise.
9428 (usaddv8qi3): Likewise.
9429 (usaddv4hi3): Likewise.
9430 (usaddv2si3): Likewise.
9431 (sub<mode>3_iwmmxt): Likewise.
9432 (sssubv8qi3): Likewise.
9433 (sssubv4hi3): Likewise.
9434 (sssubv2si3): Likewise.
9435 (ussubv8qi3): Likewise.
9436 (ussubv4hi3): Likewise.
9437 (ussubv2si3): Likewise.
9438 (mulv4hi3_iwmmxt): Likewise.
9439 (smulv4hi3_highpart): Likewise.
9440 (umulv4hi3_highpart): Likewise.
9441 (iwmmxt_wmacs): Likewise.
9442 (iwmmxt_wmacsz): Likewise.
9443 (iwmmxt_wmacu): Likewise.
9444 (iwmmxt_wmacuz): Likewise.
9445 (iwmmxt_clrdi): Likewise.
9446 (iwmmxt_clrv8qi): Likewise.
9447 (iwmmxt_clr4hi): Likewise.
9448 (iwmmxt_clr2si): Likewise.
9449 (iwmmxt_uavgrndv8qi3): Likewise.
9450 (iwmmxt_uavgrndv4hi3): Likewise.
9451 (iwmmxt_uavgv8qi3): Likewise.
9452 (iwmmxt_uavgv4hi3): Likewise.
9453 (iwmmxt_tinsrb): Likewise.
9454 (iwmmxt_tinsrh): Likewise.
9455 (iwmmxt_tinsrw): Likewise.
9456 (iwmmxt_textrmub): Likewise.
9457 (iwmmxt_textrmsb): Likewise.
9458 (iwmmxt_textrmuh): Likewise.
9459 (iwmmxt_textrmsh): Likewise.
9460 (iwmmxt_textrmw): Likewise.
9461 (iwmxxt_wshufh): Likewise.
9462 (eqv8qi3): Likewise.
9463 (eqv4hi3): Likewise.
9464 (eqv2si3): Likewise.
9465 (gtuv8qi3): Likewise.
9466 (gtuv4hi3): Likewise.
9467 (gtuv2si3): Likewise.
9468 (gtv8qi3): Likewise.
9469 (gtv4hi3): Likewise.
9470 (gtv2si3): Likewise.
9471 (smax<mode>3_iwmmxt): Likewise.
9472 (umax<mode>3_iwmmxt): Likewise.
9473 (smin<mode>3_iwmmxt): Likewise.
9474 (umin<mode>3_iwmmxt): Likewise.
9475 (iwmmxt_wpackhss): Likewise.
9476 (iwmmxt_wpackwss): Likewise.
9477 (iwmmxt_wpackdss): Likewise.
9478 (iwmmxt_wpackhus): Likewise.
9479 (iwmmxt_wpackwus): Likewise.
9480 (iwmmxt_wpackdus): Likewise.
9481 (iwmmxt_wunpckihb): Likewise.
9482 (iwmmxt_wunpckihh): Likewise.
9483 (iwmmxt_wunpckihw): Likewise.
9484 (iwmmxt_wunpckilb): Likewise.
9485 (iwmmxt_wunpckilh): Likewise.
9486 (iwmmxt_wunpckilw): Likewise.
9487 (iwmmxt_wunpckehub): Likewise.
9488 (iwmmxt_wunpckehuh): Likewise.
9489 (iwmmxt_wunpckehuw): Likewise.
9490 (iwmmxt_wunpckehsb): Likewise.
9491 (iwmmxt_wunpckehsh): Likewise.
9492 (iwmmxt_wunpckehsw): Likewise.
9493 (iwmmxt_wunpckelub): Likewise.
9494 (iwmmxt_wunpckeluh): Likewise.
9495 (iwmmxt_wunpckeluw): Likewise.
9496 (iwmmxt_wunpckelsb): Likewise.
9497 (iwmmxt_wunpckelsh): Likewise.
9498 (iwmmxt_wunpckelsw): Likewise.
9499 (ror<mode>3): Likewise.
9500 (ashr<mode>3_iwmmxt): Likewise.
9501 (lshr<mode>3_iwmmxt): Likewise.
9502 (ashl<mode>3_iwmmxt): Likewise.
9503 (ror<mode>3_di): Likewise.
9504 (ashr<mode>3_di): Likewise.
9505 (lshr<mode>3_di): Likewise.
9506 (ashl<mode>3_di): Likewise.
9507 (iwmmxt_wmadds): Likewise.
9508 (iwmmxt_wmaddu): Likewise.
9509 (iwmmxt_tmia): Likewise.
9510 (iwmmxt_tmiaph): Likewise.
9511 (iwmmxt_tmiabb): Likewise.
9512 (iwmmxt_tmiatb): Likewise.
9513 (iwmmxt_tmiabt): Likewise.
9514 (iwmmxt_tmiatt): Likewise.
9515 (iwmmxt_tmovmskb): Likewise.
9516 (iwmmxt_tmovmskh): Likewise.
9517 (iwmmxt_tmovmskw): Likewise.
9518 (iwmmxt_waccb): Likewise.
9519 (iwmmxt_wacch): Likewise.
9520 (iwmmxt_waccw): Likewise.
9521 (iwmmxt_waligni): Likewise.
9522 (iwmmxt_walignr): Likewise.
9523 (iwmmxt_walignr0): Likewise.
9524 (iwmmxt_walignr1): Likewise.
9525 (iwmmxt_walignr2): Likewise.
9526 (iwmmxt_walignr3): Likewise.
9527 (iwmmxt_wsadb): Likewise.
9528 (iwmmxt_wsadh): Likewise.
9529 (iwmmxt_wsadbz): Likewise.
9530 (iwmmxt_wsadhz): Likewise.
9531 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
9532 (iwmmxt_wabsdiffb): Likewise.
9533 (iwmmxt_wabsdiffh): Likewise.
9534 (iwmmxt_wabsdiffw): Likewise.
9535 (iwmmxt_waddsubhx): Likewise
9536 (iwmmxt_wsubaddhx): Likewise.
9537 (addc<mode>3): Likewise.
9538 (iwmmxt_avg4): Likewise.
9539 (iwmmxt_avg4r): Likewise.
9540 (iwmmxt_wmaddsx): Likewise.
9541 (iwmmxt_wmaddux): Likewise.
9542 (iwmmxt_wmaddsn): Likewise.
9543 (iwmmxt_wmaddun): Likewise.
9544 (iwmmxt_wmulwsm): Likewise.
9545 (iwmmxt_wmulwum): Likewise.
9546 (iwmmxt_wmulsmr): Likewise.
9547 (iwmmxt_wmulumr): Likewise.
9548 (iwmmxt_wmulwsmr): Likewise.
9549 (iwmmxt_wmulwumr): Likewise.
9550 (iwmmxt_wmulwl): Likewise.
9551 (iwmmxt_wqmulm): Likewise.
9552 (iwmmxt_wqmulwm): Likewise.
9553 (iwmmxt_wqmulmr): Likewise.
9554 (iwmmxt_wqmulwmr): Likewise.
9555 (iwmmxt_waddbhusm): Likewise.
9556 (iwmmxt_waddbhusl): Likewise.
9557 (iwmmxt_wqmiabb): Likewise.
9558 (iwmmxt_wqmiabt): Likewise.
9559 (iwmmxt_wqmiatb): Likewise.
9560 (iwmmxt_wqmiatt): Likewise.
9561 (iwmmxt_wqmiabbn): Likewise.
9562 (iwmmxt_wqmiabtn): Likewise.
9563 (iwmmxt_wqmiatbn): Likewise.
9564 (iwmmxt_wqmiattn): Likewise.
9565 (iwmmxt_wmiabb): Likewise.
9566 (iwmmxt_wmiabt): Likewise.
9567 (iwmmxt_wmiatb): Likewise.
9568 (iwmmxt_wmiatt): Likewise.
9569 (iwmmxt_wmiabbn): Likewise.
9570 (iwmmxt_wmiabtn): Likewise.
9571 (iwmmxt_wmiatbn): Likewise.
9572 (iwmmxt_wmiattn): Likewise.
9573 (iwmmxt_wmiawbb): Likewise.
9574 (iwmmxt_wmiawbt): Likewise.
9575 (iwmmxt_wmiawtb): Likewise.
9576 (iwmmxt_wmiawtt): Likewise.
9577 (iwmmxt_wmiawbbn): Likewise.
9578 (iwmmxt_wmiawbtn): Likewise.
9579 (iwmmxt_wmiawtbn): Likewise.
9580 (iwmmxt_wmiawttn): Likewise.
9581 (iwmmxt_wmerge): Likewise.
9582 (iwmmxt_tandc<mode>3): Likewise.
9583 (iwmmxt_torc<mode>3): Likewise.
9584 (iwmmxt_torvsc<mode>3): Likewise.
9585 (iwmmxt_textrc<mode>3): Likewise.
9586 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
9587 (wmmxt_pack): Likewise.
9588 (wmmxt_mult_c1): Likewise.
9589 (wmmxt_mult_c2): Likewise.
9590 (wmmxt_alu_c1): Likewise.
9591 (wmmxt_alu_c2): Likewise.
9592 (wmmxt_alu_c3): Likewise.
9593 (wmmxt_transfer_c1): Likewise.
9594 (wmmxt_transfer_c2): Likewise.
9595 (wmmxt_transfer_c3): Likewise.
9596 (marvell_f_iwmmxt_wstr): Likewise.
9597 (marvell_f_iwmmxt_wldr): Likewise.
9599 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
9601 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
9603 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
9606 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
9607 * lra-constraints.c (need_for_split_p): Check call used hard regs
9608 living through calls.
9610 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
9611 call used regs for call insn.
9613 2013-06-28 Jakub Jelinek <jakub@redhat.com>
9616 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
9617 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
9619 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
9621 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
9623 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
9625 * lra-constraints.c (need_for_split_p): Check call used hard regs
9626 living through calls.
9628 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
9631 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
9632 to tie with any other modes. Eliminate Altivec vector mode tests,
9633 since these are a subset of ALTIVEC or VSX vector modes. Simplify
9634 code, to return 0 if testing MODE2 for a condition, if we've
9635 already tested MODE1 for the same condition.
9637 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
9639 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
9642 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
9644 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
9645 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
9647 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
9649 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
9651 (aarch64_symbolic_constant_p): Remove.
9652 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
9653 static. Fix line length and white space.
9654 (aarch64_symbolic_constant_p): Remove.
9655 * config/aarch64/predicates.md (aarch64_valid_symref):
9656 Use aarch64_classify_symbol_expression.
9658 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9660 * config/arm/constraints.md (Ts): New constraint.
9661 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
9663 (compare_scc): Use "Ts" constraint for operand 0.
9664 (ior_scc_scc): Likewise.
9665 (and_scc_scc): Likewise.
9666 (and_scc_scc_nodom): Likewise.
9667 (ior_scc_scc_cmp): Likewise for operand 7.
9668 (and_scc_scc_cmp): Likewise.
9669 * config/arm/thumb2.md (thumb2_movsi_insn):
9670 Add alternatives for 16-bit encodings.
9671 (thumb2_movhi_insn): Likewise.
9672 (thumb2_movsicc_insn): Likewise.
9673 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
9674 (thumb2_negscc): Use "Ts" constraint.
9675 Move mvn instruction outside cond_exec block.
9676 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
9677 for 16-bit encodings.
9679 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9681 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
9683 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
9684 (mulsi3subsi): Likewise.
9685 (mulsidi3adddi): Likewise.
9686 (mulsidi3_v6): Likewise.
9687 (umulsidi3_v6): Likewise.
9688 (umulsidi3adddi_v6): Likewise.
9689 (smulsi3_highpart_v6): Likewise.
9690 (umulsi3_highpart_v6): Likewise.
9691 (mulhisi3tb): Likewise.
9692 (mulhisi3bt): Likewise.
9693 (mulhisi3tt): Likewise.
9694 (maddhisi4): Likewise.
9695 (maddhisi4tb): Likewise.
9696 (maddhisi4tt): Likewise.
9697 (maddhidi4): Likewise.
9698 (maddhidi4tb): Likewise.
9699 (maddhidi4tt): Likewise.
9700 (zeroextractsi_compare0_scratch): Likewise.
9701 (insv_zero): Likewise.
9702 (insv_t2): Likewise.
9703 (anddi_notzesidi_di): Likewise.
9704 (anddi_notsesidi_di): Likewise.
9705 (andsi_notsi_si): Likewise.
9706 (iordi_zesidi_di): Likewise.
9707 (xordi_zesidi_di): Likewise.
9708 (andsi_iorsi3_notsi): Likewise.
9710 (smax_m1): Likewise.
9712 (not_shiftsi): Likewise.
9713 (unaligned_loadsi): Likewise.
9714 (unaligned_loadhis): Likewise.
9715 (unaligned_loadhiu): Likewise.
9716 (unaligned_storesi): Likewise.
9717 (unaligned_storehi): Likewise.
9718 (extv_reg): Likewise.
9719 (extzv_t2): Likewise.
9721 (udivsi3): Likewise.
9722 (arm_zero_extendhisi2addsi): Likewise.
9723 (arm_zero_extendqisi2addsi): Likewise.
9724 (compareqi_eq0): Likewise.
9725 (arm_extendhisi2_v6): Likewise.
9726 (arm_extendqisi2addsi): Likewise.
9727 (arm_movt): Likewise.
9728 (thumb2_ldrd): Likewise.
9729 (thumb2_ldrd_base): Likewise.
9730 (thumb2_ldrd_base_neg): Likewise.
9731 (thumb2_strd): Likewise.
9732 (thumb2_strd_base): Likewise.
9733 (thumb2_strd_base_neg): Likewise.
9734 (arm_negsi2): Add alternative for 16-bit encoding.
9735 (arm_one_cmplsi2): Likewise.
9737 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9739 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
9740 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
9741 (movdfcc): Likewise.
9742 * config/arm/vfp.md (*thumb2_movsf_vfp):
9743 Disable predication for arm_restrict_it.
9744 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
9745 (*thumb2_movdfcc_vfp): Likewise.
9746 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
9747 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
9748 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
9749 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
9750 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
9751 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
9752 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
9753 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
9754 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
9755 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
9756 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
9757 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
9758 Disable predication for arm_restrict_it.
9760 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
9762 * config/i386/bmiintrin.h (_bextr_u32): New.
9763 (_bextr_u64): Ditto.
9765 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
9767 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
9768 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
9769 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
9770 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
9771 * config/mips/n32-elf.h: ...this new file.
9773 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
9776 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
9777 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
9779 2013-06-27 Catherine Moore <clm@codesourcery.com>
9781 * config/mips/mips-tables.opt: Regenerate.
9782 * config/mips/mips-cpus.def: Add m14ke and m14kec.
9783 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
9784 * doc/invoke.texi: Add -m14kc.
9786 2013-06-27 Jakub Jelinek <jakub@redhat.com>
9789 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
9790 constraints of operand 1 and 2.
9793 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
9794 to match RTL canonicalization. Swap predicates and
9795 constraints of operand 1 and 2.
9797 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
9799 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
9800 Process OP_INOUT regs for splitting too.
9802 2013-06-27 Jakub Jelinek <jakub@redhat.com>
9804 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
9805 decl before the loop, initialize to NULL.
9806 (vectorizable_load): Initialize ptr_incr to NULL.
9808 2013-06-27 Martin Jambor <mjambor@suse.cz>
9811 * ipa-ref.h (ipa_maybe_record_reference): Declare.
9812 * ipa-ref.c (ipa_maybe_record_reference): New function.
9813 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
9814 * ipa-cp.c (create_specialized_node): Record potential references from
9816 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
9818 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
9820 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
9821 parameter 'mode' of type 'enum machine_mode mode'; change to pass
9822 'mode' to force_reg.
9823 (aarch64_add_offset): Update calls to aarch64_force_temporary.
9824 (aarch64_expand_mov_immediate): Likewise.
9826 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
9828 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
9829 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
9831 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9833 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
9834 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
9835 (struct machine_function): Add tbegin_p.
9836 (s390_canonicalize_comparison): Fold CC mode compares to
9837 conditional jump if possible.
9838 (s390_emit_jump): Return the emitted jump.
9839 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
9840 Handle CCRAWmode compares.
9841 (s390_option_override): Default to -mhtm if available.
9842 (s390_reg_clobbered_rtx): Handle floating point regs as well.
9843 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
9844 FPRs instead of df_regs_ever_live_p.
9845 (s390_optimize_nonescaping_tx): New function.
9846 (s390_init_frame_layout): Extend clobbered_regs array to cover
9848 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
9849 (s390_expand_tbegin): New function.
9850 (enum s390_builtin): New enum definition.
9851 (code_for_builtin): New array definition.
9852 (s390_init_builtins): New function.
9853 (s390_expand_builtin): New function.
9854 (TARGET_INIT_BUILTINS): Define.
9855 (TARGET_EXPAND_BUILTIN): Define.
9856 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
9857 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
9858 (s390_alc_comparison): Likewise.
9859 * config/s390/s390-modes.def: Add CCRAWmode.
9860 * config/s390/s390.h (processor_flags): Add PF_TX.
9861 (TARGET_CPU_HTM): Define macro.
9862 (TARGET_HTM): Define macro.
9863 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
9864 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
9865 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
9866 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
9867 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
9868 (TBEGIN_MASK, TBEGINC_MASK): New constants.
9869 ("*cc_to_int"): Move up.
9870 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
9871 constants other than 0.
9872 ("*ccraw_to_int"): New insn and splitter definition.
9873 ("tbegin", "tbegin_nofloat", "tbegin_retry")
9874 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
9875 ("tx_assist"): New expander.
9876 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
9877 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
9878 * config/s390/s390.opt: Add -mhtm option.
9879 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
9880 * config/s390/htmxlintrin.h: New file.
9881 * config/s390/htmintrin.h: New file.
9882 * config/s390/s390intrin.h: New file.
9883 * doc/extend.texi: Document htm builtins.
9884 * config.gcc: Add the new header files to extra_headers.
9886 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
9888 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
9889 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
9891 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
9892 Pat Haugen <pthaugen@us.ibm.com>
9893 Peter Bergner <bergner@vnet.ibm.com>
9895 * config/rs6000/power8.md: New.
9896 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
9897 setting for power8 entry.
9898 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
9899 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
9900 test for Power4/Power5 only.
9901 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
9903 (force_new_group): Adjust comment.
9904 * config/rs6000/rs6000.md: Include power8.md.
9906 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
9908 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
9909 * config/arm/arm-protos.h (arm_max_conditional_execute): New
9911 (tune_params): Update comment.
9912 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
9913 (arm_max_conditional_execute): New function.
9914 (thumb2_final_prescan_insn): Use max_insn_skipped and
9915 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
9917 2013-06-25 Jakub Jelinek <jakub@redhat.com>
9919 PR tree-optimization/57705
9920 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
9921 SSA_NAME step, provided that it is not defined inside the loop.
9922 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
9923 (get_initial_def_for_induction): Handle SSA_NAME IV step.
9925 2013-06-25 Martin Jambor <mjambor@suse.cz>
9928 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
9929 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
9931 (ipa_note_param_call): Initialize member_ptr flag.
9932 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
9933 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
9934 (ipa_write_indirect_edge_info): Stream member_ptr flag.
9935 (ipa_read_indirect_edge_info): Likewise.
9937 2013-06-25 Richard Biener <rguenther@suse.de>
9940 * passes.c (init_optimization_passes): Move pass_fold_builtins
9941 and pass_dce earlier with -Og.
9943 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
9945 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
9946 <BIT_FIELD_REF>: Remove trailing TAB.
9947 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
9950 2013-06-24 Martin Jambor <mjambor@suse.cz>
9952 PR tree-optimization/57358
9953 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
9954 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
9955 (ipa_analyze_params_uses): Generate pessimistic info when true.
9957 2013-06-24 Martin Jambor <mjambor@suse.cz>
9959 PR tree-optimization/57539
9960 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
9961 global.inlined_to of the new node to it. All callers changed.
9962 * ipa-inline-transform.c (clone_inlined_nodes): New variable
9963 inlining_into, pass it to cgraph_clone_node.
9964 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
9965 ipa_free_edge_args_substructures.
9966 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
9967 rdesc linked list. Do not assert rdesc edges have inlined caller.
9968 Assert we have found an rdesc in the rdesc list.
9970 2013-06-24 Richard Biener <rguenther@suse.de>
9972 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
9973 (pointer_set_lookup): Declare.
9974 (class pointer_map): New template class implementing a
9975 generic pointer to T map.
9976 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
9977 pointer_map<T>::contains, pointer_map<T>::insert,
9978 pointer_map<T>::traverse): New functions.
9979 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
9980 (pointer_set_lookup): New function.
9981 (pointer_set_contains): Use pointer_set_lookup.
9982 (pointer_set_insert): Likewise.
9983 (insert_aux): Remove.
9984 (struct pointer_map_t): Embed a pointer_set_t.
9985 (pointer_map_create): Adjust.
9986 (pointer_map_destroy): Likewise.
9987 (pointer_map_contains): Likewise.
9988 (pointer_map_insert): Likewise.
9989 (pointer_map_traverse): Likewise.
9990 * tree-streamer.h (struct streamer_tree_cache_d): Use a
9991 pointer_map<unsigned> instead of a pointer_map_t.
9992 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
9993 (streamer_tree_cache_lookup): Likewise.
9994 (streamer_tree_cache_create): Likewise.
9995 (streamer_tree_cache_delete): Likewise.
9996 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
9997 pointer_map<unsigned> instead of a pointer_map_t.
9998 (lto_init_tree_ref_encoder): Adjust.
9999 (lto_destroy_tree_ref_encoder): Likewise.
10000 * lto-section-out.c (lto_output_decl_index): Likewise.
10001 (lto_record_function_out_decl_state): Likewise.
10002 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
10004 2013-06-24 Richard Biener <rguenther@suse.de>
10006 PR tree-optimization/57488
10007 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
10009 2013-06-24 Alan Modra <amodra@gmail.com>
10011 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
10012 (gen_easy_altivec_constant): Likewise.
10013 * config/rs6000/predicates.md (easy_vector_constant_add_self,
10014 easy_vector_constant_msb): Likewise.
10016 2013-06-23 Jakub Jelinek <jakub@redhat.com>
10019 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
10020 add missing return true.
10022 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
10025 * config/sh/predicates.md (general_extend_operand): Invoke
10026 general_movsrc_operand for memory operands.
10027 (general_movsrc_operand): Allow reg+reg addressing, do not use
10028 general_operand for memory operands.
10030 2013-06-23 Sriraman Tallam <tmsriram@google.com>
10032 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
10033 when current target options does not apply.
10034 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
10035 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
10036 * config/i386/bmiintrin.h: Pass appropriate target
10037 attributes to header.
10038 * config/i386/mmintrin.h: Ditto.
10039 * config/i386/nmmintrin.h: Ditto.
10040 * config/i386/avx2intrin.h: Ditto.
10041 * config/i386/fxsrintrin.h: Ditto.
10042 * config/i386/tbmintrin.h: Ditto.
10043 * config/i386/xsaveintrin.h: Ditto.
10044 * config/i386/f16cintrin.h: Ditto.
10045 * config/i386/xtestintrin.h: Ditto.
10046 * config/i386/xsaveoptintrin.h: Ditto.
10047 * config/i386/bmi2intrin.h: Ditto.
10048 * config/i386/lzcntintrin.h: Ditto.
10049 * config/i386/smmintrin.h: Ditto.
10050 * config/i386/wmmintrin.h: Ditto.
10051 * config/i386/x86intrin.h: Remove all header include guards.
10052 * config/i386/prfchwintrin.h: Ditto.
10053 * config/i386/pmmintrin.h: Ditto.
10054 * config/i386/tmmintrin.h: Ditto.
10055 * config/i386/xmmintrin.h: Ditto.
10056 * config/i386/popcntintrin.h: Ditto.
10057 * config/i386/rdseedintrin.h: Ditto.
10058 * config/i386/ammintrin.h: Ditto.
10059 * config/i386/emmintrin.h: Ditto.
10060 * config/i386/immintrin.h: Remove all header include guards.
10061 * config/i386/fma4intrin.h: Ditto.
10062 * config/i386/lwpintrin.h: Ditto.
10063 * config/i386/xopintrin.h: Ditto.
10064 * config/i386/ia32intrin.h: Ditto.
10065 * config/i386/avxintrin.h: Ditto.
10066 * config/i386/rtmintrin.h: Ditto.
10067 * config/i386/fmaintrin.h: Ditto.
10068 * config/i386/mm3dnow.h: Ditto.
10070 2013-06-22 Sriraman Tallam <tmsriram@google.com>
10072 * common/config/i386/i386-common.c: Handle LZCNT.
10074 2013-06-22 Andi Kleen <ak@linux.intel.com>
10076 * doc/extend.texi: Use __atomic_store_n instead of
10077 __atomic_store in HLE example.
10079 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
10081 * config/sh/sh.c: Remove <cstdlib> workaround.
10083 2013-06-21 Andi Kleen <ak@linux.intel.com>
10085 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
10087 2013-06-21 Andi Kleen <ak@linux.intel.com>
10089 * doc/extend.texi: Document that __atomic_clear and
10090 __atomic_test_and_set should only be used with bool.
10092 2013-06-20 Jan Hubicka <jh@suse.cz>
10094 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
10095 types_same_for_odr.
10096 * tree.c (decls_same_for_odr): New function.
10097 (same_for_edr): New function.
10098 (types_same_for_odr): New function.
10099 (get_binfo_at_offset): Use it.
10100 * tree.h (types_same_for_odr): Declare.
10102 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
10103 Jason Merrill <jason@redhat.com>
10105 * system.h: Include <cstdlib> as well as <stdlib.h>.
10107 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
10110 * config/i386/i386.c (construct_container): Report error if
10111 long double is used with disabled x87 float returns.
10113 2013-06-20 Jan Hubicka <jh@suse.cz>
10115 * lto-cgraph.c (input_symtab): Do not set cgraph state.
10117 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
10119 PR rtl-optimization/57425
10120 PR rtl-optimization/57569
10121 * alias.c (write_dependence_p): Remove parameters mem_mode and
10122 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
10123 Changed all callers.
10124 (canon_anti_dependence): Get comments and semantics in sync.
10125 Add parameter mem_canonicalized. Changed all callers.
10126 * rtl.h (canon_anti_dependence): Update prototype.
10128 2013-06-20 Richard Biener <rguenther@suse.de>
10130 * data-streamer-in.c (streamer_read_uhwi): Optimize single
10131 byte case, inline streamer_read_uchar and defer section
10134 2013-06-20 Richard Biener <rguenther@suse.de>
10136 PR tree-optimization/57584
10137 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
10138 SSA names into the expanded expression that take part in
10139 abnormal coalescing.
10141 2013-06-19 Sharad Singhai <singhai@google.com>
10143 * gcov.c (print_usage): Handle new option.
10144 (process_args): Ditto.
10145 (get_gcov_intermediate_filename): New function.
10146 (output_intermediate_file): New function.
10147 (output_gcov_file): New function
10148 (generate_results): Handle new option.
10149 (release_function): Relase demangled name.
10150 (read_graph_file): Handle demangled name.
10151 (output_lines): Ditto.
10152 * doc/gcov.texi: Document gcov intermediate format.
10154 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
10157 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
10158 (lra_emit_add): Use the functions. Add comment about Y as an
10161 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
10164 * collect2.c (collect_atexit): New.
10165 (collect_exit): Delete.
10166 (main): Register collect_atexit with atexit.
10167 (collect_wait): Change collect_exit to exit.
10169 * collect2.h (collect_exit): Delete.
10170 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
10172 2013-06-19 Wei Mi <wmi@google.com>
10174 PR rtl-optimization/57518
10175 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
10176 if regno is used in paradoxical subreg.
10177 (update_equiv_regs): Check pdx_subregs[regno] before
10178 set a reg to be equivalent with a mem.
10180 2013-06-19 Matthias Klose <doko@ubuntu.com>
10183 * file-find.h (find_a_file): Add a mode parameter.
10184 * file-find.c (find_a_file): Likewise.
10185 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
10186 with X_OK for the executables.
10187 * collect2.c (main): Call find_a_file with X_OK.
10189 2013-06-19 Steve Ellcey <sellcey@mips.com>
10192 * config/mips/mips.md (casesi_internal_mips16_<mode>):
10193 Use NEXT_INSN instead of next_real_insn.
10195 2013-06-19 Jan Hubicka <jh@suse.cz>
10197 * cgraph.h (const_value_known_p): Replace by ...
10198 (ctor_for_folding): .. this one.
10199 * cgraphunit.c (process_function_and_variable_attributes): Use it.
10200 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
10201 * expr.c (expand_expr_real_1): Likewise.
10202 (string_constant): Likewise.
10203 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
10204 * ipa.c (process_references): Likewise.
10205 (symtab_remove_unreachable_nodes): Likewise.
10206 * ipa-inline-analysis.c (param_change_prob): Likewise.
10207 * gimple-fold.c (canonicalize_constructor_val): Likewise.
10208 (get_base_constructor): Likwise.
10209 * varpool.c (varpool_remove_node): Likewise.
10210 (varpool_remove_initializer): LIkewise.
10211 (dump_varpool_node): LIkwise.
10212 (const_value_known_p): Rewrite to ...
10213 (ctor_for_folding): ... this one.
10215 2013-06-19 Jakub Jelinek <jakub@redhat.com>
10218 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
10219 PERSONALITY in $PATH derived prefixes.
10221 2013-06-19 Jeff Law <law@redhat.com>
10223 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
10226 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
10227 (simplify_bitwise_binary): Use it to simpify certain binary ops on
10230 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
10232 * config/arm/vfp.md: Move VFP instruction classification documentation
10234 * config/arm/arm.md: ... here. Update instruction classification
10237 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
10239 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
10240 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
10241 pattern. Use more efficient sequences on ARMv5 and Thumb2.
10243 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
10246 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
10247 with NEXT_INSN. Use tablejump_p to check for jump table data
10250 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
10253 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
10254 that now in C++ the value is correct per the C++ standards.
10256 2013-06-19 Richard Biener <rguenther@suse.de>
10258 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
10259 for global context.
10261 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10264 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10267 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
10268 with next_active_insn.
10270 2013-06-18 Sriraman Tallam <tmsriram@google.com>
10272 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
10273 functions are inlined during failures to flag an error.
10274 * tree-inline.c (expand_call_inline): Allow the error to be flagged
10275 in early inline pass.
10277 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
10279 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
10282 2013-06-18 Julian Brown <julian@codesourcery.com>
10284 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
10285 Permit virtual register pre-reload if !strict.
10286 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
10288 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
10290 * config/arm/neon.md (movmisalign<mode>): Use
10291 neon_perm_struct_or_reg_operand instead of
10292 neon_struct_or_register_operand.
10293 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
10294 neon_permissive_struct_operand instead of neon_struct_operand.
10295 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
10296 neon_vector_mem_operand.
10297 * config/arm/predicates.md (neon_struct_operand): Adjust call to
10298 neon_vector_mem_operand.
10299 (neon_permissive_struct_operand): New.
10300 (neon_struct_or_register_operand): Rename to...
10301 (neon_perm_struct_or_reg_operand): This. Adjust call to
10302 neon_vector_mem_operand.
10304 2013-06-18 Richard Biener <rguenther@suse.de>
10306 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
10307 * lto-streamer.h: Include pointer-set.h.
10308 (struct lto_decl_slot): Remove.
10309 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
10310 Remove next_index entry.
10311 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
10312 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
10313 (lto_init_tree_ref_encoder): Adjust.
10314 (lto_destroy_tree_ref_encoder): Likewise.
10315 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
10316 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
10317 (lto_output_decl_index): Adjust.
10318 (lto_new_out_decl_state): Likewise.
10319 (lto_record_function_out_decl_state): Likewise.
10320 * lto-streamer-out.c (copy_function): Likewise.
10322 2013-06-18 Richard Biener <rguenther@suse.de>
10324 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
10325 * cgraphunit.c: Include cfgloop.h.
10326 (init_lowered_empty_function): Initialize the loop tree.
10327 (assemble_thunk): Insert new BBs into loops.
10329 2013-06-18 Richard Biener <rguenther@suse.de>
10331 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
10332 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
10333 the map from cache entry to cache index optional.
10334 (streamer_tree_cache_replace_tree): Adjust accordingly.
10335 (streamer_tree_cache_append): Likewise.
10336 (streamer_tree_cache_delete): Likewise.
10337 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
10338 streamer cache map from cache entry to cache index.
10339 * lto-streamer-out.c (create_output_block): Adjust.
10341 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
10343 * config/arm/arm.md (attribute "insn"): Move multiplication and
10344 division attributes to...
10345 (attribute "type"): ... here. Remove mult.
10346 (attribute "mul32"): New attribute.
10347 (attribute "mul64"): Add umaal.
10348 (*arm_mulsi3): Update attributes.
10349 (*arm_mulsi3_v6): Likewise.
10350 (*thumb_mulsi3): Likewise.
10351 (*thumb_mulsi3_v6): Likewise.
10352 (*mulsi3_compare0): Likewise.
10353 (*mulsi3_compare0_v6): Likewise.
10354 (*mulsi_compare0_scratch): Likewise.
10355 (*mulsi_compare0_scratch_v6): Likewise.
10356 (*mulsi3addsi): Likewise.
10357 (*mulsi3addsi_v6): Likewise.
10358 (*mulsi3addsi_compare0): Likewise.
10359 (*mulsi3addsi_compare0_v6): Likewise.
10360 (*mulsi3addsi_compare0_scratch): Likewise.
10361 (*mulsi3addsi_compare0_scratch_v6): Likewise.
10362 (*mulsi3subsi): Likewise.
10363 (*mulsidi3adddi): Likewise.
10364 (*mulsi3addsi_v6): Likewise.
10365 (*mulsidi3adddi_v6): Likewise.
10366 (*mulsidi3_nov6): Likewise.
10367 (*mulsidi3_v6): Likewise.
10368 (*umulsidi3_nov6): Likewise.
10369 (*umulsidi3_v6): Likewise.
10370 (*umulsidi3adddi): Likewise.
10371 (*umulsidi3adddi_v6): Likewise.
10372 (*smulsi3_highpart_nov6): Likewise.
10373 (*smulsi3_highpart_v6): Likewise.
10374 (*umulsi3_highpart_nov6): Likewise.
10375 (*umulsi3_highpart_v6): Likewise.
10376 (mulhisi3): Likewise.
10377 (*mulhisi3tb): Likewise.
10378 (*mulhisi3bt): Likewise.
10379 (*mulhisi3tt): Likewise.
10380 (maddhisi4): Likewise.
10381 (*maddhisi4tb): Likewise.
10382 (*maddhisi4tt): Likewise.
10383 (maddhidi4): Likewise.
10384 (*maddhidi4tb): Likewise.
10385 (*maddhidi4tt): Likewise.
10386 (divsi3): Likewise.
10387 (udivsi3): Likewise.
10388 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
10389 (thumb2_mulsi_short_compare0): Likewise.
10390 (thumb2_mulsi_short_compare0_scratch): Likewise.
10391 * config/arm/arm1020e.md (1020mult1): Update attribute change.
10392 (1020mult2): Likewise.
10393 (1020mult3): Likewise.
10394 (1020mult4): Likewise.
10395 (1020mult5): Likewise.
10396 (1020mult6): Likewise.
10397 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
10399 (cortex_a15_mult64): Likewise.
10400 (cortex_a15_sdiv): Likewise.
10401 (cortex_a15_udiv): Likewise.
10402 * config/arm/arm1026ejs.md (mult1): Update attribute change.
10408 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
10409 (pj4_ir_div): Likewise.
10410 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
10411 (11_mult2): Likewise.
10412 (11_mult3): Likewise.
10413 (11_mult4): Likewise.
10414 (11_mult5): Likewise.
10415 (11_mult6): Likewise.
10416 (11_mult7): Likewise.
10417 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
10418 (cortex_a8_mla): Likewise.
10419 (cortex_a8_mull): Likewise.
10420 (cortex_a8_smulwy): Likewise.
10421 (cortex_a8_smlald): Likewise.
10422 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
10423 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
10424 (cortex_r4_mul_3): Likewise.
10425 (cortex_r4_mla_4): Likewise.
10426 (cortex_r4_mla_3): Likewise.
10427 (cortex_r4_smlald): Likewise.
10428 (cortex_r4_mull): Likewise.
10429 (cortex_r4_sdiv): Likewise.
10430 (cortex_r4_udiv): Likewise.
10431 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
10432 (cortex_a7_idiv): Likewise.
10433 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
10434 (9_mult2): Likewise.
10435 (9_mult3): Likewise.
10436 (9_mult4): Likewise.
10437 (9_mult5): Likewise.
10438 (9_mult6): Likewise.
10439 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
10440 (cortex_a53_sdiv): Likewise.
10441 (cortex_a53_udiv): Likewise.
10442 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
10443 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
10444 (mp626_mult2): Likewise.
10445 (mp626_mult3): Likewise.
10446 (mp626_mult4): Likewise.
10447 * config/arm/fa526.md (526_mult1): Update attribute change.
10448 (526_mult2): Likewise.
10449 * config/arm/arm-generic.md (mult): Update attribute change.
10450 (mult_ldsched_strongarm): Likewise.
10451 (mult_ldsched): Likewise.
10452 (multi_cycle): Likewise.
10453 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
10454 * config/arm/fa606te.md (606te_mult1): Update attribute change.
10455 (606te_mult2): Likewise.
10456 (606te_mult3): Likewise.
10457 (606te_mult4): Likewise.
10458 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
10459 (cortex_a9_mac16): Likewise.
10460 (cortex_a9_multiply): Likewise.
10461 (cortex_a9_mac): Likewise.
10462 (cortex_a9_multiply_long): Likewise.
10463 * config/arm/fa626te.md (626te_mult1): Update attribute change.
10464 (626te_mult2): Likewise.
10465 (626te_mult3): Likewise.
10466 (626te_mult4): Likewise.
10468 2013-06-18 Richard Biener <rguenther@suse.de>
10471 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
10473 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10476 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
10477 with next_active_insn.
10479 2013-06-18 Alan Modra <amodra@gmail.com>
10481 * config/rs6000/rs6000.h (enum data_align): New.
10482 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
10483 (DATA_ABI_ALIGNMENT): Define.
10484 (CONSTANT_ALIGNMENT): Correct comment.
10485 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
10486 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
10488 2013-06-17 David Malcolm <dmalcolm@redhat.com>
10490 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
10491 ATTRIBUTE_UNUSED marking.
10493 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
10495 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
10496 alternative and update.
10497 (aarch64_dup_lanedi): Delete.
10498 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
10499 * config/aarch64/aarch64-simd-builtins.def: Update.
10501 2013-06-17 Richard Biener <rguenther@suse.de>
10503 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
10504 (lto_input_scc): Declare.
10505 (lto_input_tree_1): Likewise.
10506 (struct lto_stats_d): Add num_tree_bodies_output and
10507 num_pickle_refs_output.
10508 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
10509 (lto_read_tree_1): Split out from ...
10510 (lto_read_tree): ... this.
10511 (lto_input_scc): New function.
10512 (lto_input_tree_1): Split out from ...
10513 (lto_input_tree): ... this. Handle LTO_tree_scc.
10514 (lto_data_in_create): Create the streamer cache without hashes.
10515 * lto-streamer-out.c (create_output_block): Create the streamer
10516 cache with hashes when not doing WPA.
10517 (lto_write_tree_1): Split out from ...
10518 (lto_write_tree): ... this.
10519 (get_symbol_initial_value): New function.
10520 (lto_output_tree_1): Split out from ...
10521 (lto_output_tree): ... this. Write trees as series of SCCs
10522 using a DFS walk via DFS_write_tree.
10523 (struct sccs, struct scc_entry): New types.
10524 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
10525 (DFS_write_tree_body): New function.
10526 (DFS_write_tree): Likewise.
10527 (hash_tree): Likewise.
10528 (scc_entry_compare): Likewise.
10529 (hash_scc): Likewise.
10530 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
10531 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
10532 TREE_CHAIN as regular reference.
10533 (streamer_read_integer_cst): Remove.
10534 (streamer_get_pickled_tree): Adjust.
10535 * tree-streamer-out.c (streamer_write_chain): Disable streaming
10536 of DECL_EXTERNALs in BLOCK_VARS for now.
10537 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
10539 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
10540 Add hash value argument and record that if hashes are recorded
10542 (streamer_tree_cache_insert_1): Adjust.
10543 (streamer_tree_cache_insert): Likewise.
10544 (streamer_tree_cache_insert_at): Rename to ...
10545 (streamer_tree_cache_replace_tree): ... this and adjust.
10546 (streamer_tree_cache_append): Adjust.
10547 (record_common_node): Likewise.
10548 (streamer_tree_cache_create): Add argument whether to
10549 record hash values together with trees.
10550 (streamer_tree_cache_delete): Adjust.
10551 * tree-streamer.h (struct streamer_tree_cache_d): Add
10553 (streamer_read_integer_cst): Remove.
10554 (streamer_tree_cache_insert): Adjust.
10555 (streamer_tree_cache_append): Likewise.
10556 (streamer_tree_cache_insert_at): Rename to ...
10557 (streamer_tree_cache_replace_tree): ... this and adjust.
10558 (streamer_tree_cache_create): Add argument whether to record hashes.
10559 (streamer_tree_cache_get): Rename to ...
10560 (streamer_tree_cache_get_tree): ... this.
10561 (streamer_tree_cache_get_hash): New function.
10562 * tree.c (cache_integer_cst): New function.
10563 * tree.h (cache_integer_cst): Declare.
10564 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
10565 * lto-symtab.c (lto_varpool_replace_node): Only release
10566 DECL_INITIAL of non-prevailing decls.
10567 * varpool.c (varpool_remove_initializer): Do not release
10568 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
10570 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
10572 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
10573 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
10574 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
10575 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
10576 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
10577 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
10578 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
10579 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
10580 instead of TARGET_64BIT.
10581 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
10582 Require ISA_HAS_<D>DIV.
10584 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
10586 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
10587 (mips*-*-linux*): Move default with_llsc setting to where other
10589 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
10591 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
10592 Likewise. Remove default with_tune setting. Move default float
10593 setting to its own block. Handle with_llsc in the same block as above.
10595 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
10597 PR rtl-optimization/57425
10598 PR rtl-optimization/57569
10599 * alias.c (write_dependence_p): Add new parameters mem_mode,
10600 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
10601 Changed all callers.
10602 (canon_anti_dependence): New function.
10603 * cse.c (check_dependence): Use canon_anti_dependence.
10604 * cselib.c (cselib_invalidate_mem): Likewise.
10605 * rtl.h (canon_anti_dependence): Declare.
10607 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
10609 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
10610 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
10611 ".set mips3" for 64-bit targets.
10613 2013-06-15 Dehao Chen <dehao@google.com>
10615 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
10616 * gimple-low.c (gimple_check_call_matching_types): Likewise.
10617 (gimple_check_call_args): Likewise.
10618 * value-prof.c (check_ic_target): Likewise.
10619 * ipa-inline.c (early_inliner): Likewise.
10620 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
10621 * cgraph.c (cgraph_create_edge_1): Likewise.
10622 (cgraph_make_edge_direct): Likewise.
10624 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
10627 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
10628 rs6000_output_move_128bit to handle emitting quad memory
10629 operations. Set attribute length to 8 bytes.
10631 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
10633 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
10635 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
10636 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
10637 (aarch64_<su>mlsl<mode>): Likewise.
10639 2013-06-14 Mike Stump <mikestump@comcast.net>
10641 * Makefile.in (TARGET_H): Add insn-codes.h.
10643 2013-06-14 Alan Modra <amodra@gmail.com>
10645 PR middle-end/57134
10646 PR middle-end/57586
10647 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
10648 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
10649 bitfield expansion when EXPAND_MEMORY.
10650 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
10652 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
10654 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
10655 test for clearing quad memory on 32-bit later.
10657 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
10659 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
10660 (fold_negate_expr): Likewise.
10661 (fold_real_zero_addition_p): Handle vectors.
10662 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
10664 2013-06-14 Alan Modra <amodra@gmail.com>
10666 * varasm.c (force_const_mem): Revert 2013-06-07 change.
10668 2013-06-13 Jan Hubicka <jh@suse.cz>
10670 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
10671 Local comdats are not externally visible.
10672 * symtab.c (dump_symtab_base): Dump externally visible.
10673 (verify_symtab_base): Verify back links in the symtab hash.
10675 2013-06-13 Bin Cheng <bin.cheng@arm.com>
10677 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
10678 CONVERT_EXPR as equal nodes.
10680 2013-06-13 Bin Cheng <bin.cheng@arm.com>
10682 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
10684 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
10686 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
10687 Generalize to complex and vector.
10688 * tree.c (build_all_ones_cst): New function.
10689 * tree.h (build_all_ones_cst): Declare it.
10691 2013-06-13 Alan Modra <amodra@gmail.com>
10693 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
10694 * config/rs6000/rs6000.md (signbittf2): New insn.
10695 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
10696 (abstf2_internal, cmptf_internal2): Likewise.
10697 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
10699 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
10700 Pat Haugen <pthaugen@us.ibm.com>
10701 Peter Bergner <bergner@vnet.ibm.com>
10703 * config/rs6000/rs6000.c (emit_load_locked): Add support for
10704 power8 byte, half-word, and quad-word atomic instructions.
10705 (emit_store_conditional): Likewise.
10706 (rs6000_expand_atomic_compare_and_swap): Likewise.
10707 (rs6000_expand_atomic_op): Likewise.
10709 * config/rs6000/sync.md (larx): Add new modes for power8.
10711 (AINT): New mode iterator to include TImode as well as normal
10712 integer modes on power8.
10713 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
10714 that VSX registers are not considered. Use AINT mode iterator
10715 instead of INT1 to allow inclusion of quad word atomic operations
10717 (load_locked<mode>): Likewise.
10718 (store_conditional<mode>): Likewise.
10719 (atomic_compare_and_swap<mode>): Likewise.
10720 (atomic_exchange<mode>): Likewise.
10721 (atomic_nand<mode>): Likewise.
10722 (atomic_fetch_<fetchop_name><mode>): Likewise.
10723 (atomic_nand_fetch<mode>): Likewise.
10724 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
10726 (ATOMIC): On power8, add QImode, HImode modes.
10727 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
10728 modes that promote to SImode.
10729 (load_lockedti): Convert TImode arguments to PTImode, so that we
10730 get a guaranteed even/odd register pair.
10731 (load_lockedpti): Likewise.
10732 (store_conditionalti): Likewise.
10733 (store_conditionalpti): Likewise.
10735 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
10736 atomic load/store instructions.
10739 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
10741 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
10743 (insn_count): New attribute, with most cases extracted from...
10744 (length): ...here. Redefine most cases in terms of insn_count.
10745 (single_insn): Delete.
10746 (can_delay): Use insn_count to check for single instructions.
10747 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
10748 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
10749 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
10750 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
10751 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
10752 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
10753 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
10754 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
10755 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
10756 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
10757 rather than "length".
10758 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
10759 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
10760 Use "insn_count" rather than "length".
10761 * config/mips/mips-dsp.md
10762 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
10763 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
10766 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
10768 PR tree-optimization/57361
10769 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
10771 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
10773 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
10775 (aarch64_simd_combine<mode>): New instruction expansion.
10776 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
10777 function prototype.
10778 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
10779 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
10781 2013-06-12 Jan Hubicka <jh@suse.cz>
10783 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
10784 decl has when in streaming stage.
10785 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
10786 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
10788 2013-06-12 Roland Stigge <stigge@antcom.de>
10791 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
10793 2013-06-12 Jakub Jelinek <jakub@redhat.com>
10795 PR tree-optimization/57537
10796 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
10797 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
10799 2013-06-12 Richard Biener <rguenther@suse.de>
10801 * data-streamer.h (streamer_write_char_stream): CSE
10802 obs->current_pointer.
10803 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
10804 streamer_write_char_stream manually and optimize the resulting loop.
10805 (streamer_write_hwi_stream): Likewise.
10807 2013-06-12 Jan Hubicka <jh@suse.cz>
10809 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
10810 * cgraph.h (varpool_create_empty_node): Declare.
10811 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
10813 * symtab.c (symtab_unregister_node): Be lax about missin entries
10815 (symtab_get_node): Update comment.
10816 * varpool.c (varpool_create_empty_node): Break out from ...
10817 (varpool_node_for_decl): ... here.
10818 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
10820 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
10822 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
10823 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
10824 part. Use straight-line flow at the end.
10825 <COMPONENT_REF>: Remove superfluous else.
10826 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
10828 2013-06-12 Jakub Jelinek <jakub@redhat.com>
10831 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
10832 target hook even for !TREE_PUBLIC decls. If no resolution info
10833 is available, return false for common and external decls.
10835 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
10837 * config/rl78/constraints.md (U): New constraint.
10838 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
10841 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
10844 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
10845 to allow returning address to AT_PLATFORM name.
10847 2013-06-11 Jan Hubicka <jh@suse.cz>
10849 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
10850 * cgraph.h (symtab_node_base): Add weakref flag.
10851 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
10852 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
10853 (output_weakrefs): Use weakref flag.
10854 * fold-const.c (simple_operand_p): Handle WEAK.
10855 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
10856 * ipa.c (varpool_externally_visible_p): Drop weakref.
10857 (function_and_variable_visibility): Update comment; fix weakref
10858 sanity checks; do not clear DECL_WEAK on them.
10859 * lto-cgraph.c (lto_output_node): update.
10860 (lto_output_varpool_node): Update.
10861 (input_overwrite_node): Update.
10862 (input_node): Update.
10863 (input_varpool_node): Update.
10864 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
10865 (lto_symtab_merge_symbols): Add sanity check.
10866 (lto_symtab_prevailing_decl): Do not special case weakrefs.
10867 * passes.c (rest_of_decl_compilation): Set static flag, too.
10868 * symtab.c (dump_symtab_base): Dump weakref.
10869 (verify_symtab_base): Sanity check weakrefs.
10870 (symtab_make_decl_local): Remove duplicated code.
10871 (symtab_alias_ultimate_target): Simplify.
10872 * varpool.c (varpool_create_variable_alias): Set weakref flag.
10874 2013-06-11 Tom de Vries <tom@codesourcery.com>
10876 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
10877 sequence_vect == NULL.
10879 2013-06-11 DJ Delorie <dj@redhat.com>
10881 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
10882 (rl78_unwind_word_mode): New.
10884 2013-06-11 David Malcolm <dmalcolm@redhat.com>
10886 * final.c (debug_prefix_maps): Make static.
10888 2013-06-11 David Malcolm <dmalcolm@redhat.com>
10890 * function.c (initial_trampoline): Remove stray copy.
10892 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
10894 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
10896 2013-06-11 Martin Jambor <mjambor@suse.cz>
10898 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
10899 within bounds at the beginning of the function.
10901 2013-06-11 Alan Modra <amodra@gmail.com>
10903 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
10905 (get_named_section): Don't NULL !DECL_P decl.
10907 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
10909 * doc/invoke.texi (core-avx2): Document.
10911 (atom): Updated with MOVBE.
10913 2013-06-11 Richard Biener <rguenther@suse.de>
10915 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
10917 2013-06-11 Anton Blanchard <anton@samba.org>
10919 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
10920 correct shift value in little-endian mode.
10922 2013-06-11 Jakub Jelinek <jakub@redhat.com>
10925 * varasm.c (get_variable_align): Move #endif to the right place.
10927 2013-06-10 Cary Coutant <ccoutant@google.com>
10929 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
10930 for hash so that hash table traversal order is deterministic.
10932 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
10933 Pat Haugen <pthaugen@us.ibm.com>
10934 Peter Bergner <bergner@vnet.ibm.com>
10936 * config/rs6000/vector.md (GPR move splitter): Do not split moves
10937 of vectors in GPRS if they are direct moves or quad word load or
10940 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
10942 (direct_move_p): Likewise.
10943 (quad_load_store_p): Likewise.
10945 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
10946 classes into bins based on the physical register type.
10947 (reg_class_to_reg_type): Likewise.
10948 (IS_STD_REG_TYPE): Likewise.
10949 (IS_FP_VECT_REG_TYPE): Likewise.
10950 (reload_fpr_gpr): Arrays to determine what insn to use if we can
10951 use direct move instructions.
10952 (reload_gpr_vsx): Likewise.
10953 (reload_vsx_gpr): Likewise.
10954 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
10955 information that is a simplification of register classes. Also
10956 precalculate direct move reload helpers.
10957 (direct_move_p): New function to return true if the operation can
10958 be done as a direct move instruciton.
10959 (quad_load_store_p): New function to return true if the operation
10960 is a quad memory operation.
10961 (rs6000_legitimize_address): If quad memory, only allow register
10962 indirect for TImode addresses.
10963 (rs6000_legitimate_address_p): Likewise.
10964 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
10965 (rs6000_reload_register_type): Likewise.
10966 (register_to_reg_type): Return register type.
10967 (rs6000_secondary_reload_simple_move): New helper function for
10968 secondary reload and secondary memory needed to identify anything
10969 that is a simple move, and does not need reloading.
10970 (rs6000_secondary_reload_direct_move): New helper function for
10971 secondary reload to identify cases that can be done with several
10972 instructions via the direct move instructions.
10973 (rs6000_secondary_reload_move): New helper function for secondary
10974 reload to identify moves between register types that can be done.
10975 (rs6000_secondary_reload): Add support for quad memory operations
10976 and for direct move.
10977 (rs6000_secondary_memory_needed): Likewise.
10978 (rs6000_debug_secondary_memory_needed): Change argument names.
10979 (rs6000_output_move_128bit): New function to return the move to
10980 use for 128-bit moves, including knowing about the various
10981 limitations of quad memory operations.
10983 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
10984 memory operations. call rs6000_output_move_128bit for the actual
10985 instruciton(s) to generate.
10986 (vsx_movti_64bit): Likewise.
10988 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
10989 (UNSPEC_P8V_MTVSRWZ): Likewise.
10990 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
10991 (UNSPEC_P8V_MTVSRD): Likewise.
10992 (UNSPEC_P8V_XXPERMDI): Likewise.
10993 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
10994 (UNSPEC_FUSION_GPR): Likewise.
10995 (FMOVE128_GPR): New iterator for direct move.
10996 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
10997 (f32_sv): Likewise.
10998 (f32_dm): Likewise.
10999 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
11000 loads and direct move instructions.
11001 (zero_extendsidi2_lfiwzx): Likewise.
11002 (extendsidi2_lfiwax): Likewise.
11003 (extendsidi2_nocell): Likewise.
11004 (floatsi<mode>2_lfiwax): Likewise.
11005 (lfiwax): Likewise.
11006 (floatunssi<mode>2_lfiwzx): Likewise.
11007 (lfiwzx): Likewise.
11008 (fix_trunc<mode>_stfiwx): Likewise.
11009 (fixuns_trunc<mode>_stfiwx): Likewise.
11010 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
11011 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
11012 (parity<mode>2_cmpb): Set length/type attr.
11013 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
11014 for 'mr.' to fast_compare.
11015 (bpermd_<mode>): Change type attr to popcnt.
11016 (p8_fmrgow_<mode>): New insns for power8 direct move support.
11017 (p8_mtvsrwz_1): Likewise.
11018 (p8_mtvsrwz_2): Likewise.
11019 (reload_fpr_from_gpr<mode>): Likewise.
11020 (p8_mtvsrd_1): Likewise.
11021 (p8_mtvsrd_2): Likewise.
11022 (p8_xxpermdi_<mode>): Likewise.
11023 (reload_vsx_from_gpr<mode>): Likewise.
11024 (reload_vsx_from_gprsf): Likewise.
11025 (p8_mfvsrd_3_<mode>): LIkewise.
11026 (reload_gpr_from_vsx<mode>): Likewise.
11027 (reload_gpr_from_vsxsf): Likewise.
11028 (p8_mfvsrd_4_disf): Likewise.
11029 (multi-word GPR splits): Do not split direct moves or quad memory
11032 2013-06-10 David Malcolm <dmalcolm@redhat.com>
11034 * tree-into-ssa.c (interesting_blocks): Make static.
11036 2013-06-10 Jakub Jelinek <jakub@redhat.com>
11039 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
11040 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
11041 Use DATA_ABI_ALIGNMENT for that case instead if defined.
11042 (get_variable_align): New function.
11043 (get_variable_section, emit_bss, emit_common,
11044 assemble_variable_contents, place_block_symbol): Use
11045 get_variable_align instead of DECL_ALIGN.
11046 (assemble_noswitch_variable): Add align argument, use it
11047 instead of DECL_ALIGN.
11048 (assemble_variable): Adjust caller. Use get_variable_align
11049 instead of DECL_ALIGN.
11050 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
11052 (DATA_ABI_ALIGNMENT): Define.
11053 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
11054 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
11055 opt is false, only return the psABI mandated alignment increase.
11056 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
11057 (DATA_ABI_ALIGNMENT): ... this.
11058 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
11059 (DATA_ABI_ALIGNMENT): ... this.
11060 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
11061 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
11062 (DATA_ABI_ALIGNMENT): ... this.
11063 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
11064 * doc/tm.texi: Regenerated.
11066 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
11068 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
11069 cmp_code to construct REG_EQUAL note.
11071 2013-06-09 Jakub Jelinek <jakub@redhat.com>
11074 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
11075 that operands[2] doesn't overlap with operands[0].
11077 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
11078 Jan Hubicka <jh@suse.cz>
11080 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
11081 hack to mark symbols as used.
11083 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
11085 PR rtl-optimization/57559
11086 * lra-constraints.c (process_alt_operands): Don't discourage
11087 memory with known offset for offsetable memory constraint.
11088 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
11090 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
11092 * varasm.c (struct oc_local_state): Reorder fields.
11093 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
11094 and adjust accordingly.
11095 (output_constructor): Reorder initialization code and adjust call to
11096 output_constructor_bitfield.
11098 2013-06-07 Jan Hubicka <jh@suse.cz>
11100 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
11102 2013-06-07 David Malcolm <dmalcolm@redhat.com>
11104 * tree-object-size.c (unknown): Make const.
11106 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11108 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
11109 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
11110 for last alternative in the cpu_facility attribute.
11112 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11115 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
11116 (xordi3): Change operand 2 constraint to arm_xordi_operand.
11117 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
11118 * config/arm/constraints.md (Dg): New constraint.
11119 * config/arm/neon.md (xordi3_neon): Remove.
11120 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
11121 * config/arm/predicates.md (arm_xordi_operand): New predicate.
11123 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11125 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
11126 Clean up alternatives.
11128 2013-06-07 Alan Modra <amodra@gmail.com>
11130 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
11133 2013-06-07 Alan Modra <amodra@gmail.com>
11135 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
11137 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11139 * config/arm/constraints.md (Df): New constraint.
11140 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
11141 Correct length attribute for last two alternatives.
11143 2013-06-07 Alan Modra <amodra@gmail.com>
11145 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
11146 override user -mfp-in-toc.
11147 (offsettable_ok_by_alignment): Consider just the current access
11148 rather than the whole object, unless BLKmode. Handle
11149 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
11150 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
11151 for -mcmodel=medium.
11152 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
11153 override user -mfp-in-toc or -msum-in-toc. Default to
11154 -mno-fp-in-toc for -mcmodel=medium.
11156 2013-06-06 DJ Delorie <dj@redhat.com>
11158 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
11159 TARGET_VALID_POINTER_MODE.
11161 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
11162 Pat Haugen <pthaugen@us.ibm.com>
11163 Peter Bergner <bergner@vnet.ibm.com>
11165 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
11166 Document new power8 builtins.
11168 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
11169 condition code register, to allow 128-bit logical operations to be
11170 done in the VSX or GPR registers.
11171 (nor<mode>3): Use the canonical form for nor.
11172 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
11173 vclz*, and vpopcnt* vector instructions.
11174 (nand<mode>3): Likewise.
11175 (orc<mode>3): Likewise.
11176 (clz<mode>2): LIkewise.
11177 (popcount<mode>2): Likewise.
11179 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
11180 that only the GPRs are recognized.
11182 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
11183 support for new power8 builtins.
11185 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
11187 (xscvdpspn): Likewise.
11193 (vpopcnt): Likewise.
11194 (vpopcntb): Likewise.
11195 (vpopcnth): Likewise.
11196 (vpopcntw): Likewise.
11197 (vpopcntd): Likewise.
11199 (vmrgew): Likewise.
11200 (vmrgow): Likewise.
11202 (eqv_v16qi3): Likewise.
11203 (eqv_v8hi3): Likewise.
11204 (eqv_v4si3): Likewise.
11205 (eqv_v2di3): Likewise.
11206 (eqv_v4sf3): Likewise.
11207 (eqv_v2df3): Likewise.
11209 (nand_v16qi3): Likewise.
11210 (nand_v8hi3): Likewise.
11211 (nand_v4si3): Likewise.
11212 (nand_v2di3): Likewise.
11213 (nand_v4sf3): Likewise.
11214 (nand_v2df3): Likewise.
11216 (orc_v16qi3): Likewise.
11217 (orc_v8hi3): Likewise.
11218 (orc_v4si3): Likewise.
11219 (orc_v2di3): Likewise.
11220 (orc_v4sf3): Likewise.
11221 (orc_v2df3): Likewise.
11223 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
11224 allow power8 quad mode in 64-bit.
11225 (rs6000_builtin_vectorized_function): Add support to vectorize
11226 ISA 2.07 count leading zeros, population count builtins.
11227 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
11228 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
11229 (builtin_function_type): Add vgbbd builtin function which takes an
11231 (altivec_expand_vec_perm_const): Add support for new power8 merge
11234 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
11235 that does not include TImdoe for use with 32-bit.
11236 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
11238 (UNSPEC_VSX_CVDPSPN): Likewise.
11239 (vsx_xscvdpspn): Likewise.
11240 (vsx_xscvspdpn): Likewise.
11241 (vsx_xscvdpspn_scalar): Likewise.
11242 (vsx_xscvspdpn_directmove): Likewise.
11243 (vsx_and<mode>3): Split logical operations into 32-bit and
11244 64-bit. Add support to do logical operations on TImode as well as
11245 VSX vector types. Allow logical operations to be done in either
11246 VSX registers or in general purpose registers in 64-bit mode. Add
11247 splitters if GPRs were used. For AND, add clobber of CCmode to
11248 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
11250 (vsx_and<mode>3_32bit): Likewise.
11251 (vsx_and<mode>3_64bit): Likewise.
11252 (vsx_ior<mode>3): Likewise.
11253 (vsx_ior<mode>3_32bit): Likewise.
11254 (vsx_ior<mode>3_64bit): Likewise.
11255 (vsx_xor<mode>3): Likewise.
11256 (vsx_xor<mode>3_32bit): Likewise.
11257 (vsx_xor<mode>3_64bit): Likewise.
11258 (vsx_one_cmpl<mode>2): Likewise.
11259 (vsx_one_cmpl<mode>2_32bit): Likewise.
11260 (vsx_one_cmpl<mode>2_64bit): Likewise.
11261 (vsx_nor<mode>3): Likewise.
11262 (vsx_nor<mode>3_32bit): Likewise.
11263 (vsx_nor<mode>3_64bit): Likewise.
11264 (vsx_andc<mode>3): Likewise.
11265 (vsx_andc<mode>3_32bit): Likewise.
11266 (vsx_andc<mode>3_64bit): Likewise.
11267 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
11268 and xxlorc instructions.
11269 (vsx_eqv<mode>3_64bit): Likewise.
11270 (vsx_nand<mode>3_32bit): Likewise.
11271 (vsx_nand<mode>3_64bit): Likewise.
11272 (vsx_orc<mode>3_32bit): Likewise.
11273 (vsx_orc<mode>3_64bit): Likewise.
11275 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
11277 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
11279 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
11280 (p8_vmrgow): Likewise.
11281 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
11282 GPRs to be split under VSX.
11283 (p8v_clz<mode>2): Add power8 count leading zero support.
11284 (p8v_popcount<mode>2): Add power8 population count support.
11285 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
11288 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
11291 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
11293 (vec_nand): Likewise.
11294 (vec_vclz): Likewise.
11295 (vec_vclzb): Likewise.
11296 (vec_vclzd): Likewise.
11297 (vec_vclzh): Likewise.
11298 (vec_vclzw): Likewise.
11299 (vec_vgbbd): Likewise.
11300 (vec_vmrgew): Likewise.
11301 (vec_vmrgow): Likewise.
11302 (vec_vpopcnt): Likewise.
11303 (vec_vpopcntb): Likewise.
11304 (vec_vpopcntd): Likewise.
11305 (vec_vpopcnth): Likewise.
11306 (vec_vpopcntw): Likewise.
11308 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
11310 PR rtl-optimization/57468
11311 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
11314 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
11316 PR rtl-optimization/57459
11317 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
11318 type when setting live regs.
11320 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
11322 * config/s390/s390.opt (mlra): New option.
11323 * config/s390/s390.c (s390_decompose_address): Check displacement
11324 for all registers for LRA.
11325 (s390_secondary_reload): Don't used secondary reloads for LRA.
11326 (s390_lra_p): New function.
11327 (TARGET_LRA_P): Define.
11328 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
11329 of attribute cpu_facility to zarch for the last alternative.
11330 (*cmpmem_short): Ditto.
11332 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
11334 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
11335 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
11336 (arm_expand_prologue): Likewise.
11338 2013-06-06 Teresa Johnson <tejohnson@google.com>
11341 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
11342 as this is now done by redirect_edge_and_branch_force.
11343 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
11344 barriers, and fix interaction with splitting.
11345 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
11346 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
11347 reflect changes made in the routine.
11348 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
11349 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
11350 since this is called in cfglayout mode, and replace partition fixup
11351 with assert as that is now done by force_nonfallthru_and_redirect.
11352 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
11353 already be marked with region crossing note.
11354 (insert_section_boundary_note): Make non-static, gate on flag
11355 has_bb_partition, rewrite to also check for multiple partitions.
11356 (rest_of_handle_reorder_blocks): Remove call to
11357 insert_section_boundary_note, now done later during free_cfg.
11358 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
11359 * bb-reorder.h (insert_section_boundary_note): Declare.
11360 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
11361 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
11362 invoke insert_section_boundary_note.
11363 (try_redirect_by_replacing_jump): Remove unnecessary
11364 check for region crossing note.
11365 (fixup_partition_crossing): New function.
11366 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
11367 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
11368 in non-cfglayout mode.
11369 (force_nonfallthru_and_redirect): Fixup partition boundaries,
11370 remove old code that tried to do this. Emit barrier correctly
11371 when we are in cfglayout mode.
11372 (last_bb_in_partition): New function.
11373 (rtl_split_edge): Correctly fixup partition boundaries.
11374 (commit_one_edge_insertion): Remove old code that tried to
11375 fixup region crossing edge since this is now handled in
11376 split_block, and set up insertion point correctly since
11377 block may now end in a jump.
11378 (verify_hot_cold_block_grouping): Guard against checking when not in
11379 linearized RTL mode.
11380 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
11382 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
11383 rtl_verify_flow_info, so not called in cfglayout mode.
11384 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
11385 (fixup_reorder_chain): Remove old code that attempted to fixup region
11386 crossing note as this is now handled in force_nonfallthru_and_redirect.
11387 (duplicate_insn_chain): Don't duplicate switch section notes.
11388 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
11390 * basic-block.h (emit_barrier_after_bb): Declare.
11392 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11394 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
11395 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
11396 arm_usatsihi): Adjust alternatives for arm_restrict_it.
11398 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11400 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
11402 * config/arm/ldmstm.md: Regenerate.
11404 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11406 * config/arm/sync.md (atomic_loaddi_1):
11407 Disable predication for arm_restrict_it.
11408 (arm_load_exclusive<mode>): Likewise.
11409 (arm_load_exclusivesi): Likewise.
11410 (arm_load_exclusivedi): Likewise.
11411 (arm_load_acquire_exclusive<mode>): Likewise.
11412 (arm_load_acquire_exclusivesi): Likewise.
11413 (arm_load_acquire_exclusivedi): Likewise.
11414 (arm_store_exclusive<mode>): Likewise.
11415 (arm_store_exclusive<mode>): Likewise.
11416 (arm_store_release_exclusivedi): Likewise.
11417 (arm_store_release_exclusive<mode>): Likewise.
11419 2013-06-06 Richard Biener <rguenther@suse.de>
11421 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
11423 (lto_tag_is_tree_code_p): Adjust.
11424 (lto_tag_is_gimple_code_p): Likewise.
11425 (lto_gimple_code_to_tag): Likewise.
11426 (lto_tag_to_gimple_code): Likewise.
11427 (lto_tree_code_to_tag): Likewise.
11428 (lto_tag_to_tree_code): Likewise.
11429 * data-streamer.h (streamer_write_hwi_in_range): Use
11430 uhwi streaming to stream the normalized range.
11431 (streamer_read_hwi_in_range): Likewise.
11433 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11435 * config/arm/arm.md (enabled_for_depr_it): New attribute.
11436 (predicable_short_it): Likewise.
11437 (predicated): Likewise.
11438 (enabled): Handle above.
11439 (define_cond_exec): Set predicated attribute to yes.
11441 2013-06-05 Mike Stump <mikestump@comcast.net>
11443 * gdbinit.in (__FUNCTION__): Add.
11445 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
11447 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
11448 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
11450 2013-06-05 Jan Hubicka <jh@suse.cz>
11452 * varasm.c (mark_decl_referenced): Revert the removal until targets
11455 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
11457 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
11458 instead of mark_decl_referenced.
11460 2013-06-05 Jan Hubicka <jh@suse.cz>
11462 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
11463 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
11464 and symtab_used_from_object_file_p.
11465 (cgraph_make_node_local_1): Clear forced_by_abi.
11466 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
11467 * cgraph.h (symtab_node_base): Add forced_by_abi.
11468 (decide_is_variable_needed): Remove.
11469 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
11470 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
11471 (decide_is_symbol_needed): ... this one; handle symbols in general;
11472 always analyze virtuals; honnor forced_by_abi.
11473 (cgraph_finalize_function): Update.
11474 (varpool_finalize_decl): Update.
11475 (symbol_defined_and_needed): Remove.
11476 (analyze_functions): Update.
11477 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
11478 output_refs, input_overwrite_node): Handle forced_by_abi.
11479 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
11480 (address_taken_from_non_vtable_p): ... this one.
11481 (comdat_can_be_unshared_p_1): New function.
11482 (cgraph_comdat_can_be_unshared_p): Rename to ...
11483 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
11484 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
11485 (function_and_variable_visibility): Clear forced_by_abi as needed.
11486 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
11487 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
11488 * symtab.c (dump_symtab_base): Dump forced_by_abi.
11489 * varpool.c (decide_is_variable_needed): Remove.
11491 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11493 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
11494 (arm_option_override): Override arm_restrict_it where appropriate.
11495 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
11496 * config/arm/arm.opt (mrestrict-it): New command-line option.
11497 * doc/invoke.texi: Document -mrestrict-it.
11499 2013-06-05 David Malcolm <dmalcolm@redhat.com>
11501 * tsan.c (tsan_atomic_table): Make const.
11503 2013-06-05 Richard Biener <rguenther@suse.de>
11505 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
11506 index associated with the tree we are supposed to replace.
11507 * tree-streamer-out.c (pack_ts_base_value_fields): Output
11508 TREE_ASM_WRITTEN as zero for everything but SSA names.
11510 2013-06-05 David Malcolm <dmalcolm@redhat.com>
11512 * tree-ssa-structalias.c (call_stmt_vars): Make static.
11514 2013-06-04 Jan Hubicka <jh@suse.cz>
11516 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
11517 (input_node, input_varpool_node): Handle correctly external same
11519 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
11520 nodes at ltrans stage.
11522 2013-06-04 Jan Hubicka <jh@suse.cz>
11524 * ipa-inline.c (update_caller_keys): Fix availability test.
11525 (update_callee_keys): Likewise.
11526 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
11527 to follow ELF standard.
11529 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
11531 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
11532 (mips64r5900el-*-elf*): New configurations.
11533 * config/mips/mips-cpus.def (r5900): New processor.
11534 * config/mips/mips-tables.opt: Regenerate.
11535 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
11536 (mips_issue_rate): Handle PROCESSOR_R5900.
11537 (mips_reorg_process_insns): Force reorder mode for the R5900.
11538 * config/mips/mips.h (TARGET_MIPS5900): Define.
11539 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
11541 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
11543 * config/mips/mips.md (processor): Add r5900.
11544 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
11546 2013-06-04 Ian Bolton <ian.bolton@arm.com>
11548 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
11549 into function to generate MOVI instruction.
11550 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
11551 (aarch64_preferred_simd_mode): Turn into wrapper.
11552 (aarch64_output_scalar_simd_mov_immediate): New function.
11553 * config/aarch64/aarch64-protos.h: Add prototype for above.
11555 2013-06-04 Ian Bolton <ian.bolton@arm.com>
11557 * config/aarch64/aarch64.c (simd_immediate_info): Remove
11558 element_char member.
11559 (sizetochar): Return signed char.
11560 (aarch64_simd_valid_immediate): Remove elchar and other
11561 unnecessary variables.
11562 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
11563 Calculate element_char as required.
11564 * config/aarch64/aarch64-protos.h: Update and move prototype
11565 for aarch64_output_simd_mov_immediate.
11566 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
11569 2013-06-04 Ian Bolton <ian.bolton@arm.com>
11571 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
11572 information completed by aarch64_simd_valid_immediate.
11573 (aarch64_legitimate_constant_p): Update arguments.
11574 (aarch64_simd_valid_immediate): Work with struct rather than many
11576 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
11577 (aarch64_simd_make_constant): Update arguments.
11578 (aarch64_output_simd_mov_immediate): Work with struct rather than
11579 many pointers. Output immediate directly rather than as operand.
11580 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
11582 * config/aarch64/constraints.md (Dn): Update arguments.
11584 2013-06-04 Ian Bolton <ian.bolton@arm.com>
11586 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
11588 (aarch64_simd_immediate_valid_for_move): Remove.
11589 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
11590 (aarch64_simd_make_constant): Update call.
11591 (aarch64_output_simd_mov_immediate): Update call.
11592 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
11594 * config/aarch64/constraints.md (Dn): Update call.
11596 2013-06-04 Ian Bolton <ian.bolton@arm.com>
11598 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
11599 return type to bool for prototype.
11600 (aarch64_legitimate_constant_p): Check for true instead of not -1.
11601 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
11602 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
11604 2013-06-04 Catherine Moore <clm@codesourcery.com>
11606 * config/mips/mips.opt (meva): New.
11607 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
11608 (ASM_SPEC): Handle -meva.
11609 * doc/invoke.texi (meva): Document.
11611 2013-06-04 Alan Modra <amodra@gmail.com>
11613 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
11616 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11618 * rtl.def: Add extra fourth optional field to define_cond_exec.
11619 * gensupport.c (process_one_cond_exec): Process attributes from
11621 * doc/md.texi: Document fourth field in define_cond_exec.
11623 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
11625 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
11626 out the processing order as in store_bit_field_1.
11628 2013-06-04 Jan Hubicka <jh@suse.cz>
11630 PR middle-end/57500
11631 * cgraphunit.c (cgraph_process_same_body_aliases): Create
11632 non-VAR_DECL node if it does not exist yet.
11634 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
11636 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
11637 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
11638 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
11639 target_cpu_default setting.
11641 2013-06-03 Teresa Johnson <tejohnson@google.com>
11643 * dumpfile.c (opt_info_switch_p): Change -fopt-info
11644 default to -fopt-info=optimized instead of all.
11645 * doc/invoke.texi: Ditto.
11646 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
11647 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
11648 (execute_vect_slp): Emit BB vectorization success under
11649 MSG_OPTIMIZED_LOCATIONS.
11650 * tree-vect-slp.c (vect_slp_transform_bb): Change
11651 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
11652 * tree-vect-loop.c (vect_transform_loop): Ditto.
11654 2013-06-03 Jason Merrill <jason@redhat.com>
11657 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
11658 Use TARGET_EXPR for C++.
11660 2013-06-03 Jakub Jelinek <jakub@redhat.com>
11662 PR rtl-optimization/57268
11663 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
11664 if DEBUG_INSN_P (insn).
11667 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
11669 PR rtl-optimization/57268
11670 * sched-deps.c (sched_analyze_2): Flush dependence lists if
11671 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
11673 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
11675 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
11676 (ix86_avoid_lea_for_addr): Likewise.
11677 (exact_dependency_1): Likewise.
11678 (ix86_adjust_cost): Likewise.
11679 (swap_top_of_ready_list): Fix formatting and !reload_completed check
11681 (do_reorder_for_imul): Fix typo, formatting and
11682 !reload_completed check removed.
11683 (ix86_sched_reorder): Fix typo and formatting.
11684 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
11687 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
11689 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
11691 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
11693 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
11694 <STRING_CST>: Likewise.
11695 <VECTOR_CST>: Likewise.
11697 2013-06-01 Janus Weil <janus@gcc.gnu.org>
11698 Mikael Morin <mikael@gcc.gnu.org>
11700 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
11701 * config.in: Regenerated.
11702 * configure: Regenerated.
11704 2013-06-01 Jan Hubicka <jh@suse.cz>
11706 PR middle-end/57366
11707 * cgraphunit.c (compile): When weakref is not supported,
11708 set up transparent aliases before final output pass.
11709 * varasm.c (assemble_alias): Do not try to do it here.
11711 2013-06-01 Jan Hubicka <jh@suse.cz>
11713 PR middle-end/57467
11714 * passes.c (for_per_function): Skip unanalyzed functions.
11716 2013-06-01 Jan Hubicka <jh@suse.cz>
11718 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
11719 (lto_symtab_merge_symbols_1): ... this one.
11720 (lto_symtab_merge_cgraph_nodes): Rename to ...
11721 (lto_symtab_merge_symbols): ... this one; simplify.
11722 * cgraph.c (same_body_aliases_done): Rename to ...
11723 (cpp_implicit_aliases_done): ... this one.
11724 (cgraph_create_function_alias): Update.
11725 (cgraph_same_body_alias): Update.
11726 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
11727 (verify_edge_corresponds_to_fndecl): Simplify.
11728 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
11729 (cgraph_node): Remove same_body_alias.
11730 (varpool_node): Remove alias_of and extra_name_alias.
11731 (same_body_aliases_done): Rename to ..
11732 (cpp_implicit_aliases_done): ... this one.
11733 (symtab_alias_ultimate_target): Add default parameter.
11734 (symtab_resolve_alias): New function.
11735 (fixup_same_cpp_alias_visibility): Declare.
11736 (cgraph_function_node): Add default parameter.
11737 (cgraph_node_asm_name): Likewise.
11738 (cgraph_function_or_thunk_node): Add default parameter; do
11739 not ICE when it is NULL.
11740 (varpool_variable_node): Likewise.
11741 * tree-emutls.c (create_emultls_var): Update.
11742 (ipa_lower_emutls): Update.
11743 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
11744 (cgraph_reset_node): Reset alias info.
11745 (cgraph_finalize_function): Update.
11746 (fixup_same_cpp_alias_visibility): Move to symtab.c.
11747 (analyze_function): Simplify.
11748 (cgraph_process_same_body_aliases): Simplify.
11749 (analyze_functions): Fixup same body aliases.
11750 (handle_alias_pairs): Simplify.
11751 (assemble_thunk): Update.
11752 (assemble_thunks_and_aliases): Update.
11753 (output_weakrefs): Rewrite.
11754 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
11755 (lto_output_varpool_node): Likewise.
11756 (compute_ltrans_boundary): Remve assert.
11757 (get_alias_symbol): New functoin.
11758 (input_node): Rewrite alias handling.
11759 (input_varpool_node): Likewise.
11760 * ipa-pure-const.c (propagate_pure_const): Fix formating.
11761 * ipa.c (process_references): Handle weakrefs correctly.
11762 (symtab_remove_unreachable_nodes): Likewise.
11763 * trans-mem.c (get_cg_data): Update.
11764 (ipa_tm_create_version_alias): Update.
11765 (ipa_tm_execute): Update.
11766 * symtab.c (dump_symtab_base): Dump aliases.
11767 (verify_symtab_base): Verify aliases.
11768 (symtab_node_availability): New function.
11769 (symtab_alias_ultimate_target): Simplify.
11770 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
11771 handle all the fixup cases.
11772 (symtab_resolve_alias): New function.
11773 * passes.c (ipa_write_summaries): Handle weakrefs.
11774 * varpool.c (varpool_analyze_node): Simplify.
11775 (assemble_aliases): Update.
11776 (varpool_create_variable_alias): Simplify.
11777 (varpool_extra_name_alias): Simplify.
11778 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
11779 (lto_symtab_merge_symbols): ... this one.
11781 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
11784 PR rtl-optimization/57268
11785 * sched-deps.c (sched_analyze_2): Flush dependence lists if
11786 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
11788 2013-06-01 Tobias Burnus <burnus@net-b.de>
11790 Partially reverted:
11791 2013-05-31 Tobias Burnus <burnus@net-b.de>
11793 PR middle-end/57073
11794 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
11797 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
11799 PR rtl-optimization/57268
11800 * sched-deps.c (sched_analyze_2): Flush dependence lists if
11801 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
11803 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
11805 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
11806 unordered comparison operators when -fno-trapping-math is in effect
11808 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
11809 and implement unordered comparison operators properly on the e500.
11811 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
11813 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
11814 for constant scalar integers.
11815 (simplify_relational_operation_1): Likewise.
11817 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
11819 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
11820 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
11823 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
11824 Igor Zamyatin <igor.zamyatin@intel.com>
11826 Silvermont (SLM) architecture performance tuning.
11827 * config/i386/i386.h (enum ix86_tune_indices): Add
11828 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
11829 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
11831 * config/i386/i386.c (initial_ix86_tune_features)
11832 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
11833 (ix86_lea_outperforms): Handle Silvermont tuning.
11834 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
11836 (ix86_use_lea_for_mov): Likewise.
11837 (ix86_avoid_lea_for_addr): Likewise.
11838 (ix86_lea_for_add_ok): Likewise.
11839 (exact_dependency_1): New function.
11840 (exact_store_load_dependency): Likewise.
11841 (ix86_adjust_cost): Handle Silvermont tuning.
11842 (do_reoder_for_imul): Likewise.
11843 (swap_top_of_ready_list): New function.
11844 (ix86_sched_reorder): Changed to handle Silvermont tuning.
11846 * config/i386/i386.md (peepholes that split memory operand in fp
11849 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
11851 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
11852 Remove un-necessary braces.
11854 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
11856 * config/aarch64/aarch64.c (aarch64_classify_symbol):
11857 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
11859 2013-05-31 Tobias Burnus <burnus@net-b.de>
11861 PR middle-end/57073
11862 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
11864 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11867 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
11868 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
11869 * config/arm/neon.md (iordi3_neon): Remove.
11870 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
11871 * config/arm/predicates.md (imm_for_neon_logic_operand):
11872 Move to earlier in the file.
11873 (neon_logic_op2): Likewise.
11874 (arm_iordi_operand_neon): New predicate.
11876 2013-05-31 Richard Biener <rguenther@suse.de>
11878 PR tree-optimization/57478
11879 PR tree-optimization/57453
11880 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
11883 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
11885 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
11886 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
11888 2013-05-30 Tobias Burnus <burnus@net-b.de>
11889 Thomas Koenig <tkoenig@gcc.gnu.org>
11891 PR middle-end/57073
11892 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
11893 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
11895 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
11897 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
11899 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
11901 * target.def (register_usage_leveling_p): New hook.
11902 * targhooks.c (default_register_usage_leveling_p): New.
11903 * targhooks.h (default_register_usage_leveling_p): New prototype.
11904 * lra-assigns.c (register_usage_leveling_p): Use the hook.
11905 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
11906 * doc/tm.texi: Update.
11907 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
11909 2013-05-30 Ian Bolton <ian.bolton@arm.com>
11911 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
11912 (*insv_reg<mode>): New define_insn.
11914 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
11916 PR rtl-optimization/57439
11917 * postreload.c (move2add_valid_value_p): Check that we have
11918 a zero subreg_regno_offset when accessing the register in
11919 the requested mode.
11921 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
11922 Igor Zamyatin <igor.zamyatin@intel.com>
11924 Silvermont (SLM) architecture pipeline model, tuning and
11926 * config.gcc: Add slm config options and target.
11928 * config/i386/slm.md: New.
11930 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
11932 * config/i386/i386-c.c (ix86_target_macros_internal): New case
11934 (ix86_target_macros_internal): Likewise.
11936 * config/i386/i386.c (slm_cost): New cost.
11937 (m_SLM): New macro flag.
11938 (initial_ix86_tune_features): Set m_SLM.
11939 (x86_accumulate_outgoing_args): Likewise.
11940 (x86_arch_always_fancy_math_387): Likewise.
11941 (processor_target_table): Add slm cost.
11942 (cpu_names): Add slm cpu name.
11943 (x86_option_override_internal): Set SLM ISA.
11944 (ix86_issue_rate): New case PROCESSOR_SLM.
11945 (ia32_multipass_dfa_lookahead): Likewise.
11946 (fold_builtin_cpu): Add slm.
11948 * config/i386/i386.h (TARGET_SLM): New target macro.
11949 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
11950 (processor_type): Add PROCESSOR_SLM.
11952 * config/i386/i386.md (cpu): Add new value "slm".
11953 (slm.md): Include slm.md.
11955 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
11956 Zhenqiang Chen <zhenqiang.chen@linaro.org>
11958 * config/arm/arm-protos.h: Add and update function protos.
11959 * config/arm/arm.c (use_simple_return_p): New added.
11960 (thumb2_expand_return): Check simple_return flag.
11961 * config/arm/arm.md: Add simple_return and conditional simple_return.
11962 * config/arm/iterators.md: Add iterator for return and simple_return.
11964 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
11966 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
11967 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
11968 (arm_emit_vfp_multi_reg_pop): Likewise.
11969 (thumb2_emit_ldrd_pop): Likewise.
11970 (arm_expand_epilogue): Add misc REG_CFA notes.
11971 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
11973 2013-05-29 Lawrence Crowl <crowl@google.com>
11975 * config/arm/t-arm: Update for below.
11977 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
11978 Change type to hash_table. Update dependent calls and types.
11980 * config/i386/t-cygming: Update for below.
11982 * config/i386/t-interix: Update for below.
11984 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
11985 Change type to hash_table. Update dependent calls and types.
11986 (i386_find_on_wrapper_list::wrappers): Likewise.
11988 * config/ia64/t-ia64: Update for below.
11990 * config/ia64/ia64.c (bundle_state_table):
11991 Change type to hash_table. Update dependent calls and types.
11993 * config/mips/mips.c (mips_reorg_process_insns::htab):
11994 Change type to hash_table. Update dependent calls and types.
11996 * config/sol2.c (solaris_comdat_htab):
11997 Change type to hash_table. Update dependent calls and types.
11999 * config/t-sol2: Update for above.
12001 2013-05-29 Teresa Johnson <tejohnson@google.com>
12003 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
12004 functions are not yet marked as defined.
12006 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
12007 Pat Haugen <pthaugen@us.ibm.com>
12008 Peter Bergner <bergner@vnet.ibm.com>
12010 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
12014 (vrotl<mode>3): Likewise.
12015 (vashl<mode>3): Likewise.
12016 (vlshr<mode>3): Likewise.
12017 (vashr<mode>3): Likewise.
12019 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
12020 support for power8 V2DI builtins.
12022 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
12023 power8 V2DI builtins.
12024 (vupkhsw): Likewise.
12025 (vupklsw): Likewise.
12026 (vaddudm): Likewise.
12027 (vminsd): Likewise.
12028 (vmaxsd): Likewise.
12029 (vminud): Likewise.
12030 (vmaxud): Likewise.
12031 (vpkudum): Likewise.
12032 (vpksdss): Likewise.
12033 (vpkudus): Likewise.
12034 (vpksdus): Likewise.
12039 (vsubudm): Likewise.
12040 (vcmpequd): Likewise.
12041 (vcmpgtsd): Likewise.
12042 (vcmpgtud): Likewise.
12043 (vcmpequd_p): Likewise.
12044 (vcmpgtsd_p): Likewise.
12045 (vcmpgtud_p): Likewise.
12046 (vupkhsw): Likewise.
12047 (vupklsw): Likewise.
12048 (vaddudm): Likewise.
12049 (vmaxsd): Likewise.
12050 (vmaxud): Likewise.
12051 (vminsd): Likewise.
12052 (vminud): Likewise.
12053 (vpksdss): Likewise.
12054 (vpksdus): Likewise.
12055 (vpkudum): Likewise.
12056 (vpkudus): Likewise.
12061 (vsubudm): Likewise.
12063 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
12064 support for power8 V2DI instructions.
12066 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
12067 power8 V2DI instructions. Combine pack and unpack insns to use an
12068 iterator for each mode. Check whether a particular mode supports
12069 Altivec instructions instead of just checking TARGET_ALTIVEC.
12070 (UNSPEC_VPKUWUM): Likewise.
12071 (UNSPEC_VPKSHSS): Likewise.
12072 (UNSPEC_VPKSWSS): Likewise.
12073 (UNSPEC_VPKUHUS): Likewise.
12074 (UNSPEC_VPKSHUS): Likewise.
12075 (UNSPEC_VPKUWUS): Likewise.
12076 (UNSPEC_VPKSWUS): Likewise.
12077 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
12078 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
12079 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
12080 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
12081 (UNSPEC_VUPKHSB): Likewise.
12082 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
12083 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
12084 (UNSPEC_VUPKHSH): Likewise.
12085 (UNSPEC_VUPKLSB): Likewise.
12086 (UNSPEC_VUPKLSH): Likewise.
12088 (VI_char): Likewise.
12089 (VI_scalar): Likewise.
12090 (VI_unit): Likewise.
12092 (VP_small): Likewise.
12093 (VP_small_lc): Likewise.
12094 (VU_char): Likewise.
12095 (add<mode>3): Likewise.
12096 (altivec_vaddcuw): Likewise.
12097 (altivec_vaddu<VI_char>s): Likewise.
12098 (altivec_vadds<VI_char>s): Likewise.
12099 (sub<mode>3): Likewise.
12100 (altivec_vsubcuw): Likewise.
12101 (altivec_vsubu<VI_char>s): Likewise.
12102 (altivec_vsubs<VI_char>s): Likewise.
12103 (altivec_vavgs<VI_char>): Likewise.
12104 (altivec_vcmpbfp): Likewise.
12105 (altivec_eq<mode>): Likewise.
12106 (altivec_gt<mode>): Likewise.
12107 (altivec_gtu<mode>): Likewise.
12108 (umax<mode>3): Likewise.
12109 (smax<mode>3): Likewise.
12110 (umin<mode>3): Likewise.
12111 (smin<mode>3): Likewise.
12112 (altivec_vpkuhum): Likewise.
12113 (altivec_vpkuwum): Likewise.
12114 (altivec_vpkshss): Likewise.
12115 (altivec_vpkswss): Likewise.
12116 (altivec_vpkuhus): Likewise.
12117 (altivec_vpkshus): Likewise.
12118 (altivec_vpkuwus): Likewise.
12119 (altivec_vpkswus): Likewise.
12120 (altivec_vpks<VI_char>ss): Likewise.
12121 (altivec_vpks<VI_char>us): Likewise.
12122 (altivec_vpku<VI_char>us): Likewise.
12123 (altivec_vpku<VI_char>um): Likewise.
12124 (altivec_vrl<VI_char>): Likewise.
12125 (altivec_vsl<VI_char>): Likewise.
12126 (altivec_vsr<VI_char>): Likewise.
12127 (altivec_vsra<VI_char>): Likewise.
12128 (altivec_vsldoi_<mode>): Likewise.
12129 (altivec_vupkhsb): Likewise.
12130 (altivec_vupkhs<VU_char>): Likewise.
12131 (altivec_vupkls<VU_char>): Likewise.
12132 (altivec_vupkhsh): Likewise.
12133 (altivec_vupklsb): Likewise.
12134 (altivec_vupklsh): Likewise.
12135 (altivec_vcmpequ<VI_char>_p): Likewise.
12136 (altivec_vcmpgts<VI_char>_p): Likewise.
12137 (altivec_vcmpgtu<VI_char>_p): Likewise.
12138 (abs<mode>2): Likewise.
12139 (vec_unpacks_hi_v16qi): Likewise.
12140 (vec_unpacks_hi_v8hi): Likewise.
12141 (vec_unpacks_lo_v16qi): Likewise.
12142 (vec_unpacks_hi_<VP_small_lc>): Likewise.
12143 (vec_unpacks_lo_v8hi): Likewise.
12144 (vec_unpacks_lo_<VP_small_lc>): Likewise.
12145 (vec_pack_trunc_v8h): Likewise.
12146 (vec_pack_trunc_v4si): Likewise.
12147 (vec_pack_trunc_<mode>): Likewise.
12149 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
12151 (vec_vmaxsd): Likewise.
12152 (vec_vmaxud): Likewise.
12153 (vec_vminsd): Likewise.
12154 (vec_vminud): Likewise.
12155 (vec_vpksdss): Likewise.
12156 (vec_vpksdus): Likewise.
12157 (vec_vpkudum): Likewise.
12158 (vec_vpkudus): Likewise.
12159 (vec_vrld): Likewise.
12160 (vec_vsld): Likewise.
12161 (vec_vsrad): Likewise.
12162 (vec_vsrd): Likewise.
12163 (vec_vsubudm): Likewise.
12164 (vec_vupkhsw): Likewise.
12165 (vec_vupklsw): Likewise.
12167 2013-05-29 Jan Hubicka <jh@suse.cz>
12169 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
12170 flags; reorder rest of fields in more consistent way.
12171 (varpool_node): Remove analyzed, finalized and alias.
12172 (cgraph_ndoe): Likewise.
12173 (symtab_alias_ultimate_target): New function.
12174 (cgraph_function_node): Move offline.
12175 (cgraph_reset_node): Declare.
12176 (cgraph_comdat_can_be_unshared_p): Remove.
12177 (varpool_remove_initializer): Declare.
12178 (varpool_first_defined_variable, varpool_next_defined_variable
12179 cgraph_first_defined_function, cgraph_next_defined_function): Update.
12180 (cgraph_function_with_gimple_body_p): Update.
12181 (varpool_all_refs_explicit_p): Update.
12182 (symtab_alias_target): New function.
12183 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
12184 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
12185 (cgraph_function_or_thunk_node): Simplify using
12186 symtab_alias_ultimate_target.
12187 (varpool_variable_node): Likewise.
12188 * cgraph.c (cgraph_create_function_alias): Update.
12189 (cgraph_add_thunk): Update.
12190 (cgraph_remove_node): Update.
12191 (dump_cgraph_node): Do not dump removed flags.
12192 (cgraph_function_body_availability): Update.
12193 (cgraph_propagate_frequency): Update.
12194 (verify_cgraph_node): Check sanity of local flag.
12195 (cgraph_function_node): Move here from cgraph.h; revamp for
12196 cgraph_function_or_thunk_node.
12197 * lto-symtab.c (lto_varpool_replace_node): Update.
12198 (lto_symtab_resolve_can_prevail_p): Update.
12199 (lto_symtab_merge_cgraph_nodes): Update.
12200 * ipa-cp.c (determine_versionability, initialize_node_lattices,
12201 propagate_constants_accross_call, devirtualization_time_bonus,
12202 ipcp_propagate_stage): Update.
12203 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
12204 * ipa-inline-transform.c (clone_inlined_nodes,
12205 preserve_function_body_p): Update.
12206 * ipa-reference.c (propagate): Update.
12207 (write_node_summary_p): Update.
12208 * toplev.c (wrapup_global_declaration_2): Update.
12209 * cgraphunit.c (cgraph_analyze_function): Rename to ...
12210 (analyze_function) ... this one.
12211 (cgraph_process_new_functions): Update.
12212 (cgraph_reset_node): Export.
12213 (cgraph_finalize_function): Update.
12214 (cgraph_add_new_function): Update.
12215 (process_function_and_variable_attributes): Update.
12216 (varpool_finalize_decl): Update.
12217 (symbol_finalized): Remove.
12218 (symbol_finalized_and_needed): Rename to ...
12219 (symbol_defined_and_needed): ... update.
12220 (cgraph_analyze_functions): Update.
12221 (handle_alias_pairs): Update.
12222 (mark_functions_to_output): Update.
12223 (assemble_thunk): Update.
12224 (output_in_order): Update.
12225 (output_weakrefs): Update.
12226 (finalize_compilation_unit): Update.
12227 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
12228 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
12229 input_node, input_varpool_node): Update.
12230 * dbxout.c (dbxout_expand_expr): Update.
12231 * cgraphclones.c (cgraph_clone_node): Update.
12232 (cgraph_copy_node_for_versioning): Update.
12233 (cgraph_materialize_clone): Update.
12234 (cgraph_materialize_all_clones): Update.
12235 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
12236 propagate_pure_const, propagate_nothrow): Update.
12237 * lto-streamer-out.c (lto_output, write_symbol): Update.
12238 * ipa-utils.c (ipa_reverse_postorder): Update.
12239 * ipa-inline.c (can_inline_edge_p): Update.
12240 (update_caller_keys, ipa_inline): Update.
12241 * dwarf2out.c (reference_to_unused,
12242 premark_types_used_by_global_vars_helper): Update.
12243 * tree-eh.c (tree_could_trap_p): Update.
12244 * ipa-split.c (consider_split, execute_split_functions): Update.
12245 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
12246 has_addr_references_p): Update; move ahead in file for better
12248 (process_references): Simplify.
12249 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
12250 bodies are removed.
12251 (cgraph_comdat_can_be_unshared_p): Make static.
12252 (cgraph_externally_visible_p): Update.
12253 (varpool_externally_visible_p): Update.
12254 (function_and_variable_visibility): Update.
12255 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
12256 ipa_tm_mark_force_output_node): Update.
12257 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
12258 estimate_edge_devirt_benefit, inline_generate_summary,
12259 inline_write_summary): Update.
12260 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
12261 * ipa-prop.c (ipa_compute_jump_functions): Update.
12262 (ipa_print_node_params, ipa_prop_read_section,
12263 ipa_update_after_lto_read, read_replacements_section): Update.
12264 * varasm.c (mark_decl_referenced): Update.
12265 (assemble_alias, dump_tm_clone_pairs): Update.
12266 * tree-inline.c (copy_bb): Update.
12267 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
12269 * symtab.c (dump_symtab_base): Print new flags.
12270 (verify_symtab_base): Verify new flags.
12271 (symtab_alias_ultimate_target): New function.
12272 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
12273 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
12275 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
12277 * i386.c (ix86_get_function_versions_dispatcher,
12278 ix86_generate_version_dispatcher_body): Update.
12279 (fold_builtin_cpu): Use varpool_add_new_variable.
12280 * varpool.c (varpool_remove_initializer): Break out from ...
12281 (varpool_remove_node): ... this one.
12282 (dump_varpool_node, varpool_node_for_asm,
12283 cgraph_variable_initializer_availability, varpool_analyze_node,
12284 varpool_assemble_decl, varpool_remove_unreferenced_decls,
12285 varpool_finalize_named_section_flags, varpool_create_variable_alias):
12288 2013-05-29 Jan Hubicka <jh@suse.cz>
12290 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
12292 2013-05-29 Easwaran Raman <eraman@google.com>
12294 PR tree-optimization/57442
12295 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
12296 when control exits the main loop.
12298 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
12300 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
12302 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
12303 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
12304 * rx/t-rx: Add rx100 under multi library matches option for nofpu
12307 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12309 PR tree-optimization/57441
12310 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
12311 Don't limit size of incr_vec to number of candidates.
12313 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
12315 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
12316 and mips16 directories.
12317 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
12318 (MULTILIB_DIRNAMES): Ditto.
12319 (MULTILIB_EXCEPTIONS): Add new exceptions.
12320 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
12321 (MULTILIB_DIRNAMES): Ditto.
12322 (MULTILIB_EXCEPTIONS): Add new exceptions.
12324 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
12325 Marcus Shawcroft <marcus.shawcroft@arm.com>
12327 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
12328 SYMBOL_TINY_ABSOLUTE.
12329 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
12330 SYMBOL_TINY_ABSOLUTE.
12331 (aarch64_expand_mov_immediate): Likewise.
12332 (aarch64_classify_symbol): Likewise.
12333 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
12334 Permit SYMBOL_TINY_ABSOLUTE.
12335 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
12337 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
12338 Marcus Shawcroft <marcus.shawcroft@arm.com>
12340 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
12341 Refactor if/switch. Replace gcc_assert with if.
12343 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
12345 * config/i386/i386.c (initial_ix86_tune_features): Enable
12346 FP Reassociation for AMD bdver1 and bdver2.
12348 2013-05-29 Martin Jambor <mjambor@suse.cz>
12350 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
12351 and IMAGPART_EXPR do not occur within other handled_components.
12353 2013-05-29 Richard Biener <rguenther@suse.de>
12355 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
12356 access on whether the use is in the BB we currently try to
12358 (vect_bb_vectorization_profitable_p): Pass the BB we currently
12359 vectorize to vect_bb_slp_scalar_cost.
12361 2013-05-29 Richard Biener <rguenther@suse.de>
12363 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
12364 computing scalar cost offsetted by stmts that are kept live
12366 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
12367 for computation of scalar cost.
12369 2013-05-28 Steve Ellcey <sellcey@mips.com>
12371 * config/mips/mips-cpus.def (mips32r2): Change processor type.
12373 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
12375 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
12376 array notation built-in reduction functions.
12377 * doc/passes.texi (Passes): Added documentation about changes done
12379 * doc/invoke.texi (C Dialect Options): Added documentation about
12380 the -fcilkplus flag.
12381 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
12382 (BUILTINS_DEF): Depend on cilkplus.def.
12383 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
12384 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
12385 * cilkplus.def: New file.
12387 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
12389 PR rtl-optimization/57439
12390 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
12392 2013-05-28 Easwaran Raman <eraman@google.com>
12394 PR tree-optimization/57337
12395 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
12396 (find_insert_point): Correctly identify the insertion point
12397 when two statements with the same UID is compared.
12399 2013-05-28 Richard Biener <rguenther@suse.de>
12401 PR tree-optimization/56787
12402 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
12403 from the list of data references.
12404 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
12406 (vect_analyze_loop_operations): Likewise.
12407 (vect_transform_loop): Remove clobbers.
12409 2013-05-28 Martin Jambor <mjambor@suse.cz>
12411 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
12412 and REALPART_EXPRs have scalar type.
12414 2013-05-28 Richard Biener <rguenther@suse.de>
12416 PR tree-optimization/57411
12417 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
12419 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
12420 virtual operand propagation.
12422 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
12424 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
12425 destination register for bmasksi_vis.
12426 (vector_init_bshuffle): Likewise.
12427 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
12429 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
12431 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
12432 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
12433 mode if the instruction isn't available in the original mode.
12434 * config/sparc/sparc.opt (mfix-ut699): New option.
12435 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
12436 (divdf3): Turn into expander.
12437 (divdf3_nofix): New insn.
12438 (divdf3_fix): Likewise.
12439 (divsf3): Disable if -mfix-ut699.
12440 (sqrtdf2): Turn into expander.
12441 (sqrtdf2_nofix): New insn.
12442 (sqrtdf2_fix): Likewise.
12443 (sqrtsf2): Disable if -mfix-ut699.
12445 2013-05-27 Richard Biener <rguenther@suse.de>
12447 PR middle-end/57412
12448 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
12449 block for the new loop.
12451 2013-05-27 Richard Biener <rguenther@suse.de>
12453 PR tree-optimization/57343
12454 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
12455 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
12456 (number_of_iterations_cond): Do not build the folded tree.
12458 2013-05-27 Richard Biener <rguenther@suse.de>
12461 PR middle-end/57381
12462 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
12463 OEP_CONSTANT_ADDRESS_OF retained.
12465 PR tree-optimization/57417
12466 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
12467 for unchanged base.
12468 (set_ssa_val_to): Compare addresses using
12469 get_addr_base_and_unit_offset.
12471 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
12473 PR rtl-optimization/56833
12474 * postreload.c (move2add_record_mode): New function.
12475 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
12476 (move2add_use_add2_insn): Use move2add_record_sym_value.
12477 (move2add_use_add3_insn): Likewise.
12478 (reload_cse_move2add): Use move2add_valid_value_p and
12479 move2add_record_mode. Invalidate call-clobbered and REG_INC
12480 affected regs by setting reg_mode to VOIDmode.
12481 (move2add_note_store): Don't pretend the inside of a SUBREG is
12482 the actual destination. Invalidate single/leading registers by
12483 setting reg_mode to VOIDmode.
12484 Use move2add_record_sym_value, move2add_valid_value_p and
12485 move2add_record_mode.
12487 2013-05-27 Richard Biener <rguenther@suse.de>
12489 PR tree-optimization/57396
12490 * tree-affine.c (double_int_constant_multiple_p): Properly
12491 return false for val == 0 and div != 0.
12493 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
12495 * config/mips/mips.h: Use #elif in preprocessor conditions.
12497 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
12500 * config/mips/constraints.md (kl): New constraint.
12501 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
12502 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
12503 constraint for operand 0. Split after CSE for MIPS16. Emit a move
12504 from LO for MIPS16.
12505 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
12507 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
12510 * config/mips/mips.c (mips_can_inline_p): New function.
12511 (TARGET_CAN_INLINE_P): Define.
12513 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
12515 * sched-int.h (ds_t, dw_t): Make unsigned int.
12516 Fix documentation that describes how all the ds_t bits are used.
12517 Reserve the last bit for delayed-branch scheduling.
12518 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
12519 (BITS_PER_DEP_WEAK): Fix definition and documentation.
12520 (gen_dep_weak_1): Remove prototype.
12521 * sched-deps.c (get_dep_weak_1): Make static.
12522 * target.def (speculate_insn, needs_block_p, gen_spec_check,
12523 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
12524 * doc/tm.texi: Regenerate.
12525 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
12527 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
12530 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
12532 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
12533 Sandra Loosemore <sandra@codesourcery.com>
12535 * config.gcc (powerpc-*): Allow native for with-cpu.
12537 2013-05-24 Jeff Law <law@redhat.com>
12539 PR tree-optimization/57124
12540 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
12541 conversion feeding a condition if the range has an overflow
12542 if -fstrict-overflow. Add warnings for when we do make the
12545 2013-05-24 Dehao Chen <dehao@google.com>
12547 * tree-cfg.c (locus_discrim_map): Fix the typo.
12548 (locus_discrim_hasher): Likewise.
12549 (locus_discrim_hasher::hash): Likewise.
12550 (locus_discrim_hasher::equal): Likewise.
12552 2013-05-24 Martin Jambor <mjambor@suse.cz>
12554 PR tree-optimization/57294
12555 * cgraph.h (ipa_record_stmt_references): Declare.
12556 * cgraphbuild.c (ipa_record_stmt_references): New function.
12557 (build_cgraph_edges): Use ipa_record_stmt_references.
12558 (rebuild_cgraph_edges): Likewise.
12559 (cgraph_rebuild_references): Likewise.
12560 * ipa-prop.c (ipa_modify_call_arguments): Discard references
12561 associated with the old statement and build references from the
12562 newly built statements.
12563 * ipa-ref.c (ipa_remove_stmt_references): New function.
12564 * ipa-ref.h (ipa_remove_stmt_references): Declare.
12566 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
12568 * lra-constraints.c (emit_spill_move): Use smaller mode for
12570 (check_and_process_move): Consider mem-reg moves for secondary
12572 (curr_insn_transform): Don't lose insns emitted before for
12573 secondary memory moves.
12574 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
12575 reg set up in the current insn.
12577 2013-05-24 Dehao Chen <dehao@google.com>
12579 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
12581 (locus_descrim_hasher::equal): Likewise.
12582 (build_gimple_cfg): New discriminator assignment algorithm.
12583 (make_edges): Likewise.
12584 (next_discriminator_for_locus): Likewise.
12585 (same_line_p): Likewise.
12586 (assign_discriminators): Likewise.
12587 (make_cond_expr_edges): Likewise.
12588 (make_gimple_switch_edges): Likewise.
12589 (make_goto_expr_edges): Likewise.
12590 (make_gimple_asm_edges): Likewise.
12592 2013-05-24 Ian Bolton <ian.bolton@arm.com>
12594 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
12595 X format specifier to only display bottom 16 bits.
12596 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
12597 immediate to match for operand 2, since it will be masked.
12599 2013-05-24 Richard Biener <rguenther@suse.de>
12601 PR tree-optimization/57287
12602 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
12603 all SSA names that occur in abnormal PHIs.
12605 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
12608 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
12609 that index is not negative.
12611 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
12613 PR rtl-optimization/55177
12614 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
12615 (simplify_byte_swapping_operation): New.
12616 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
12617 (simplify_relational_operation_1): Deal with BSWAP.
12619 2013-05-23 Richard Henderson <rth@redhat.com>
12622 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
12623 (ix86_reorg): Call it.
12625 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
12628 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
12629 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
12630 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
12632 2013-05-23 Christian Bruel <christian.bruel@st.com>
12635 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
12637 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
12638 Marcus Shawcroft <marcus.shawcroft@arm.com>
12640 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
12641 * config/aarch64/constraints.md (Usa): Remove.
12642 * doc/md.texi (AArch64 Usa): Remove.
12644 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
12645 Marcus Shawcroft <marcus.shawcroft@arm.com>
12647 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
12648 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
12649 * config/aarch64/predicates.md (aarch64_const_address): Remove.
12650 (aarch64_mov_operand): Use aarch64_mov_operand_p.
12652 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
12654 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
12655 instruction (AdvSIMD).
12656 * config/aarch64/aarch64-builtins.c
12657 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
12658 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
12660 2013-05-23 Martin Jambor <mjambor@suse.cz>
12662 PR middle-end/57347
12663 * tree.h (contains_bitfld_component_ref_p): Declare.
12664 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
12665 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
12667 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
12668 not access a bit-field. Assert all final offsets are byte-aligned.
12670 2013-05-23 Richard Biener <rguenther@suse.de>
12672 PR tree-optimization/57380
12673 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
12674 least one invariant or re-used load.
12675 * passes.c (init_optimization_passes): Move pass_phiprop before
12678 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
12680 * config/aarch64/aarch64-simd.md
12681 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
12683 2013-05-23 Richard Biener <rguenther@suse.de>
12685 PR middle-end/57381
12686 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
12687 OEP_CONSTANT_ADDRESS_OF retained.
12689 2013-05-23 Jakub Jelinek <jakub@redhat.com>
12691 PR middle-end/57344
12692 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
12693 don't lower unit. Handle unit not being always BITS_PER_WORD.
12695 2013-05-23 Richard Biener <rguenther@suse.de>
12697 PR rtl-optimization/57341
12698 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
12699 instead of true_dependence.
12701 2013-05-22 David Malcolm <dmalcolm@redhat.com>
12703 * bb-reorder.c (branch_threshold): Make const.
12704 (exec_threshold): Ditto.
12706 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
12707 Pat Haugen <pthaugen@us.ibm.com>
12708 Peter Bergner <bergner@vnet.ibm.com>
12710 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
12711 documentation for the power8 crypto builtins.
12713 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
12715 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
12716 macros for defining power8 builtin functions.
12717 (BU_P8V_AV_2): Likewise.
12718 (BU_P8V_AV_P): Likewise.
12719 (BU_P8V_VSX_1): Likewise.
12720 (BU_P8V_OVERLOAD_1): Likewise.
12721 (BU_P8V_OVERLOAD_2): Likewise.
12722 (BU_CRYPTO_1): Likewise.
12723 (BU_CRYPTO_2): Likewise.
12724 (BU_CRYPTO_3): Likewise.
12725 (BU_CRYPTO_OVERLOAD_1): Likewise.
12726 (BU_CRYPTO_OVERLOAD_2): Likewise.
12727 (XSCVSPDP): Fix typo, point to the correct instruction.
12728 (VCIPHER): Add power8 crypto builtins.
12729 (VCIPHERLAST): Likewise.
12730 (VNCIPHER): Likewise.
12731 (VNCIPHERLAST): Likewise.
12732 (VPMSUMB): Likewise.
12733 (VPMSUMH): Likewise.
12734 (VPMSUMW): Likewise.
12735 (VPERMXOR_V2DI): Likewise.
12736 (VPERMXOR_V4SI: Likewise.
12737 (VPERMXOR_V8HI: Likewise.
12738 (VPERMXOR_V16QI: Likewise.
12739 (VSHASIGMAW): Likewise.
12740 (VSHASIGMAD): Likewise.
12741 (VPMSUM): Likewise.
12742 (VPERMXOR): Likewise.
12743 (VSHASIGMA): Likewise.
12745 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
12746 __CRYPTO__ if the crypto instructions are available.
12747 (altivec_overloaded_builtins): Add support for overloaded power8
12750 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
12751 support for power8 crypto builtins.
12752 (builtin_function_type): Likewise.
12753 (altivec_init_builtins): Add support for builtins that take vector
12754 long long (V2DI) arguments.
12756 * config/rs6000/crypto.md: New file, define power8 crypto
12759 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
12760 Pat Haugen <pthaugen@us.ibm.com>
12761 Peter Bergner <bergner@vnet.ibm.com>
12763 * doc/invoke.texi (Option Summary): Add power8 options.
12764 (RS/6000 and PowerPC Options): Likewise.
12766 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
12767 constraints.md instead of rs6000.h. Reorder w* constraints. Add
12768 wm, wn, wr documentation.
12770 * config/rs6000/constraints.md (wm): New constraint for VSX
12771 registers if direct move instructions are enabled.
12772 (wn): New constraint for no registers.
12773 (wq): New constraint for quad word even GPR registers.
12774 (wr): New constraint if 64-bit instructions are enabled.
12775 (wv): New constraint if power8 vector instructions are enabled.
12776 (wQ): New constraint for quad word memory locations.
12778 * config/rs6000/predicates.md (const_0_to_15_operand): New
12779 constraint for 0..15 for crypto instructions.
12780 (gpc_reg_operand): If VSX allow registers in VSX registers as well
12781 as GPR and floating point registers.
12782 (int_reg_operand): New predicate to match only GPR registers.
12783 (base_reg_operand): New predicate to match base registers.
12784 (quad_int_reg_operand): New predicate to match even GPR registers
12785 for quad memory operations.
12786 (vsx_reg_or_cint_operand): New predicate to allow vector logical
12787 operations in both GPR and VSX registers.
12788 (quad_memory_operand): New predicate for quad memory operations.
12789 (reg_or_indexed_operand): New predicate for direct move support.
12791 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
12792 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
12793 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
12794 (POWERPC_MASKS): Add power8 options.
12795 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
12798 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
12799 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
12801 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
12802 (-mpower8-fusion): New power8 options.
12803 (-mpower8-fusion-sign): Likewise.
12804 (-mpower8-vector): Likewise.
12805 (-mcrypto): Likewise.
12806 (-mdirect-move): Likewise.
12807 (-mquad-memory): Likewise.
12809 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
12810 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
12812 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
12813 (rs6000_debug_vector_unit): Add p8_vector.
12814 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
12815 definitions. Also print fusion state.
12816 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
12817 (rs6000_builtin_mask_calculate): Add power8 builtin support.
12818 (rs6000_option_override_internal): Add support for power8.
12819 (rs6000_common_init_builtins): Add debugging for skipped builtins
12820 if -mdebug=builtin.
12821 (rs6000_adjust_cost): Add power8 support.
12822 (rs6000_issue_rate): Likewise.
12823 (insn_must_be_first_in_group): Likewise.
12824 (insn_must_be_last_in_group): Likewise.
12825 (force_new_group): Likewise.
12826 (rs6000_register_move_cost): Likewise.
12827 (rs6000_opt_masks): Likewise.
12829 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
12830 power8 capable assembler, default to power7 options.
12831 (TARGET_DIRECT_MOVE): Likewise.
12832 (TARGET_CRYPTO): Likewise.
12833 (TARGET_P8_VECTOR): Likewise.
12834 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
12835 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
12836 (VECTOR_MEM_P8_VECTOR_P): Likewise.
12837 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
12838 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
12839 (TARGET_XSCVDPSPN): Likewise.
12840 (TARGET_XSCVSPDPN): Likewsie.
12841 (TARGET_SYNC_HI_QI): Likewise.
12842 (TARGET_SYNC_TI): Likewise.
12843 (MASK_CRYPTO): Likewise.
12844 (MASK_DIRECT_MOVE): Likewise.
12845 (MASK_P8_FUSION): Likewise.
12846 (MASK_P8_VECTOR): Likewise.
12847 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
12848 temporary used by some of the direct move instructions to get two FP
12849 temporary registers does not force creation of a stack frame.
12850 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
12851 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
12852 that any VSX registers are tieable, even if they are also an
12853 Altivec vector mode.
12854 (r6000_reg_class_enum): Add wm, wr, wv constraints.
12855 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
12856 (RS6000_BTM_CRYPTO): Likewise.
12857 (RS6000_BTM_COMMON): Likewise.
12859 * config/rs6000/rs6000.md (cpu attribute): Add power8.
12860 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
12861 (enum rs6000_vector): Add power8 vector support.
12863 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
12867 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
12868 (any_sibcall_could_use_r3): this and handle indirect calls.
12869 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
12871 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12873 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
12875 2013-05-22 Richard Biener <rguenther@suse.de>
12877 PR middle-end/57349
12878 * profile.c (branch_prob): Do not split blocks that are
12879 abnormally receiving from ECF_RETURNS_TWICE functions.
12881 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
12883 * recog.c (offsettable_address_addr_space_p): Fix calculation of
12884 address mode. Move pointer mode initialization to the same place.
12886 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
12888 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
12889 while it has any effect.
12891 2013-05-21 Easwaran Raman <eraman@google.com>
12893 PR tree-optimization/57322
12894 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
12895 UID of the statement added to the BB to be 1.
12897 2013-05-21 Jakub Jelinek <jakub@redhat.com>
12899 PR tree-optimization/57331
12900 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
12901 of conversion from pointer type to integral type with integer.
12903 2013-05-21 Martin Jambor <mjambor@suse.cz>
12906 * ipa-prop.c (ipa_read_node_info): Process param_used and
12907 controlled_uses in the same order as when writing.
12909 2013-05-21 Magnus Granberg <baldrick@free.fr>
12912 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
12914 2013-05-21 Richard Biener <rguenther@suse.de>
12916 PR tree-optimization/57318
12917 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
12918 estimate stmts with side-effects as likely eliminated.
12920 2013-05-21 Richard Biener <rguenther@suse.de>
12922 PR tree-optimization/57330
12923 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
12924 preserve the call stmts fntype.
12926 2013-05-21 Richard Biener <rguenther@suse.de>
12928 PR tree-optimization/57303
12929 * tree-ssa-sink.c (statement_sink_location): Improve killing
12930 stmt detection and properly handle self-assignments.
12932 2013-05-21 Christian Bruel <christian.bruel@st.com>
12934 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
12935 spanning registers. LEAF_REG_REMAP is supported only for contiguous
12936 registers. Set register size out of the PARALLEL loop.
12938 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
12941 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
12942 (*fmasf4, *fmasf4_media): New insns.
12944 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
12946 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
12947 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
12948 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
12949 (mips_idiv_insns): Update the comments to say that the returned
12950 instruction counts are in units of BASE_INSN_LENGTH.
12951 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
12952 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
12953 using 2 rather than 4 as the length of indirect MIPS16 and
12954 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
12955 length of a NOP. Don't divide MIPS16 lengths by 2.
12956 (mips16_split_long_branches): Assume a branch is long if the
12957 length is greater than 4 rather than 8.
12958 * config/mips/mips.md (length): Give MIPS16 lengths directly,
12959 rather than multiplying them by 2. Multiply instruction counts
12960 by BASE_INSN_LENGTH rather than 4.
12961 (*jump_mips16, tls_get_tp_mips16_<mode>)
12962 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
12964 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
12966 * config/mips/mips.md (extended_mips16): Remove branch case.
12967 (length): Remove duplicated extended_mips16 test.
12969 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
12971 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
12973 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
12975 * recog.h (Recog_data): Rename to...
12976 (recog_data_d): ...this.
12977 (recog_data): Update accordingly.
12978 * recog.c (recog_data): Likewise.
12979 * reload.c (save_recog_data): Likewise.
12980 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
12981 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
12983 2013-05-17 Julian Brown <julian@codesourcery.com>
12985 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
12986 found in a REG_EQUAL note, invalidate it.
12988 2013-05-17 Easwaran Raman <eraman@google.com>
12990 * tree-ssa-reassoc.c (find_insert_point): New function.
12991 (insert_stmt_after): Likewise.
12992 (get_def_stmt): Likewise.
12993 (ensure_ops_are_available): Likewise.
12994 (not_dominated_by): Likewise.
12995 (rewrite_expr_tree): Do not move statements beyond what is
12996 necessary. Remove call to swap_ops_for_binary_stmt...
12997 (reassociate_bb): ... and move it here.
12998 (build_and_add_sum): Assign UIDs for new statements.
12999 (linearize_expr): Likewise.
13000 (do_reassoc): Renumber gimple statement UIDs.
13002 2013-05-17 Jan Hubicka <jh@suse.cz>
13004 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
13006 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
13007 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
13009 (output_weakrefs): Update.
13011 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
13012 Martin Jambor <mjambor@suse.cz>
13014 PR middle-end/57276
13015 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
13016 value that corresponds to the given aggval is found in values vector.
13018 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
13020 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
13021 sse, sse2, sse3, ssse3 and sse4a flags to options.
13023 2013-05-17 David Malcolm <dmalcolm@redhat.com>
13025 * gengtype-state.c: (s_expr_writer): New class, to handle
13026 prettifying of output layout of s-expressions.
13027 (state_writer): New class, to write out gtype.state.
13028 (state_written_type_count): Move this variable into member data of
13030 (s_expr_writer::s_expr_writer): New code: constructor for new class
13031 (state_writer::state_writer(): ditto
13032 (s_expr_writer::write_new_line): New function
13033 (s_expr_writer::write_any_indent): ditto
13034 (s_expr_writer::begin_s_expr): ditto
13035 (s_expr_writer::end_s_expr): ditto
13036 (write_state_fileloc): convert to method of state_writer...
13037 (state_writer:: write_state_fileloc): ...and use methods of
13038 s_expr_writer to write indentation into the gtype.state output file
13039 to visually represent the hierarchical structure of the list
13041 (write_state_fields): ditto, renaming to...
13042 (state_writer::write_state_fields)
13043 (write_state_a_string): ditto, renaming to...
13044 (state_writer::write_state_a_string)
13045 (write_state_string_option): ditto, renaming to...
13046 (state_writer::write_state_string_option)
13047 (write_state_type_option): ditto, renaming to...
13048 (state_writer::write_state_type_option)
13049 (write_state_nested_option): ditto, renaming to...
13050 (state_writer::write_state_nested_option)
13051 (write_state_option): ditto, renaming to...
13052 (state_writer::write_state_option)
13053 (write_state_options): ditto, renaming to...
13054 (state_writer::write_state_options)
13055 (write_state_lang_bitmap): ditto, renaming to...
13056 (state_writer::write_state_lang_bitmap)
13057 (write_state_version): ditto, renaming to...
13058 (state_writer::write_state_version)
13059 (write_state_scalar_type): ditto, renaming to...
13060 (state_writer::write_state_scalar_type)
13061 (write_state_string_type): ditto, renaming to...
13062 (state_writer::write_state_string_type)
13063 (write_state_undefined_type): ditto, renaming to...
13064 (state_writer::write_state_undefined_type)
13065 (write_state_struct_union_type): ditto, renaming to...
13066 (state_writer::write_state_struct_union_type)
13067 (write_state_struct_type): ditto, renaming to...
13068 (state_writer::write_state_struct_type)
13069 (write_state_user_struct_type): ditto, renaming to...
13070 (state_writer::write_state_user_struct_type)
13071 (write_state_lang_struct_type): ditto, renaming to...
13072 (state_writer::write_state_lang_struct_type)
13073 (write_state_param_struct_type): ditto, renaming to...
13074 (state_writer::write_state_param_struct_type)
13075 (write_state_pointer_type): ditto, renaming to...
13076 (state_writer::write_state_pointer_type)
13077 (write_state_array_type): ditto, renaming to...
13078 (state_writer::write_state_array_type)
13079 (write_state_gc_used): ditto, renaming to...
13080 (state_writer::write_state_gc_used)
13081 (write_state_common_type_content): ditto, renaming to...
13082 (state_writer::write_state_common_type_content)
13083 (write_state_type): ditto, renaming to...
13084 (state_writer::write_state_type)
13085 (write_state_pair_list): ditto, renaming to...
13086 (state_writer::write_state_pair_list)
13087 (write_state_pair): ditto, renaming to...
13088 (state_writer::write_state_pair)
13089 (write_state_typedefs): ditto, renaming to...
13090 (state_writer::write_state_typedefs)
13091 (write_state_structures): ditto, renaming to...
13092 (state_writer::write_state_structures)
13093 (write_state_param_structs): ditto, renaming to...
13094 (state_writer::write_state_param_structs)
13095 (write_state_variables): ditto, renaming to...
13096 (state_writer::write_state_variables)
13097 (write_state_srcdir): ditto, renaming to...
13098 (state_writer::write_state_srcdir)
13099 (write_state_files_list): ditto, renaming to...
13100 (state_writer::write_state_files_list)
13101 (write_state_languages): ditto, renaming to...
13102 (state_writer::write_state_languages)
13103 (write_state): create a state_writer instance and use it when
13104 writing out the state file
13106 2013-05-17 Mike Stump <mikestump@comcast.net>
13108 PR rtl-optimization/57304
13109 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
13110 accessing DF_REF_REAL_LOC.
13112 2013-05-17 Jakub Jelinek <jakub@redhat.com>
13114 PR rtl-optimization/57281
13115 PR rtl-optimization/57300
13116 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
13117 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
13118 what the other splitter did if the registers are dead.
13120 2013-05-17 Richard Biener <rguenther@suse.de>
13122 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
13125 2013-05-17 Jakub Jelinek <jakub@redhat.com>
13127 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
13130 2013-05-17 Marek Polacek <polacek@redhat.com>
13132 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
13133 length when doing non-zero store of storing '\0' to '\0'.
13135 2013-05-17 Jakub Jelinek <jakub@redhat.com>
13137 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
13138 vect_external_def oprnd1 with loop_vinfo, try to emit
13139 optional cast, negation and and stmts on the loop preheader
13140 edge instead of into the pattern def seq.
13142 PR tree-optimization/57051
13143 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
13144 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
13146 2013-05-16 Nick Clifton <nickc@redhat.com>
13148 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
13149 (rl78_is_naked_func): New function.
13150 (rl78_expand_prologue): Skip prologue generation for naked functions.
13151 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
13152 * doc/extend.texi (naked): Add RL78 to the list of processors
13153 that supports this attribute.
13155 2013-05-16 Jeff Law <law@redhat.com>
13157 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
13159 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
13161 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
13162 cache parameters using detect_caches_amd also for CYRIX,
13163 NSC and TM2 signatures.
13165 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
13166 Dzianis Kahanovich <mahatma@eu.by>
13170 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
13171 VIA/Centaur processors and determine their cache parameters
13172 using detect_caches_amd.
13174 2013-05-16 Teresa Johnson <tejohnson@google.com>
13176 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
13177 (rtl_verify_edges): New function.
13178 (rtl_verify_bb_insns): Ditto.
13179 (rtl_verify_bb_pointers): Ditto.
13180 (rtl_verify_bb_insn_chain): Ditto.
13181 (rtl_verify_fallthru): Ditto.
13182 (rtl_verify_bb_layout): Ditto.
13183 (rtl_verify_flow_info_1): Outline checks into new functions.
13184 (rtl_verify_flow_info): Ditto.
13186 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
13188 * cfghooks.c (copy_bbs): Add update_dominance argument.
13189 * cfghooks.h (copy_bbs): Update prototype.
13190 * tree-cfg.c (gimple_duplicate_sese_region):
13191 Add update_dominance argument.
13192 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
13193 * tree-ssa-loop-ch.c (copy_loop_headers): Update
13194 gimple_duplicate_sese_region call.
13195 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
13196 Update copy_bbs call.
13197 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
13198 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
13200 2013-05-16 Jakub Jelinek <jakub@redhat.com>
13202 * tree-vectorizer.h (NUM_PATTERNS): Increment.
13203 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
13204 vect_recog_rotate_pattern.
13205 (vect_recog_rotate_pattern): New function.
13207 2013-05-16 Jason Merrill <jason@redhat.com>
13209 * Makefile.in (LLINKER): New variable.
13210 (mostlyclean): Remove link mutex.
13211 * configure.ac: Handle --enable-link-mutex.
13212 * lock-and-run.sh: New script.
13214 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
13217 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
13220 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
13222 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
13224 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
13226 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
13227 * config/arm/arm.c (next_consecutive_mem): New function.
13228 (gen_movmem_ldrd_strd): Likewise.
13229 * config/arm/arm.md (movmemqi): Update condition and code.
13230 (unaligned_loaddi, unaligned_storedi): New patterns.
13232 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
13234 * config.gcc: Obsolete *-*-solaris2.9*.
13235 * doc/install.texi (Specific, *-*-solaris2*): Document it.
13237 2013-05-16 Richard Biener <rguenther@suse.de>
13239 * passes.c (init_optimization_passes): Move pass_parallelize_loops
13240 earlier, after GRAPHITE transforms and IV canonicalization.
13242 2013-05-16 Jakub Jelinek <jakub@redhat.com>
13244 * omp-low.c (extract_omp_for_data): For collapsed loops,
13245 if at least one of the loops is known at compile time to
13246 iterate zero times, set count to 0.
13247 (expand_omp_regimplify_p): New function.
13248 (expand_omp_for_generic): For collapsed loops, if at least
13249 one of the loops isn't known to iterate at least once,
13250 add runtime check with setting count to 0.
13251 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
13252 For unsigned types if it isn't known at compile time that
13253 the loop will iterate at least once, add runtime check to bypass
13254 the whole loop if initial condition isn't true.
13256 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
13258 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
13260 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
13262 PR middle-end/57286
13263 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
13264 transformations to avoid an infinite loop.
13266 2013-05-16 Marek Polacek <polacek@redhat.com>
13268 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
13270 2013-05-15 Leif Ekblad <leif@rdos.net>
13272 * config/i386/i386.c (ix86_decompose_address): Use
13273 DEFAULT_TLS_SEG_REG to access TLS segment register.
13274 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
13275 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
13276 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
13278 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
13281 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
13282 sibling calls to functions that would normally be lazily bound,
13283 unless $gp is call-clobbered.
13285 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
13287 * config/i386/i386.c (ix86_option_override_internal): Update
13288 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
13289 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
13290 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
13291 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
13293 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
13295 2013-05-15 Andreas Schwab <schwab@suse.de>
13297 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
13298 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
13301 2013-05-15 Teresa Johnson <tejohnson@google.com>
13303 * loop-unroll.c (report_unroll_peel): Check decision before
13304 emitting unroll/peel message.
13306 2013-05-15 Teresa Johnson <tejohnson@google.com>
13308 * function.h (has_bb_partition): New rtl_data flag.
13309 (bb_reorder_complete): Ditto.
13310 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
13311 instead of flag_reorder_blocks_and_partition.
13312 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
13313 with some enhancements.
13314 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
13315 * bb-reorder.c (connect_traces): Check for has_bb_partition
13316 instead of flag_reorder_blocks_and_partition.
13317 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
13318 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
13319 verify_hot_cold_block_grouping.
13320 (partition_hot_cold_basic_blocks): Set has_bb_partition.
13322 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
13325 * config/arm/predicates.md (call_insn_operand): New predicate.
13326 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
13327 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
13328 if insn is not a tail call.
13329 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
13331 * config/arm/arm.h (enum reg_class): New caller save register class.
13332 (REG_CLASS_NAMES): Likewise.
13333 (REG_CLASS_CONTENTS): Likewise.
13334 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
13337 2013-05-15 Richard Biener <rguenther@suse.de>
13339 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
13340 of MSG_OPTIMIZED_LOCATIONS.
13341 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
13342 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
13344 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
13345 of MSG_OPTIMIZED_LOCATIONS.
13346 (execute_vect_slp): Likewise.
13347 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
13348 (vect_create_cond_for_alias_checks): Likewise.
13349 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
13350 (vect_recog_widen_mult_pattern): Likewise.
13351 (vect_recog_widen_sum_pattern): Likewise.
13352 (vect_recog_over_widening_pattern): Likewise.
13353 (vect_recog_widen_shift_pattern): Likewise.
13354 (vect_recog_vector_vector_shift_pattern): Likewise.
13355 (vect_recog_divmod_pattern): Likewise.
13356 (vect_recog_mixed_size_cond_pattern): Likewise.
13357 (vect_recog_bool_pattern): Likewise.
13358 (vect_pattern_recog_1): Likewise.
13360 2013-05-15 Martin Jambor <mjambor@suse.cz>
13362 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
13363 non-functions to builtin_unreachable.
13364 * ipa-inline-transform.c (inline_call): Do not assert estimates were
13365 correct when new direct edges were discovered.
13367 2013-05-15 Martin Jambor <mjambor@suse.cz>
13369 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
13370 header, print symbol order instead of node uid, print more information
13371 about indirect edge targets.
13372 (ipa_make_edge_direct_to_target): Print symbol order instead of node
13374 (ipa_make_edge_direct_to_target): Likewise.
13375 (remove_described_reference): Likewise.
13376 (propagate_controlled_uses): Likewise.
13377 (ipa_print_node_params): Also print symbol order.
13378 (ipcp_transform_function): Print symbol order instead of node uids.
13379 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
13380 (cgraph_get_create_real_symbol_node): Likewise.
13381 * ipa-cp.c (print_lattice): Likewise.
13382 (print_all_lattices): Likewise.
13383 (determine_versionability): Likewise.
13384 (initialize_node_lattices): Likewise.
13385 (estimate_local_effects): Likewise.
13386 (update_profiling_info): Likewise.
13387 (create_specialized_node): Likewise.
13388 (perhaps_add_new_callers): Likewise.
13389 (decide_about_value): Likewise.
13390 (decide_whether_version_node): Likewise.
13391 (identify_dead_nodes): Likewise.
13392 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
13393 (dump_inline_summary): Likewise.
13394 (estimate_node_size_and_time): Likewise.
13395 (inline_analyze_function): Likewise.
13396 * ipa-inline.c (report_inline_failed_reason): Likewise.
13397 (want_early_inline_function_p): Likewise.
13398 (edge_badness): Likewise.
13399 (update_edge_key): Likewise.
13400 (inline_small_functions): Likewise. Add dumping of order to two other
13402 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
13403 instead of node uids.
13404 (propagate_pure_const): Likewise.
13405 (propagate_pure_const): Likewise.
13406 * ipa-utils.c (dump_cgraph_node_set): Likewise.
13407 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
13408 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
13410 * tree-pretty-print.c (dump_function_header): Likewise.
13411 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
13412 Print symbol order instead of node uids.
13414 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13416 * config/s390/s390.c (s390_register_move_cost): Don't impose the
13417 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
13419 2013-05-15 Richard Biener <rguenther@suse.de>
13421 PR tree-optimization/57275
13422 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
13423 return value for fail to do runtime alias checks for gather loads.
13425 2013-05-15 Jan Hubicka <jh@suse.cz>
13429 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
13430 weakrefs are not external.
13431 (lto_symtab_merge_decls): Fix thinko when dealing with
13432 non-lto_symtab decls.
13433 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
13434 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
13435 * varpool.c (dump_varpool_node): Dump more flags.
13437 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
13439 * config/i386/i386.c (processor_alias_table): Add instruction
13440 FSGSBASE for AMD bdver3 architecture.
13442 2013-05-14 Jakub Jelinek <jakub@redhat.com>
13444 * tree.c (warn_deprecated_use): Print file:line using locus color.
13445 * diagnostic.c (diagnostic_report_current_module): Print file:line
13446 and file:line:column using locus color.
13448 2013-05-14 Mike Stump <mikestump@comcast.net>
13450 * gdbinit.in: Add __null.
13452 2013-05-14 Mike Stump <mikestump@comcast.net>
13454 * recog.h: Rename struct recog_data to Recog_data.
13455 * recog.c: Likewise.
13456 * reload.c (can_reload_into): Likewise.
13457 * config/picochip/picochip.c: Likewise.
13459 2013-05-14 Mike Stump <mikestump@comcast.net>
13461 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
13463 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
13465 * resource.h (struct resources): Remove unch_memory member.
13466 (CLEAR_RESOURCE): Don't clear unch_memory.
13467 * resource.c (mark_referenced_resources): Don't set it.
13468 (mark_set_resources): Likewise.
13469 (mark_target_live_regs): Don't clear it.
13470 (init_resource_info): Likewise.
13471 * reorg.c (resource_conflicts_p): Don't compare it.
13472 (redundant_insn): Don't set it.
13474 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
13476 * emit-rtl.c (next_label): Remove unused function.
13477 (skip_consecutive_labels, link_cc0_insns): Move to ...
13478 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
13479 only place where these functions are used, and make them static.
13481 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
13483 * fold-const.c (fold_negate_expr): Handle vectors.
13484 (fold_truth_not_expr): Make it static.
13485 (fold_invert_truthvalue): New static function.
13486 (invert_truthvalue_loc): Handle vectors. Do not call
13487 fold_truth_not_expr directly.
13488 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
13489 <TRUTH_NOT_EXPR>: Do not cast to boolean.
13490 (fold_comparison): Handle vector constants.
13491 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
13492 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
13493 * tree.h (fold_truth_not_expr): Remove declaration.
13495 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
13497 * config/aarch64/aarch64-simd.md
13498 (aarch64_vcond_internal<mode>): Rename to...
13499 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
13500 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
13501 float modes. Clarify all iterator modes.
13502 (vcond<mode><mode>): Use new name for vcond expanders.
13503 (vcond<v_cmp_result><mode>): Likewise.
13504 (vcondu<mode><mode>: Likewise.
13505 * config/aarch64/iterators.md (VDQF_COND): New.
13507 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
13510 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
13511 variable for the shift amount. Check that we shift by non-negative
13514 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
13517 * config/arm/arm.h (EPILOGUE_USES): Only return true
13518 for LR_REGNUM after epilogue_completed.
13520 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
13522 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
13523 is error_mark_node.
13525 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
13528 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
13529 and Solaris 11+/x86 with gld.
13530 * configure: Regenerate.
13532 2013-05-14 Jakub Jelinek <jakub@redhat.com>
13534 * expmed.c (expand_shift_1): Canonicalize rotates by
13535 constant bitsize / 2 to bitsize - 1.
13536 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
13537 case ROTATERT>: Likewise.
13540 2013-05-10 Jakub Jelinek <jakub@redhat.com>
13542 * config/i386/i386.md (rotateinv): New code attr.
13543 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
13544 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
13545 roll $31, %eax, etc.
13547 2013-05-14 Richard Biener <rguenther@suse.de>
13549 PR middle-end/57235
13550 * tree-eh.c (sink_clobbers): Give up for successors with
13551 multiple predecessors and no virtual uses.
13553 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
13555 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
13556 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
13558 2013-05-14 Jakub Jelinek <jakub@redhat.com>
13560 PR middle-end/57251
13561 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
13562 the case when both op0 and op1 have VOIDmode.
13564 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
13566 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
13567 in multiply-accumulate mode.
13569 2013-05-13 Guozhi Wei <carrot@google.com>
13571 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
13573 2013-05-13 Kai Tietz <ktietz@redhat.com>
13576 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
13577 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
13578 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
13579 * config/i386/i386.c (ix86_option_override_internal): Likewise.
13580 (ix86_expand_prologue): Likewise.
13581 (ix86_expand_split_stack_prologue): Likewise.
13582 (legitimate_pic_address_disp_p): Likewise.
13583 (legitimize_pic_address): Likewise.
13584 (legitimize_tls_address): Likewise.
13585 (legitimize_pe_coff_symbol): Likewise.
13586 (output_pic_addr_const): Likewise.
13587 (construct_plt_address): Likewise.
13588 (ix86_expand_call): Likewise.
13589 (x86_output_mi_thunk): Likewise.
13590 (x86_function_profiler): Likewise.
13592 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
13594 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
13595 similar switch cases.
13596 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
13597 (aarch64_simd_mov_to_<mode>low): Delete.
13598 (aarch64_simd_mov_to_<mode>high): Delete.
13599 (move_lo_quad_<mode>): Add w<-r alternative.
13600 (aarch64_simd_move_hi_quad_<mode>): Likewise.
13601 (aarch64_simd_mov_from_*): Update type attribute.
13602 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
13605 2013-05-13 Jan Hubicka <jh@suse.cz>
13607 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
13608 * config/i386/i386.c (ix86_compute_frame_layout,
13609 ix86_expand_epilogue, emit_i387_cw_initialization,
13610 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
13611 ix86_local_alignment): Fix use of size/speed predicates.
13613 2013-05-13 Jakub Jelinek <jakub@redhat.com>
13615 PR tree-optimization/45216
13616 PR tree-optimization/57157
13617 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
13618 the (-Y) & (B - 1) variant if OP is |.
13619 * expmed.c (expand_shift_1): For rotations by const0_rtx just
13620 return shifted. Use (-op1) & (prec - 1) as other_amount
13621 instead of prec - op1.
13623 2013-05-13 Martin Jambor <mjambor@suse.cz>
13625 PR middle-end/42371
13626 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
13627 (ipa_constant_data): New type.
13628 (ipa_jump_func): Use ipa_constant_data to hold information about
13629 constant jump functions.
13630 (ipa_get_jf_constant): Adjust to jump function type changes.
13631 (ipa_get_jf_constant_rdesc): New function.
13632 (ipa_param_descriptor): New field controlled_uses.
13633 (ipa_get_controlled_uses): New function.
13634 (ipa_set_controlled_uses): Likewise.
13635 * ipa-ref.h (ipa_find_reference): Declare.
13636 * ipa-prop.c (ipa_cst_ref_desc): New type.
13637 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
13639 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
13640 New parameter cs. Adjust all callers.
13641 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
13642 (remove_described_reference): New function.
13643 (jfunc_rdesc_usable): Likewise.
13644 (try_make_edge_direct_simple_call): Decrement controlled use count,
13645 attempt to remove reference if it hits zero.
13646 (combine_controlled_uses_counters): New function.
13647 (propagate_controlled_uses): Likewise.
13648 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
13649 (ipa_edge_duplication_hook): Duplicate reference descriptions.
13650 (ipa_print_node_params): Print described use counter.
13651 (ipa_write_jump_function): Adjust to jump function type changes.
13652 (ipa_read_jump_function): New parameter CS, pass it to
13653 ipa_set_jf_constant. Adjust caller.
13654 (ipa_write_node_info): Stream controlled use count
13655 (ipa_read_node_info): Likewise.
13656 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
13658 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
13659 count. Remove cloning-added reference if it reaches zero.
13660 * ipa-ref.c (ipa_find_reference): New function.
13662 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
13664 * config/i386/i386.c (processor_target_table): Modified default
13665 alignment values for AMD BD and BT architectures.
13667 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
13669 * tree-vect-generic.c (uniform_vector_p): Move ...
13670 * tree.c (uniform_vector_p): ... here.
13671 * tree.h (uniform_vector_p): Declare it.
13672 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
13675 2013-05-13 Jakub Jelinek <jakub@redhat.com>
13677 PR tree-optimization/57230
13678 * tree-ssa-strlen.c (handle_char_store): Record length for
13679 array store from STRING_CST.
13681 PR tree-optimization/57230
13682 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
13685 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
13687 * config/epiphany/epiphany.c (epiphany_init): Check size of
13688 NUM_MODES_FOR_MODE_SWITCHING.
13689 (epiphany_expand_prologue):
13690 Remove CONFIG_REGNUM initial value handling code.
13691 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
13692 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
13693 (emit_set_fp_mode, epiphany_mode_after): Likewise.
13694 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
13695 Don't return 1 for FP_MODE_NONE.
13696 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
13697 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
13698 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
13699 * config/epiphany/epiphany.md (save_config): New pattern.
13701 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
13703 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
13705 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
13707 * config/i386/i386.md (memory): Handle sseishft1.
13708 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
13709 (*vec_extractv2di_1): Ditto.
13711 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
13713 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
13716 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
13718 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
13719 Add mthumb/march=armv7-a multilib.
13720 Add mthumb/march=armv7-r multilib.
13721 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
13723 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
13725 * config/v850/t-rtems: Add more multilibs.
13727 2013-05-10 Richard Biener <rguenther@suse.de>
13729 PR tree-optimization/57214
13730 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
13731 not propagate from SSA names that occur in abnormal PHI nodes.
13733 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
13735 * stor-layout.c (element_precision): New function.
13736 * machmode.h (element_precision): Declare it.
13737 * tree.c (build_minus_one_cst): New function.
13738 (element_precision): Likewise.
13739 * tree.h (build_minus_one_cst): Declare new function.
13740 (element_precision): Likewise.
13741 * fold-const.c (operand_equal_p): Use element_precision.
13742 (fold_binary_loc): Handle vector types.
13743 * convert.c (convert_to_integer): Use element_precision.
13744 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
13747 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
13749 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
13750 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
13751 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
13752 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
13753 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
13754 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
13755 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
13756 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
13757 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
13758 (Uuw8): New constraints.
13759 (Usb4): Move into alphabetical order.
13760 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
13761 (sd8_operand, ub8_operand, uw8_operand): New predicates.
13762 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
13763 previously unnamed patterns.
13764 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
13765 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
13766 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
13767 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
13768 of set_attr_alternative/if_then_else. Use extended_mips16 instead
13769 of specific lengths.
13771 2013-05-10 Jakub Jelinek <jakub@redhat.com>
13773 * config/i386/i386.md (rotateinv): New code attr.
13774 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
13775 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
13776 roll $31, %eax, etc.
13778 PR tree-optimization/45216
13779 PR tree-optimization/57157
13780 * tree-ssa-forwprop.c (simplify_rotate): New function.
13781 (ssa_forward_propagate_and_combine): Call it.
13783 2013-05-10 Richard Biener <rguenther@suse.de>
13785 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
13786 disable peeling when we version for aliasing.
13787 (vector_alignment_reachable_p): Honor explicit user alignment.
13788 (vect_supportable_dr_alignment): Likewise.
13789 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
13790 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
13791 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
13792 then peeling to arrange for the cost-model check to come first.
13794 2013-05-10 Alan Modra <amodra@gmail.com>
13796 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
13797 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
13798 * configure: Regenerate.
13800 2013-05-10 Alan Modra <amodra@gmail.com>
13803 * varasm.c (default_elf_select_section): Move !DECL_P check..
13804 (get_named_section): ..to here before calling get_section_name.
13806 (default_section_type_flags): Add DECL_P check.
13807 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
13808 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
13810 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
13812 * config/epiphany/epiphany.c (epiphany_expand_prologue):
13813 When using gen_stack_adjust_str with a register offset, add a
13814 REG_FRAME_RELATED_EXPR note.
13816 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
13818 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
13819 (*vec_extractv4si_zext_mem): Ditto.
13820 (*vec_extractv2di): Add 0->x and x->x alternatives.
13821 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
13822 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
13824 2013-05-09 Jason Merrill <jason@redhat.com>
13826 N3639 C++1y VLA support
13827 * gimplify.c (gimplify_vla_decl): Don't touch an existing
13830 * tree.c (build_constructor_va): New.
13831 * tree.h: Declare it.
13833 2013-05-09 Martin Jambor <mjambor@suse.cz>
13836 * gimple-fold.c (canonicalize_constructor_val): Call
13837 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
13839 2013-05-09 Jan Hubicka <jh@suse.cz>
13840 Richard Biener <rguenther@suse.de>
13843 * symtab.c (symtab_make_decl_local): Do not add private names.
13845 2013-05-09 Jan Hubicka <jh@suse.cz>
13848 * symtab.c (insert_to_assembler_name_hash): Handle clones.
13849 (unlink_from_assembler_name_hash): Likewise.
13850 (symtab_prevail_in_asm_name_hash, symtab_register_node,
13851 symtab_unregister_node, symtab_initialize_asm_name_hash,
13852 change_decl_assembler_name): Update.
13854 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
13856 * config/aarch64/aarch64.md: New movtf split.
13857 (*movtf_aarch64): Update.
13858 (aarch64_movdi_tilow): Handle TF modes and rename to
13859 aarch64_movdi_<mode>low.
13860 (aarch64_movdi_tihigh): Handle TF modes and rename to
13861 aarch64_movdi_<mode>high
13862 (aarch64_movtihigh_di): Handle TF modes and rename to
13863 aarch64_mov<mode>high_di
13864 (aarch64_movtilow_di): Handle TF modes and rename to
13865 aarch64_mov<mode>low_di
13866 (aarch64_movtilow_tilow): Remove spurious whitespace.
13867 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
13869 (aarch64_print_operand): Update.
13871 2013-05-09 Alan Modra <amodra@gmail.com>
13873 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
13875 * configure: Regenerate.
13877 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
13879 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
13880 splitter preparation statements.
13881 * config/i386/sse.md (*vec_extract* splitters): Ditto.
13882 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
13885 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13887 * gimple-ssa-strength-reduction.c (count_candidates): Change
13888 return value to int.
13889 (analyze_candidates_and_replace): Change type of length to int.
13891 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
13893 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
13894 (*vec_extract<mode>): Use VI12_128 mode iterator.
13895 (*vec_extract<mode>_mem): Ditto.
13896 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
13899 2013-05-08 Diego Novillo <dnovillo@google.com>
13904 2012-08-17 Diego Novillo <dnovillo@google.com>
13907 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
13908 * config.in: Regenerate.
13909 * configure: Regenerate.
13910 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
13912 2013-05-08 Jan Hubicka <jh@suse.cz>
13915 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
13916 * cgraph.h (symtab_node_base): Add unique_name.
13917 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
13918 input_overwrite_node, input_varpool_node): Stream unique_name.
13919 * cgraphclones.c (cgraph_create_virtual_clone,
13920 cgraph_function_versioning): Set unique_name.
13921 * ipa.c (function_and_variable_visibility): Set unique_name.
13923 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13925 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
13926 (alloc_cand_and_find_basis): Restrict conditional candidate
13927 processing to CAND_MULTs.
13929 2013-05-08 Jan Hubicka <jh@suse.cz>
13932 lto-symtab.c (lto_symtab_symbol_p): New function.
13933 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
13934 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
13935 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
13936 Skip static symbols.
13938 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
13940 PR tree-optimization/57200
13941 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
13942 Only call inform if the preceding warning_at returns true.
13944 2013-05-07 Han Shen <shenhan@google.com>
13946 * cfgexpand.c (record_or_union_type_has_array_p): New function.
13947 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
13948 * common.opt (fstack-protector-strong): New option.
13949 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
13950 * doc/invoke.texi (Optimization Options): Document
13951 "-fstack-protector-strong".
13952 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
13954 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
13956 * config/mips/mips.c (mips_machine_reorg2): Return 0.
13958 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
13960 * ira.c (update_equiv_regs): Add insn having equiv memory even if
13961 it is not lhs of the insn.
13962 (setup_reg_equiv): Remove insn having equiv memory which it is not
13964 * lra-constraints.c (process_address): Try to improve generation
13965 code for address base + disp.
13966 (lra_constraints): Make correct the code for checking insn setting
13967 up backward equivalence. Remove insn only if it is in the init
13969 * lra-eliminations.c (update_reg_eliminate): Change return value.
13970 (lra_eliminate): Use the result.
13972 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
13974 * config/i386/sse.md (ssescalarnummask): New mode attribute.
13975 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
13976 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
13977 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
13978 register target operands.
13979 (*vec_extractv8hi_sse2): New pattern.
13980 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
13981 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
13982 (*vec_extract<mode>_mem): New insn and split pattern.
13984 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
13986 * config/arm/arm.c (arm_asan_shadow_offset): New function.
13987 (TARGET_ASAN_SHADOW_OFFSET): Define.
13988 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
13989 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
13991 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13993 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
13994 (incr_vec_index): Return -1 if increment not found.
13995 (create_add_on_incoming_edge): Assert if increment not found.
13996 (record_increment): Limit number of increments recorded.
13997 (all_phi_incrs_profitable): Return false if an increment not found.
13998 (replace_profitable_candidates): Don't process increments that were
14000 (analyze_candidates_and_replace): Limit size of incr_vec.
14002 2013-05-07 Richard Biener <rguenther@suse.de>
14004 * calls.c (special_function_p): setjmp-like functions are leaf.
14005 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
14006 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
14008 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
14010 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
14011 (aarch64_simd_mov<mode>): New expander.
14012 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
14013 (aarch64_simd_mov_to_<mode>high): Likewise.
14014 (aarch64_simd_mov_from_<mode>low): Likewise.
14015 (aarch64_simd_mov_from_<mode>high): Likewise.
14016 (aarch64_dup_lane<mode>): Update.
14017 (aarch64_dup_lanedi): New instruction pattern.
14018 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
14019 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
14021 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14023 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
14024 (replace_mult_candidate): Remove unnecessary argument; remove
14025 unnecessary parameter from call to introduce_cast_before_cand.
14026 (replace_unconditional_candidate): Remove unnecessary parameter
14027 from call to replace_mult_candidate.
14028 (replace_conditional_candidate): Likewise.
14029 (insert_initializers): Use make_temp_ssa_name.
14030 (introduce_cast_before_cand): Remove unnecessary argument; use
14031 make_temp_ssa_name.
14032 (replace_one_candidate): Remove unnecessary argument; remove
14033 unnecessary parameter from calls to introduce_cast_before_cand.
14034 (replace_profitable_candidates): Remove unnecessary parameters
14035 from calls to replace_one_candidate.
14037 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14039 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
14040 phi def as possibly hiding a basis for a CAND_ADD whose operands
14041 have been commuted in the analysis.
14042 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
14044 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
14046 * config/aarch64/aarch64.md
14047 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
14048 shift value between 0-4.
14050 2013-05-07 Richard Biener <rguenther@suse.de>
14052 * double-int.h (rshift): New overload.
14053 * double-int.c (rshift): New function.
14054 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
14055 (create_reference_ops_from_ref): Remove.
14056 (vn_reference_insert): Use shared ops for constructing the
14057 reference and copy it.
14059 2013-05-07 Richard Biener <rguenther@suse.de>
14061 PR middle-end/57190
14062 * tree-eh.c (sink_clobbers): Properly propagate
14063 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
14065 2013-05-07 Jakub Jelinek <jakub@redhat.com>
14067 PR tree-optimization/57149
14068 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
14069 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
14070 collect_phi_def_edges, execute_late_warn_uninitialized): Use
14071 uninit_undefined_value_p instead of ssa_undefined_value_p.
14074 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
14075 for modifier == EXPAND_INITIALIZER.
14077 2013-05-07 Anton Blanchard <anton@samba.org>
14079 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
14080 for powerpc64 little endian.
14081 * configure: Regenerate.
14083 2013-05-06 Graham Stott <grahams@btinternet.com>
14085 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
14086 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
14087 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
14088 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
14090 2013-05-06 Graham Stott <grahams@btinternet.com>
14092 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
14093 codes which allow non-lvalues.
14095 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
14097 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
14098 components are all 1s.
14099 (integer_minus_onep): New function.
14100 * tree.h (integer_minus_onep): Declare it.
14101 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
14102 integer_minus_onep instead of integer_all_onesp.
14104 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
14107 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
14108 variations of these patterns.
14110 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
14112 * config/i386/i386.md (isa): Add x64_sse4 member.
14113 (enabled): Handle x64_sse4.
14114 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
14115 instruction for 64bit SSE4_1 targets. Update insn attributes.
14116 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
14117 instruction for SSE4_1 targets. Update insn attributes.
14118 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
14119 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
14120 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
14122 (*vec_extractv4si): Rename from *sse4_1_pextrd.
14123 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
14124 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
14126 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
14129 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
14131 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
14133 * final.c (do_assembler_dialects): Don't handle curly braces and
14134 vertical bar escaped by % as dialect delimiters.
14135 (output_asm_insn): Print curly braces and vertical bar if escaped
14136 by % and ASSEMBLER_DIALECT defined.
14137 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
14138 * doc/tm.texi: Regenerated.
14140 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
14142 * config/mips/mips.c: Include tree-pass.h.
14143 (mips_reorg): Split in pre- and post-dbr_schedule parts.
14144 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
14145 (pass_mips_machine_reorg2): New machine specific pass.
14146 (insert_pass_mips_machine_reorg2): New pass plugin definition.
14147 (mips_option_override): Register the new pass.
14148 * rtl.h (cleanup_barriers): Remove prototype.
14149 (dbr_schedule): Likewise.
14150 * jump.c (cleanup_barriers): Make static.
14151 * reorg.c (dbr_schedule): Likewise.
14153 2013-05-06 Richard Biener <rguenther@suse.de>
14155 PR tree-optimization/57185
14156 * tree-parloops.c (add_field_for_reduction): Handle anonymous
14157 SSA names properly.
14159 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
14162 * config/i386/i386.c (add_parameter_dependencies): Add dependence
14163 between "first_arg" and "insn", not "last" and "insn".
14165 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14167 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
14168 (find_candidates_in_block): Re-enable slsr_process_phi.
14169 (create_phi_basis): Fix double counting of candidate adjustment.
14171 2013-05-06 Richard Biener <rguenther@suse.de>
14173 PR middle-end/57147
14174 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
14175 the edge is also fallthru, preserve it and just clear the
14177 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
14178 also complex, preserve that and just clear the fallthru flag.
14179 * tree-inline.c (update_ssa_across_abnormal_edges): Also
14180 update virtual operands.
14182 2013-05-06 Alan Modra <amodra@gmail.com>
14184 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
14185 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
14186 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
14187 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
14188 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
14189 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
14191 2013-05-06 Alan Modra <amodra@gmail.com>
14193 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
14194 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
14195 (DEFAULT_ASM_ENDIAN): Define.
14196 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
14197 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
14198 Update -K PIC clause from sysv4.h.
14199 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
14200 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
14202 2013-05-06 Alan Modra <amodra@gmail.com>
14204 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
14205 twice for little-endian.
14206 (ashrdi3_no_power, ashrdi3): Support little-endian.
14208 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
14211 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
14212 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
14214 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
14215 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
14216 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
14219 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
14220 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
14222 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
14223 * config/pa/pa.opt: Make mbig-switch a no-op.
14224 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
14225 (CASE_VECTOR_MODE): Always return SImode.
14226 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
14227 for the !TARGET_BIG_SWITCH case.
14228 * config/pa/pa-linux.h: Likewise.
14229 * config/pa/pa-openbsd.h: Likewise.
14230 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
14231 * config/pa/pa.md (short_jump): Remove define_insn.
14232 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
14233 (casesi0): Remove define_insn.
14234 (type): Remove btable_branch.
14235 (pa_combine_type): Likewise.
14236 (in_nullified_branch_delay): Likewise.
14237 (in_call_delay): Likewise.
14238 (define_delay): Likewise.
14239 (define_insn_reservation "Z3"): Likewise.
14240 (define_insn_reservation "Z4"): Likewise.
14241 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
14242 (pa_adjust_insn_length): Remove adjustment for btable branches.
14243 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
14246 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
14248 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
14249 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
14250 Add m->r,x alternatives.
14251 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
14252 splitters using SWI48x mode iterator.
14253 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
14254 TARGET_64BIT. Add m->x alternative.
14255 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
14256 Add o->x alternative. Enable for TARGET_SSE.
14257 (sse_storeq): Remove expander.
14258 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
14259 with memory input operand.
14260 (*vec_extractv2di_1 splitter): New.
14261 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
14262 * config/i386/i386.md (ssevecmodelower): New mode attribute.
14264 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
14266 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
14267 (INT_LOWPART): Delete.
14268 (extract_MB): Adjust.
14269 (extract_ME): Adjust.
14270 (print_operand): Adjust.
14272 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
14274 * config/rs6000/predicates.md (reg_or_add_cint_operand,
14275 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
14276 (reg_or_logical_cint_operand, easy_fp_constant,
14277 logical_const_operand): Delete "CONST_DOUBLE" case.
14278 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
14279 "HOST_BITS_PER_WIDE_INT == 64" test.
14280 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
14281 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
14282 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
14283 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
14285 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
14286 CONST_DOUBLE DImode/VOIDmode case.
14287 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
14288 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
14289 CONST_DOUBLE VOIDmode case.
14290 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
14291 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
14292 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
14293 Delete CONST_DOUBLE case.
14294 (splitters for mov FMOVE64 const_double): Delete
14295 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
14296 "HOST_BITS_PER_WIDE_INT >= 64" test.
14297 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
14299 (mov DI const_double): Delete.
14301 2013-05-04 Jakub Jelinek <jakub@redhat.com>
14303 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
14304 on op shows all bits zero in mode of a lowpart subreg, return zero.
14306 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
14309 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
14310 to save TFmode registers and DImode to save TImode registers for
14311 caller save operations.
14312 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
14313 mark being partially clobbered since they only use the first
14316 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
14317 and TDmode only use the upper 64-bits of each VSX register.
14319 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14321 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
14322 (find_candidates_in_block): Disable slsr_process_phi.
14324 2013-05-03 Guozhi Wei <carrot@google.com>
14326 * coverage.c (coverage_obj_init): Move the construction of gcov
14328 (build_init_ctor): ... here.
14330 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14332 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
14333 (slsr_cand_d): Redefine def_phi.
14334 (stride_status, phi_adjust_status, count_phis_status): New enums.
14335 (find_phi_def): New.
14336 (find_basis_for_base_expr): New.
14337 (find_basis_for_candidate): Handle hidden bases.
14338 (alloc_cand_and_find_basis): Handle phi candidates.
14339 (slsr_process_phi): New.
14340 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
14341 (create_mul_imm_cand): Likewise.
14342 (create_add_ssa_cand): Exclude phi base candidates.
14343 (create_add_imm_cand): Likewise.
14344 (slsr_process_cast): Likewise.
14345 (slsr_process_copy): Likewise.
14346 (find_candidates_in_block): Handle phi candidates.
14347 (dump_candidate): Likewise.
14348 (unconditional_cands): Delete.
14349 (unconditional_cands_with_known_stride_p): Delete.
14350 (phi_dependent_cand_p): New.
14351 (cand_increment): Handle phi-dependent candidates.
14352 (replace_dependent): Delete.
14353 (replace_mult_candidate): New.
14354 (replace_unconditional_candidate): New.
14355 (incr_vec_index): Move to avoid forward reference.
14356 (create_add_on_incoming_edge): New.
14357 (create_phi_basis): New.
14358 (replace_dependents): Delete.
14359 (replace_conditional_candidate): New.
14360 (phi_add_costs): New.
14361 (replace_uncond_cands_and_profitable_phis): New.
14362 (record_increment): Handle phi adjustments.
14363 (record_phi_increments): New.
14364 (record_increments): Handle phi adjustments.
14365 (phi_incr_cost): New.
14366 (lowest_cost_path): Handle phis.
14367 (total_savings): Likewise.
14368 (analyze_increments): Likewise.
14369 (ncd_with_phi): New.
14370 (ncd_of_cand_and_phis): New.
14371 (nearest_common_dominator_for_cands): Handle phi increments.
14372 (all_phi_incrs_profitable): New.
14373 (replace_profitable_candidates): Handle phi-dependent candidates.
14374 (analyze_candidates_and_replace): Likewise.
14376 2013-05-03 Teresa Johnson <tejohnson@google.com>
14379 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
14380 do not exceed REG_BR_PROB_BASE.
14382 2013-05-03 Jeff Law <law@redhat.com>
14384 PR tree-optimization/57144
14385 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
14386 operand of the condition will bit into the new type when eliminating
14387 a cast feeding a condition.
14389 2013-05-03 Jakub Jelinek <jakub@redhat.com>
14391 PR rtl-optimization/57130
14392 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
14393 of COMPARE as in_code to the recursive call if needed.
14395 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
14397 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
14398 (enabled): Handle new members.
14399 * config/i386/sse.md (*vec_concatv2si): Merge from
14400 *vec_concatv2si_sse2 and vec_concatv2si_sse.
14401 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
14403 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
14405 PR tree-optimization/57027
14406 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
14407 for fnms opportunity, check we got the prerequisite kind
14408 of tree / gimple before using accessor functions.
14410 2013-05-03 Richard Biener <rguenther@suse.de>
14412 * double-int.h (lshift): New overload without precision
14413 and arith argument.
14414 (operator *=, operator +=, operator -=): Move ...
14415 * double-int.c (operator *=, operator +=, operator -=): ... here
14416 and implement more efficiently.
14417 (mul_double_with_sign): Remove.
14418 (lshift_double): Adjust to take unsinged shift argument, push
14419 dispatching code to callers.
14420 (mul_double_wide_with_sign): Add early out for callers that
14421 are not interested in high parts or overflow.
14422 (lshift): New function.
14423 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
14424 dispatch code here.
14425 (lrotate, rrotate): Use logical shifts.
14426 * expr.c (get_inner_reference): Use lshift.
14427 * fixed-value.c (do_fixed_divide): Likewise.
14428 * tree-dfa.c (get_ref_base_and_extent): Likewise.
14429 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
14430 (indirect_refs_may_alias_p): Likewise.
14431 (stmt_kills_ref_p_1): Likewise.
14433 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
14435 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
14437 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
14439 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
14440 scalar form of FABD instruction.
14442 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
14444 * lra-constraints.c (process_alt_operands): Add checking alt
14445 number to choose the best alternative.
14447 2013-05-02 Richard Biener <rguenther@suse.de>
14449 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
14450 bitmap and its handling.
14451 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
14453 2013-05-02 Richard Biener <rguenther@suse.de>
14455 PR middle-end/57140
14456 * tree-inline.c (copy_loops): Properly handle removed loops.
14457 (copy_cfg_body): Mark destination loops for fixup if source
14458 loops needed fixup.
14460 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
14463 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
14464 generating simple_return for naked functions.
14466 2013-05-02 Martin Jambor <mjambor@suse.cz>
14468 PR middle-end/56988
14469 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
14470 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
14472 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
14473 ipa_agg_replacement_value structures.
14474 (known_aggs_to_agg_replacement_list): Likewise.
14475 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
14476 (read_agg_replacement_chain): Likewise.
14477 (ipcp_transform_function): Also check that by_ref flags match.
14479 2013-05-02 Richard Biener <rguenther@suse.de>
14481 * graphds.h (struct graph): Add obstack member.
14482 * graphds.c (new_graph): Initialize obstack and allocate
14484 (add_edge): Allocate edge from the obstack.
14485 (free_graph): Free the obstack instead of all edges and vertices.
14487 2013-05-02 Teresa Johnson <tejohnson@google.com>
14489 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
14491 * cfg.c (update_bb_profile_for_threading): Ditto.
14492 * tree-inline.c (copy_bb): Ditto.
14493 (copy_edges_for_bb): Ditto.
14494 (initialize_cfun): Ditto.
14495 (copy_cfg_body): Ditto.
14496 (expand_call_inline): Ditto.
14497 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
14498 (estimate_node_size_and_time): Ditto.
14499 (inline_merge_summary): Ditto.
14500 * cgraphclones.c (cgraph_clone_edge): Ditto.
14501 (cgraph_clone_node): Ditto.
14502 * sched-rgn.c (compute_dom_prob_ps): Ditto.
14503 (compute_trg_info): Ditto.
14505 2013-05-02 Ian Bolton <ian.bolton@arm.com>
14507 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
14508 S reg when fp attribute set.
14509 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
14511 2013-05-02 Ian Bolton <ian.bolton@arm.com>
14513 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
14515 (*and_one_cmplsi3_compare0_uxtw): Likewise.
14516 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
14517 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
14519 2013-05-02 Richard Biener <rguenther@suse.de>
14521 * tree-scalar-evolution.c (scev_info_hasher): Remove.
14522 (struct instantiate_cache_entry): New type.
14523 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
14524 (struct instantiate_cache_type): New type.
14525 (set_instantiated_value, get_instantiated_value): Remove.
14526 (get_instantiated_value_entry): New function.
14527 (instantiate_scev_name): Use the new cache and adjust.
14528 (instantiate_scev_poly): Adjust.
14529 (instantiate_scev_binary): Likewise.
14530 (instantiate_array_ref): Likewise.
14531 (instantiate_scev_convert): Likewise.
14532 (instantiate_scev_not): Likewise.
14533 (instantiate_scev_3): Likewise.
14534 (instantiate_scev_2): Likewise.
14535 (instantiate_scev_r): Likewise.
14536 (instantiate_scev): Likewise.
14537 (resolve_mixers): Likewise.
14539 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
14542 * lra-constraints.c (best_small_class_operands_num): Remove.
14543 (process_alt_operands): Remove small_class_operands_num. Take
14544 small classes operands into losers and only if the operand is not
14545 matched. Modify debugging output.
14546 (curr_insn_transform): Remove best_small_class_operands_num.
14549 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14551 * config/aarch64/aarch64-builtins.c
14552 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
14553 * config/aarch64/aarch64-simd-builtins.def
14554 (reduc_splus_): Add new modes.
14555 (reduc_uplus_): New.
14556 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
14557 (reduc_uplus_v4sf): Likewise.
14558 (reduc_splus_v4sf): Likewise.
14559 (aarch64_addv<mode>): Likewise.
14560 (reduc_uplus_<mode>): Likewise.
14561 (reduc_splus_<mode>): Likewise.
14562 (aarch64_addvv2di): Likewise.
14563 (reduc_uplus_v2di): Likewise.
14564 (reduc_splus_v2di): Likewise.
14565 (aarch64_addvv2si): Likewise.
14566 (reduc_uplus_v2si): Likewise.
14567 (reduc_splus_v2si): Likewise.
14568 (reduc_<sur>plus_<mode>): New.
14569 (reduc_<sur>plus_v2di): Likewise.
14570 (reduc_<sur>plus_v2si): Likewise.
14571 (reduc_<sur>plus_v4sf): Likewise.
14572 (aarch64_addpv4sf): Likewise.
14573 * config/aarch64/arm_neon.h
14574 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
14575 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
14576 add UNSPEC_SADDV, UNSPEC_UADDV.
14578 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
14580 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14582 * config/aarch64/arm_neon.h
14583 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
14585 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14587 * config/aarch64/aarch64-builtins
14588 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
14590 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14592 * config/aarch64/aarch64-simd-builtins.def
14593 (reduc_smax_): New.
14594 (reduc_smin_): Likewise.
14595 (reduc_umax_): Likewise.
14596 (reduc_umin_): Likewise.
14597 (reduc_smax_nan_): Likewise.
14598 (reduc_smin_nan_): Likewise.
14601 (smax): Update for V2SF, V4SF and V2DF modes.
14604 (smin_nan): Likewise.
14605 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
14606 (<su><maxmin><mode>3): ...This, refactor.
14607 (s<maxmin><mode>3): New.
14608 (<maxmin_uns><mode>3): Likewise.
14609 (reduc_<maxmin_uns>_<mode>): Refactor.
14610 (reduc_<maxmin_uns>_v4sf): Likewise.
14611 (reduc_<maxmin_uns>_v2si): Likewise.
14612 (aarch64_<fmaxmin><mode>: Remove.
14613 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
14615 (vmin<q>_f<32,64>): Likewise.
14616 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
14618 (su): Add mappings for smax, smin, umax, umin.
14620 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
14621 (FMAXMIN): Rename as...
14622 (FMAXMIN_UNS): ...This.
14624 (fmaxminv): Likewise.
14625 (fmaxmin): Likewise.
14627 (maxmin_uns_op): Likewise.
14629 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14631 * config/aarch64/arm_neon.h
14632 (vac<ge, gt><sd>_f<32, 64>): Rename to...
14633 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
14634 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
14636 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14638 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
14639 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
14641 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14643 * config/aarch64/aarch64-simd.md
14644 (vcond<mode>_internal): Handle special cases for constant masks.
14645 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
14646 (vcondu<mode><mode>): Likewise.
14647 (vcond<v_cmp_result><mode>): New.
14649 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14651 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
14652 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
14653 * config/aarch64/aarch64-simd-builtins.def
14654 (cmeq): Update to BUILTIN_VALLDI.
14659 * config/aarch64/arm_neon.h
14660 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
14661 to builtins or C as appropriate.
14663 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14665 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
14667 (cmhi): Rename to...
14669 * config/aarch64/aarch64-simd.md
14670 (simd_mode): Add SF.
14671 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
14672 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
14673 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
14674 (cstore<mode>_neg): ...This.
14675 * config/aarch64/iterators.md
14677 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
14678 (COMPARISONS): New.
14679 (UCOMPARISONS): Likewise.
14680 (optab): Add missing comparisons.
14686 (VCMP_S): Likewise.
14687 (VCMP_U): Likewise.
14688 (V_cmp_result): Add DF, SF modes.
14689 (v_cmp_result): Likewise.
14691 (vmtype): Likewise.
14692 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
14694 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
14696 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
14697 define_insn to define_insn_and_split.
14698 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
14699 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
14700 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
14701 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
14702 (thumb2_negscc): Likewise.
14704 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
14706 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
14708 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
14710 * config/arm/thumb2.md: Remove trailing whitespaces.
14712 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
14714 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
14715 Use gen_int_mode rather than GEN_INT.
14717 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
14719 * value-prof.c (stream_in_histogram_value): Remove the strayed
14722 2013-04-30 Richard Biener <rguenther@suse.de>
14724 PR middle-end/57122
14725 * cfghooks.c (split_edge): Properly check for the loop latch edge.
14727 2013-04-30 Richard Biener <rguenther@suse.de>
14729 PR middle-end/57107
14730 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
14732 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
14734 PR rtl-optimization/56957
14735 PR rtl-optimization/57105
14736 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
14737 variable. Use just INSN_UID for determining whether an insn
14738 should be only disconnected from the insn stream.
14739 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
14741 2013-04-30 Jakub Jelinek <jakub@redhat.com>
14743 PR tree-optimization/57104
14744 * tsan.c (instrument_expr): Don't instrument accesses to
14745 DECL_HARD_REGISTER VAR_DECLs.
14747 2013-04-30 Richard Biener <rguenther@suse.de>
14749 * function.h (loops_for_fn): New inline function.
14750 (set_loops_for_fn): Likewise.
14751 * cfgloop.h (place_new_loop): Add struct function parameter.
14752 (get_loop): Likewise.
14753 (get_loops): Likewise.
14754 (number_of_loops): Likewise.
14755 (fel_next): Adjust.
14756 (fel_init): Likewise.
14757 * cfg.c (get_loop_copy): Adjust.
14758 * cfgloop.c (flow_loops_dump): Likewise.
14759 (record_loop_exits): Likewise.
14760 (verify_loop_structure): Likewise.
14761 * cfgloopanal.c (mark_irreducible_loops): Likewise.
14762 (estimate_reg_pressure_cost): Likewise.
14763 (mark_loop_exit_edges): Likewise.
14764 * cfgloopmanip.c (place_new_loop): Likewise.
14765 (add_loop): Likewise.
14766 (duplicate_loop): Likewise.
14767 * graph.c (draw_cfg_nodes): Likewise.
14768 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
14769 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
14770 (extract_affine_chrec): Likewise.
14771 (build_scop_iteration_domain): Likewise.
14772 * graphite.c (graphite_initialize): Likewise.
14773 * ira-build.c (create_loop_tree_nodes): Likewise.
14774 (more_one_region_p): Likewise.
14775 (rebuild_regno_allocno_maps): Likewise.
14776 (mark_loops_for_removal): Likewise.
14777 (mark_all_loops_for_removal): Likewise.
14778 (remove_unnecessary_regions): Likewise.
14779 (ira_build): Likewise.
14780 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
14781 * loop-init.c (fix_loop_structure): Likewise.
14782 (gate_rtl_move_loop_invariants): Likewise.
14783 (gate_rtl_unswitch): Likewise.
14784 (gate_rtl_unroll_and_peel_loops): Likewise.
14785 (rtl_doloop): Likewise.
14786 * lto-streamer-in.c (input_cfg): Likewise.
14787 * lto-streamer-out.c (output_cfg): Likewise.
14788 * modulo-sched.c (sms_schedule): Likewise.
14789 * predict.c (tree_estimate_probability): Likewise.
14790 (tree_estimate_probability_driver): Likewise.
14791 (estimate_loops): Likewise.
14792 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
14793 (move_sese_region_to_fn): Likewise.
14794 (debug_loop_num): Likewise.
14795 * tree-chrec.c (chrec_evaluate): Likewise.
14796 (hide_evolution_in_other_loops_than_loop): Likewise.
14797 (chrec_component_in_loop_num): Likewise.
14798 (reset_evolution_in_loop): Likewise.
14799 (evolution_function_is_invariant_rec_p): Likewise.
14800 * tree-if-conv.c (main_tree_if_conversion): Likewise.
14801 * tree-inline.c (copy_loops): Likewise.
14802 (copy_cfg_body): Likewise.
14803 (tree_function_versioning): Likewise.
14804 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
14805 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
14807 (add_to_evolution_1): Likewise.
14808 (scev_const_prop): Likewise.
14809 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
14810 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
14811 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
14812 (tree_ssa_lim_initialize): Likewise.
14813 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
14814 (verify_loop_closed_ssa): Likewise.
14815 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
14816 (tree_ssa_loop_im): Likewise.
14817 (tree_ssa_loop_unswitch): Likewise.
14818 (tree_vectorize): Likewise.
14819 (check_data_deps): Likewise.
14820 (tree_ssa_loop_ivcanon): Likewise.
14821 (tree_ssa_loop_bounds): Likewise.
14822 (tree_complete_unroll): Likewise.
14823 (tree_complete_unroll_inner): Likewise.
14824 (tree_parallelize_loops): Likewise.
14825 (tree_ssa_loop_prefetch): Likewise.
14826 (tree_ssa_loop_ivopts): Likewise.
14827 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
14828 * tree-vectorizer.c (vectorize_loops): Likewise.
14830 2013-04-29 Mike Frysinger <vapier@gentoo.org>
14832 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
14833 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
14834 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
14835 with EABI_LINK_SPEC.
14837 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
14840 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
14843 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
14846 * lra-constraints.c (process_alt_operands): Discourage a bit more
14847 using memory for pseudos. Print cost dump for alternatives.
14848 Modify cost values for conflicts with early clobbers.
14849 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
14851 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
14854 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
14856 2013-04-29 Ian Bolton <ian.bolton@arm.com>
14858 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
14859 from/to S register.
14860 (movdi_aarch64): Support LDR/STR from/to D register.
14862 2013-04-29 Ian Bolton <ian.bolton@arm.com>
14864 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
14865 or higher by default.
14867 2013-04-29 Richard Biener <rguenther@suse.de>
14869 PR middle-end/57075
14870 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
14871 even if not adding abnormal edges for calls that can make
14874 2013-04-29 Richard Biener <rguenther@suse.de>
14876 PR middle-end/57103
14877 * tree-cfg.c (move_stmt_op): Fix condition under which to update
14879 (move_stmt_r): Remove redundant checking.
14881 2013-04-29 Teresa Johnson <tejohnson@google.com>
14884 * basic-block.h (apply_scale): New function.
14885 (apply_probability): Use apply_scale.
14886 * gimple-streamer-in.c (input_bb): Ditto.
14887 * lto-streamer-in.c (input_cfg): Ditto.
14888 * lto-cgraph.c (merge_profile_summaries): Ditto.
14889 * tree-optimize.c (execute_fixup_cfg): Ditto.
14890 * tree-inline.c (copy_bb): Update comment to use apply_scale.
14891 (copy_edges_for_bb): Ditto.
14892 (copy_cfg_body): Ditto.
14894 2013-04-29 Tom de Vries <tom@codesourcery.com>
14896 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
14897 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
14898 (tail_merge_optimize): Handle current_loops == NULL.
14900 2013-04-26 Jeff Law <law@redhat.com>
14902 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
14903 (simplify_cond_using_ranges): Generalize code to simplify
14904 COND_EXPRs where one argument is a constant and the other
14905 is an SSA_NAME created by an integral type conversion.
14907 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14909 * config/arm/arm.md (store_minmaxsi): Use only when
14910 optimize_insn_for_size_p.
14912 2013-04-29 Christian Bruel <christian.bruel@st.com>
14915 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
14917 2013-04-29 Richard Biener <rguenther@suse.de>
14919 PR middle-end/57089
14920 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
14921 loop tree make sure to schedule a fixup for the child as well.
14922 (expand_omp_for_generic): Properly add loops.
14923 (expand_omp_for_static_nochunk): Likewise.
14924 (expand_omp_for_static_chunk): Likewise.
14925 (expand_omp_for): For the degenerate case fixup loops.
14926 (expand_omp_sections): Fix default bb placement in loops.
14927 (expand_omp_atomic_pipeline): Properly add loops.
14929 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14931 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
14933 2013-04-29 Tom de Vries <tom@codesourcery.com>
14935 * tree-ssa-tail-merge.c: Update header comment.
14937 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14939 * config/aarch64/arm_neon.h
14940 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
14941 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
14942 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
14943 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
14944 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
14945 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
14946 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
14947 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
14949 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14951 * config/aarch64/aarch64-simd.md
14952 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
14953 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
14954 fix_trunc, fixuns_trunc.
14955 (ftrunc<VDQF:mode>2): New.
14956 * config/aarch64/iterators.md (optab): Add fix, fixuns.
14957 (fix_trunc_optab): New.
14959 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14961 * config/aarch64/aarch64-builtins.c
14962 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
14963 iceilf, lround, iroundf.
14965 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
14968 * config/i386/i386.h (enum ix86_tune_indices)
14969 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
14970 New, split from X86_TUNE_INTER_UNIT_MOVES.
14971 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
14972 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
14973 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
14974 (TARGET_INTER_UNIT_MOVES): Remove.
14975 * config/i386/i386.c (initial_ix86_tune_features): Update.
14976 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
14977 (ix86_expand_convert_uns_didf_sse): Use
14978 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
14979 (ix86_expand_vector_init_one_nonzero): Ditto.
14980 (ix86_expand_vector_init_interleave): Ditto.
14981 (inline_secondary_memory_needed): Return true for moves from SSE class
14982 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
14983 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
14984 * config/i386/constraints.md (Yi, Ym): Depend on
14985 TARGET_INTER_UNIT_MOVES_TO_VEC.
14986 (Yj, Yn): New constraints.
14987 * config/i386/i386.md (*movdi_internal): Change constraints of
14988 operand 1 from Yi to Yj and from Ym to Yn.
14989 (*movsi_internal): Ditto.
14990 (*movdf_internal): Ditto.
14991 (*movsf_internal): Ditto.
14992 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
14993 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
14994 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
14995 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
14996 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
14997 * config/i386/sse.md (movdi_to_sse): Ditto.
14998 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
14999 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
15000 TARGET_INTER_UNIT_MOVES.
15001 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
15002 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
15003 instead of TARGET_INTER_UNIT_MOVES.
15004 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
15005 operand 1 from Yi to Yj and from Ym to Yn.
15007 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
15009 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
15010 (float_truncate_hi_): Likewise.
15011 (float_extend_lo_): Likewise.
15012 (float_truncate_lo_): Likewise.
15013 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
15014 (aarch64_float_extend_lo_v2df): Likewise.
15015 (vec_unpacks_hi_v4sf): Likewise.
15016 (aarch64_float_truncate_lo_v2sf): Likewise.
15017 (aarch64_float_truncate_hi_v4sf): Likewise.
15018 (vec_pack_trunc_v2df): Likewise.
15019 (vec_pack_trunc_df): Likewise.
15021 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
15023 * config/aarch64/aarch64-builtins.c
15024 (aarch64_fold_builtin): Fold float conversions.
15025 * config/aarch64/aarch64-simd-builtins.def
15026 (floatv2si, floatv4si, floatv2di): New.
15027 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
15028 * config/aarch64/aarch64-simd.md
15029 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
15030 * config/aarch64/iterators.md (FLOATUORS): New.
15031 (optab): Add float, floatuns.
15032 (su_optab): Likewise.
15034 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
15036 * config/aarch64/aarch64-builtins.c
15037 (aarch64_builtin_vectorized_function): Use new names for
15039 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
15040 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
15041 (fcvtzu): Split as...
15042 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
15043 (fcvtas): Split as...
15044 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
15045 (fcvtau): Split as...
15046 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
15047 (fcvtps): Split as...
15048 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
15049 (fcvtpu): Split as...
15050 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
15051 (fcvtms): Split as...
15052 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
15053 (fcvtmu): Split as...
15054 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
15055 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
15056 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
15057 (lfrintnusf, lfrintnudf): Likewise.
15058 * config/aarch64/aarch64-simd.md
15059 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
15061 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
15062 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
15063 (fcvt_pattern): Likewise.
15065 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
15067 * config/aarch64/aarch64-simd.md
15068 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
15069 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
15071 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
15073 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
15074 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
15075 (vrnd<a,m,n,p>_f32): Implement using builtins.
15076 (vrnd<i,x><q>_f<32, 64>): New.
15078 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
15080 * config/aarch64/aarch64-builtins.c
15081 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
15082 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
15083 (frintz): Rename to...
15085 (frintp): Rename to...
15087 (frintm): Rename to...
15089 (frinti): Rename to...
15090 (nearbyint): ...this.
15091 (frintx): Rename to...
15093 (frinta): Rename to...
15095 * config/aarch64/aarch64-simd.md
15096 (aarch64_frint<frint_suffix><mode>): Delete.
15097 (<frint_pattern><mode>2): Convert to insn.
15098 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
15099 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
15100 (frint_pattern): Likewise.
15101 (frint_suffix): Likewise.
15103 2013-04-29 Richard Biener <rguenther@suse.de>
15105 PR tree-optimization/57081
15106 * loop-init.c: Include tree-flow.h.
15107 (loop_optimizer_finalize): Free number of iteration estimates.
15108 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
15110 2013-04-29 Jakub Jelinek <jakub@redhat.com>
15112 PR tree-optimization/57083
15113 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
15114 non-singleton shift count range, zero extend low_bound for uns case.
15116 * config/i386/predicates.md (general_vector_operand): New predicate.
15117 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
15118 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
15119 if they aren't nonimmediate operands. If their original values
15120 satisfy const_vector_equal_evenodd_p, don't shift them.
15121 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
15122 predicates. For the SSE4.1 case force operands[{1,2}] into registers
15123 if not nonimmediate_operand.
15124 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
15125 instead of register_operand.
15126 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
15128 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
15130 * stor-layout.c (finalize_size_functions): Allocate a structure and
15131 reset cfun before dumping the functions.
15133 2013-04-27 Jakub Jelinek <jakub@redhat.com>
15135 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
15138 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
15139 use xop_pmacsdqh if uns_p.
15140 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
15141 the immediate rotate count.
15143 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
15145 * rtl.h (struct rtx_def): Add comment for field jump.
15146 (LRA_SUBREG_P): New macro.
15147 * recog.c (register_operand): Check LRA_SUBREG_P.
15148 * lra.c (lra): Add note at the end of RTL code. Align non-empty
15150 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
15151 (lra_final_code_change): Skip subreg change for operators.
15152 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
15153 if there are no operand changes.
15154 * lra-constraints.c (curr_insn_set): New.
15155 (match_reload): Set LRA_SUBREG_P.
15156 (emit_spill_move): Ditto.
15157 (check_and_process_move): Use curr_insn_set. Process only single
15158 set insns. Don't initialize sec_mem_p and change_p.
15159 (simplify_operand_subreg): Use LRA_SUBREG_P.
15160 (reg_in_class_p): New function.
15161 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
15162 of #ifdef. Add code to remove cycling.
15163 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
15164 non-null disp. Reload inner instead of disp when base and index
15165 are null. Try to put lo_sum into register.
15166 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
15167 (check_and_process_move): Move code for move cost check to
15168 simple_move_p. Remove equiv_substitution.
15169 (simple_move_p): New function.
15170 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
15171 curr_insn_set. Call check_and_process_move only for single set
15172 insns. Use the new function. Move call of check_and_process_move
15173 after operand equiv substitution and address process.
15175 2013-04-26 Jakub Jelinek <jakub@redhat.com>
15178 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
15179 with nonlocal goto receivers or returns twice calls, ignore
15180 unininitialized values from abnormal edges to nl goto receiver
15181 or returns twice call.
15183 2013-04-26 Jakub Jelinek <jakub@redhat.com>
15185 PR tree-optimization/57051
15186 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
15187 and VEC_RSHIFT_EXPR if shift count is a multiple of element
15190 2013-04-26 Richard Biener <rguenther@suse.de>
15192 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
15193 (expand_omp_taskreg): Likewise. Mark loops for fixup.
15194 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
15195 (fixup_loop_arrays_after_move): New function.
15196 (move_sese_region_to_fn): Properly outline the loop tree parts
15197 of the SESE region.
15199 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
15201 * config/i386/i386.md (type, unit): Fix long lines.
15203 2013-04-26 Richard Biener <rguenther@suse.de>
15205 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
15206 (lto-streamer-out.o): Likewise.
15207 * cfgloop.c (init_loops_structure): Export, add struct function
15208 argument and adjust.
15209 (flow_loops_find): Adjust.
15210 * cfgloop.h (enum loop_estimation): Add EST_LAST.
15211 (init_loops_structure): Declare.
15212 * lto-streamer-in.c: Include cfgloop.h.
15213 (input_cfg): Input the loop tree.
15214 * lto-streamer-out.c: Include cfgloop.h.
15215 (output_cfg): Output the loop tree.
15216 (output_struct_function_base): Do not drop PROP_loops.
15218 2013-03-26 Richard Biener <rguenther@suse.de>
15220 * tree-cfg.c (execute_build_cfg): Build the loop tree.
15221 (pass_build_cfg): Provide PROP_loops.
15222 (move_sese_region_to_fn): Remove loops that are outlined into fn
15224 * tree-inline.c: Include cfgloop.h.
15225 (initialize_cfun): Do not drop PROP_loops.
15226 (copy_loops): New function.
15227 (copy_cfg_body): Copy loop structure.
15228 (tree_function_versioning): Initialize destination loop tree.
15229 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
15230 (pass_parallelize_loops): Do IL verification.
15231 * loop-init.c (loop_optimizer_init): Fixup loops if required.
15232 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
15233 the CFG make sure we fixup loops as well.
15234 * tree-ssa-tail-merge.c: Include cfgloop.h.
15235 (replace_block_by): When merging loop latches mark loops for fixup.
15236 * lto-streamer-out.c (output_struct_function_base): Drop
15237 PROP_loops for now.
15238 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
15239 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
15240 * ipa-split.c: Include cfgloop.h.
15241 (split_function): Add the new return block to the loop tree root.
15242 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
15243 whether we have removed the forwarder block.
15244 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
15245 * cfgloop.h (place_new_loop): Declare.
15246 * cfgloopmanip.c (place_new_loop): Export.
15247 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
15248 (tree-switch-conversion.o): Likewise.
15249 (tree-complex.o): Likewise.
15250 (tree-inline.o): Likewise.
15251 (tree-ssa-tailmerge.o): Likewise.
15252 (ipa-split.o): Likewise.
15253 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
15254 (tree-ssa-copy.o): Likewise.
15255 * tree-switch-conversion.c: Include cfgloop.h
15256 (process_switch): If we emit a bit-test cascade, schedule loops
15258 * tree-complex.c: Include cfgloop.h.
15259 (expand_complex_div_wide): Properly add new basic-blocks to loops.
15260 * asan.c: Include cfgloop.h.
15261 (create_cond_insert_point): Properly add new basic-blocks to
15262 loops, schedule loop fixup.
15263 * cfgloop.c (verify_loop_structure): Check that looks are not
15265 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
15267 (expand_omp_for_generic): Likewise.
15268 (expand_omp_sections): Likewise.
15269 (expand_omp_atomic_pipeline): Schedule loops for fixup.
15270 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
15271 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
15272 is initialized, not when loops are present.
15273 * tree-parloops.c (parallelize_loops): Remove checking here.
15274 * passes.c (init_optimization_passes): Schedule a copy-propagation
15275 pass before complete unrolling of inner loops.
15277 2013-04-26 Jakub Jelinek <jakub@redhat.com>
15279 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
15280 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
15281 (colorize_init): Add argument to _WIN32 version.
15282 * toplev.c: Include diagnostic-color.h.
15283 (process_options): Default to -fdiagnostics-color=auto if
15284 GCC_COLORS env var is in the environment.
15285 * common.opt (fdiagnostics-color=): Add Var and Init.
15286 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
15287 env var is in the environment, the default is auto rather than never.
15289 * diagnostic.h (file_name_as_prefix): Add context argument.
15290 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
15291 the string as locus.
15292 * langhooks.c (lhd_print_error_function): Adjust caller.
15294 2013-04-25 Lawrence Crowl <crowl@google.com>
15296 * var-tracking.c (shared_hash_def::htab):
15297 Change type to hash_table. Update dependent calls and types.
15299 2013-04-25 Lawrence Crowl <crowl@google.com>
15301 * Makefile.in: Update as needed below.
15303 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
15304 Move declaration to after the type's method definitons.
15306 * attribs.c (htab_t scoped_attributes::attribute_hash):
15307 Change type to hash_table. Update dependent calls and types.
15309 * bitmap.c (htab_t bitmap_desc_hash):
15310 Change type to hash_table. Update dependent calls and types.
15312 * cselib.c (htab_t cselib_hash_table):
15313 Change type to hash_table. Update dependent calls and types.
15315 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
15316 (hash_string_slot_node): Move implementation into lto-streamer.h
15317 struct string_slot_hasher.
15318 (eq_string_slot_node): Likewise.
15320 * data-streamer-out.c: Update output_block::string_hash_table
15321 dependent calls and types.
15323 * dwarf2cfi.c (htab_t trace_index):
15324 Change type to hash_table. Update dependent calls and types.
15326 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
15327 Change type to hash_table. Update dependent calls and types.
15328 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
15329 (htab_t optimize_external_refs::map): Likewise.
15330 (htab_t output_comp_unit::extern_map): Likewise.
15331 (htab_t output_comdat_type_unit::extern_map): Likewise.
15332 (htab_t output_macinfo::macinfo_htab): Likewise.
15333 (htab_t optimize_location_lists::htab): Likewise.
15334 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
15336 * except.c (htab_t ehspec_hash_type):
15337 Change type to hash_table. Update dependent calls and types.
15338 (assign_filter_values::ttypes): Likewise.
15339 (assign_filter_values::ehspec): Likewise.
15340 (sjlj_assign_call_site_values::ar_hash): Likewise.
15341 (convert_to_eh_region_ranges::ar_hash): Likewise.
15343 * gcse.c (htab_t pre_ldst_table):
15344 Change type to hash_table. Update dependent calls and types.
15346 * ggc-common.c (htab_t saving_htab):
15347 Change type to hash_table. Update dependent calls and types.
15348 (htab_t loc_hash): Likewise.
15349 (htab_t ptr_hash): Likewise.
15350 (call_count): Rename ggc_call_count.
15351 (call_alloc): Rename ggc_call_alloc.
15352 (loc_descriptor): Rename make_loc_descriptor.
15353 (add_statistics): Rename ggc_add_statistics.
15355 * ggc-common.c (saving_htab):
15356 Change type to hash_table. Update dependent calls and types.
15358 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
15359 (push_gimplify_context): Likewise.
15360 (pop_gimplify_context): Likewise.
15361 (struct gimple_temp_hash_elt): Added.
15362 (struct gimplify_hasher): Likewise.
15363 (struct gimplify_ctx.temp_htab):
15364 Change type to hash_table. Update dependent calls and types.
15366 * gimple-fold.c: Include gimplify-ctx.h.
15368 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
15369 Change type to hash_table. Update dependent calls and types.
15370 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
15371 avoid potential global name collision.
15373 * gimplify.c: Include gimplify-ctx.h.
15374 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
15375 (htab_t gimplify_ctx::temp_htab):
15376 Update dependent calls and types for new type hash_table.
15377 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
15378 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
15380 * gimplify-ctx.h: New.
15381 (struct gimple_temp_hash_elt): Move from gimplify.c.
15382 (class gimplify_hasher): New.
15383 (struct gimplify_ctx): Move from gimple.h.
15384 (htab_t gimplify_ctx::temp_htab):
15385 Change type to hash_table. Update dependent calls and types.
15387 * graphite-clast-to-gimple.c: Include graphite-htab.h.
15388 (htab_t ivs_params::newivs_index):
15389 Change type to hash_table. Update dependent calls and types.
15390 (htab_t ivs_params::params_index): Likewise.
15391 (htab_t print_generated_program::params_index): Likewise.
15392 (htab_t gloog::newivs_index): Likewise.
15393 (htab_t gloog::params_index): Likewise.
15395 * graphite.c: Include graphite-htab.h.
15396 4htab_t graphite_transform_loops::bb_pbb_mapping):
15397 Change type to hash_table. Update dependent calls and types.
15399 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
15400 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
15401 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
15403 * graphite-dependences.c: Include graphite-htab.h.
15404 (loop_is_parallel_p): Change hash table type of parameter.
15406 * graphite-htab.h: New.
15407 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
15408 (extern find_pbb_via_hash): Move from graphite-poly.h.
15409 (extern loop_is_parallel_p): Move from graphite-poly.h.
15410 (extern get_loop_body_pbbs): Move from graphite-poly.h.
15412 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
15413 (extern loop_is_parallel_p): Move to graphite-htab.h.
15414 (extern get_loop_body_pbbs): Move to graphite-htab.h.
15416 * haifa-sched.c (htab_t delay_htab):
15417 Change type to hash_table. Update dependent calls and types.
15418 (htab_t delay_htab_i2): Likewise.
15420 * ira-color.c (htab_t allocno_hard_regs_htab):
15421 Change type to hash_table. Update dependent calls and types.
15423 * ira-costs.c (htab_t cost_classes_htab):
15424 Change type to hash_table. Update dependent calls and types.
15426 * loop-invariant.c (htab_t merge_identical_invariants::eq):
15427 Change type to hash_table. Update dependent calls and types.
15429 * loop-iv.c (htab_t bivs):
15430 Change type to hash_table. Update dependent calls and types.
15432 * loop-unroll.c (htab_t opt_info::insns_to_split):
15433 Change type to hash_table. Update dependent calls and types.
15434 (htab_t opt_info::insns_with_var_to_expand): Likewise.
15436 * lto-streamer.h (struct string_slot): Move from data-streamer.h
15437 (struct string_slot_hasher): New.
15438 (htab_t output_block::string_hash_table):
15439 Change type to hash_table. Update dependent calls and types.
15441 * lto-streamer-in.c (freeing_string_slot_hasher): New.
15442 (htab_t file_name_hash_table):
15443 Change type to hash_table. Update dependent calls and types.
15445 * lto-streamer-out.c: Update output_block::string_hash_table dependent
15448 * lto-streamer.c (htab_t tree_htab):
15449 Change type to hash_table. Update dependent calls and types.
15451 * omp-low.c: Include gimplify-ctx.h.
15453 * passes.c (htab_t name_to_pass_map):
15454 Change type to hash_table. Update dependent calls and types.
15455 (pass_traverse): Rename to passes_pass_traverse.
15457 * plugin.c (htab_t event_tab):
15458 Change type to hash_table. Update dependent calls and types.
15460 * postreload-gcse.c (htab_t expr_table):
15461 Change type to hash_table. Update dependent calls and types.
15462 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
15464 * sese.c (debug_rename_map_1): Make extern.
15465 (htab_t copy_bb_and_scalar_dependences::rename_map):
15466 Change type to hash_table. Update dependent calls and types.
15468 * sese.h (extern debug_rename_map): Move to .c file.
15470 * store-motion.c (htab_t store_motion_mems_table):
15471 Change type to hash_table. Update dependent calls and types.
15473 * trans-mem.c (htab_t tm_new_mem_hash):
15474 Change type to hash_table. Update dependent calls and types.
15476 * tree-browser.c (htab_t TB_up_ht):
15477 Change type to hash_table. Update dependent calls and types.
15479 * tree-cfg.c (htab_t discriminator_per_locus):
15480 Change type to hash_table. Update dependent calls and types.
15482 * tree-complex.c: Include tree-hasher.h
15483 (htab_t complex_variable_components):
15484 Change type to hash_table. Update dependent calls and types.
15486 * tree-eh.c (htab_t finally_tree):
15487 Change type to hash_table. Update dependent calls and types.
15489 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
15490 struct int_tree_hasher.
15491 (extern int_tree_map_eq): Likewise.
15492 (uid_decl_map_hash): Removed.
15493 (extern decl_tree_map_eq): Likewise.
15495 * tree-hasher.h: New.
15496 (struct int_tree_hasher): New.
15497 (typedef int_tree_htab_type): New.
15499 * tree-inline.c: Include gimplify-ctx.h.
15501 * tree-mudflap.c: Include gimplify-ctx.h.
15503 * tree-parloops.c: Include tree-hasher.h.
15504 (htab_t eliminate_local_variables_stmt::decl_address):
15505 Change type to hash_table. Update dependent calls and types.
15506 (htab_t separate_decls_in_region::decl_copies): Likewise.
15508 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
15509 Change type to hash_table. Update dependent calls and types.
15511 * tree-sra.c (candidates):
15512 Change type to hash_table. Update dependent calls and types.
15514 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
15516 (int_tree_map_hash): Likewise.
15518 * tree-ssa-dom.c (htab_t avail_exprs):
15519 Change type to hash_table. Update dependent calls and types.
15521 * tree-ssa-live.c (var_map_base_init::tree_to_index):
15522 Change type to hash_table. Update dependent calls and types.
15524 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
15525 Change type to hash_table. Update dependent calls and types.
15527 * tree-ssa-phiopt.c (seen_ssa_names):
15528 Change type to hash_table. Update dependent calls and types.
15530 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
15531 Change type to hash_table. Update dependent calls and types.
15533 * tree-ssa-uncprop.c (equiv):
15534 Change type to hash_table. Update dependent calls and types.
15536 2013-04-25 Jakub Jelinek <jakub@redhat.com>
15538 PR rtl-optimization/57003
15539 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
15540 call note_stores with kill_clobbered_value callback again after
15541 killing regs_invalidated_by_call.
15543 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
15545 * config/aarch64/aarch64-simd.md
15546 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
15547 (aarch64_simd_bsl<mode>): Likewise.
15548 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
15550 2013-04-25 Marek Polacek <polacek@redhat.com>
15552 PR tree-optimization/57066
15553 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
15555 2013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
15557 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
15559 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
15561 * config/aarch64/aarch64-builtins.c
15562 (aarch64_fold_builtin): New.
15563 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
15564 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
15565 * config/aarch64/aarch64-simd-builtins.def (abs): New.
15566 * config/aarch64/arm_neon.h
15567 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
15569 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
15570 Tejas Belagod <tejas.belagod@arm.com>
15572 * config/aarch64/aarch64-builtins.c
15573 (aarch64_gimple_fold_builtin): New.
15574 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
15575 * config/aarch64/aarch64-simd-builtins.def (addv): New.
15576 * config/aarch64/aarch64-simd.md (addpv4sf): New.
15577 (addvv4sf): Update.
15578 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
15580 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
15582 * config/aarch64/aarch64.md
15583 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
15585 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
15587 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
15588 (*ngcsi_uxtw): New pattern.
15590 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15591 Julian Brown <julian@codesourcery.com>
15593 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
15594 (TB_DREG): Add T_V4HF.
15595 (v4hf_UP): New macro.
15596 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
15597 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
15598 Handle initialisation of V4HF. Adjust initialisation of reinterpret
15600 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
15601 (arm_vector_mode_supported_p): Handle V4HF.
15602 (arm_mangle_map): Handle V4HFmode.
15603 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
15604 * config/arm/arm_neon_builtins.def: Add entries for
15605 vcvtv4hfv4sf, vcvtv4sfv4hf.
15606 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
15607 (neon_vcvtv4hfv4sf): Likewise.
15608 * config/arm/neon-gen.ml: Handle half-precision floating point
15610 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
15611 * config/arm/arm_neon.h: Regenerate.
15612 * config/arm/neon.ml (type elts): Add F16.
15613 (type vectype): Add T_float16x4, T_floatHF.
15614 (type vecmode): Add V4HF.
15615 (type features): Add Requires_FP_bit feature.
15616 (elt_width): Handle F16.
15617 (elt_class): Likewise.
15618 (elt_of_class_width): Likewise.
15619 (mode_of_elt): Refactor.
15620 (type_for_elt): Handle F16, fix error messages.
15621 (vectype_size): Handle T_float16x4.
15622 (vcvt_sh): New function.
15623 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
15624 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
15625 (string_of_mode): Handle V4HF.
15626 * doc/arm-neon-intrinsics.texi: Regenerate.
15628 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
15630 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
15631 format specifier in 'X' case.
15633 2013-04-25 Alan Modra <amodra@gmail.com>
15636 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
15637 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
15638 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
15639 Repeat for many other rotate/shift and mask patterns using subregs.
15640 Name lshiftrt insns.
15641 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
15642 on WORDS_BIG_ENDIAN.
15644 2013-04-25 Alan Modra <amodra@gmail.com>
15646 * config.gcc: Support little-endian powerpc-linux targets.
15647 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
15648 (LINK_OS_LINUX_SPEC): Define.
15649 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
15650 Preserve MASK_LITTLE_ENDIAN.
15651 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
15652 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
15653 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
15654 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
15655 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
15656 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
15657 Correct fp word order for little-endian. Don't shift toc entries
15658 smaller than a word for little-endian.
15659 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
15660 (bswapdi2 splits): Correct low-part subreg for little-endian.
15661 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
15662 low/high where such is correct only for be.
15663 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
15664 little-endian for -mcall-aixdesc.
15666 2013-04-25 Alan Modra <amodra@gmail.com>
15668 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
15669 replace_equiv_address_nv.
15671 2013-04-25 Alan Modra <amodra@gmail.com>
15673 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
15675 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
15678 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
15679 * rtl.h (struct rtx_def): ...
15681 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
15683 PR rtl-optimizations/57046
15684 * lra-constraints (split_reg): Set up lra_risky_transformations_p
15685 for multi-reg splits.
15687 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
15689 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
15691 2013-04-24 Sterling Augustine <saugustine@google.com>
15693 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
15694 (comp_dir_string, debug_str_dwo_section): New.
15695 (DEBUG_STR_DWO_SECTION): Rename to ...
15696 (DEBUG_DWO_STR_SECTION): ... this.
15697 (DEBUG_NORM_STR_SECTION): Delete.
15698 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
15699 (DEBUG_STR_DWO_SECTION_FLAGS): New.
15700 (find_AT_string): Move most logic to ...
15701 (find_AT_string_in_table): ... here. New.
15702 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
15703 add_skeleton_AT_string. Delete logic.
15704 (output_skeleton_debug_sections): Remove call to
15705 add_top_level_skeleton_die_attrs.
15706 (add_comp_dir_attribute): Move logic to comp_dir_string.
15707 (dwarf2out_init): Initialize debug_str_dwo_section.
15708 (output_indirect_string): Call find_string_form.
15709 (output_indirect_strings): Rewrite.
15710 (prune_unused_types): Empty skeleton_debug_str_hash.
15711 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
15712 (dwarf2out_finish): Call output_indirect_strings.
15714 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
15716 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
15718 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
15720 * rtl.h (struct rtx_def): Add comment for field jump.
15721 (LRA_SUBREG_P): New macro.
15722 * recog.c (register_operand): Check LRA_SUBREG_P.
15723 * lra.c (lra): Add note at the end of RTL code. Align non-empty
15725 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
15726 (lra_final_code_change): Skip subreg change for operators.
15727 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
15728 if there are no operand changes.
15729 * lra-constraints.c (curr_insn_set): New.
15730 (match_reload): Set LRA_SUBREG_P.
15731 (emit_spill_move): Ditto.
15732 (check_and_process_move): Use curr_insn_set. Process only single
15733 set insns. Don't initialize sec_mem_p and change_p.
15734 (simplify_operand_subreg): Use LRA_SUBREG_P.
15735 (reg_in_class_p): New function.
15736 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
15737 of #ifdef. Add code to remove cycling.
15738 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
15739 non-null disp. Reload inner instead of disp when base and index
15740 are null. Try to put lo_sum into register.
15741 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
15742 (check_and_process_move): Move code for move cost check to
15743 simple_move_p. Remove equiv_substitution.
15744 (simple_move_p): New function.
15745 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
15746 curr_insn_set. Call check_and_process_move only for single set
15747 insns. Use the new function. Move call of check_and_process_move
15748 after operand equiv substitution and address process.
15750 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
15752 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
15753 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
15754 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
15756 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
15758 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
15760 2013-04-24 Marek Polacek <polacek@redhat.com>
15762 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
15763 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
15764 (select_loops_exit_conditions): Likewise.
15765 (number_of_iterations_for_all_loops): Likewise.
15766 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
15767 (scev_analysis): Likewise.
15769 2013-04-02 Catherine Moore <clm@codesourcery.com>
15770 Chao-ying Fu <fu@mips.com>
15772 * config/mips/micromips.md (jraddiusp): New pattern.
15773 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
15774 instruction if possible.
15776 2013-04-24 Alan Modra <amodra@gmail.com>
15778 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
15780 2013-04-24 Julian Brown <julian@codesourcery.com>
15781 Chung-Lin Tang <cltang@codesourcery.com>
15783 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
15784 dependency behavior in enumeration type DIE generation. Add TODO note
15785 to comments about future DW_FORM_sdata/udata re-work of related code.
15787 2013-04-23 Lawrence Crowl <crowl@google.com>
15789 * Makefile.in: Update as needed below.
15791 * hash-table.h (class hash_table):
15792 Correct many methods with parameter types compare_type to the correct
15793 value_type. (Correct code was unlikely to notice the change.)
15794 (hash_table::elements_with_deleted) New.
15795 (class hashtable::iterator): New.
15796 (hashtable::begin()): New.
15797 (hashtable::end()): New.
15798 (FOR_EACH_HASH_TABLE_ELEMENT): New.
15800 * statistics.c (statistics_hashes):
15801 Change type to hash_table. Update dependent calls and types.
15803 * tree-into-ssa.c (var_infos):
15804 Change type to hash_table. Update dependent calls and types.
15806 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
15807 Change type to hash_table. Update dependent calls and types.
15809 * tree-ssa-loop-im.c (struct mem_ref.refs):
15810 Change type to hash_table. Update dependent calls and types.
15812 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
15813 Change type to hash_table. Update dependent calls and types.
15815 * tree-ssa-sccvn.c (vn_tables_s::nary):
15816 Change type to hash_table. Update dependent calls and types.
15817 (vn_tables_s::phis): Likewise.
15818 (vn_tables_s::references): Likewise.
15820 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
15821 (vn_reference_eq): Update parameter and return types.
15823 * tree-ssa-structalias.c (pointer_equiv_class_table):
15824 Change type to hash_table. Update dependent calls and types.
15825 (location_equiv_class_table): Likewise.
15827 * tree-vect-data-refs.c: Consequential changes for making
15828 peeling a hash_table.
15830 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
15831 (destroy_loop_vec_info): Dependent hash_table update.
15833 * tree-vectorizer.h (peeling_htab):
15834 Change type to hash_table. Update dependent calls and types.
15836 2013-04-23 Shiva Chen <shiva0217@gmail.com>
15838 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
15839 to check the register content is equal or not.
15840 * lra-constraints.c (match_reload): Use lra_assign_reg_val
15841 to assign register content record.
15842 * lra-eliminations.c (update_reg_eliminate): Use
15843 lra_update_reg_val_offset to update register content offset.
15844 * lra-int.h (struct lra_reg): Add offset member.
15845 (lra_reg_val_equal_p): New static inline function.
15846 (lra_update_reg_val_offset): New static inline function.
15847 (lra_assign_reg_val): New static inline function.
15848 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
15849 to assign register content record.
15850 (initialize_lra_reg_info_element): Initial offset to zero.
15852 2013-04-23 Catherine Moore <clm@codesourcery.com>
15854 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
15855 operands. Record compression.
15857 2013-04-23 Xinliang David Li <davidxl@google.com>
15859 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
15861 2013-04-23 Richard Biener <rguenther@suse.de>
15863 PR middle-end/57036
15864 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
15865 parameter, only add abnormal goto edges from the copied body
15866 if the call could perform abnormal gotos.
15867 (copy_cfg_body): Adjust.
15869 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
15871 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
15873 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
15875 * coretypes.h (gimple_stmt_iterator): Add struct to make
15878 2013-04-23 Richard Biener <rguenther@suse.de>
15880 PR tree-optimization/57026
15881 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
15882 from SSA names occuring in abnormal PHI nodes.
15884 2013-04-22 Andi Kleen <ak@linux.intel.com>
15886 * lto/lto.c (print_lto_report_1): Fix LTO report names.
15888 2013-04-22 Andi Kleen <ak@linux.intel.com>
15890 * lto/lto.c (print_lto_report_1): Declare early.
15891 (read_cgraph_and_symbols): Call print_lto_report_1 early.
15893 2013-04-22 Andi Kleen <ak@linux.intel.com>
15895 * common.opt (-flto-report-wpa): Add.
15896 * doc/invoke.texi (-flto-report-wpa): Add.
15897 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
15900 2013-04-22 Xinliang David Li <davidxl@google.com>
15902 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
15903 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
15904 * Makefile.in: New dependency
15906 David Daney <ddaney.cavm@gmail.com>
15908 * configure.ac (gcc_cv_as_micromips_support): Use the
15909 --fatal-warnings option.
15910 * configure: Regenerate.
15912 2013-04-22 Marek Polacek <polacek@redhat.com>
15915 * tsan.c (instrument_expr): Don't instrument expression
15916 in case its size is zero.
15918 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
15922 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
15924 * config/alpha/alpha.c (TARGET_LRA_P): New define.
15926 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
15928 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
15929 (gimple_stmt_iterator): New typedef.
15930 * gimple.h (gimple_stmt_iterator): Rename to...
15931 (gimple_stmt_iterator_d): ... This.
15932 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
15933 trees be valid for GIMPLE and GENERIC.
15934 (TARGET_GIMPLE_FOLD_BUILTIN): New.
15935 * gimple-fold.c (gimple_fold_call): Call target hook
15936 gimple_fold_builtin.
15937 * hooks.c (hook_bool_gsiptr_false): New.
15938 * hooks.h (hook_bool_gsiptr_false): New.
15939 * target.def (fold_stmt): New.
15940 * doc/tm.texi: Regenerate.
15942 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
15945 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
15946 a set sp if no stack realignment.
15948 2013-04-22 Nick Clifton <nickc@redhat.com>
15950 * config.gcc (tilegx-linux): Extend extra_objs rather than
15952 (tilepro-linux): Likewise.
15954 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
15956 * config/aarch64/aarch64-builtins.c
15958 (CF0, CF1, CF2, CF3, CF4, CF10): New.
15959 (VAR<1-12>): Add MAP parameter.
15960 (BUILTIN_*): Likewise.
15961 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
15962 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
15963 (aarch64_ushl_n<mode>): Likewise.
15964 (aarch64_sshr_n<mode>): Likewise.
15965 (aarch64_ushr_n<mode>): Likewise.
15966 (aarch64_<maxmin><mode>): Likewise.
15967 (aarch64_sqrt<mode>): Likewise.
15968 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
15969 (vshr<q>_n_*): Likewise.
15971 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
15973 * config/aarch64/aarch64-builtins.c
15974 (aarch64_simd_builtin_type_mode): Handle SF types.
15976 (BUILTIN_GPF): Define.
15977 (aarch64_init_simd_builtins): Handle SF types.
15978 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
15979 (frecps): Likewise.
15980 (frecpx): Likewise.
15981 * config/aarch64/aarch64-simd.md
15982 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
15983 (aarch64_frecpe<mode>): New.
15984 (aarch64_frecps<mode>): Likewise.
15985 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
15986 (v8type): Add frecp<esx>.
15987 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
15988 (aarch64_frecps<mode>): Likewise.
15989 * config/aarch64/iterators.md (FRECP): New.
15990 (frecp_suffix): Likewise.
15991 * config/aarch64/arm_neon.h
15992 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
15994 2013-04-22 Christian Bruel <christian.bruel@st.com>
15997 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
15998 (REG_CLASS_NAMES): Idem.
15999 (REG_CLASS_CONTENTS): Idem.
16000 (REGCLASS_HAS_FP_REG): Idem.
16001 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
16002 (sh_conditional_register_usage): Idem.
16004 2013-04-21 Jeff Law <law@redhat.com>
16006 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
16007 (ssa_forward_propagate_and_combine): Use it.
16009 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
16011 * lra.c: Update the flow chart diagram.
16013 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
16015 PR rtl-optimization/56847
16016 * lra-constraints.c (process_alt_operands): Discourage alternative
16017 with non-matche doffsettable memory constraint fro memory with
16020 2013-04-19 Richard Biener <rguenther@suse.de>
16022 PR tree-optimization/56982
16023 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
16025 * gimplify.c (gimplify_call_expr): Notice special calls.
16026 (gimplify_modify_expr): Likewise.
16027 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
16028 abnormal control flow receivers.
16029 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
16030 in the same way as cfun->has_nonlocal_labels.
16031 (gimple_purge_dead_abnormal_call_edges): Likewise.
16032 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
16033 receivers start a basic-block.
16035 2013-04-19 Richard Biener <rguenther@suse.de>
16037 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
16039 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
16040 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
16041 (SLP_TREE_LOAD_PERMUTATION): Add.
16042 (vect_transform_slp_perm_load): Adjust prototype.
16043 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
16044 (vect_free_slp_instance): Likewise.
16045 (vect_create_new_slp_node): Likewise.
16046 (vect_supported_slp_permutation_p): Remove.
16047 (vect_slp_rearrange_stmts): Adjust.
16048 (vect_supported_load_permutation_p): Likewise. Inline
16049 vect_supported_slp_permutation_p here.
16050 (vect_analyze_slp_instance): Compute load permutations per
16051 slp node instead of per instance.
16052 (vect_get_slp_defs): Adjust.
16053 (vect_transform_slp_perm_load): Likewise.
16054 (vect_schedule_slp_instance): Remove redundant code.
16055 (vect_schedule_slp): Remove hack for PR56270, add it ...
16056 * tree-vect-stmts.c (vectorizable_load): ... here, do not
16057 CSE loads for SLP. Adjust.
16059 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
16061 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
16062 spelling in two comments.
16064 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
16067 * config/arm/arm.c (load_multiple_sequence): Require SP
16068 as base register for loads if SP is in the register list.
16070 2013-04-19 Martin Jambor <mjambor@suse.cz>
16072 PR tree-optimization/56718
16073 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
16074 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
16075 and made public. Adjusted all callers.
16076 (ipa_intraprocedural_devirtualization): New function.
16077 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
16078 (ipa_intraprocedural_devirtualization): Likewise.
16079 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
16081 2013-04-19 Richard Biener <rguenther@suse.de>
16083 PR tree-optimization/57000
16084 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
16086 2013-04-19 Terry Guo <terry.guo@arm.com>
16088 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
16090 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
16091 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
16092 (cortex_m4_fmacs): Use new reservations.
16093 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
16095 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
16097 PR rtl-optimization/56999
16098 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
16100 (lra_coalesce): Remove split_origin_bitmap and related code.
16101 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
16102 ranges if necessary.
16104 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
16106 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
16108 (ix86_expand_call): Remove clobbered_registers array and use
16109 x86_64_ms_sysv_extra_clobbered_registers instead.
16110 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
16112 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
16114 * config/i386/i386.md (*call_rex64_ms_sysv): Use
16115 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
16116 (*call_value_rex64_ms_sysv): Ditto.
16118 2013-04-18 Cary Coutant <ccoutant@google.com>
16120 * dwarf2out.c (output_pubnames): Check die_perennial_p of
16121 parent instead of die_mark.
16123 2013-04-18 Diego Novillo <dnovillo@google.com>
16125 * gimple.c (create_gimple_tmp): New.
16126 (get_expr_type): New.
16127 (build_assign): New.
16128 (build_type_cast): New.
16129 * gimple.h (enum ssa_mode): Define.
16130 (gimple_seq_set_location): New.
16131 * asan.c (build_check_stmt): Change some gimple_build_* calls
16132 to use build_assign and build_type_cast.
16134 2013-04-18 Richard Biener <rguenther@suse.de>
16136 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
16137 handle negative step. Remove redundant checks.
16138 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
16139 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
16140 for negative step and grouped loads fail to vectorize.
16142 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
16144 * emit-rtl.c (reset_insn_used_flags): New function.
16145 (reset_all_used_flags): Use it.
16146 (verify_insn_sharing): New function.
16147 (verify_rtl_sharing): Fix verification for SEQUENCEs.
16149 2013-04-18 Jakub Jelinek <jakub@redhat.com>
16151 PR tree-optimization/56984
16152 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
16153 and (x >> M) >= N don't register any assertion if N << M is the
16156 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
16158 * lower-subreg.c (resolve_simple_move): If called self-recursive,
16159 do not delete_insn insns that have not yet been emitted, only
16160 unlink them with remove_insn.
16161 * df-scan.c (df_insn_delete): Revert r197492.
16163 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
16165 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
16166 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
16168 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
16170 * config/arm/arm.md (movsicc_insn): Convert define_insn into
16171 define_insn_and_split.
16172 (and_scc,ior_scc,negscc): Likewise.
16173 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
16175 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
16177 * config/arm/arm.c (use_return_insn): Return 0 for targets that
16178 can benefit from using a sequence of LDRD instructions in epilogue
16179 instead of a single LDM instruction.
16181 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
16184 * doc/extend.texi: Fix typo.
16186 2013-04-17 Richard Biener <rguenther@suse.de>
16188 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
16189 (vect_build_slp_tree): ... here.
16190 (vect_build_slp_tree_1): Compute which stmts of the SLP group
16191 match. Remove special-casing of mismatched complex loads.
16192 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
16193 re-try the match with swapped commutative operands.
16194 (vect_supported_load_permutation_p): Remove special-casing of
16195 mismatched complex loads.
16196 (vect_analyze_slp_instance): Adjust.
16198 2013-04-17 Richard Biener <rguenther@suse.de>
16200 PR rtl-optimization/56921
16201 * cfgloop.h (struct loop): Add simple_loop_desc member.
16202 (struct niter_desc): Mark with GTY(()).
16203 (simple_loop_desc): Do not use aux field but simple_loop_desc.
16204 * loop-iv.c (get_simple_loop_desc): Likewise.
16205 (free_simple_loop_desc): Likewise.
16208 2013-04-16 Richard Biener <rguenther@suse.de>
16210 PR rtl-optimization/56921
16211 * loop-init.c (pass_rtl_move_loop_invariants): Add
16212 TODO_do_not_ggc_collect to todo_flags_finish.
16213 (pass_rtl_unswitch): Same.
16214 (pass_rtl_unroll_and_peel_loops): Same.
16215 (pass_rtl_doloop): Same.
16217 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
16219 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
16220 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
16221 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
16223 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
16224 * tree-streamer.c (record_common_node): Adjust reference in comment.
16226 2013-04-17 Terry Guo <terry.guo@arm.com>
16228 * config/arm/cortex-m4.md: Add a new bypass.
16230 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16232 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
16234 (*subs_<optab><mode>_multp2): New pattern.
16235 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
16236 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
16238 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16240 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
16241 (*subs_mul_imm_<mode>): New pattern.
16243 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
16246 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
16247 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
16248 (vsx_movti_32bit): Change j->wa to O->wa.
16250 2013-04-16 Richard Biener <rguenther@suse.de>
16252 PR rtl-optimization/56921
16253 * loop-init.c (pass_rtl_move_loop_invariants): Add
16254 TODO_do_not_ggc_collect to todo_flags_finish.
16255 (pass_rtl_unswitch): Same.
16256 (pass_rtl_unroll_and_peel_loops): Same.
16257 (pass_rtl_doloop): Same.
16259 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
16261 * config/arm/arm.c (emit_multi_reg_push): New declaration
16262 for an existing function.
16263 (arm_emit_strd_push): New function.
16264 (arm_expand_prologue): Used here.
16265 (arm_emit_ldrd_pop): New function.
16266 (arm_expand_epilogue): Used here.
16267 (arm_get_frame_offsets): Update condition.
16268 (arm_emit_multi_reg_pop): Add a special case for load of a single
16269 register with writeback.
16271 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
16273 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
16276 2013-04-16 Richard Biener <rguenther@suse.de>
16278 PR tree-optimization/56756
16279 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
16280 (first_mem_ref_loc): New.
16281 (execute_sm): Place the load temporarily before a previous
16282 access instead of in the latch edge to ensure its SSA dependencies
16283 are defined at points dominating the load.
16285 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
16287 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
16288 correct fix by moving header and footer insn to the footer of
16289 the merged basic block. Clear BB_END of the merged-away block.
16291 PR middle-end/43631
16292 * emit-rtl.c (make_note_raw): New function.
16293 (link_insn_into_chain): New static inline function.
16294 (add_insn): Use it.
16295 (add_insn_before, add_insn_after): Factor insn chain linking code...
16296 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
16297 using link_insn_into_chain.
16298 (note_outside_basic_block_p): New helper function for emit_note_after
16299 and emit_note_before.
16300 (emit_note_after): Use nobb variant of add_insn_after if the note
16301 should not be contained in a basic block.
16302 (emit_note_before): Use nobb variant of add_insn_before if the note
16303 should not be contained in a basic block.
16304 (emit_note_copy): Use make_note_raw.
16305 (emit_note): Likewise.
16306 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
16307 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
16308 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
16309 the moved barrier the tail of the basic block it follows.
16310 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
16312 2013-04-15 Jakub Jelinek <jakub@redhat.com>
16314 PR tree-optimization/56962
16315 * gimple-ssa-strength-reduction.c (record_increment): Only set
16316 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
16317 either rhs1 or rhs2 is equal to c->base_expr.
16319 2013-04-15 Richard Biener <rguenther@suse.de>
16321 PR tree-optimization/56933
16322 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
16324 (GROUP_READ_WRITE_DEPENDENCE): Remove.
16325 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
16326 * tree-vect-data-refs.c (vect_analyze_group_access): Move
16327 dependence check ...
16328 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
16330 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
16331 GROUP_READ_WRITE_DEPENDENCE.
16333 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16335 * emit-rtl.c (reset_all_used_flags): New function.
16336 (verify_rtl_sharing): Call reset_all_used_flags before and after
16337 performing the checks.
16339 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16341 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
16342 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
16343 * config/arm/constraints.md (De): New constraint.
16344 * config/arm/neon.md (anddi3_neon): Delete.
16345 (neon_vand<mode>): Expand to standard anddi3 pattern.
16346 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
16347 Move earlier in the file.
16348 (neon_inv_logic_op2): Likewise.
16349 (arm_anddi_operand_neon): New predicate.
16351 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16353 * configure.ac (gcc_cv_ld_as_needed): Set
16354 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
16355 Use -z ignore, -z record on *-*-solaris2*.
16356 (HAVE_LD_AS_NEEDED): Update comment.
16357 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
16358 * configure: Regenerate.
16359 * config.in: Regenerate.
16360 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
16361 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
16362 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
16363 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
16364 equivalents. Fix markup.
16365 * doc/tm.texi: Regenerate.
16367 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
16369 * config/i386/i386.opt: New option mstack-protector-guard=.
16370 * config/i386/i386-opts.h: Add enum stack_protector_guard.
16371 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
16372 TARGET_SSP_TLS_GUARD.
16373 * config/i386/i386.c (ix86_option_override_internal): Set
16374 ix86_stack_protector_guard.
16375 * config/i386/i386.md (stack_protect_set): Enable for
16376 TARGET_SSP_TLS_GUARD only.
16377 (stack_protect_set_<mode>): Ditto.
16378 (stack_protect_test): Ditto.
16379 (stack_protect_test_<mode>): Ditto.
16380 * doc/invoke.texi (i386 Option): Document.
16382 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
16385 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
16386 (S_MODES): Set H_MODE bit.
16387 (SF_MODES): Set only S_MODE and SF_MODE bits.
16388 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
16389 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
16390 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
16391 <MODE_FLOAT>: Likewise.
16393 2013-04-15 Joey Ye <joey.ye@arm.com>
16395 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
16397 2013-04-15 Joey Ye <joey.ye@arm.com>
16399 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
16401 (thumb_far_jump_used_p): Count instruction size and set
16404 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
16406 * reorg.c (fill_simple_delay_slots): Reindent block of code.
16407 * resource.c (mark_target_live_regs): Reformat conditional block.
16409 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
16411 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
16412 notes, they are emitted only just before final.
16413 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
16415 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
16417 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
16418 * cfgrtl.c (delete_insn): Call it here instead.
16419 * lra-spills.c (lra_final_code_change): Use delete_insn.
16420 * haifa-sched.c (sched_remove_insn): Likewise.
16421 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
16422 returning to the nop pool.
16423 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
16424 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
16426 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
16428 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
16429 * doc/tm.texi: Regenerated.
16431 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
16433 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
16436 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
16438 * df-core.c (df_find_def): Compare register numbers.
16439 (df_find_use): Likewise.
16441 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
16444 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
16445 lra_in_progress for return.
16447 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
16449 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
16450 define_insn into define_insn_and_split and emit movsicc patterns.
16452 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
16454 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
16456 2013-04-12 Richard Biener <rguenther@suse.de>
16458 * tree-pass.h (TODO_do_not_ggc_collect): New.
16459 * passes.c (execute_one_ipa_transform_pass): Honor
16460 TODO_do_not_ggc_collect.
16461 (execute_one_pass): Likewise.
16464 2013-04-10 Richard Biener <rguenther@suse.de>
16466 * passes.c (init_optimization_passes): Remove reload pass.
16467 * ira.c (do_reload): Merge into ...
16469 (rest_of_handle_reload): Remove.
16470 (pass_reload): Likewise.
16471 * config/i386/i386.c (ix86_option_override): Refer to ira instead
16472 of reload for vzeroupper pass placement.
16474 2013-04-12 Jakub Jelinek <jakub@redhat.com>
16476 PR tree-optimization/56918
16477 PR tree-optimization/56920
16478 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
16479 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
16480 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
16481 use wide_mul_with_sign method.
16483 2013-04-12 Richard Biener <rguenther@suse.de>
16485 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
16486 not be considered a gimple constant.
16488 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
16490 * fold-const.c (const_binop): Handle vector shifts by a scalar.
16491 (fold_binary_loc): Call const_binop also for mixed vector-scalar
16494 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
16495 Jakub Jelinek <jakub@redhat.com>
16497 * opts.c: Include diagnostic-color.h.
16498 (common_handle_option): Handle OPT_fdiagnostics_color_.
16499 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
16500 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
16501 (diagnostic-color.o): New.
16502 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
16503 (diagnostic_color_rule): New enum.
16504 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
16505 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
16506 the location string.
16507 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
16508 either NULL, or color kind.
16509 * diagnostic-color.c: New file.
16510 * diagnostic-color.h: New file.
16511 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
16513 * doc/invoke.texi (-fdiagnostics-color): Document.
16514 * pretty-print.h (pp_show_color): Define.
16515 (struct pretty_print_info): Add show_color field.
16516 * diagnostic.c: Include diagnostic-color.h.
16517 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
16518 macros. Colorize error:, warning: etc. strings and also the location
16520 (diagnostic_show_locus): Colorize the caret line.
16521 * pretty-print.c: Include diagnostic-color.h.
16522 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
16523 inside of %< %> quotes or quoted through q format modifier.
16525 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16527 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
16529 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16531 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
16532 code in CC_NZ mode.
16533 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
16536 2013-04-11 Marek Polacek <polacek@redhat.com>
16538 PR tree-optimization/48184
16539 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
16541 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
16543 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
16544 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
16545 (skip_simple_arithmetic): Tidy up.
16546 * tree.h (skip_simple_constant_arithmetic): Declare.
16548 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16550 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
16552 2013-04-11 Richard Biener <rguenther@suse.de>
16554 * tree-vect-loop.c (get_initial_def_for_induction): Properly
16555 generate vector constants.
16557 2013-04-11 Richard Biener <rguenther@suse.de>
16559 PR tree-optimization/56878
16560 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
16561 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
16563 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
16564 Prefer to align the DR with the most invariant base address.
16566 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
16568 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
16571 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
16573 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
16574 floating-point vector comparisons against 0.
16576 2013-04-11 Jakub Jelinek <jakub@redhat.com>
16578 PR tree-optimization/56899
16579 * fold-const.c (extract_muldiv_1): Apply distributive law
16580 only if TYPE_OVERFLOW_WRAPS (ctype).
16582 2013-04-11 Bin Cheng <bin.cheng@arm.com>
16585 * ira-costs.c (scan_one_insn): Check whether the source rtx of
16586 loading has side effect.
16588 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
16590 * config/sparc/sparc.c: Include tree-pass.h.
16591 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
16592 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
16593 head of file. Change return type. Split off gate function.
16594 (sparc_gate_work_around_errata): New function.
16595 (pass_work_around_errata): New pass definition.
16596 (insert_pass_work_around_errata) New pass insert definition to
16597 insert pass_work_around_errata just after delayed-branch scheduling.
16598 (sparc_option_override): Insert the pass.
16599 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
16601 2013-04-10 David S. Miller <davem@davemloft.net>
16603 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
16604 or -mcpu=hypersparc.
16606 * target.def (cstore_mode): New hook.
16607 * target.h: Include insn-codes.h
16608 * targhooks.c: Likewise.
16609 (default_cstore_mode): New function.
16610 * targhooks.h: Declare it.
16611 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
16612 * doc/tm.texi: Rebuild.
16613 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
16614 target hook, rather than inspecting the insn_data.
16615 * config/sparc/sparc.c (sparc_cstore_mode): New function.
16616 (TARGET_CSTORE_MODE): Redefine.
16617 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
16619 * config/sparc/predicates.md (cstore_result_operand): New special
16621 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
16622 Use it for operand 0.
16623 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
16624 (*snesi_special): Likewise.
16625 (*snesi_zero): Likewise.
16626 (*seqsi_zero): Likewise.
16627 (*sltu_insn): Likewise.
16628 (*sgeu_insn): Likewise.
16629 (*seqdi_special): Make operand 0 and comparison operation be of
16631 (*snedi_special): Likewise.
16632 (*snedi_special_vis3): Likewise.
16633 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
16634 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
16635 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
16636 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
16637 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
16638 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
16639 (*sltu_extend_sp64): Likewise.
16640 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
16641 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
16642 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
16643 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
16644 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
16646 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
16648 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
16649 (aarch64_start_file): Use the new function.
16651 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
16652 Jason Merrill <jason@redhat.com>
16654 * common.opt: Add -gdwarf.
16655 * opts.c (common_handle_option): Handle it.
16656 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
16658 2013-04-10 Richard Biener <rguenther@suse.de>
16660 * passes.c (execute_todo): Do not call ggc_collect conditional here.
16661 (execute_one_ipa_transform_pass): But unconditionally here.
16662 (execute_one_pass): And here.
16663 (init_optimization_passes): Remove reload pass.
16664 * tree-pass.h (TODO_ggc_collect): Remove.
16665 (pass_reload): Likewise.
16666 * ira.c (do_reload): Merge into ...
16668 (rest_of_handle_reload): Remove.
16669 (pass_reload): Likewise.
16670 * config/i386/i386.c (ix86_option_override): Refer to ira instead
16671 of reload for vzeroupper pass placement.
16672 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
16673 and todo_flags_finish of all passes.
16675 2013-04-10 Richard Biener <rguenther@suse.de>
16677 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
16678 first_const_oprnd field, rename first_def_type to first_op_type.
16679 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
16680 (vect_get_and_check_slp_defs): Always use the type of the
16681 operand. Allow mixed vect_external_def, vect_constant_def types.
16682 (vect_get_constant_vectors): Handle mixed vect_external_def,
16683 vect_constant_def types.
16685 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
16687 PR tree-optimization/55524
16688 * tree-ssa-math-opts.c
16689 (convert_mult_to_fma): Don't use an fms construct
16690 when we don't have an fms operation, but fnma, and it looks
16691 likely that we'll be able to use the latter.
16693 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
16695 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
16697 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
16698 inline fail caused by overwritable functions.
16700 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
16702 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
16703 unnecessary bits in the constant power of two case.
16705 2013-04-10 Richard Biener <rguenther@suse.de>
16707 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
16708 broken code swapping operands.
16709 (vect_build_slp_tree): Do not compute load permutations here.
16710 (vect_analyze_slp_instance): Compute load permutations here,
16711 after building the SLP tree.
16713 2013-04-09 Christian Bruel <christian.bruel@st.com>
16715 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
16716 of next/prev_real_insn.
16718 2013-04-09 Jan Hubicka <jh@suse.cz>
16720 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
16721 Drop aliased parameter.
16722 (function_and_variable_visibility): Do not handle alias pairs.
16723 * cgraph.c (varpool_externally_visible_p): Update prototype.
16724 * varpool.c (varpool_add_new_variable): Update.
16726 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16728 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
16730 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
16732 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
16734 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
16736 2013-04-09 Marek Polacek <polacek@redhat.com>
16738 PR tree-optimization/48762
16739 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
16741 2013-04-09 Richard Biener <rguenther@suse.de>
16743 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
16745 (vect_build_slp_tree): Likewise.
16746 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
16747 calculating the cost of a SLP instance.
16748 (vect_analyze_slp_instance): Use it from here, after building
16751 2013-04-09 Jakub Jelinek <jakub@redhat.com>
16753 PR middle-end/56883
16754 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
16755 expand_omp_for_static_chunk): Use simple_p = true in
16756 force_gimple_operand_gsi calls when assigning to addressable decls.
16758 2013-04-09 Jeff Law <law@redhat.com>
16760 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
16761 when the boolean was created by converting a wider object which
16762 had a boolean range.
16764 2013-04-09 Richard Biener <rguenther@suse.de>
16766 * tree-vectorizer.h (slp_void_p): Remove.
16767 (slp_tree): Typedef before _slp_tree declaration.
16768 (struct _slp_tree): Use a vector of slp_tree as children.
16769 (vect_get_place_in_interleaving_chain): Remove.
16770 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
16772 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
16774 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
16775 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
16776 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
16777 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
16778 Use slp_node instead of slp_void_p and adjust.
16780 2013-04-09 Richard Biener <rguenther@suse.de>
16782 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
16783 work that is not necessary.
16785 2013-04-09 Jakub Jelinek <jakub@redhat.com>
16787 PR tree-optimization/56854
16788 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
16789 forward into clobber stmts if it would change MEM_REF lhs into
16792 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
16794 * tree.c (type_hash_lookup, type_hash_add): Make static.
16795 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
16797 2013-04-09 Richard Biener <rguenther@suse.de>
16799 * tree.h (unsave_expr_now): Remove.
16800 * tree-inline.c (mark_local_for_remap_r): Remove.
16801 (unsave_expr_1): Likewise.
16802 (unsave_r): Likewise.
16803 (unsave_expr_now): Likewise.
16804 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
16805 (propagate_tree_value): Likewise.
16807 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
16809 * doc/rtl.texi (sequence): Rewrite documentation to match the
16810 current use of SEQUENCE rtl objects.
16811 * rtl.def (SEQUENCE): Likewise.
16813 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
16814 Update documentation.
16815 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
16816 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
16818 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
16820 2013-04-08 Teresa Johnson <tejohnson@google.com>
16822 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
16823 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
16825 (estimate_edge_size_and_time): Add comment to suggest using rounding
16827 (estimate_node_size_and_time): Ditto.
16828 (remap_edge_change_prob): Use helper rounding divide methods.
16829 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
16830 (gimple_mod_pow2_value_transform): Ditto.
16831 (gimple_mod_subtract_transform): Ditto.
16832 (gimple_ic_transform): Ditto.
16833 (gimple_stringops_transform): Ditto.
16834 * stmt.c (conditional_probability): Ditto.
16835 (emit_case_dispatch_table): Ditto.
16836 * lto-cgraph.c (merge_profile_summaries): Ditto.
16837 * tree-optimize.c (execute_fixup_cfg): Ditto.
16838 * cfgcleanup.c (try_forward_edges): Ditto.
16839 * cfgloopmanip.c (scale_loop_profile): Ditto.
16841 (duplicate_loop_to_header_edge): Ditto.
16842 (lv_adjust_loop_entry_edge): Ditto.
16843 * tree-vect-loop.c (vect_transform_loop): Ditto.
16844 * profile.c (compute_branch_probabilities): Ditto.
16845 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
16846 * lto-streamer-in.c (input_cfg): Ditto.
16847 * gimple-streamer-in.c (input_bb): Ditto.
16848 * ipa-cp.c (update_profiling_info): Ditto.
16849 (update_specialized_profile): Ditto.
16850 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
16851 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
16853 * sched-rgn.c (compute_dom_prob_ps): Ditto.
16854 (compute_trg_info): Ditto.
16855 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
16856 (purge_dead_edges): Ditto.
16857 * loop-unswitch.c (unswitch_loop): Ditto.
16858 * cgraphclones.c (cgraph_clone_edge): Ditto.
16859 (cgraph_clone_node): Ditto.
16860 * tree-inline.c (copy_bb): Ditto.
16861 (copy_edges_for_bb): Ditto.
16862 (initialize_cfun): Ditto.
16863 (copy_cfg_body): Ditto.
16864 (expand_call_inline): Ditto.
16866 2013-04-08 Kai Tietz <ktietz@redhat.com>
16868 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
16869 TARGET_CYGWIN64 by TARGET_64BIT.
16871 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
16873 * config/epiphany/epiphany.md (GPR_1): New constant.
16874 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
16875 * config/epiphany/epiphany.c (gen_compare_reg):
16876 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
16877 is already in place.
16878 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
16879 Don't require being called during rtl expansion; If y operlaps r0,
16881 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
16882 (epiphany_expand_epilogue): Likewise.
16884 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
16885 Don't use CC_FPmode for ORDERED / UNORDERED.
16886 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
16888 * config/epiphany/constraints.md (CnL): New constraint.
16889 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
16890 * config/epiphany/predicates.md (add_operand): Allow 1024.
16892 * config/epiphany/epiphany.md (logical_op): New code iterator.
16893 (op_mnc): New code attribute.
16894 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
16895 (mov_f+1, mov_f+2): New peephole2 patterns.
16897 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
16898 (cstoresi4): Also allow re-use of zero result when doing a NE
16899 comparison to a non-zero operand.
16900 Use (clobber (scratch)) for first insn if the gpr output is not needed.
16902 * config/epiphany/epiphany.md (<insn_opname>v2si3):
16903 Use gen_addsi3_i / gen_subsi3_i.
16905 2013-04-08 Jakub Jelinek <jakub@redhat.com>
16909 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
16910 contain anything but clobbers, at most one __builtin_stack_restore,
16911 optionally debug stmts and final resx, and if it has at least one
16912 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
16913 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
16914 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
16915 which isn't defaut definition, remove them.
16916 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
16917 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
16918 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
16919 with MEM_REF LHS with SSA_NAME address.
16921 2013-04-08 Jeff Law <law@redhat.com>
16923 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
16925 2013-04-08 Richard Biener <rguenther@suse.de>
16927 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
16929 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
16930 determined vector type.
16931 (vect_analyze_data_refs): Likewise.
16932 (vect_get_new_vect_var): Adjust.
16933 (vect_create_destination_var): Preserve SSA name versions.
16934 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
16935 not dump anything here.
16937 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
16939 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
16940 Add member lr_slot_known.
16941 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
16943 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
16944 Remove code that sets lr_slot_offset according to what a previous
16945 version of epiphany_emit_save_restore used to do.
16946 (epiphany_emit_save_restore): When doing an lr save or restore,
16947 set/verify lr_slot_known and lr_slot_offset.
16949 2013-04-08 Xinyu Qi <xyqi@marvell.com>
16952 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
16955 2013-04-08 Richard Biener <rguenther@suse.de>
16957 * alias.c (find_base_term): Fix thinko in previous change.
16959 2013-04-08 Jakub Jelinek <jakub@redhat.com>
16961 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
16962 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
16963 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
16964 if possible to compute val.
16965 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
16966 For QImode integers don't require anything about precision. Use
16967 const_with_all_bytes_same to find out if the constant doesn't have
16968 repeated bytes in it.
16970 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16972 * config/s390/s390.c (s390_expand_insv): Only accept insertions
16975 2013-04-08 Marek Polacek <polacek@redhat.com>
16977 PR rtl-optimization/48182
16978 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
16981 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
16984 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
16985 nuses, make sure we have a label.
16987 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16990 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
16991 (rs6000_emit_swdiv_low_precision): Remove.
16992 (rs6000_emit_swdiv): Rewrite to handle between one and four
16993 iterations of Newton-Raphson generally; modify required number of
16994 iterations for some cases.
16995 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
16997 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
16999 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
17000 set-but-unused variable.
17002 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
17003 basic blocks of released function bodies garbage-collectable.
17005 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
17006 (struct rtl_opt_pass): Add TODO_df_finish.
17008 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
17010 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17012 * config/arm/constraints.md (q): New constraint.
17013 * config/arm/ldrdstrd.md: New file.
17014 * config/arm/arm.md (ldrdstrd.md) New include.
17015 (arm_movdi): Use "q" instead of "r" constraint
17016 for double-word memory access.
17017 (movdf_soft_insn): Likewise.
17018 * config/arm/vfp.md (movdi_vfp): Likewise.
17019 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
17020 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
17021 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
17022 (mem_ok_for_ldrd_strd): Likewise.
17023 (output_move_double): Update assertion.
17025 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17027 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
17029 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17031 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
17032 define_insn_and_split.
17033 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
17035 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17037 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
17038 define_insn_and_split.
17039 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
17040 (shiftsi3_compare): New pattern.
17041 (rrx): New pattern.
17042 * config/arm/unspecs.md (UNSPEC_RRX): New.
17044 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17046 * config/arm/arm.md (negdi_extendsidi): New pattern.
17047 (negdi_zero_extendsidi): Likewise.
17049 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17051 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
17052 define_insn_and_split.
17053 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
17054 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
17056 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17058 * config/arm/arm.md (arm_subdi3): Convert define_insn into
17059 define_insn_and_split.
17060 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
17061 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
17063 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17065 * config/arm/arm.md (subsi3_carryin): New pattern.
17066 (subsi3_carryin_const): Likewise.
17067 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
17068 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
17070 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17072 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
17074 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17076 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
17077 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
17079 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17081 * config/arm/arm.c (arm_expand_builtin): Change fcode
17082 type to unsigned int.
17084 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17086 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
17088 2013-04-04 Ian Lance Taylor <iant@google.com>
17090 * doc/standards.texi (Standards): The Go frontend supports the Go 1
17093 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
17095 PR middle-end/56729
17096 * df-scan.c (df_insn_delete): Disable failing assert.
17098 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17100 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
17101 New function prototype.
17102 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
17103 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
17104 (arm_builtin_vectorized_function): New function.
17106 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17108 * config/arm/arm_neon_builtins.def: New file.
17109 * config/arm/arm.c (neon_builtin_data): Move contents to
17110 arm_neon_builtins.def.
17111 (enum arm_builtins): Include neon builtin definitions.
17112 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
17113 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
17115 2013-04-04 Marek Polacek <polacek@redhat.com>
17117 PR tree-optimization/48186
17118 * predict.c (maybe_hot_frequency_p): Return false if
17119 HOT_BB_FREQUENCY_FRACTION is 0.
17120 (cgraph_maybe_hot_edge_p): Likewise.
17122 2013-04-04 Richard Biener <rguenther@suse.de>
17124 PR tree-optimization/56826
17125 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
17128 2013-04-04 Richard Biener <rguenther@suse.de>
17130 PR tree-optimization/56213
17131 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
17132 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
17134 2013-04-04 Richard Biener <rguenther@suse.de>
17136 PR tree-optimization/56837
17137 * tree-loop-distribution.c (classify_partition): For non-zero
17138 values require that the value has the same precision as its
17139 mode to be useful as memset value.
17141 2013-04-03 Nick Clifton <nickc@redhat.com>
17143 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
17144 (fmssf4): Use fmsf.s on E3V5 architectures.
17145 (fnmasf4): Use fnmaf.s on E3V5 architectures.
17146 (fnmssf4): Use fnmsf.s on E3V5 architectures.
17148 2013-04-03 Jeff Law <law@redhat.com>
17150 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
17151 (lra-eliminations.o): Likewise.
17153 2013-04-03 Teresa Johnson <tejohnson@google.com>
17155 * gcov-io.c (compute_working_sets): Moved most of body of old
17156 compute_working_sets here from profile.c.
17157 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
17158 (gcov_working_set_t): Moved typedef here from basic-block.h
17159 (compute_working_set): Declare.
17160 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
17161 (get_working_sets): Renamed from compute_working_set,
17162 replace most of body with call to new compute_working_sets.
17163 (get_exec_counts): Replace call to compute_working_sets
17164 to get_working_sets.
17165 * profile.h (get_working_sets): Renamed from compute_working_set.
17166 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
17167 to get_working_sets.
17168 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
17169 * gcov-dump.c (dump_working_sets): New function.
17171 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
17173 * hwint.c (sext_hwi, zext_hwi): New functions.
17174 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
17175 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
17176 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
17177 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
17178 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
17179 (sext_hwi, zext_hwi): New functions.
17181 2013-04-03 Jeff Law <law@redhat.com>
17183 PR tree-optimization/56799
17184 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
17185 back test for widening conversion erroneously dropped in prior change.
17187 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17190 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
17191 instead of next_real_insn.
17193 2013-04-03 Marek Polacek <polacek@redhat.com>
17196 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
17198 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17201 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
17203 (thumb1_output_casesi): Likewise.
17204 (thumb2_output_casesi): Likewise.
17206 2013-04-03 Richard Biener <rguenther@suse.de>
17208 PR tree-optimization/56817
17209 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
17211 (tree_unroll_loops_completely_1): ... new function to manually
17212 walk the loop tree, properly defering outer loops of unrolled
17213 loops to later iterations.
17215 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
17217 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
17218 (vectorizable_load): Likewise.
17219 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
17220 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
17222 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
17224 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
17227 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
17229 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
17231 2013-04-03 Bin Cheng <bin.cheng@arm.com>
17233 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
17235 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
17237 PR tree-optimization/56790
17238 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
17241 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
17243 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
17245 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
17246 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
17249 2013-04-03 Jakub Jelinek <jakub@redhat.com>
17252 * tree.h (force_folding_builtin_constant_p): New decl.
17253 * builtins.c (force_folding_builtin_constant_p): New variable.
17254 (fold_builtin_constant_p): Fold immediately also if
17255 force_folding_builtin_constant_p.
17257 2013-04-03 Richard Biener <rguenther@suse.de>
17259 PR tree-optimization/56812
17260 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
17261 DRs of the same interleaving chain are independent.
17263 2013-04-02 Jason Merrill <jason@redhat.com>
17265 * gdbinit.in (pbb): Use debug fn.
17267 2013-04-02 Lawrence Crowl <crowl@google.com>
17269 * sese.h (struct ivtype_map_elt_s): Remove unused.
17270 (extern debug_ivtype_map): Remove unused.
17271 (extern eq_ivtype_map_elts): Remove unused.
17272 * sese.c (debug_ivtype_map): Removed unused.
17273 (debug_ivtype_map_1): Removed unused.
17274 (debug_ivtype_elt): Remove unused.
17275 (eq_ivtype_map_elts): Remove unused.
17278 2013-04-02 Kai Tietz <ktietz@redhat.com>
17281 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
17282 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
17283 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
17285 (legitimize_pe_coff_symbol): Likewise.
17286 (is_imported_p): New helper-function.
17287 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
17288 for Windows x64 targets.
17289 (ix86_expand_prologue): Optimize for pe-coff targets.
17290 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
17291 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
17292 medium/large code-model.
17293 (legitimize_pic_address): Likewise.
17294 (legitimize_tls_address): Likewise.
17295 (ix86_expand_call): Likewise.
17296 (x86_output_mi_thunk): Likewise.
17297 (get_dllimport_decl): Add new beimport argument.
17298 (construct_plt_address): Don't assert for x64 pe-coff targets.
17299 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
17301 (SYMBOL_FLAG_STUBVAR): New macro.
17302 (SYMBOL_REF_STUBVAR_P): Likewise.
17303 * config/i386/winnt.c (stub_list): New structure.
17304 (stub_head): New local variable.
17305 (i386_pe_record_stub): New function.
17306 (i386_pe_file_end): Emit refptr-stubs.
17308 2013-04-02 Jakub Jelinek <jakub@redhat.com>
17310 PR rtl-optimization/56745
17311 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
17312 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
17315 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
17316 and both of them are MEM_REFs, just compare first argument for
17317 equality and attempt to deal even with differing offsets.
17320 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
17321 of gimple_clobber_p to be MEM_REF.
17322 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
17323 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
17324 after gimplification.
17325 * asan.c (get_mem_ref_of_assignment): Don't instrument
17326 gimple_clobber_p stmts.
17327 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
17328 gimple_clobber_p stmt if they have MEM_REF lhs and
17329 are dead because of another gimple_clobber_p stmt.
17330 * tree-ssa-live.c (clear_unused_block_pointer): Treat
17331 gimple_clobber_p stmts like debug stmts.
17332 (remove_unused_locals): Remove clobbers with MEM_REF lhs
17333 that refer to unused VAR_DECLs or uninitialized values.
17334 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
17335 gimple_clobber_p stmts if they refer to removed parameters.
17336 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
17339 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
17341 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
17342 using SWI48 mode attribute.
17344 2013-04-02 Wei Mi <wmi@google.com>
17346 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
17347 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
17348 *<rotate_insn><mode>3_mask in i386.md.
17350 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
17352 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
17354 2013-04-02 Richard Biener <rguenther@suse.de>
17356 PR tree-optimization/56778
17357 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
17358 Runtime alias tests are not supported for gather loads.
17359 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
17360 stmts referenced from SSA operands before updating SSA form.
17362 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
17363 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17365 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
17366 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
17367 * config/arm/cortex-a53.md: New file.
17368 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
17369 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
17370 * config/arm/arm.c (arm_issue_rate): Likewise.
17371 * config/arm/arm-tune.md: Regenerate
17372 * config/arm/arm-tables.opt: Regenerate.
17373 * config/arm/arm-cores.def: Add cortex-a53.
17375 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
17377 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
17380 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
17382 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
17383 scalar load/store operations using B/H registers.
17384 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
17386 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
17388 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
17390 * config/aarch64/aarch64.c
17391 (aarch64_simd_scalar_immediate_valid_for_move): New.
17392 * config/aarch64/aarch64-protos.h
17393 (aarch64_simd_scalar_immediate_valid_for_move): New.
17394 * config/aarch64/constraints.md (Dh, Dq): New.
17395 * config/aarch64/iterators.md (hq): New.
17397 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
17399 * reorg.c (get_branch_condition): Deal with conditional returns.
17400 (fill_simple_delay_slots): Remove dead code dealing with jumps.
17402 2013-04-01 Wei Mi <wmi@google.com>
17404 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
17405 Truncate operand 2 using %b asm operand modifier.
17406 (*<shift_insn><mode>3_mask): Ditto.
17407 (*<rotate_insn><mode>3_mask): Ditto.
17409 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
17411 PR middle-end/56798
17412 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
17414 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
17416 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
17418 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
17420 2013-03-30 Lawrence Crowl <crowl@google.com>
17422 * dse.c (clear_alias_sets): Remove never set.
17423 (disqualified_clear_alias_sets): Remove never set.
17424 (clear_alias_mode_pool): Remove never set.
17425 (dse_step0): Remove condition that is never true.
17426 (canon_address): Remove condition that is never true.
17427 (dse_step7): Remove condition that is never true.
17428 (rest_of_handle_dse): Remove condition that is never true.
17429 (rest_of_handle_dse::did_global): Remove never read from above.
17430 (dse_step2_spill): Remove never called from above.
17431 (dse_step5_spill): Remove never called from above.
17433 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
17435 * doc/md.texi (Standard Names) <casesi>: Update documentation for
17436 JUMP_TABLE_DATA changes.
17437 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
17438 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
17439 (Insns) <jump_table_data>: New entry.
17440 * doc/tm.texi: Regenerate.
17442 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
17444 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
17445 for table jump at the end of a basic block using tablejump_p.
17446 * targhooks.c (default_invalid_within_doloop): Likewise.
17447 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
17448 target hook implementation that is identical to the default hook.
17449 (rs6000_invalid_within_doloop): Remove.
17451 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
17452 unused variable from tablejump_p call.
17454 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
17455 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
17456 (INSN_DELETED_P): Likewise.
17457 (emit_jump_table_data): New prototype.
17458 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
17459 after 4th as unused.
17460 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
17461 * sched-vis.c (print_insn): Likewise.
17462 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
17463 insn for compatibility with back ends that use next_active_insn to
17464 identify jump table data.
17465 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
17466 (remove_insn): Likewise.
17467 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
17469 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
17470 (emit_jump_table_data): New function.
17472 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
17473 basic block, a JUMP_TABLE_DATA never is.
17474 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
17475 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
17476 off from code handling real insns.
17477 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
17478 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
17479 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
17480 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
17481 is not a NONDEBUG_INSN_P.
17482 * ira-costs.c (scan_one_insn): Likewise.
17483 * jump.c (mark_all_labels): Likewise.
17484 (mark_jump_label_1): Likewise.
17485 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
17486 * lra.c (get_insn_freq): Expect all insns reaching here to be in
17488 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
17489 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
17490 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
17491 JUMP_TABLE_DATA_P insns.
17492 (calculate_elim_costs_all_insns): Likewise.
17493 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
17494 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
17495 (delete_output_reload): Code style fixups.
17496 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
17497 insn flags on this non-insn.
17498 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
17499 as scheduling barriers, for pre-change compatibility.
17500 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
17501 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
17503 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
17504 redundant JUMP_TABLE_DATA_P test.
17505 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
17506 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
17507 (frv_for_each_packet): Likewise.
17508 * config/i386/i386.c (min_insn_size): Likewise.
17509 (ix86_avoid_jump_mispredicts): Likewise.
17510 * config/m32r/m32r.c (m32r_is_insn): Likewise.
17511 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
17512 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
17513 (mips16_insn_length): Robustify.
17514 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
17515 (mips16_split_long_branches): Likewise.
17516 * config/pa/pa.c (pa_combine_instructions): Likewise.
17517 * config/rs6000/rs6000.c (get_next_active_insn): Treat
17518 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
17519 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
17520 as contributing to pool range lengths.
17521 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
17522 Remove redundant JUMP_TABLE_DATA_P test.
17523 (sh_loop_align): Likewise.
17524 (split_branches): Likewise.
17525 (sh_insn_length_adjustment): Likewise.
17526 * config/spu/spu.c (get_branch_target): Likewise.
17528 2013-03-29 Jan Hubicka <jh@suse.cz>
17530 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
17531 gcov streaming; stream hot bb threshold to ltrans.
17532 * predict.c (get_hot_bb_threshold): Break out from ....
17533 (maybe_hot_count_p): ... here.
17534 (set_hot_bb_threshold): New function.
17535 * lto-section-in.c (lto_section_name): Add profile.
17536 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
17537 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
17538 and data-streamer.h
17539 (histogram_entry): New structure.
17540 (histogram, histogram_pool): New global vars.
17541 (histogram_hash): New structure.
17542 (histogram_hash::hash): New method.
17543 (histogram_hash::equal): Likewise.
17544 (account_time_size): New function.
17545 (cmp_counts): New function.
17546 (dump_histogram): New function.
17547 (ipa_profile_generate_summary): New function.
17548 (ipa_profile_write_summary): New function.
17549 (ipa_profile_read_summary): New function.
17550 (ipa_profile): Decide on threshold.
17551 (pass_ipa_profile): Add ipa_profile_write_summary and
17552 ipa_profile_read_summary.
17553 * Makefile.in (ipa.o): Update dependencies.
17554 * lto-streamer.h (LTO_section_ipa_profile): New section.
17556 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
17558 * tree.h (VAR_P): New.
17560 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
17563 * doc/invoke.texi ([-fwhole-program]): Fix typo.
17565 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
17567 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
17568 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
17569 (control_flow_insn_p): Likewise.
17570 * cfgrtl.c (duplicate_insn_chain): Likewise.
17571 * final.c (get_attr_length_1): Likewise.
17572 (shorten_branches): Likewise.
17573 (final_scan_insn): Likewise.
17574 * function.c (instantiate_virtual_regs): Likewise.
17575 * gcse.c (insert_insn_end_basic_block): Likewise.
17576 * ira-costs.c (scan_one_insn): Likewise.
17577 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
17578 * lra.c (check_rtl): Likewise.
17579 * reload1.c (elimination_costs_in_insn): Likewise.
17580 * reorg.c (follow_jumps): Likewise.
17582 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
17583 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
17584 (thumb_far_jump_used_p): Likewise.
17585 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
17586 (workaround_speculation): Likewise.
17587 (add_sched_insns_for_speculation): Likewise.
17588 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
17589 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
17590 (frv_for_each_packet): Likewise.
17591 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
17592 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
17593 (final_emit_insn_group_barriers): Likewise.
17594 * config/m32r/m32r.c (m32r_is_insn): Likewise.
17595 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
17596 (mips16_insn_length): Likewise.
17597 * config/pa/pa.c (pa_reorg): Likewise.
17598 (pa_combine_instructions): Likewise.
17599 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
17600 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
17601 (sh_reorg): Likewise.
17602 (split_branches): Likewise.
17603 * config/spu/spu.c (get_branch_target): Likewise.
17605 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
17608 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
17610 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
17611 Fix declaration name.
17613 2013-03-28 Lawrence Crowl <crowl@google.com>
17615 * graphds.h (struct graph.indicies): Remove unused.
17616 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
17617 (SCOP_ORIGINAL_PDDRS): Remove unused.
17618 * sese.h (extern insert_loop_close_phis): Removed unused.
17619 (extern insert_guard_phis): Removed unused.
17620 (extern ivtype_map_elt_info): Removed unused.
17621 (new_ivtype_map_elt): Removed unused.
17622 * sese.c (ivtype_map_elt_info): Removed unused.
17624 2013-03-28 Lawrence Crowl <crowl@google.com>
17626 * Makefile.in: Add several missing include dependences.
17628 (test-dump.o): New. This object is not added to any executable,
17629 but is present for ad-hoc testing.
17631 (debug (const bitmap_head_def &)): New.
17632 (debug (const bitmap_head_def *)): New.
17634 (extern debug (const bitmap_head_def &)): New.
17635 (extern debug (const bitmap_head_def *)): New.
17637 (debug (edge_def &)): New.
17638 (debug (edge_def *)): New.
17640 (debug (basic_block_def &)): New.
17641 (debug (basic_block_def *)): New.
17643 (dump_node (const_tree, int, FILE *)): Correct source file.
17645 (debug (die_struct &)): New.
17646 (debug (die_struct *)): New.
17648 (extern debug (die_struct &)): New.
17649 (extern debug (die_struct *)): New.
17650 * gimple-pretty-print.c
17651 (debug (gimple_statement_d &)): New.
17652 (debug (gimple_statement_d *)): New.
17653 * gimple-pretty-print.h
17654 (extern debug (gimple_statement_d &)): New.
17655 (extern debug (gimple_statement_d *)): New.
17657 (debug (ira_allocno_copy &)): New.
17658 (debug (ira_allocno_copy *)): New.
17659 (debug (ira_allocno &)): New.
17660 (debug (ira_allocno *)): New.
17662 (extern debug (ira_allocno_copy &)): New.
17663 (extern debug (ira_allocno_copy *)): New.
17664 (extern debug (ira_allocno &)): New.
17665 (extern debug (ira_allocno *)): New.
17667 (debug (live_range &)): New.
17668 (debug (live_range *)): New.
17670 (debug (lra_live_range &)): New.
17671 (debug (lra_live_range *)): New.
17673 (debug (lra_live_range &)): New.
17674 (debug (lra_live_range *)): New.
17676 (debug (omega_pb_d &)): New.
17677 (debug (omega_pb_d *)): New.
17679 (extern debug (omega_pb_d &)): New.
17680 (extern debug (omega_pb_d *)): New.
17682 (debug (const rtx_def &)): New.
17683 (debug (const rtx_def *)): New.
17685 (debug_tree (tree): Move within file.
17686 (debug_raw (const tree_node &)): New.
17687 (debug_raw (const tree_node *)): New.
17688 (dump_tree_via_hooks (const tree_node *, int)): New.
17689 (debug (const tree_node &)): New.
17690 (debug (const tree_node *)): New.
17691 (debug_verbose (const tree_node &)): New.
17692 (debug_verbose (const tree_node *)): New.
17693 (debug_head (const tree_node &)): New.
17694 (debug_head (const tree_node *)): New.
17695 (debug_body (const tree_node &)): New.
17696 (debug_body (const tree_node *)): New.
17697 (debug_vec_tree (tree): Move and reimplement in terms of dump.
17698 (debug (vec<tree, va_gc> &)): New.
17699 (debug (vec<tree, va_gc> *)): New.
17701 (extern debug (const rtx_def &)): New.
17702 (extern debug (const rtx_def *)): New.
17704 (debug_raw (simple_bitmap_def &)): New.
17705 (debug_raw (simple_bitmap_def *)): New.
17706 (debug (simple_bitmap_def &)): New.
17707 (debug (simple_bitmap_def *)): New.
17709 (extern debug (simple_bitmap_def &)): New.
17710 (extern debug (simple_bitmap_def *)): New.
17711 (extern debug_raw (simple_bitmap_def &)): New.
17712 (extern debug_raw (simple_bitmap_def *)): New.
17714 (debug (vinsn_def &)): New.
17715 (debug (vinsn_def *)): New.
17716 (debug_verbose (vinsn_def &)): New.
17717 (debug_verbose (vinsn_def *)): New.
17718 (debug (expr_def &)): New.
17719 (debug (expr_def *)): New.
17720 (debug_verbose (expr_def &)): New.
17721 (debug_verbose (expr_def *)): New.
17722 (debug (vec<rtx> &)): New.
17723 (debug (vec<rtx> *)): New.
17725 (extern debug (vinsn_def &)): New.
17726 (extern debug (vinsn_def *)): New.
17727 (extern debug_verbose (vinsn_def &)): New.
17728 (extern debug_verbose (vinsn_def *)): New.
17729 (extern debug (expr_def &)): New.
17730 (extern debug (expr_def *)): New.
17731 (extern debug_verbose (expr_def &)): New.
17732 (extern debug_verbose (expr_def *)): New.
17733 (extern debug (vec<rtx> &)): New.
17734 (extern debug (vec<rtx> *)): New.
17736 (_list_iter_cond_expr): Make inline instead of static.
17738 (debug (sreal &)): New.
17739 (debug (sreal *)): New.
17741 (extern debug (sreal &)): New.
17742 (extern debug (sreal *)): New.
17744 (extern debug_raw (const tree_node &)): New.
17745 (extern debug_raw (const tree_node *)): New.
17746 (extern debug (const tree_node &)): New.
17747 (extern debug (const tree_node *)): New.
17748 (extern debug_verbose (const tree_node &)): New.
17749 (extern debug_verbose (const tree_node *)): New.
17750 (extern debug_head (const tree_node &)): New.
17751 (extern debug_head (const tree_node *)): New.
17752 (extern debug_body (const tree_node &)): New.
17753 (extern debug_body (const tree_node *)): New.
17754 (extern debug (vec<tree, va_gc> &)): New.
17755 (extern debug (vec<tree, va_gc> *)): New.
17757 (debug (struct loop &)): New.
17758 (debug (struct loop *)): New.
17759 (debug_verbose (struct loop &)): New.
17760 (debug_verbose (struct loop *)): New.
17761 * tree-dump.c: Add header dependence.
17763 (extern debug (struct loop &)): New.
17764 (extern debug (struct loop *)): New.
17765 (extern debug_verbose (struct loop &)): New.
17766 (extern debug_verbose (struct loop *)): New.
17768 (debug (data_reference &)): New.
17769 (debug (data_reference *)): New.
17770 (debug (vec<data_reference_p> &)): New.
17771 (debug (vec<data_reference_p> *)): New.
17772 (debug (vec<ddr_p> &)): New.
17773 (debug (vec<ddr_p> *)): New.
17775 (extern debug (data_reference &)): New.
17776 (extern debug (data_reference *)): New.
17777 (extern debug (vec<data_reference_p> &)): New.
17778 (extern debug (vec<data_reference_p> *)): New.
17779 (extern debug (vec<ddr_p> &)): New.
17780 (extern debug (vec<ddr_p> *)): New.
17782 (debug (pt_solution &)): New.
17783 (debug (pt_solution *)): New.
17785 (extern debug (pt_solution &)): New.
17786 (extern debug (pt_solution *)): New.
17788 (debug (_var_map &)): New.
17789 (debug (_var_map *)): New.
17790 (debug (tree_live_info_d &)): New.
17791 (debug (tree_live_info_d *)): New.
17793 (extern debug (_var_map &)): New.
17794 (extern debug (_var_map *)): New.
17795 (extern debug (tree_live_info_d &)): New.
17796 (extern debug (tree_live_info_d *)): New.
17798 2013-03-28 Jan Hubicka <jh@suse.cz>
17800 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
17802 2013-03-28 Ian Bolton <ian.bolton@arm.com>
17804 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
17805 record only when desired or required.
17807 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
17809 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
17810 *vec_extractv2di_1_rex64. Use x64 isa attribute.
17812 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17814 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
17815 (*andsi3_compare0_uxtw): New pattern.
17816 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
17817 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
17819 2013-03-28 Jan Hubicka <jh@suse.cz>
17821 * data-streamer-in.c (streamer_read_gcov_count): New function.
17822 * gimple-streamer-out.c: Include value-prof.h.
17823 (output_gimple_stmt): Output histogram.
17824 (output_bb): Use streamer_write_gcov_count.
17825 * value-prof.c: Include data-streamer.h
17826 (dump_histogram_value): Add HIST_TYPE_MAX.
17827 (stream_out_histogram_value): New function.
17828 (stream_in_histogram_value): New function.
17829 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
17830 (stream_out_histogram_value, stream_in_histogram_value): Declare.
17831 * data-streamer-out.c (streamer_write_gcov_count): New function.
17832 (streamer_write_gcov_count_stream): New function.
17833 * lto-cgraph.c (lto_output_edge): Update counter streaming.
17834 (lto_output_node): Likewise.
17835 (input_node, input_edge): Likewise.
17836 * lto-streamer-out.c (output_cfg): Update streaming.
17837 * lto-streamer-in.c (input_cfg): Likewise.
17838 * data-streamer.h (streamer_write_gcov_count,
17839 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
17840 * gimple-streamer-in.c: Include value-prof.h
17841 (input_gimple_stmt): Input histograms.
17842 (input_bb): Update profile streaming.
17844 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
17846 * genmodes.c (emit_max_int): New function.
17847 (emit_insn_modes_h): Added call to emit_max_function.
17848 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
17850 * machmode.def: Fixed comment.
17852 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
17854 * combine.c (try_combine): Removed useless assert.
17855 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
17857 2013-03-28 Marek Polacek <polacek@redhat.com>
17858 Richard Biener <rguenther@suse.de>
17860 PR tree-optimization/56695
17861 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
17862 build signed result of a vector comparison.
17863 * tree-cfg.c (verify_gimple_comparison): Check that a result
17864 of a vector comparison has signed type.
17866 2013-03-28 Richard Biener <rguenther@suse.de>
17868 PR tree-optimization/37021
17869 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
17870 do not restrict gaps between groups.
17871 * tree-vect-stmts.c (vectorizable_load): Properly account for
17872 a gap between groups.
17874 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
17876 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
17877 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
17880 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
17882 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
17883 * doc/extend.texi (Named Address Spaces): Ditto.
17884 (Variable Attributes): Ditto.
17886 2013-03-27 Kai Tietz <ktietz@redhat.com>
17888 * config.build: Add support for cygwin x64 target.
17889 * config.gcc: Likewise.
17890 * config.host: Likewise.
17891 * configure.ac: Likewise
17892 * configure: Regenerated.
17894 2013-03-27 Kai Tietz <ktietz@redhat.com>
17896 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
17897 * config/i386/t-cygwin-w64: New file.
17898 * config/i386/cygwin-w64.h: New file.
17899 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
17900 and add support for x64-cygwin target.
17901 (CPP_SPEC): Likewise.
17902 (CXX_WRAP_SPEC_LIST): Undefine before define.
17903 (LIBGCJ_SONAME): Use 15 as version.
17905 2013-03-27 Richard Biener <rguenther@suse.de>
17907 PR tree-optimization/56716
17908 * tree-ssa-structalias.c (perform_var_substitution): Adjust
17909 dumping for ref nodes.
17911 2013-03-27 Martin Jambor <mjambor@suse.cz>
17913 PR tree-optimization/55334
17914 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
17915 restricted pointers to arrays.
17917 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
17919 * Makefile.in (.SUFFIXES): Add .cc.
17920 (.c.o): Apply same recipe for implicit rule .cc.o.
17922 2013-03-27 Richard Biener <rguenther@suse.de>
17924 PR tree-optimization/37021
17925 * tree-vect-data-refs.c (vect_check_strided_load): Allow
17926 REALPART/IMAGPART_EXPRs around the supported refs.
17927 * tree-ssa-structalias.c (find_func_aliases): Assume that
17928 floating-point values are not used to transfer pointers.
17930 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
17932 * target.def (TARGET_HAS_IFUNC_P): New target hook.
17933 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
17934 * doc/tm.texi: Regenerate.
17935 * targhooks.h (default_has_ifunc_p): New.
17936 * targhooks.c (default_has_ifunc_p): Ditto.
17937 * config/linux-protos.h: New file.
17938 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
17939 hook for linux which disables support of indirect functions in android.
17940 * config/linux-android.c: New file.
17941 * config/t-linux-android.c: Ditto.
17942 * config.gcc: Added new object file linux-android.o.
17943 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
17944 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
17945 * varasm.c (do_assemble_alias): Likewise.
17946 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
17947 doesn't support indirect functions.
17948 * configure: Regenerate.
17950 2013-03-27 Bin Cheng <bin.cheng@arm.com>
17953 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
17954 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
17957 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17959 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
17961 2013-03-27 Terry Guo <terry.guo@arm.com>
17963 * config/arm/arm-cores.def: Added core cortex-r7.
17964 * config/arm/arm-tune.md: Regenerated.
17965 * config/arm/arm-tables.opt: Regenerated.
17966 * doc/invoke.texi: Added entry for core cortex-r7.
17968 2013-03-27 Walter Lee <walt@tilera.com>
17970 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
17971 double-decrement of next_scratch_regno.
17973 2013-03-27 Walter Lee <walt@tilera.com>
17975 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
17977 (insn_v1mulus): Ditto.
17978 (insn_v2muls): Ditto.
17980 2013-03-27 Walter Lee <walt@tilera.com>
17982 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
17983 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
17985 2013-03-27 Walter Lee <walt@tilera.com>
17987 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
17988 (*sibcall_value): Ditto.
17990 2013-03-27 Walter Lee <walt@tilera.com>
17992 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
17993 (insn_mnz_v8qi): ... this ...
17994 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
17996 (insn_v<n>mnz): Replaced by ...
17997 (insn_v1mnz): ... this ...
17998 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
18000 (insn_mz_<mode>): Replaced by ...
18001 (insn_mz_v8qi): ... this ...
18002 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
18004 (insn_v<n>mz): Replaced by ...
18005 (insn_v1mz): ... this ...
18006 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
18009 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
18011 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
18013 2013-03-26 Roland McGrath <mcgrathr@google.com>
18015 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
18016 than fprintf with a non-constant, non-format string.
18018 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
18020 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
18021 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
18022 operand 0 predicate.
18023 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
18024 attribute. Use general_x64nomem_operand as operand 1 predicate.
18025 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
18026 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
18027 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
18028 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
18029 (mov<mode>_insv_1): Remove expander. Merge insn with
18030 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
18031 Use general_x64nomem_operand as operand 1 predicate.
18032 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
18033 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
18034 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
18035 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
18036 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
18037 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
18038 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
18039 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
18040 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
18041 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
18042 (general_x64nomem_operand): Ditto.
18044 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
18046 * config/rtems.opt: Add -pthread option.
18048 2013-03-26 Richard Biener <rguenther@suse.de>
18050 * alias.c (find_base_term): Avoid redundant and not used recursion.
18051 (base_alias_check): Get the initial base term from the caller.
18052 (true_dependence_1): Compute and pass base terms to base_alias_check.
18053 (write_dependence_p): Likewise.
18054 (may_alias_p): Likewise.
18056 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
18058 * config/aarch64/aarch64.c (aarch64_classify_address): Support
18059 PC-relative load in SI modes and above only.
18061 2013-03-26 Xinyu Qi <xyqi@marvell.com>
18063 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
18064 * config/arm/iwmmxt.md (WCGR0): Update.
18065 (WCGR1, WCGR2, WCGR3): Likewise.
18067 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
18069 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
18070 Use x64 and nox64 isa attributes.
18072 2013-03-26 Richard Biener <rguenther@suse.de>
18074 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
18075 alignment computations and rely on get_object_alignment_1
18076 for the !TYPE_P case.
18077 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
18079 2013-03-26 Walter Lee <walt@tilera.com>
18081 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
18082 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
18084 2013-03-25 Jeff Law <law@redhat.com>
18086 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
18087 check for INTEGRAL_TYPE_P that was missing due to checking in
18088 wrong version of prior patch.
18090 2013-03-25 Walter Lee <walt@tilera.com>
18092 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
18093 TILEGX_INSN_SHUFFLEBYTES1.
18094 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
18096 (tilegx_builtins): Ditto.
18097 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
18099 2013-03-25 Walter Lee <walt@tilera.com>
18101 * config/tilegx/tilegx.md (floatsisf2): New pattern.
18102 (floatunssisf2): New pattern.
18103 (floatsidf2): New pattern.
18104 (floatunssidf2): New pattern.
18106 2013-03-25 Walter Lee <walt@tilera.com>
18108 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
18109 tests for constraint J, K, N, P.
18111 2013-03-25 Walter Lee <walt@tilera.com>
18113 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
18114 Use indirect/pcrel encoding.
18115 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
18118 2013-03-25 Steve Ellcey <sellcey@mips.com>
18120 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
18121 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
18122 * config/mips/mips.c (mips_option_override): Set IMADD default.
18123 * config/mips/mips.h (PTF_AVOID_IMADD): New.
18124 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
18125 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
18126 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
18127 * doc/invoke.texi (-mimadd/-mno-imadd): New.
18129 2013-03-25 Jeff Law <law@redhat.com>
18131 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
18132 slightly to avoid creating and folding useless trees. Simplify
18133 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
18135 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
18137 * config/i386/i386.md (*zero_extendsidi2): Merge with
18138 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
18139 * config/i386/predicates.md (x86_64_zext_operand): Rename from
18140 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
18141 targets. Clarify comment.
18143 2013-03-25 Martin Jambor <mjambor@suse.cz>
18145 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
18146 pass-through jump functions differently.
18147 (ipa_read_jump_function): Likewise. Also use setter functions to set
18150 2013-03-25 Martin Jambor <mjambor@suse.cz>
18152 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
18153 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
18155 (ipa_get_indirect_edge_target): New function.
18156 (devirtualization_time_bonus): New parameter known_aggs, pass it to
18157 ipa_get_indirect_edge_target. Update all callers.
18158 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
18159 ipa_get_indirect_edge_target_1 instead of calling
18160 ipa_get_indirect_edge_target.
18161 (create_specialized_node): Pass aggvlas to
18162 ipcp_discover_new_direct_edges.
18164 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18166 * config/arm/arm.md (f_sels, f_seld): New types.
18167 (*cmov<mode>): New pattern.
18168 * config/arm/predicates.md (arm_vsel_comparison_operator): New
18171 2013-03-25 Kai Tietz <ktietz@redhat.com>
18173 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
18174 POSIX-printf for mingw-hosted builds.
18176 2013-03-25 Richard Biener <rguenther@suse.de>
18178 PR middle-end/56694
18179 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
18180 must-not-throw stmt location.
18182 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18184 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
18185 Emit load-acquire versions when acq is true.
18186 (arm_emit_store_exclusive): Add rel parameter.
18187 Emit store-release versions when rel is true.
18188 (arm_split_compare_and_swap): Use acquire-release instructions
18190 of barriers when appropriate.
18191 (arm_split_atomic_op): Likewise.
18192 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
18193 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
18194 (VUNSPEC_SLX): Likewise.
18195 (VUNSPEC_LDA): Likewise.
18196 (VUNSPEC_STL): Likewise.
18197 * config/arm/sync.md (atomic_load<mode>): New pattern.
18198 (atomic_store<mode>): Likewise.
18199 (arm_load_acquire_exclusive<mode>): Likewise.
18200 (arm_load_acquire_exclusivesi): Likewise.
18201 (arm_load_acquire_exclusivedi): Likewise.
18202 (arm_store_release_exclusive<mode>): Likewise.
18204 2013-03-25 Catherine Moore <clm@codesourcery.com>
18206 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
18207 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
18208 * config/mip/predicates.md (lwsp_swsp_operand,
18209 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
18210 sb16_operand, db4_operand, db7_operand, ib3_operand,
18211 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
18212 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
18213 andi16_operand): New predicates.
18214 * config/mips/mips.md (compression): New attribute.
18215 (enabled): New attribute.
18216 (length): Consider compression in computing length.
18217 (shift_compression): New code attribute.
18218 (*add<mode>3): New operands. Record compression.
18219 (sub<mode>3): Likewise.
18220 (one_cmpl<mode>2): Likewise.
18221 (*and<mode>3): Likewise.
18222 (*ior<mode>3): Likewise.
18223 (unnamed pattern for xor): Likewise.
18224 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
18225 (*<optab><mode>3): Likewise.
18226 (*mov<mode>_internal: Likewise.
18227 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
18228 (mips_unsigned_immediate_p): New.
18229 (umips_lwsp_swsp_address_p): New.
18230 (m16_based_address_p): New.
18231 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
18232 (mips_unsigned_immediate_p): New prototype.
18233 (lwsp_swsp_address_p): New prototype.
18234 (m16_based_address_p): New prototype.
18235 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
18236 (mips_signed_immediate_p): New function.
18237 (m16_based_address_p): New function.
18238 (lwsp_swsp_address_p): New function.
18239 (mips_print_operand_punctuation): Recognize short delay slot insns
18240 for microMIPS.add<mode>3"
18242 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18245 * config/arm/iterators.md (v_cmp_result): New mode attribute.
18246 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
18248 2013-03-25 Richard Biener <rguenther@suse.de>
18250 PR tree-optimization/56689
18251 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
18254 2013-03-25 Richard Biener <rguenther@suse.de>
18256 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
18258 (memory_references): Likewise.
18259 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
18260 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
18261 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
18262 (gather_mem_refs_in_loops): Fold into ...
18263 (analyze_memory_references): ... this. Move initialization
18264 to tree_ssa_lim_initialize.
18265 (fill_always_executed_in): Rename to ...
18266 (fill_always_executed_in_1): ... this.
18267 (fill_always_executed_in): Move contains_call computation to
18268 this new function from ...
18269 (tree_ssa_lim_initialize): ... here.
18270 (tree_ssa_lim): Call fill_always_executed_in.
18272 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
18274 * postreload.c (reload_combine): Fix code detecting returns.
18276 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
18278 * function.c (emit_use_return_register_into_block): On cc0 targets,
18279 do not emit the sequence between cc0 setter and user.
18281 2013-03-25 Kai Tietz <ktietz@redhat.com>
18283 * config/i386/predicates.md (local_symbolic_operand): Interpret
18284 dll-imported symbols as none-local.
18286 2013-03-25 Richard Biener <rguenther@suse.de>
18288 * tree-ssa-loop-im.c (struct depend): Remove.
18289 (struct lim_aux_data): Make depends a vec of gimples.
18290 (free_lim_aux_data): Adjust.
18291 (add_dependency): Likewise.
18292 (set_level): Likewise.
18294 2013-03-25 Richard Biener <rguenther@suse.de>
18296 PR middle-end/56434
18297 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
18298 the pointer returned by calls with ECF_MALLOC set.
18300 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
18302 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
18304 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
18306 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
18307 using MMXMODE mode iterator.
18308 (*move<mode>_internal): Merge with *movv2sf_internal and
18309 *movv2sf_internal_rex64 using MMXMODE mode iterator.
18311 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
18313 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
18314 (record_last_mem_set_info): Likewise.
18316 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
18317 of XNEWVEC followed by memset.
18318 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
18320 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
18322 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
18323 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
18324 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
18325 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
18326 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
18327 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
18328 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
18329 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
18330 BARRIER_P instead of GET_CODE.
18332 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
18334 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
18335 inaccuracy in the probing code.
18337 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
18338 (ctrapdi4): Likewise.
18340 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
18342 * calls.c (expand_call): Add missing guard to code handling return
18343 of non-BLKmode structures in MSB.
18344 * function.c (expand_function_end): Likewise.
18346 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
18348 * combine.c (try_combine): Adjust comment. Do not add the set of
18349 insn #0 if the destination indirectly is set or dies in insn #2.
18350 Tidy up code to distribute a new note.
18352 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
18354 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
18355 also for alternatives 16 and 17.
18357 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
18359 * config/i386/sse.md (*mov<mode>_internal): Merge with
18360 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
18361 Emit insn template depending on type attribute. Use
18362 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
18363 movd instead of movq mnemonic for interunit moves. Rewrite mode
18364 attribute calculation. Remove unit attribute calculation.
18365 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
18366 Set prefix_data16 attribute for DImode ssemov types.
18367 Use Ym instead of y for SSE-MMX conversion alternatives.
18368 Reorder operand constraints.
18370 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
18372 * df.h (df_insn_delete): Adjust prototype.
18373 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
18374 and let it decide whether mark the basic block dirty.
18375 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
18376 * df-scan.c (df_insn_info_delete): New helper function, split
18377 off from df_insn_delete.
18378 (df_scan_free_bb_info): Use it.
18379 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
18381 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
18382 that the insn is actually an insn and it has a non-NULL basic block.
18383 Do not mark basic block dirty if only deleting a DEBUG_INSN.
18385 2013-03-22 Richard Biener <rguenther@suse.de>
18387 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
18389 (mem_ref_alloc): Do not allocate them.
18390 (refs_independent_p): Do not query or maintain a cache.
18392 2013-03-22 Richard Biener <rguenther@suse.de>
18394 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
18395 (gather_mem_refs_in_loops): Do not compute it.
18396 (analyze_memory_references): Do not allocate it.
18397 (tree_ssa_lim_finalize): Do not free it.
18398 (for_all_locs_in_loop): Do not query all_refs_in_loop.
18400 2013-03-22 Richard Biener <rguenther@suse.de>
18402 * is-a.h (as_a): Use gcc_checking_assert.
18404 2013-03-22 Ian Bolton <ian.bolton@arm.com>
18406 * config/aarch64/aarch64.c (aarch64_print_operand): New
18407 format specifier for printing a constant in hex.
18408 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
18409 format specifier for printing second operand.
18411 2013-03-22 Richard Biener <rguenther@suse.de>
18413 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
18415 (gather_mem_refs_in_loops): Perform store accumulation here.
18416 (create_vop_ref_mapping_loop): Remove.
18417 (create_vop_ref_mapping): Likewise.
18418 (analyze_memory_references): Initialize refs_stored_in_loop.
18419 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
18420 (record_indep_loop): Remove.
18421 (record_dep_loop): New function.
18422 (ref_indep_loop_p_1): Adjust to only walk over references
18423 in the loop, not its subloops.
18424 (ref_indep_loop_p): Rename to ...
18425 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
18426 maintaining a more fine-grained cache.
18427 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
18428 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
18430 2013-03-22 Richard Biener <rguenther@suse.de>
18432 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
18433 (struct mem_ref): Make accesses_in_loop a vec of a vec of
18434 aggregate mem_ref_loc.
18435 (free_mem_ref_locs): Inline into ...
18436 (memref_free): ... this and adjust.
18437 (mem_ref_alloc): Adjust.
18438 (mem_ref_locs_alloc): Remove.
18439 (record_mem_ref_loc): Adjust.
18440 (get_all_locs_in_loop): Rewrite into ...
18441 (for_all_locs_in_loop): ... this iterator.
18442 (rewrite_mem_ref_loc): New functor.
18443 (rewrite_mem_refs): Use for_all_locs_in_loop.
18444 (sm_set_flag_if_changed): New functor.
18445 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
18446 (ref_always_accessed): New functor.
18447 (ref_always_accessed_p): Use for_all_locs_in_loop.
18449 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
18451 * tree-pass.h (PROP_gimple_lvec): New.
18452 * passes.c (dump_properties): Handle PROP_gimple_lvec.
18453 (init_optimization_passes): Move pass_lower_vector.
18454 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
18456 (pass_lower_vector): Provide PROP_gimple_lvec.
18457 (pass_lower_vector_ssa): Likewise.
18458 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
18460 2013-03-21 Mark Wielaard <mjw@redhat.com>
18462 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
18464 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
18466 * config/i386/i386.md (*movdi_internal): Disparage slightly
18467 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
18468 conversion alternatives.
18470 2013-03-21 Jakub Jelinek <jakub@redhat.com>
18472 PR middle-end/48087
18473 * diagnostic.def (DK_WERROR): New kind.
18474 * diagnostic.h (werrorcount): Define.
18475 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
18476 promoted to DK_ERROR, increment DK_WERROR counter instead of
18478 * toplev.c (toplev_main): Call print_ignored_options even if
18479 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
18480 even if just werrorcount is non-zero.
18483 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
18485 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
18486 (string_cst_pool_decl): New function.
18487 (optimize_one_addr_into_implicit_ptr): New function.
18488 (resolve_addr_in_expr): Optimize DWARF location expression
18489 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
18490 which doesn't live in memory, but has DW_AT_location or
18491 DW_AT_const_value, or refers to a string literal, into
18492 DW_OP_GNU_implicit_pointer.
18493 (optimize_location_into_implicit_ptr): New function.
18494 (resolve_addr): If removing DW_AT_location of a variable because
18495 it was DW_OP_addr of address of the variable, but the variable doesn't
18496 live in memory, try to emit const value attribute for the initializer.
18498 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
18500 * tree.h (VECTOR_TYPE_P): New macro.
18501 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
18502 TYPE_MODE): Use it.
18503 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
18504 VEC_COND_EXPR cannot be lvalues.
18505 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
18507 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
18509 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
18510 Restrict the transformation to equal modes.
18512 2013-03-21 Richard Biener <rguenther@suse.de>
18514 PR tree-optimization/39326
18515 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
18516 (MEM_ANALYZABLE): Adjust.
18517 (record_mem_ref_loc): Move bitmap ops ...
18518 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
18519 unanalyzable refs, do not record locations for it.
18520 (analyze_memory_references): Allocate ref zero as shared
18522 (refs_independent_p): Do not test for unanalyzed mems here.
18523 (ref_indep_loop_p_1): Special-case disambiguation against
18524 the unanalyzed ref.
18525 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
18527 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
18529 * config/arm/arm-protos.h (tune_params): Add
18530 prefer_neon_for_64bits field.
18531 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
18532 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
18533 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
18534 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
18535 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
18536 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
18537 (arm_option_override): Handle -mneon-for-64bits new option.
18538 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
18539 (prefer_neon_for_64bits): Declare new variable.
18540 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
18541 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
18542 (arch_enabled): Handle new arch types. Remove support for onlya8
18544 (one_cmpldi2): Use new arch names.
18545 (zero_extend<mode>di2, extend<mode>di2): Ditto.
18546 * config/arm/arm.opt (mneon-for-64bits): Add option.
18547 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
18548 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
18549 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
18551 * doc/invoke.texi (-mneon-for-64bits): Document.
18553 2013-03-21 Richard Biener <rguenther@suse.de>
18555 PR tree-optimization/39326
18556 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
18557 (sort_bbs_in_loop_postorder_cmp): New function.
18558 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
18560 2013-03-21 Richard Biener <rguenther@suse.de>
18562 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
18563 (vect_insert_into_interleaving_chain): Likewise.
18564 (vect_drs_dependent_in_basic_block): Inline ...
18565 (vect_slp_analyze_data_ref_dependence): ... here. New function,
18567 (vect_analyze_data_ref_dependence): ... here. Simplify.
18568 (vect_check_interleaving): Simplify.
18569 (vect_analyze_data_ref_dependences): Likewise. Split out ...
18570 (vect_slp_analyze_data_ref_dependences): ... this new function.
18571 (dr_group_sort_cmp): New function.
18572 (vect_analyze_data_ref_accesses): Compute data-reference groups
18573 here instead of in vect_analyze_data_ref_dependence. Use
18574 a more efficient algorithm.
18575 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
18576 vect_slp_analyze_data_ref_dependences. Call
18577 vect_analyze_data_ref_accesses earlier.
18578 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
18579 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
18580 (vect_slp_analyze_data_ref_dependences): New prototype.
18582 2013-03-21 Richard Biener <rguenther@suse.de>
18584 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
18585 ref is stored in the loop.
18586 (find_refs_for_sm): Walk only over all stores.
18587 (store_motion_loop): Allocate from lim_bitmap_obstack.
18588 (store_motion): Likewise.
18590 2013-03-21 Richard Biener <rguenther@suse.de>
18592 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
18593 Update virtual SSA form.
18595 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
18597 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
18598 * configure: Regenerate.
18599 * config.in: Regenerate.
18600 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
18601 if !HAVE_LD_EH_FRAME_CIEV3.
18603 2013-03-21 Richard Biener <rguenther@suse.de>
18605 * tree-cfg.c (verify_expr_no_block): New function.
18606 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
18607 nor DECL_VALUE_EXPR have locations with associated blocks.
18608 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
18609 (clear_unused_block_pointer): Remove code dealing with
18610 blocks in DECL_DEBUG_EXPR locations.
18612 2013-03-21 Richard Biener <rguenther@suse.de>
18614 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
18615 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
18616 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
18617 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
18618 instead of DECL_DEBUG_EXPR_IS_FROM.
18619 * gimplify.c (gimplify_modify_expr): Likewise.
18620 * tree-cfg.c (verify_expr_location_1): Likewise.
18621 * tree-complex.c (create_one_component_var): Likewise.
18622 * tree-sra.c (create_access_replacement): Likewise.
18623 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
18624 (clear_unused_block_pointer): Likewise.
18625 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
18626 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
18627 * var-tracking.c (var_debug_decl): Likewise.
18628 (track_expr_p): Likewise.
18629 * tree-inline.c (add_local_variables): Likewise. Set
18630 DECL_HAS_DEBUG_EXPR_P after copying it.
18631 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
18632 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
18634 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
18637 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
18638 * configure: Regenerate.
18639 * config.in: Regenerate.
18640 * config/i386/i386.md (*movdf_internal): Use
18641 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
18642 movd instead of movq mnemonic for interunit moves.
18643 (*movdi_internal): Ditto.
18645 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
18647 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
18648 (abd<mode>_3): New pattern.
18649 (aba<mode>_3): New pattern.
18650 (fabd<mode>_3): New pattern.
18652 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
18654 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
18655 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
18656 occurrence of REGISTER_PREFIX as its empty string.
18658 2013-03-20 Jeff Law <law@redhat.com>
18660 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
18661 addititional equivalences for equality comparisons between an SSA_NAME
18662 and a constant where the SSA_NAME was set from a widening conversion.
18664 2013-03-20 Walter Lee <walt@tilera.com>
18666 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
18668 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
18670 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
18671 depending on type attribute.
18672 (*movti_internal): Ditto.
18673 (*movtf_internal): Ditto.
18674 (*movxf_internal): Ditto.
18675 (*movdf_internal): Ditto.
18676 (*movsf_internal): Ditto.
18678 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
18680 * config/i386/i386.md (*movti_internal): Set prefix attribute to
18681 maybe_vex for sselog1 and ssemov types.
18682 (*movdi_internal): Reorder operand constraints.
18683 (*movsi_internal): Ditto. Set prefix attribute to
18684 maybe_vex for sselog1 and ssemov types.
18685 (*movtf_internal): Set prefix attribute to maybe_vex
18686 for sselog1 and ssemov types.
18687 (*movdf_internal): Ditto. Set prefix_data16 attribute for
18688 DImode ssemov types. Reorder operand constraints.
18689 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
18690 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
18691 attribute for SImode ssemov types. Reorder operand constraints.
18693 2013-03-20 Martin Jambor <mjambor@suse.cz>
18695 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
18696 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
18698 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
18700 * config/rs6000/predicates.md (indexed_address, update_address_mem
18701 update_indexed_address_mem): New predicates.
18702 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
18703 attribute for load/store instructions.
18704 * config/rs6000/dfp.md (movsd_store): Likewise.
18705 (movsd_load): Likewise.
18706 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
18707 (unnamed HI->DI extend define_insn): Likewise.
18708 (unnamed SI->DI extend define_insn): Likewise.
18709 (unnamed QI->SI extend define_insn): Likewise.
18710 (unnamed QI->HI extend define_insn): Likewise.
18711 (unnamed HI->SI extend define_insn): Likewise.
18712 (unnamed HI->SI extend define_insn): Likewise.
18713 (extendsfdf2_fpr): Likewise.
18714 (movsi_internal1): Likewise.
18715 (movsi_internal1_single): Likewise.
18716 (movhi_internal): Likewise.
18717 (movqi_internal): Likewise.
18718 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
18719 attribute for load/store instructions.
18720 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
18722 (mov<mode>_softfloat): Likewise.
18723 (mov<mode>_hardfloat32): Likewise.
18724 (mov<mode>_hardfloat64): Likewise.
18725 (mov<mode>_softfloat64): Likewise.
18726 (movdi_internal32): Likewise.
18727 (movdi_internal64): Likewise.
18728 (probe_stack_<mode>): Likewise.
18730 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
18732 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
18733 floating point, and decimal floating point to reload iterator.
18735 * config/rs6000/constraints.md (wl constraint): New constraints to
18736 return FLOAT_REGS if certain options are used to reduce the number
18737 of separate patterns that exist in the file.
18738 (wx constraint): Likewise.
18739 (wz constraint): Likewise.
18741 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
18742 -mdebug=reg, print wg, wl, wx, and wz constraints.
18743 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
18744 Initialize the reload functions for 64-bit binary/decimal floating
18746 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
18747 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
18748 create the buffer on the stack to overcome not having a 32-bit
18750 (rs6000_emit_move): Likewise.
18751 (rs6000_secondary_memory_needed_rtx): Likewise.
18752 (rs6000_alloc_sdmode_stack_slot): Likewise.
18753 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
18754 via xxlxor, just like DFmode 0.0.
18756 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
18757 define as 1 if we are running on a power7 or newer.
18758 (enum r6000_reg_class_enum): Add new constraints.
18760 * config/rs6000/dfp.md (movsd): Delete, combine with binary
18761 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
18762 with other moves by using conditional constraits (wg). Use LFIWZX
18763 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
18764 (movsd splitter): Likewise.
18765 (movsd_hardfloat): Likewise.
18766 (movsd_softfloat): Likewise.
18768 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
18769 binary and decimal floating point moves.
18770 (fmove_ok): New attributes to combine binary and decimal floating
18771 point moves, and to combine power6x (mfpgpr) moves along normal
18773 (real_value_to_target): Likewise.
18774 (f32_lr): Likewise.
18775 (f32_lm): Likewise.
18776 (f32_li): Likewise.
18777 (f32_sr): Likewise.
18778 (f32_sm): Likewise.
18779 (f32_si): Likewise.
18780 (movsf): Combine binary and decimal floating point moves. Combine
18781 power6x (mfpgpr) moves with other moves by using conditional
18782 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
18783 (mov<mode> for SFmode/SDmode); Likewise.
18784 (SFmode/SDmode splitters): Likewise.
18785 (movsf_hardfloat): Likewise.
18786 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
18787 (movsf_softfloat): Likewise.
18788 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
18790 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
18791 wx and wz constraints.
18793 * config/rs6000/constraints.md (wg constraint): New constraint to
18794 return FLOAT_REGS if -mmfpgpr (power6x) was used.
18796 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
18799 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
18800 -mdebug=reg, print wg, wl, wx, and wz constraints.
18801 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
18802 Initialize the reload functions for 64-bit binary/decimal floating
18804 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
18805 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
18806 create the buffer on the stack to overcome not having a 32-bit
18808 (rs6000_emit_move): Likewise.
18809 (rs6000_secondary_memory_needed_rtx): Likewise.
18810 (rs6000_alloc_sdmode_stack_slot): Likewise.
18811 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
18812 via xxlxor, just like DFmode 0.0.
18814 * config/rs6000/dfp.md (movdd): Delete, combine with binary
18815 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
18816 with other moves by using conditional constraits (wg). Use LFIWZX
18817 and STFIWX for loading SDmode on power7.
18818 (movdd splitters): Likewise.
18819 (movdd_hardfloat32): Likewise.
18820 (movdd_softfloat32): Likewise.
18821 (movdd_hardfloat64_mfpgpr): Likewise.
18822 (movdd_hardfloat64): Likewise.
18823 (movdd_softfloat64): Likewise.
18825 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
18826 64-bit binary and decimal floating point moves.
18827 (FMOVE64X): Likewise.
18828 (movdf): Combine 64-bit binary and decimal floating point moves.
18829 Combine power6x (mfpgpr) moves with other moves by using
18830 conditional constraits (wg).
18831 (mov<mode> for DFmode/DDmode): Likewise.
18832 (DFmode/DDmode splitters): Likewise.
18833 (movdf_hardfloat32): Likewise.
18834 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
18835 (movdf_softfloat32): Likewise.
18836 (movdf_hardfloat64_mfpgpr): Likewise.
18837 (movdf_hardfloat64): Likewise.
18838 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
18839 (movdf_softfloat64): Likewise.
18840 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
18841 (reload_<mode>_load): Move to later in the file so they aren't in
18842 the middle of the floating point move insns.
18843 (reload_<mode>_store): Likewise.
18845 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
18848 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
18849 constraint if -mdebug=reg.
18850 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
18851 Enable using dd reload support if needed.
18853 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
18854 binary and decimal floating point moves in rs6000.md.
18855 (movtd_internal): Likewise.
18857 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
18858 decimal floating point moves.
18860 (movtf_internal): Likewise.
18861 (mov<mode>_internal, TDmode/TFmode): Likewise.
18862 (movtf_softfloat): Likewise.
18863 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
18865 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
18866 movdi_internal64, using wg constraint for move direct operations.
18867 (movdi_internal64): Likewise.
18869 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
18870 MODES_TIEABLE_P for selected modes. Print the numerical value of
18871 the various virtual registers. Use GPR/FPR first/last values,
18872 instead of hard coding the register numbers. Print which modes
18873 have reload functions registered.
18874 (rs6000_option_override_internal): If -mdebug=reg, trace the options
18875 settings before/after setting cpu, target and subtarget settings.
18876 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
18877 and for secondary reload failures in rs6000_secondary_reload_inner.
18878 (rs6000_secondary_reload_fail): Likewise.
18879 (rs6000_secondary_reload_inner): Likewise.
18881 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
18882 macros for first/last GPR and FPR registers.
18883 (LAST_GPR_REGNO): Likewise.
18884 (FIRST_FPR_REGNO): Likewise.
18885 (LAST_FPR_REGNO): Likewise.
18887 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
18888 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
18889 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
18890 (vcond<mode><mode>): Likewise.
18891 (vcondu<mode><mode>): Likewise.
18892 (vector_gtu<mode>): Likewise.
18893 (vector_gte<mode>): Likewise.
18894 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
18895 to prevent the compiler from converting DImode operations to TImode.
18896 (ior<mode>3): Likewise.
18897 (and<mode>3): Likewise.
18898 (one_cmpl<mode>2): Likewise.
18899 (nor<mode>3): Likewise.
18900 (andc<mode>3): Likewise.
18902 * config/rs6000/constraints.md (wt constraint): New constraint
18903 that returns VSX_REGS if TImode is allowed in VSX registers.
18905 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
18906 constant under VSX.
18908 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
18909 similar to TImode, but it is restricted to being in the GPRs.
18911 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
18912 TImode to occupy a single VSX register.
18914 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
18915 -mvsx-timode for power7/power8.
18916 (power7 cpu): Likewise.
18917 (power8 cpu): Likewise.
18919 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
18920 sure that TFmode/TDmode take up two registers if they are ever
18921 allowed in the upper VSX registers.
18922 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
18924 (rs6000_init_hard_regno_mode_ok): Likewise.
18925 (rs6000_debug_reg_global): Add debugging for PTImode and wt
18926 constraint. Print if LRA is turned on.
18927 (rs6000_option_override_internal): Give an error if -mvsx-timode
18928 and VSX is not enabled.
18929 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
18930 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
18931 to reg+offset addressing. Use PTImode when checking offset
18932 addresses for validity.
18933 (reg_offset_addressing_ok_p): Likewise.
18934 (rs6000_legitimate_offset_address_p): Likewise.
18935 (rs6000_legitimize_address): Likewise.
18936 (rs6000_legitimize_reload_address): Likewise.
18937 (rs6000_legitimate_address_p): Likewise.
18938 (rs6000_eliminate_indexed_memrefs): Likewise.
18939 (rs6000_emit_move): Likewise.
18940 (rs6000_secondary_reload): Likewise.
18941 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
18942 reloads to fpr registers to continue to use reg+offset addressing,
18943 but 64-bit reloads to altivec registers need reg+reg addressing.
18944 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
18945 it. Treat LO_SUM like a PLUS operation.
18946 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
18947 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
18948 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
18949 registers to share a register with a smaller sized type, since VSX
18950 puts scalars in the upper 64-bits.
18951 (print_operand): Add support for PTImode.
18952 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
18953 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
18954 registers, but don't have arithmetic support.
18955 (rs6000_memory_move_cost): Add test for VSX.
18956 (rs6000_opt_masks): Add -mvsx-timode.
18958 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
18961 (VSr): Use wt constraint for TImode.
18962 (VSv): Drop TImode support.
18963 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
18964 (vsx_movti_64bit): Likewise.
18965 (vsx_movti_32bit): Likewise.
18966 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
18967 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
18968 one '?' on the appropriate output constraint. Do not allow TImode
18969 logical operations on 32-bit systems.
18970 (vsx_ior<mode>3): Likewise.
18971 (vsx_xor<mode>3): Likewise.
18972 (vsx_one_cmpl<mode>2): Likewise.
18973 (vsx_nor<mode>3): Likewise.
18974 (vsx_andc<mode>3): Likewise.
18975 (vsx_concat_<mode>): Likewise.
18976 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
18978 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
18979 OPTION_MASK_VSX_TIMODE.
18980 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
18981 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
18983 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
18984 (TI2 iterator): New iterator for TImode, PTImode.
18985 (wd mode attribute): Add values for vector types.
18986 (movti_string): Replace TI move operations with operations for TImode
18987 and PTImode. Add support for TImode being allowed in VSX registers.
18988 (mov<mode>_string, TImode/PTImode): Likewise.
18989 (movti_ppc64): Likewise.
18990 (mov<mode>_ppc64, TImode/PTImode): Likewise.
18991 (TI mode splitters): Likewise.
18993 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
18996 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
18998 PR tree-optimization/56355
18999 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
19000 Also handle integers with undefined overflow.
19002 2013-03-20 Catherine Moore <clm@codesourcery.com>
19003 Maciej W. Rozycki <macro@codesourcery.com>
19004 Tom de Vries <tom@codesourcery.com>
19005 Nathan Sidwell <nathan@codesourcery.com>
19006 Iain Sandoe <iain@codesourcery.com>
19007 Nathan Froyd <froydnj@codesourcery.com>
19008 Chao-ying Fu <fu@mips.com>
19010 * doc/extend.texi: (micromips, nomicromips, nocompression):
19011 Document new function attributes.
19012 * doc/invoke.texi (minterlink-compressed, mmicromips,
19013 m14k, m14ke, m14kec): Document new options.
19014 (minterlink-mips16): Update documentation.
19015 * doc/md.texi (ZC, ZD): Document new constraints.
19016 * configure.ac (gcc_cv_as_micromips): Check if linker
19017 supports the .set micromips directive.
19018 * configure: Regenerate.
19019 * config.in: Regenerate.
19020 * config/mips/mips-tables.opt: Regenerate.
19021 * config/mips/micromips.md: New file.
19022 * constraints.md (ZC, ZD): New constraints.
19023 * config/mips/predicates.md (movep_src_register): New predicate.
19024 (movep_src_operand): New predicate.
19025 (non_volatile_mem_operand): New predicate.
19026 * config/mips/mips.md (multimem): New type.
19027 (length): Differentiate between 17-bit and 18-bit branch offsets.
19028 (MOVEP1, MOVEP2): New mode iterator.
19029 (mov_<load>l): Use ZC constraint.
19030 (mov_<load>r): Likewise.
19031 (mov_<store>l): Likewise.
19032 (mov_<store>r): Likewise.
19033 (*branch_equality<mode>_inverted): Add microMIPS support.
19034 (*branch_equality<mode>): Likewise.
19035 (*jump_absolute): Likewise.
19036 (indirect_jump_<mode>): Likewise.
19037 (tablejump_<mode>): Likewise.
19038 (<optab>_internal): Likewise.
19039 (sibcall_internal): Likewise.
19040 (sibcall_value_internal): Likewise.
19041 (prefetch): Use constraint ZD.
19042 * config/mips/mips.opt (minterlink-compressed): New option.
19043 (minterlink-mips16): Now an alias for minterlink-compressed.
19044 (mmicromips): New option.
19045 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
19046 (compare_and_swap_12): Likewise.
19047 (sync_add<mode>): Likewise.
19048 (sync_<optab>_12): Likewise.
19049 (sync_old_<optab>_12): Likewise.
19050 (sync_new_<optab>_12): Likewise.
19051 (sync_nand_12): Likewise.
19052 (sync_old_nand_12): Likewise.
19053 (sync_new_nand_12): Likewise.
19054 (sync_sub<mode>): Likewise.
19055 (sync_old_add<mode>): Likewise.
19056 (sync_old_sub<mode>): Likewise.
19057 (sync_new_add<mode>): Likewise.
19058 (sync_new_sub<mode>): Likewise.
19059 (sync_<optab><mode>): Likewise.
19060 (sync_old_<optab><mode>): Likewise.
19061 (sync_new_<optab><mode>): Likewise.
19062 (sync_nand<mode>): Likewise.
19063 (sync_old_nand<mode>): Likewise.
19064 (sync_new_nand<mode>): Likewise.
19065 (sync_lock_test_and_set<mode>): Likewise.
19066 (test_and_set_12): Likewise.
19067 (atomic_compare_and_swap<mode>): Likewise.
19068 (atomic_exchange<mode>_llsc): Likewise.
19069 (atomic_fetch_add<mode>_llsc): Likewise.
19070 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
19071 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
19072 (umips_save_restore_pattern_p): Likewise.
19073 (umips_load_store_pair_p): Likewise.
19074 (umips_output_load_store_pair): Likewise.
19075 (umips_movep_target_p): Likewise.
19076 (umips_12bit_offset_address_p): Likewise.
19077 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
19078 (mips_base_mips16): Rename this...
19079 (mips_base_compression_flags): ...to this. Update all uses.
19080 (mips_attribute_table): Add micromips, nomicromips and nocompression.
19081 (mips_mips16_decl_p): Delete.
19082 (mips_nomips16_decl_p): Delete.
19083 (mips_get_compress_on_flags): New function.
19084 (mips_get_compress_off_flags): New function.
19085 (mips_get_compress_mode): New function.
19086 (mips_get_compress_on_name): New function.
19087 (mips_get_compress_off_name): New function.
19088 (mips_insert_attributes): Support multiple compression types.
19089 (mips_merge_decl_attributes): Likewise.
19090 (umips_12bit_offset_address_p): New function.
19091 (mips_start_function_definition): Emit .set micromips directive.
19092 (mips_call_may_need_jalx_p): New function.
19093 (mips_function_ok_for_sibcall): Add microMIPS support.
19094 (mips_print_operand_punctuation): Support short delay slots and
19096 (umips_swm_mask, umips_swm_encoding): New.
19097 (umips_build_save_restore): New function.
19098 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
19099 (was_mips16_p): Remove.
19100 (old_compression_mode): New.
19101 (mips_set_compression_mode): New function.
19102 (mips_set_current_function): Add microMIPS support.
19103 (mips_option_override): Likewise.
19104 (umips_save_restore_pattern_p): New function.
19105 (umips_output_save_restore): New function.
19106 (umips_load_store_pair_p_1): New function.
19107 (umips_load_store_pair_p): New function.
19108 (umips_output_load_store_pair_1): New function.
19109 (umips_output_load_store_pair): New function.
19110 (umips_movep_target_p) New function.
19111 (mips_prepare_pch_save): Add microMIPS support.
19112 * config/mips/mips.h (TARGET_COMPRESSION): New.
19113 (TARGET_CPU_CPP_BUILTINS): Update macro
19114 to use new compression flags and to support microMIPS.
19115 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
19116 (MIPS_ARCH_FLOAT_SPEC): Likewise.
19117 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
19118 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
19119 (ASM_SPEC): Support mmicromips and mno-micromips.
19120 (M16STORE_REG_P): New macro.
19121 (MIPS_CALL): Support TARGET_MICROMIPS.
19122 (MICROMIPS_J): New macro.
19123 (mips_base_mips16): Rename this...
19124 (mips_base_compression_flags): ...to this.
19125 (UMIPS_12BIT_OFFSET_P): New macro.
19126 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
19127 (MULTILIB_DIRNAMES): Likewise.
19128 2013-03-20 Richard Biener <rguenther@suse.de>
19130 PR tree-optimization/56661
19131 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
19132 the result does not have to be distinct.
19134 2013-03-20 Richard Biener <rguenther@suse.de>
19136 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
19139 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
19140 Steven Bosscher <steven@gcc.gnu.org>
19142 PR rtl-optimization/56605
19143 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
19145 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
19148 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
19149 that require movd instead of movq.
19151 2013-03-20 Richard Biener <rguenther@suse.de>
19153 * tree-ssa-structalias.c (struct variable_info): Add pointer
19154 to the first field of an aggregate with sub-vars. Make
19155 this and the pointer to the next subfield its ID.
19156 (vi_next): New function.
19157 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
19158 storedanything_id, integer_id): Increment by one.
19159 (new_var_info, get_call_vi, lookup_call_clobber_vi,
19160 get_call_clobber_vi): Adjust.
19161 (solution_set_expand): Simplify and speedup.
19162 (solution_set_add): Inline into ...
19163 (set_union_with_increment): ... this. Adjust accordingly.
19164 (do_sd_constraint): Likewise.
19165 (do_ds_constraint): Likewise.
19166 (do_complex_constraint): Simplify.
19167 (build_pred_graph): Adjust.
19168 (solve_graph): Likewise. Simplify and speedup.
19169 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
19170 get_constraint_for_component_ref, get_constraint_for_1,
19171 first_vi_for_offset, first_or_preceding_vi_for_offset,
19172 create_function_info_for, create_variable_info_for_1,
19173 create_variable_info_for, intra_create_variable_infos): Adjust.
19174 (init_base_vars): Push NULL for ID zero.
19175 (compute_points_to_sets): Adjust.
19177 2013-03-20 Richard Biener <rguenther@suse.de>
19179 * cfgloop.c (verify_loop_structure): Streamline and avoid
19180 ICEing on corrupt loop tree.
19181 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
19184 2013-03-20 Richard Biener <rguenther@suse.de>
19186 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
19187 check whether an SSA update is needed.
19189 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
19191 * config/mips/constraints.md (T): Rename to...
19195 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
19196 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
19198 2013-03-19 Ian Bolton <ian.bolton@arm.com>
19200 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
19201 (*subsi3_carryin_uxtw): Likewise.
19203 2013-03-19 Ian Bolton <ian.bolton@arm.com>
19205 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
19206 (*rorsi3_insn_uxtw): Likewise.
19208 2013-03-19 Ian Bolton <ian.bolton@arm.com>
19210 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
19211 (*extrsi5_insn_uxtw): Likewise.
19213 2013-03-19 Richard Biener <rguenther@suse.de>
19215 PR tree-optimization/56273
19216 * passes.c (init_optimization_passes): Move second VRP after DOM.
19218 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
19220 * config/i386/i386.md (*movti_internal): Merge from
19221 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
19222 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
19223 nox64 isa attributes.
19225 2013-03-18 Richard Biener <rguenther@suse.de>
19227 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
19229 (merge_node_constraints): Likewise.
19230 (build_succ_graph): Likewise.
19231 (valid_graph_edge): Inline into single caller.
19232 (unify_nodes): Likewise. Use bitmap_set_bit return value
19234 (scc_visit): Fix formatting and variable use.
19235 (do_sd_constraint): Use gcc_checking_assert.
19236 (do_ds_constraint): Likewise.
19237 (do_complex_constraint): Likewise.
19238 (condense_visit): Likewise. Cleanup.
19239 (dump_pred_graph): New function.
19240 (perform_var_substitution): Dump the pred-graph before
19241 variable substitution.
19242 (find_equivalent_node): Use gcc_checking_assert.
19243 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
19245 2013-03-18 Richard Biener <rguenther@suse.de>
19247 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
19248 Remove cond_expr_stmt_list argument and do not gimplify the
19250 (vect_loop_versioning): Adjust.
19251 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
19252 Cleanup to use less temporaries.
19253 (vect_create_data_ref_ptr): Cleanup.
19255 2013-03-18 Jakub Jelinek <jakub@redhat.com>
19257 PR tree-optimization/56635
19258 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
19259 require types_compatible_p types.
19261 2013-03-18 Nick Clifton <nickc@redhat.com>
19263 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
19264 spurious backslash.
19266 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
19267 Add missing line to comment describing function.
19269 2013-03-18 Richard Biener <rguenther@suse.de>
19271 PR tree-optimization/56210
19272 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
19273 Handle string / character search functions.
19274 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
19276 2013-03-18 Richard Biener <rguenther@suse.de>
19278 PR middle-end/56483
19279 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
19280 and implement properly.
19281 * gimple.h (gimple_cond_single_var_p): Remove.
19283 2013-03-18 Richard Biener <rguenther@suse.de>
19285 * tree-data-ref.h (find_data_references_in_loop): Declare.
19286 * tree-data-ref.c (get_references_in_stmt): Use a stack
19287 vector pre-allocated in the callers.
19288 (find_data_references_in_stmt): Adjust.
19289 (graphite_find_data_references_in_stmt): Likewise.
19290 (create_rdg_vertices): Likewise.
19291 (find_data_references_in_loop): Export.
19292 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
19293 Compute dependences here...
19294 (vect_analyze_data_refs): ...not here. When we encounter
19295 a non-vectorizable data reference in basic-block vectorization
19296 truncate the data reference vector. Do not bother to
19297 fixup data-dependence information for gather loads.
19298 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
19299 of data references, as reported.
19301 2013-03-18 Richard Biener <rguenther@suse.de>
19303 PR tree-optimization/3713
19304 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
19305 has_constants and expr.
19306 (stmt_has_constants): Properly valueize SSA names when deciding
19307 whether the stmt has constants.
19309 2013-03-18 Richard Biener <rguenther@suse.de>
19311 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
19312 whole function when there is nothing to do.
19313 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
19314 * tree-vectorizer.c (vectorize_loops): Update virtual and
19315 loop-closed SSA once.
19316 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
19318 2013-03-18 Richard Biener <rguenther@suse.de>
19320 PR middle-end/56113
19321 * domwalk.c (bb_postorder): New global static.
19322 (cmp_bb_postorder): New function.
19323 (walk_dominator_tree): Replace scheme imposing an order for
19324 visiting dominator sons by one sorting them at the time they
19325 are pushed on the stack.
19327 2013-03-18 Richard Biener <rguenther@suse.de>
19329 PR tree-optimization/39326
19330 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
19331 (struct mem_ref): Replace mem member with ao_ref typed member.
19332 (MEM_ANALYZABLE): Adjust.
19333 (memref_eq): Likewise.
19334 (mem_ref_alloc): Likewise.
19335 (gather_mem_refs_stmt): Likewise.
19336 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
19337 (execute_sm_if_changed_flag_set): Adjust.
19338 (execute_sm): Likewise.
19339 (ref_always_accessed_p): Likewise.
19340 (refs_independent_p): Likewise.
19341 (can_sm_ref_p): Likewise.
19343 2013-03-18 Jakub Jelinek <jakub@redhat.com>
19346 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
19347 return 1 even for !unsignedp.
19349 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
19351 * config/i386/i386.md (isa): Add x64 and nox64.
19352 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
19353 (*pushtf): Enable *roF alternative for x64 isa only.
19354 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
19355 mode attribute of integer alternatives to DImode for TARGET_64BIT.
19356 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
19357 (*movtf_internal): Merge from *movtf_internal_rex64 and
19358 *movtf_internal_sse. Use x64 and nox64 isa attributes.
19359 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
19360 nox64 isa attributes.
19361 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
19362 nox64 isa attributes.
19363 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
19365 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
19367 * config/alpha/alpha.c (TARGET_LRA_P): New define.
19369 2013-03-17 Jakub Jelinek <jakub@redhat.com>
19372 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
19373 class names. Remove trailing comma after "ALL_REGS".
19375 2013-03-16 Jan Hubicka <jh@suse.cz>
19377 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
19378 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
19379 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
19380 of cgraph_get_create_node.
19381 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
19383 2013-03-16 Jason Merrill <jason@redhat.com>
19386 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
19387 with DW_AT_default_value.
19389 2013-03-16 Jakub Jelinek <jakub@redhat.com>
19391 * BASE-VER: Set to 4.9.0.
19393 2013-03-14 Andi Kleen <ak@linux.intel.com>
19396 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
19397 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
19398 Document _x* TSX intrinsics.
19400 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
19401 David Holsgrove <david.holsgrove@xilinx.com>
19403 * configure.ac: Add MicroBlaze TLS support detection.
19404 * configure: Regenerate.
19405 * config/microblaze/microblaze-protos.h
19406 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
19407 symbol_mentioned_p, label_mentioned_p): Add prototypes.
19408 * config/microblaze/microblaze.c (microblaze_address_type): Add
19409 ADDRESS_TLS and tls_reloc address types.
19410 (microblaze_address_info): Add tls_reloc.
19411 (TARGET_HAVE_TLS): Define.
19412 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
19413 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
19414 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
19415 load_tls_operand, microblaze_call_tls_get_addr,
19416 microblaze_legitimize_tls_address): New functions.
19417 (microblaze_classify_unspec): Handle UNSPEC_TLS.
19418 (get_base_reg): Use microblaze_tls_symbol_p.
19419 (microblaze_classify_address): Handle TLS.
19420 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
19421 label_mentioned_p and microblaze_tls_referenced_p.
19422 (microblaze_legitimize_address): Handle TLS.
19423 (microblaze_address_insns): Handle ADDRESS_TLS.
19424 (pic_address_needs_scratch): Handle TLS.
19425 (print_operand_address): Handle TLS.
19426 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
19427 (microblaze_expand_move): Handle TLS.
19428 (microblaze_legitimate_constant_p): Check
19429 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
19430 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
19431 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
19432 (PIC_OFFSET_TABLE_REGNUM): Set.
19433 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
19434 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
19435 (addsi3, movsi_internal2, movdf_internal): Update constraints
19436 * config/microblaze/predicates.md (arith_plus_operand): Define
19437 (move_operand): Redefine as move_src_operand,
19438 check microblaze_tls_referenced_p.
19440 2013-03-14 Ian Bolton <ian.bolton@arm.com>
19442 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
19443 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
19445 2013-03-14 Ian Bolton <ian.bolton@arm.com>
19447 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
19450 2013-03-14 Jakub Jelinek <jakub@redhat.com>
19452 PR tree-optimization/53265
19453 * common.opt (Waggressive-loop-optimizations): New option.
19454 * tree-ssa-loop-niter.c: Include tree-pass.h.
19455 (do_warn_aggressive_loop_optimizations): New function.
19456 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
19457 if number_of_latch_executions returned constant.
19458 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
19459 early. If number_of_latch_executions returned constant, set
19460 nb_iterations_upper_bound back to it.
19461 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
19463 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
19464 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
19466 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
19467 (MULTILIB_OSDIRNAMES): Set.
19468 * genmultilib: If defaultosdirname doesn't start with :: , set
19469 defaultosdirname2 instead, clear it and emit two . multilib_raw
19470 entries instead of just one.
19472 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
19474 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
19475 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
19476 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
19477 (SUBTARGET_OVERRIDE_OPTIONS): New.
19479 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
19482 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
19483 (musermode): Convert to Var(TARGET_USERMODE).
19484 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
19485 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
19486 * config/sh/sh.c (sh_option_override): Use
19487 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
19488 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
19490 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
19492 (udivsi3_i4_single, divsi3_i4_single): Use
19493 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
19495 2013-03-13 Dave Korn <dave.korn.cygwin@....>
19497 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
19500 2013-03-13 Richard Biener <rguenther@suse.de>
19502 PR tree-optimization/56608
19503 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
19504 calls when vectorizing basic-blocks.
19506 2013-03-13 Jakub Jelinek <jakub@redhat.com>
19509 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
19512 2013-03-12 Jakub Jelinek <jakub@redhat.com>
19514 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
19516 2013-03-11 Jan Hubicka <jh@suse.cz>
19519 * lto-streamer-out.c (output_symbol_p): Skip references from
19520 constructors of external variables.
19522 2013-03-11 Jan Hubicka <jh@suse.cz>
19524 PR middle-end/56571
19525 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
19527 * emit-rtl.c (verify_rtx_sharing): Likewise.
19528 (copy_insn_1): Likewise.
19529 * rtl.c (copy_rtx): Likewise.
19531 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
19534 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
19535 output_operand_lossage message.
19537 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
19540 * arm.c (shift_op): Validate RTL pattern on the fly.
19541 (arm_print_operand, case 'S'): Don't use shift_operator to validate
19544 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
19547 * config/pa/pa.md (call_value): Check for calls to powf and direct to
19548 new call patterns that clobber %fr12.
19549 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
19550 split and postreload patterns.
19551 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
19552 registers %fr12 and %fr12R as call used.
19554 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
19556 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
19557 (canon_address, record_store, replace_read, check_mem_read_rtx,
19558 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
19559 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
19560 rest_of_handle_dse): Likewise.
19562 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
19564 PR middle-end/56524
19565 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
19567 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
19568 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
19569 (save_optabs_if_changed): Replace with...
19570 (init_tree_optimization_optabs): ...this.
19571 * optabs.c (save_optabs_if_changed): Rename to...
19572 (init_tree_optimization_optabs): ...this. Take the optimization node
19573 as argument. Do nothing if the base optabs are already correct.
19574 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
19575 to recompute optabs.
19576 * function.h (function): Remove optabs field.
19577 * function.c (invoke_set_current_function_hook): Call
19578 init_tree_optimization_optabs. Use the result to initialize
19581 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
19583 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
19584 if GTMA_HAS_NO_INSTRUMENTATION.
19585 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
19586 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
19587 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
19588 * gimple-pretty-print.c (dump_gimple_transaction): Handle
19589 GTMA_HAS_NO_INSTRUMENTATION.
19591 2013-03-08 Jakub Jelinek <jakub@redhat.com>
19593 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
19596 2013-03-08 Marek Polacek <polacek@redhat.com>
19597 Jakub Jelinek <jakub@redhat.com>
19599 PR tree-optimization/56478
19600 * predict.c (is_comparison_with_loop_invariant_p): Change the
19601 type of loop_step to tree.
19602 (predict_loops): Adjust.
19603 (predict_iv_comparison): Perform the computations on double_ints.
19605 2013-03-08 Richard Biener <rguenther@suse.de>
19607 PR tree-optimization/56570
19608 * tree-cfg.c (verify_expr_location_1): Verify locations for
19610 * tree-sra.c (create_access_replacement): Strip locations
19611 from DECL_DEBUG_EXPRs.
19613 2013-03-08 Richard Biener <rguenther@suse.de>
19615 * tree-inline.c (expand_call_inline): Do not associate
19616 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
19617 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
19619 2013-03-08 Richard Biener <rguenther@suse.de>
19621 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
19622 or block changes with -Og. Fix for location / block encoding
19623 changes and PHI arguments with locations.
19625 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
19627 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
19629 (struct output_info): Likewise.
19630 (register_overhead): Remove bad gcc_assert.
19631 (bitmap_find_bit): If there is only a single bitmap element, do not
19632 count a miss as a search.
19633 (print_statistics): Update for counter type changes.
19634 (dump_bitmap_statistics): Likewise. Print headers such that they
19635 are properly lined up with the printed counters.
19637 2013-03-07 Jakub Jelinek <jakub@redhat.com>
19639 PR tree-optimization/56559
19640 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
19641 check that it has only a single use.
19643 2013-03-07 Richard Biener <rguenther@suse.de>
19645 * doc/invoke.texi (fwhole-program): Discourage use in combination
19648 2013-03-06 Jakub Jelinek <jakub@redhat.com>
19650 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
19652 PR tree-optimization/56539
19653 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
19654 instead of GSI_CONTINUE_LINKING as last argument to
19655 force_gimple_operand_gsi. Adjust function comment.
19657 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
19660 PR middle-end/56548
19661 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
19662 promoted mode, convert the result back to the original mode.
19664 2013-03-06 Richard Biener <rguenther@suse.de>
19666 PR middle-end/56294
19667 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
19668 (insert_updated_phi_nodes_compare_uids): New function.
19669 (update_ssa): Sort symbols_to_rename after UID before
19670 traversing it to insert PHI nodes.
19672 2013-03-06 Richard Biener <rguenther@suse.de>
19674 PR middle-end/50494
19675 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
19676 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
19679 2013-02-13 Richard Biener <rguenther@suse.de>
19682 * varasm.c (output_constant_def_1): Get the decl representing
19683 the constant as argument.
19684 (output_constant_def): Wrap output_constant_def_1.
19685 (make_decl_rtl): Use output_constant_def_1 with the decl
19686 representing the constant.
19687 (build_constant_desc): Optionally re-use a decl already
19688 representing the constant.
19689 (tree_output_constant_def): Adjust.
19691 2013-03-06 Joey Ye <joey.ye@arm.com>
19694 * gcc.c (convert_white_space): New function.
19695 (main): Handles white space in function name.
19697 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
19700 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
19701 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
19702 to SH_DIV_CALL_TABLE for TARGET_SH2.
19703 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
19705 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
19706 call-table options.
19708 2013-03-05 Sterling Augustine <saugustine@google.com>
19709 Cary Coutant <ccoutant@google.com>
19712 * dwarf2out.c (resolve_addr): Don't call
19713 remove_loc_list_addr_table_entries a second time for the same
19716 2013-03-05 Jakub Jelinek <jakub@redhat.com>
19719 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
19720 (avoid_complex_debug_insns): New function.
19721 (expand_debug_locations): Call it.
19723 PR rtl-optimization/56484
19724 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
19725 lifetimes of hard registers on small register class machines.
19727 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
19729 * config/microblaze/microblaze-protos.h: Rename
19730 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
19731 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
19733 (microblaze_fast_interrupt_function_p): New function.
19734 (microblaze_is_interrupt_handler): Rename to
19735 microblaze_is_interrupt_variant and add fast_interrupt check.
19736 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
19737 (save_restore_insns): Likewise.
19738 (compute_frame_size): Likewise.
19739 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
19740 (microblaze_globalize_label): Likewise.
19741 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
19742 * config/microblaze/microblaze.md: Use wrapper
19743 microblaze_is_interrupt_variant.
19745 2013-03-05 Kai Tietz <ktietz@redhat.com>
19747 * sdbout.c (sdbout_one_type): Switch to current function's section
19748 supporting cold/hot.
19750 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
19752 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
19755 2013-03-05 Jakub Jelinek <jakub@redhat.com>
19757 PR middle-end/56461
19758 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
19759 if VALGRIND_GET_VBITS is defined, temporarily make object
19760 memory all defined, and restore previous valgrind addressability
19761 and definability afterwards. Free this_object at the end.
19763 PR middle-end/56461
19764 * lra.c (lra): Call lra_clear_live_ranges if live_p,
19765 right before calling lra_create_live_ranges, also call it
19766 when clearing live_p. Only call lra_clear_live_ranges
19767 at the end if live_p.
19769 PR middle-end/56461
19770 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
19772 2013-03-05 Richard Biener <rguenther@suse.de>
19774 PR tree-optimization/56521
19775 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
19778 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
19781 * except.h (remove_unreachable_eh_regions): New prototype.
19782 * except.c (remove_eh_handler_splicer): New function, split out
19783 of remove_eh_handler.
19784 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
19785 warning about running it on many EH regions one at a time.
19786 (remove_unreachable_eh_regions_worker): New function, walk the
19787 EH tree in depth-first order and remove non-marked regions.
19788 (remove_unreachable_eh_regions): New function.
19789 * tree-eh.c (mark_reachable_handlers): New function, split out
19790 from remove_unreachable_handlers.
19791 (remove_unreachable_handlers): Use mark_reachable_handlers and
19792 remove_unreachable_eh_regions.
19793 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
19794 and remove_unreachable_eh_regions.
19796 2013-03-05 Richard Biener <rguenther@suse.de>
19798 PR middle-end/56525
19799 * loop-init.c (fix_loop_structure): Remove loops in two stages,
19800 not freeing them until the end.
19802 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19804 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
19806 2013-03-05 Richard Biener <rguenther@suse.de>
19808 PR tree-optimization/56270
19809 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
19810 of loads after scheduling an SLP instance.
19812 2013-03-05 Jakub Jelinek <jakub@redhat.com>
19814 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
19816 (check_gcc_parallelize): Run guality.exp as a separate job from
19817 vect.exp with unsorted.exp and $(dg_target_exps) separately from
19818 struct-layout-1.exp with stackalign.exp.
19820 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
19822 PR middle-end/56461
19823 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
19824 load_index sbitmap even if some bit in it isn't set.
19826 PR middle-end/56461
19827 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
19828 (discover_iteration_bound_by_body_walk): Change queues to
19829 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
19830 spelling in comment. Call safe_push on queues[bound_index] directly.
19831 Release queues[queue_index] in every iteration unconditionally.
19832 Release bounds vector.
19834 PR middle-end/56461
19835 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
19836 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
19837 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
19840 2013-03-05 Richard Biener <rguenther@suse.de>
19843 * tree-inline.c (remap_blocks_to_null): New function.
19844 (expand_call_inline): When expanding a call stmt without
19845 an associated block inline remap all callee blocks to NULL.
19847 2013-03-05 Jakub Jelinek <jakub@redhat.com>
19849 PR rtl-optimization/56494
19850 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
19851 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
19852 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
19854 PR middle-end/56461
19855 * sel-sched-ir.c (free_sched_pools): Release
19856 succs_info_pool.stack[succs_info_pool.max_top] vectors too
19857 if succs_info_pool.max_top isn't -1.
19860 * opts.c (opts_obstack, opts_concat): Moved to...
19861 * opts-common.c (opts_obstack, opts_concat): ... here.
19863 2013-03-04 Jakub Jelinek <jakub@redhat.com>
19865 PR middle-end/56461
19866 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
19868 2013-03-04 Martin Jambor <mjambor@suse.cz>
19870 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
19871 all appropriate places.
19873 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
19875 PR tree-optimization/56424
19876 * ipa-split.c (split_function): Do not set the RSO flag if result is
19877 not by reference and its type is a register type.
19879 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
19881 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
19882 (microblaze_legitimate_pic_operand): Likewise
19883 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
19884 new function microblaze_legitimate_pic_operand
19885 * config/microblaze/microblaze-protos.h
19886 (microblaze_legitimate_pic_operand): Declare.
19888 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
19890 * config/microblaze/predicates.md (call_insn_simple_operand):
19891 New predicate for supported rtx code types.
19892 * config/microblaze/microblaze.md (call_internal1): Use
19893 call_insn_simple_operand predicate.
19895 2013-03-04 Jakub Jelinek <jakub@redhat.com>
19897 PR middle-end/56461
19898 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
19899 partitions.ordered_remove.
19901 PR middle-end/56461
19902 * tree-vect-stmts.c (vectorizable_conversion): Don't call
19903 vec_oprnds0.create (1) for modifier == NONE.
19905 PR middle-end/56461
19906 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
19907 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
19908 vec_oprnds1 right before pushing anything to it for
19911 PR middle-end/56461
19912 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
19913 set nbbs to 0 instead of having separate code path.
19914 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
19915 instead of false as last argument if returning NULL.
19917 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
19919 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
19920 the attribute is now called "target" instead of "option".
19921 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
19922 * doc/tm.texi.in (Target Attributes): Likewise document the correct
19923 attribute/pragma name for TARGET_OPTION_VALID_P and
19924 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
19925 * doc/tm.texi: Regenerated.
19927 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
19929 * config/microblaze/microblaze.c:
19930 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
19931 * config/microblaze/microblaze.h: Add -mxl-reorder to
19933 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
19934 instructions emitted if TARGET_REORDER.
19935 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
19936 or 0 for -m/-mno case, but initialises as 2 to detect default use case
19939 2013-03-01 Xinliang David Li <davidxl@google.com>
19941 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
19944 2013-03-01 Jakub Jelinek <jakub@redhat.com>
19946 PR middle-end/56461
19947 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
19948 vector even when returning true. Fix up function comment formatting.
19950 PR middle-end/56461
19951 * ira-build.c (ira_loop_nodes_count): New variable.
19952 (create_loop_tree_nodes): Initialize it.
19953 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
19955 PR middle-end/56461
19956 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
19957 method on dr_chain and result_chain.
19958 * tree-vect-stmts.c (vectorizable_store): Only call
19959 result_chain.create if j == 0.
19961 PR middle-end/56461
19962 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
19963 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
19964 before overwriting it.
19966 2013-03-01 Tobias Burnus <burnus@net-b.de>
19968 * doc/extended.texi (C Extensions): Change order in @menu
19970 (Other MIPS Built-in Functions): Move last MIPS entry before
19971 "picoChip Built-in Functions".
19972 (SH Built-in Functions): Move after RX Built-in Functions.
19973 * doc/gcc.texi (Introduction): Change order in @menu
19975 * doc/md.texi (Constraints): Ditto.
19976 * gty.texi (Type Information): Ditto.
19977 (User-provided marking routines for template types): Make
19979 * doc/invoke.texi (AArch64 Options): Move before
19980 "Adapteva Epiphany Options".
19982 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
19983 Jakub Jelinek <jakub@redhat.com>
19986 * asan.c (gate_asan): Lookup no_sanitize_address instead of
19987 no_address_safety_analysis attribute.
19988 * doc/extend.texi (no_address_safety_attribute): Rename to
19989 no_sanitize_address attribute, mention no_address_safety_analysis
19990 attribute as deprecated alias.
19992 2013-02-28 Jakub Jelinek <jakub@redhat.com>
19994 PR middle-end/56461
19995 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
19996 type to vec<vec<tree> > *.
19997 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
19998 to be vec<tree> instead of vec<tree> *, set vec_defs
19999 to vNULL and call vec_defs.create (number_of_vects), adjust other
20001 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
20002 vectorizable_condition): Adjust vect_get_slp_defs callers.
20004 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
20006 * config/aarch64/aarch64.c
20007 (aarch64_float_const_representable): Remove unused variable.
20009 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
20011 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
20013 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
20015 * config/aarch64/aarch64-builtins.c
20016 (aarch64_init_simd_builtins): Make static.
20018 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
20020 * config/aarch64/aarch64.c
20021 (aarch64_simd_make_constant): Make static.
20023 2013-02-28 Martin Jambor <mjambor@suse.cz>
20025 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
20026 with no initialization to the RHS of debug statements.
20028 2013-02-28 Martin Jambor <mjambor@suse.cz>
20030 PR tree-optimization/56294
20031 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
20033 (get_access_replacement): Do not call create_access_replacement.
20034 Assert a replacement exists.
20035 (get_repl_default_def_ssa_name): Create the replacement declaration
20038 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20040 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
20041 final_end_function.
20043 2013-02-28 Marek Polacek <polacek@redhat.com>
20045 PR rtl-optimization/56466
20046 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
20047 if we're changing a loop.
20048 (peel_loops_completely): Likewise.
20050 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
20053 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
20055 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
20058 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
20059 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
20060 INTX_FTYPE_FX, FX_FTYPE_INTX.
20061 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
20063 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
20065 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
20066 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
20067 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
20068 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
20069 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
20070 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
20071 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
20072 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
20073 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
20074 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
20075 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
20076 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
20077 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
20078 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
20079 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
20080 (avrxmega6): Increase max flash segments from 5 to 6.
20081 * config/avr/t-multilib: Regenerate.
20082 * config/avr/avr-tables.opt: Regenerate.
20083 * doc/avr-mmcu.texi: Regenerate.
20085 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
20087 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
20088 (avr_device_to_arch): Rename to avr_device_to_ld.
20089 (avr_device_to_as): New prototype.
20090 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
20091 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
20092 * config/avr/driver-avr.c (avr_device_to_as): New.
20093 (avr_device_to_arch): Rename to avr_device_to_ld.
20095 2013-02-27 Jakub Jelinek <jakub@redhat.com>
20097 PR middle-end/56461
20098 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
20099 method on dr_chain and result_chain.
20101 PR middle-end/56461
20102 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
20103 pointer_set_destroy on not_executed_last_iteration.
20105 PR middle-end/56461
20106 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
20108 PR middle-end/56461
20109 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
20110 FOR_EACH_DEFINED_FUNCTION when freeing state.
20112 PR middle-end/56461
20113 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
20115 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
20118 PR middle-end/56461
20119 * ipa-cp.c (decide_whether_version_node): Call vec_free on
20120 known_aggs[i].items and release known_aggs vector.
20122 PR middle-end/56461
20123 * ipa-reference.c (propagate): Free node_info even for alias nodes.
20125 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20127 * config/microblaze/microblaze.c (microblaze_emit_compare):
20128 Use xor for EQ/NE comparisions.
20129 * config/microblaze/microblaze.md (cstoresf4): Add constraints
20130 (cbranchsf4): Adjust operator to comparison_operator.
20132 2013-02-27 Jakub Jelinek <jakub@redhat.com>
20134 PR middle-end/56461
20135 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
20137 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
20138 vec_safe_push, always update *slot.
20139 (redirect_edge_var_map_clear): Use vec_free.
20140 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
20141 (free_var_map_entry): Use vec_free.
20142 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
20143 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
20145 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
20147 PR middle-end/45472
20148 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
20149 when the may_trap_p bit of the exprs being merged differs.
20150 Reorder tests for speculativeness in the logical and operator.
20152 2013-02-27 Jakub Jelinek <jakub@redhat.com>
20154 * incpath.c (add_standard_paths): Use reconcat instead of concat
20155 where appropriate and avoid leaking memory.
20157 * opts.h: Include obstack.h.
20158 (opts_concat): New prototype.
20159 (opts_obstack): New declaration.
20160 * opts.c (opts_concat): New function.
20161 (opts_obstack): New variable.
20162 (init_options_struct): Call gcc_init_obstack on opts_obstack.
20163 (finish_options): Use opts_concat instead of concat
20164 and XOBNEWVEC instead of XNEWVEC.
20165 * opts-common.c (generate_canonical_option, decode_cmdline_option,
20166 generate_option): Likewise.
20167 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
20168 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
20171 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
20172 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
20174 2013-02-26 Jakub Jelinek <jakub@redhat.com>
20176 PR middle-end/56461
20177 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
20179 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
20181 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
20182 (arm_block_move_unaligned_straight): Likewise.
20183 (arm_adjust_block_mem): Likewise.
20185 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
20188 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
20189 temp, cond and label.
20190 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
20193 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
20194 * config/c6x/c6x.h (dbx_register_map): Update declaration.
20197 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
20198 of prologue/epilogue functions.
20201 * config/tilegx/tilegx.c (tilegx_expand_prologue):
20202 Remove unused variable cfa_offset.
20203 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
20206 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
20207 type promotion to unsigned.
20210 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
20211 for HOST_WIDE_INT of 32 bit / same size as int.
20212 (arm_block_move_unaligned_straight): Likewise.
20213 (arm_adjust_block_mem): Likewise.
20216 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
20219 2013-02-26 Marek Polacek <polacek@redhat.com>
20221 PR tree-optimization/56426
20222 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
20224 2013-02-26 Richard Biener <rguenther@suse.de>
20227 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
20228 unused variable loops.
20230 2013-02-26 Jakub Jelinek <jakub@redhat.com>
20232 PR tree-optimization/56448
20233 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
20234 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
20235 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
20236 later operands of the references, or even first operand for
20237 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
20239 PR tree-optimization/56443
20240 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
20241 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
20242 to type_for_mode langhook.
20244 2013-02-25 Matt Turner <mattst88@gmail.com>
20246 * doc/invoke.texi: Document r4700.
20248 2013-02-25 Richard Biener <rguenther@suse.de>
20250 PR tree-optimization/56175
20251 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
20253 (simplify_bitwise_binary): ... here. Also guard the conversion
20254 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
20256 2013-02-25 Catherine Moore <clm@codesourcery.com>
20259 2013-02-24 Catherine Moore <clm@codesourcery.com>
20260 Maciej W. Rozycki <macro@codesourcery.com>
20261 Tom de Vries <tom@codesourcery.com>
20262 Nathan Sidwell <nathan@codesourcery.com>
20263 Iain Sandoe <iain@codesourcery.com>
20264 Nathan Froyd <froydnj@codesourcery.com>
20265 Chao-ying Fu <fu@mips.com>
20267 * doc/extend.texi: (micromips, nomicromips, nocompression):
20268 Document new function attributes.
20269 * doc/invoke.texi (minterlink-compressed, mmicromips,
20270 m14k, m14ke, m14kec): Document new options.
20271 (minterlink-mips16): Update documentation.
20272 * doc/md.texi (ZC, ZD): Document new constraints.
20273 * configure.ac (gcc_cv_as_micromips): Check if linker
20274 supports the .set micromips directive.
20275 * configure: Regenerate.
20276 * config.in: Regenerate.
20277 * config/mips/mips-tables.opt: Regenerate.
20278 * config/mips/micromips.md: New file.
20279 * constraints.md (ZC, AD): New constraints.
20280 * config/mips/predicates.md (movep_src_register): New predicate.
20281 (movep_src_operand): New predicate.
20282 (non_volatile_mem_operand): New predicate.
20283 * config/mips/mips.md (multimem): New type.
20284 (length): Differentiate between 17-bit and 18-bit branch offsets.
20285 (MOVEP1, MOVEP2): New mode iterator.
20286 (mov_<load>l): Use ZC constraint.
20287 (mov_<load>r): Likewise.
20288 (mov_<store>l): Likewise.
20289 (mov_<store>r): Likewise.
20290 (*branch_equality<mode>_inverted): Add microMIPS support.
20291 (*branch_equality<mode>): Likewise.
20292 (*jump_absolute): Likewise.
20293 (indirect_jump_<mode>): Likewise.
20294 (tablejump_<mode>): Likewise.
20295 (<optab>_internal): Likewise.
20296 (sibcall_internal): Likewise.
20297 (sibcall_value_internal): Likewise.
20298 (prefetch): Use constraint ZD.
20299 * config/mips/mips.opt (minterlink-compressed): New option.
20300 (minterlink-mips16): Now an alias for minterlink-compressed.
20301 (mmicromips): New option.
20302 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
20303 (compare_and_swap_12): Likewise.
20304 (sync_add<mode>): Likewise.
20305 (sync_<optab>_12): Likewise.
20306 (sync_old_<optab>_12): Likewise.
20307 (sync_new_<optab>_12): Likewise.
20308 (sync_nand_12): Likewise.
20309 (sync_old_nand_12): Likewise.
20310 (sync_new_nand_12): Likewise.
20311 (sync_sub<mode>): Likewise.
20312 (sync_old_add<mode>): Likewise.
20313 (sync_old_sub<mode>): Likewise.
20314 (sync_new_add<mode>): Likewise.
20315 (sync_new_sub<mode>): Likewise.
20316 (sync_<optab><mode>): Likewise.
20317 (sync_old_<optab><mode>): Likewise.
20318 (sync_new_<optab><mode>): Likewise.
20319 (sync_nand<mode>): Likewise.
20320 (sync_old_nand<mode>): Likewise.
20321 (sync_new_nand<mode>): Likewise.
20322 (sync_lock_test_and_set<mode>): Likewise.
20323 (test_and_set_12): Likewise.
20324 (atomic_compare_and_swap<mode>): Likewise.
20325 (atomic_exchange<mode>_llsc): Likewise.
20326 (atomic_fetch_add<mode>_llsc): Likewise.
20327 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
20328 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
20329 (umips_save_restore_pattern_p): Likewise.
20330 (umips_load_store_pair_p): Likewise.
20331 (umips_output_load_store_pair): Likewise.
20332 (umips_movep_target_p): Likewise.
20333 (umips_12bit_offset_address_p): Likewise.
20334 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
20335 (mips_base_mips16): Rename this...
20336 (mips_base_compression_flags): ...to this. Update all uses.
20337 (mips_attribute_table): Add micromips, nomicromips and nocompression.
20338 (mips_mips16_decl_p): Delete.
20339 (mips_nomips16_decl_p): Delete.
20340 (mips_get_compress_on_flags): New function.
20341 (mips_get_compress_off_flags): New function.
20342 (mips_get_compress_mode): New function.
20343 (mips_get_compress_on_name): New function.
20344 (mips_get_compress_off_name): New function.
20345 (mips_insert_attributes): Support multiple compression types.
20346 (mips_merge_decl_attributes): Likewise.
20347 (umips_12bit_offset_address_p): New function.
20348 (mips_start_function_definition): Emit .set micromips directive.
20349 (mips_call_may_need_jalx_p): New function.
20350 (mips_function_ok_for_sibcall): Add microMIPS support.
20351 (mips_print_operand_punctuation): Support short delay slots and
20353 (umips_swm_mask, umips_swm_encoding): New.
20354 (umips_build_save_restore): New function.
20355 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
20356 (was_mips16_p): Remove.
20357 (old_compression_mode): New.
20358 (mips_set_compression_mode): New function.
20359 (mips_set_current_function): Add microMIPS support.
20360 (mips_option_override): Likewise.
20361 (umips_save_restore_pattern_p): New function.
20362 (umips_output_save_restore): New function.
20363 (umips_load_store_pair_p_1): New function.
20364 (umips_load_store_pair_p): New function.
20365 (umips_output_load_store_pair_1): New function.
20366 (umips_output_load_store_pair): New function.
20367 (umips_movep_target_p) New function.
20368 (mips_prepare_pch_save): Add microMIPS support.
20369 * config/mips/mips.h (TARGET_COMPRESSION): New.
20370 (TARGET_CPU_CPP_BUILTINS): Update macro
20371 to use new compression flags and to support microMIPS.
20372 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
20373 (MIPS_ARCH_FLOAT_SPEC): Likewise.
20374 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
20375 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
20376 (ASM_SPEC): Support mmicromips and mno-micromips.
20377 (M16STORE_REG_P): New macro.
20378 (MIPS_CALL): Support TARGET_MICROMIPS.
20379 (MICROMIPS_J): New macro.
20380 (mips_base_mips16): Rename this...
20381 (mips_base_compression_flags): ...to this.
20382 (UMIPS_12BIT_OFFSET_P): New macro.
20383 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
20384 (MULTILIB_DIRNAMES): Likewise.
20386 2013-02-25 Tom de Vries <tom@codesourcery.com>
20388 PR rtl-optimization/56131
20389 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
20390 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
20391 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
20393 2013-02-25 Tobias Burnus <burnus@net-b.de>
20395 * doc/invoke.texi (-fsanitize=): Move from optimization
20396 to debugging options.
20398 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
20400 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
20402 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
20403 Alexander Monakov <amonakov@ispras.ru>
20405 PR middle-end/56077
20406 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
20407 flush pending lists also on non-jumps. Adjust comment.
20409 2013-02-24 Catherine Moore <clm@codesourcery.com>
20410 Maciej W. Rozycki <macro@codesourcery.com>
20411 Tom de Vries <tom@codesourcery.com>
20412 Nathan Sidwell <nathan@codesourcery.com>
20413 Iain Sandoe <iain@codesourcery.com>
20414 Nathan Froyd <froydnj@codesourcery.com>
20415 Chao-ying Fu <fu@mips.com>
20417 * doc/extend.texi: (micromips, nomicromips, nocompression):
20418 Document new function attributes.
20419 * doc/invoke.texi (minterlink-compressed, mmicromips,
20420 m14k, m14ke, m14kec): Document new options.
20421 (minterlink-mips16): Update documentation.
20422 * doc/md.texi (ZC, ZD): Document new constraints.
20423 * configure.ac (gcc_cv_as_micromips): Check if linker
20424 supports the .set micromips directive.
20425 * configure: Regenerate.
20426 * config.in: Regenerate.
20427 * config/mips/mips-tables.opt: Regenerate.
20428 * config/mips/micromips.md: New file.
20429 * constraints.md (ZC, AD): New constraints.
20430 * config/mips/predicates.md (movep_src_register): New predicate.
20431 (movep_src_operand): New predicate.
20432 (non_volatile_mem_operand): New predicate.
20433 * config/mips/mips.md (multimem): New type.
20434 (length): Differentiate between 17-bit and 18-bit branch offsets.
20435 (MOVEP1, MOVEP2): New mode iterator.
20436 (mov_<load>l): Use ZC constraint.
20437 (mov_<load>r): Likewise.
20438 (mov_<store>l): Likewise.
20439 (mov_<store>r): Likewise.
20440 (*branch_equality<mode>_inverted): Add microMIPS support.
20441 (*branch_equality<mode>): Likewise.
20442 (*jump_absolute): Likewise.
20443 (indirect_jump_<mode>): Likewise.
20444 (tablejump_<mode>): Likewise.
20445 (<optab>_internal): Likewise.
20446 (sibcall_internal): Likewise.
20447 (sibcall_value_internal): Likewise.
20448 (prefetch): Use constraint ZD.
20449 * config/mips/mips.opt (minterlink-compressed): New option.
20450 (minterlink-mips16): Now an alias for minterlink-compressed.
20451 (mmicromips): New option.
20452 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
20453 (compare_and_swap_12): Likewise.
20454 (sync_add<mode>): Likewise.
20455 (sync_<optab>_12): Likewise.
20456 (sync_old_<optab>_12): Likewise.
20457 (sync_new_<optab>_12): Likewise.
20458 (sync_nand_12): Likewise.
20459 (sync_old_nand_12): Likewise.
20460 (sync_new_nand_12): Likewise.
20461 (sync_sub<mode>): Likewise.
20462 (sync_old_add<mode>): Likewise.
20463 (sync_old_sub<mode>): Likewise.
20464 (sync_new_add<mode>): Likewise.
20465 (sync_new_sub<mode>): Likewise.
20466 (sync_<optab><mode>): Likewise.
20467 (sync_old_<optab><mode>): Likewise.
20468 (sync_new_<optab><mode>): Likewise.
20469 (sync_nand<mode>): Likewise.
20470 (sync_old_nand<mode>): Likewise.
20471 (sync_new_nand<mode>): Likewise.
20472 (sync_lock_test_and_set<mode>): Likewise.
20473 (test_and_set_12): Likewise.
20474 (atomic_compare_and_swap<mode>): Likewise.
20475 (atomic_exchange<mode>_llsc): Likewise.
20476 (atomic_fetch_add<mode>_llsc): Likewise.
20477 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
20478 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
20479 (umips_save_restore_pattern_p): Likewise.
20480 (umips_load_store_pair_p): Likewise.
20481 (umips_output_load_store_pair): Likewise.
20482 (umips_movep_target_p): Likewise.
20483 (umips_12bit_offset_address_p): Likewise.
20484 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
20485 (mips_base_mips16): Rename this...
20486 (mips_base_compression_flags): ...to this. Update all uses.
20487 (mips_attribute_table): Add micromips, nomicromips and nocompression.
20488 (mips_mips16_decl_p): Delete.
20489 (mips_nomips16_decl_p): Delete.
20490 (mips_get_compress_on_flags): New function.
20491 (mips_get_compress_off_flags): New function.
20492 (mips_get_compress_mode): New function.
20493 (mips_get_compress_on_name): New function.
20494 (mips_get_compress_off_name): New function.
20495 (mips_insert_attributes): Support multiple compression types.
20496 (mips_merge_decl_attributes): Likewise.
20497 (umips_12bit_offset_address_p): New function.
20498 (mips_start_function_definition): Emit .set micromips directive.
20499 (mips_call_may_need_jalx_p): New function.
20500 (mips_function_ok_for_sibcall): Add microMIPS support.
20501 (mips_print_operand_punctuation): Support short delay slots and
20503 (umips_swm_mask, umips_swm_encoding): New.
20504 (umips_build_save_restore): New function.
20505 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
20506 (was_mips16_p): Remove.
20507 (old_compression_mode): New.
20508 (mips_set_compression_mode): New function.
20509 (mips_set_current_function): Add microMIPS support.
20510 (mips_option_override): Likewise.
20511 (umips_save_restore_pattern_p): New function.
20512 (umips_output_save_restore): New function.
20513 (umips_load_store_pair_p_1): New function.
20514 (umips_load_store_pair_p): New function.
20515 (umips_output_load_store_pair_1): New function.
20516 (umips_output_load_store_pair): New function.
20517 (umips_movep_target_p) New function.
20518 (mips_prepare_pch_save): Add microMIPS support.
20519 * config/mips/mips.h (TARGET_COMPRESSION): New.
20520 (TARGET_CPU_CPP_BUILTINS): Update macro
20521 to use new compression flags and to support microMIPS.
20522 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
20523 (MIPS_ARCH_FLOAT_SPEC): Likewise.
20524 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
20525 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
20526 (ASM_SPEC): Support mmicromips and mno-micromips.
20527 (M16STORE_REG_P): New macro.
20528 (MIPS_CALL): Support TARGET_MICROMIPS.
20529 (MICROMIPS_J): New macro.
20530 (mips_base_mips16): Rename this...
20531 (mips_base_compression_flags): ...to this.
20532 (UMIPS_12BIT_OFFSET_P): New macro.
20533 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
20534 (MULTILIB_DIRNAMES): Likewise.
20536 2013-02-24 Jakub Jelinek <jakub@redhat.com>
20539 * target-globals.c (save_target_globals): For init_reg_sets and
20540 target_reinit remporarily set this_fn_optabs to this_target_optabs.
20542 2013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
20544 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
20545 * config/aarch64/t-aarch64
20546 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
20548 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
20550 PR inline-asm/56148
20551 * lra-constraints.c (process_alt_operands): Reload operand
20552 conflicting with earlier clobber only if no more other conflicting
20555 2013-02-22 Jakub Jelinek <jakub@redhat.com>
20558 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
20559 if not linking a shared library.
20561 2013-02-22 Seth LaForge <sethml@google.com>
20563 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
20565 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
20567 * config/arm/arm.md (split for extendsidi): Update condition.
20568 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
20569 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
20570 (qhs_zextenddi_cstr): Likewise.
20572 2013-02-21 Jakub Jelinek <jakub@redhat.com>
20574 PR middle-end/56420
20575 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
20576 avoid signed wrapping.
20577 (expand_mult): Handle properly multiplication by
20578 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
20579 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
20580 in the compiler if coeff is HOST_WIDE_INT_MIN.
20581 (expand_divmod): Don't make ext_op1 static, change it's type to
20582 uhwi. Avoid undefined behavior in -INTVAL (op1).
20584 PR rtl-optimization/50339
20585 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
20587 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
20588 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
20589 into splitting_ashiftrt field.
20590 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
20592 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
20595 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
20597 PR middle-end/56108
20598 * trans-mem.c (execute_tm_mark): Do not expand transactions that
20599 are sure to go irrevocable.
20601 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
20603 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
20604 scalars are valid operands.
20606 2013-02-21 Martin Jambor <mjambor@suse.cz>
20608 PR tree-optimization/56310
20609 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
20610 only matching indices and non-negative final offsets.
20611 (intersect_aggregates_with_edge): Pass src_idx to
20612 agg_replacements_to_vector. Pass src_idx insstead of index to
20613 intersect_with_agg_replacements.
20615 2013-02-21 Martin Jambor <mjambor@suse.cz>
20617 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
20618 instead of hard-wired defaults.
20620 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
20622 * doc/invoke.texi (MIPS Options): Update documentation of the
20623 floating-point multiply-accumulate instruction restrictions.
20625 2013-02-21 Kostya Serebryany <kcc@google.com>
20627 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
20628 asan_shadow_offset on x86_64 linux.
20630 2013-02-21 Richard Biener <rguenther@suse.de>
20632 PR tree-optimization/56415
20634 2013-02-11 Richard Biener <rguenther@suse.de>
20636 PR tree-optimization/56273
20637 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
20640 2013-02-21 Jakub Jelinek <jakub@redhat.com>
20643 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
20646 PR inline-asm/56405
20647 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
20648 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
20650 2013-02-20 Jan Hubicka <jh@suse.cz>
20652 PR tree-optimization/56265
20653 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
20654 when target is referenced for first time.
20656 2013-02-20 Richard Biener <rguenther@suse.de>
20658 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
20659 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
20660 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
20661 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
20662 not return anything.
20663 (rename_ssa_copies): Do not remove unused locals.
20664 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
20665 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
20666 * passes.c (execute_function_todo): Do not schedule unused locals
20667 removal if cleanup_tree_cfg did something.
20668 * tree-ssa-live.c (remove_unused_locals): Dump statistics
20669 about the number of removed locals.
20671 2013-02-20 Richard Biener <rguenther@suse.de>
20673 PR tree-optimization/56398
20674 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
20676 2013-02-20 Martin Jambor <mjambor@suse.cz>
20678 PR tree-optimization/55334
20679 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
20680 restricted pointers to arrays.
20682 2013-02-20 Richard Biener <rguenther@suse.de>
20683 Jakub Jelinek <jakub@redhat.com>
20685 PR tree-optimization/56396
20686 * tree-ssa-ccp.c (n_const_val): New static variable.
20687 (get_value): Return NULL for SSA names we don't have a lattice
20689 (ccp_initialize): Initialize n_const_val.
20690 * tree-ssa-copy.c (n_copy_of): New static variable.
20691 (init_copy_prop): Initialize n_copy_of.
20692 (get_value): Return NULL_TREE for SSA names we don't have a
20695 2013-02-20 Martin Jambor <mjambor@suse.cz>
20697 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
20699 2013-02-20 Richard Biener <rguenther@suse.de>
20701 * genpreds.c (write_lookup_constraint): Do not compare first
20702 letter of the constraint again.
20704 2013-02-20 Richard Biener <rguenther@suse.de>
20706 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
20708 (get_use_iv_cost): Terminate hashtable walk when coming across
20711 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
20713 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
20714 reassociation for avx2 targets.
20716 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20718 * config/microblaze/microblaze.c: microblaze_has_clz = 0
20719 Add version check for v8.10.a to enable microblaze_has_clz
20720 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
20721 version and TARGET_PATTERN_COMPARE check
20722 * config/microblaze/microblaze.md: New clzsi2 instruction
20724 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20726 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
20727 function before branching.
20729 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
20731 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
20733 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
20735 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
20737 PR middle-end/55889
20738 * sel-sched.c: Include ira.h.
20739 (implicit_clobber_conflict_p): New function.
20740 (moveup_expr): Use it.
20741 * Makefile.in (sel-sched.o): Depend on ira.h.
20743 2013-02-19 Richard Biener <rguenther@suse.de>
20745 PR tree-optimization/56384
20746 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
20747 (vn_hash_type): Split out from ...
20748 (vn_hash_constant_with_type): ... here.
20749 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
20750 (vn_phi_eq): Compare types from vn_phi_s structure.
20751 (vn_phi_lookup): Populate vn_phi_s type.
20752 (vn_phi_insert): Likewise.
20754 2013-02-19 Jakub Jelinek <jakub@redhat.com>
20756 PR tree-optimization/56350
20757 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
20758 if haven't found reduction or nested cycle operand, rather than
20759 asserting we must find it.
20761 PR tree-optimization/56381
20762 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
20765 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
20766 Jakub Jelinek <jakub@redhat.com>
20769 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
20770 (swap_optab_enable): Same.
20771 (init_all_optabs): Use argument instead of global.
20772 * tree.h (struct tree_optimization_option): New field target_optabs.
20773 * expr.h (init_all_optabs): Add argument to prototype.
20774 (TREE_OPTIMIZATION_OPTABS): New.
20775 (save_optabs_if_changed): Protoize.
20776 * optabs.h: Declare this_fn_optabs.
20777 * optabs.c (save_optabs_if_changed): New.
20778 Declare this_fn_optabs.
20779 (init_optabs): Add argument to init_all_optabs() call.
20780 * function.c (invoke_set_current_function_hook): Handle per
20782 * function.h (struct function): New field optabs.
20783 * config/mips/mips.c (mips_set_mips16_mode): Handle when
20784 optimization_current_node has changed.
20785 * target-globals.h (save_target_globals_default_opts): Protoize.
20786 * target-globals.c (save_target_globals_default_opts): New.
20788 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20791 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
20792 registers %fr12 and %fr12R as call used.
20795 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
20796 and HImode, require all displacements to be an integer multiple of
20798 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
20799 only allow QImode and HImode when reload is in progress and strict is
20800 true. Likewise for symbolic addresses. Use base14_operand to check
20801 displacements in REG+BASE addresses.
20803 2013-02-18 Richard Biener <rguenther@suse.de>
20805 PR tree-optimization/56366
20806 * tree-vect-loop.c (get_initial_def_for_induction): Properly
20807 handle sign-conversion of outer-loop initial induction value.
20809 2013-02-18 Richard Biener <rguenther@suse.de>
20811 PR middle-end/56349
20812 * cfghooks.c (merge_blocks): If we merge a latch into another
20813 block adjust references to it.
20814 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
20815 (verify_loop_structure): Verify that a recorded latch is in fact
20818 2013-02-18 Richard Biener <rguenther@suse.de>
20820 PR tree-optimization/56321
20821 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
20822 order SSA name release and virtual operand unlinking.
20824 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20826 * config/microblaze/microblaze.md (save_stack_block): Define.
20827 (restore_stack_block): Likewise.
20829 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20831 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
20832 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
20833 * config/microblaze/microblaze.c (microblaze_option_override):
20834 Bail out early for PIC modes when target does not support PIC.
20836 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20838 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
20839 Replace with a microblaze version.
20840 (microblaze_trampoline_init): Adapt for microblaze.
20841 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
20844 2013-02-16 Jakub Jelinek <jakub@redhat.com>
20845 Dodji Seketeli <dodji@redhat.com>
20848 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
20849 (instrument_mem_region_access): Do not forget to always put
20850 instrumentation of the of 'base' and 'base + len' in a "if (len !=
20851 0) statement, even for cases where either 'base' or 'base + len'
20852 are not instrumented -- because they have been previously
20853 instrumented. Simplify the logic by putting all the statements
20854 instrument 'base + len' inside a sequence, and then insert that
20855 sequence right before the current insertion point. Then, to
20856 instrument 'base + len', just get an iterator on that statement.
20857 And do not forget to update the pointer to iterator the function
20858 received as argument.
20860 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
20862 PR rtl-optimization/56348
20863 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
20865 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
20867 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
20868 (clean_graph_dump_file): Pass base to start_graph_dump.
20870 2013-02-14 Richard Henderson <rth@redhat.com>
20873 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
20875 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
20877 * collect2-aix.h: Define F_LOADONLY.
20879 2013-02-14 Richard Biener <rguenther@suse.de>
20882 * varasm.c (output_constant_def_1): Get the decl representing
20883 the constant as argument.
20884 (output_constant_def): Wrap output_constant_def_1.
20885 (make_decl_rtl): Use output_constant_def_1 with the decl
20886 representing the constant.
20887 (build_constant_desc): Optionally re-use a decl already
20888 representing the constant.
20889 (tree_output_constant_def): Adjust.
20891 2013-02-14 Dodji Seketeli <dodji@redhat.com>
20894 * asan.c (instrument_builtin_call): Really put the length of the
20895 second source argument into src1_len.
20897 2013-02-13 Jakub Jelinek <jakub@redhat.com>
20899 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
20900 argument. If it is false, don't create edge from then_bb to
20902 (insert_if_then_before_iter): Pass true to it.
20903 (build_check_stmt): Pass false to it.
20904 (transform_statements): Flush hash table only on extended basic
20905 block boundaries, rather than at the beginning of every bb.
20906 Don't flush hash table on nonfreeing_call_p calls.
20907 * tree-flow.h (nonfreeing_call_p): New prototype.
20908 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
20910 2013-02-13 David S. Miller <davem@davemloft.net>
20912 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
20914 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
20917 * ira.c (max_regno_before_ira): Move from ...
20919 (fix_reg_equiv_init): Use max_regno_before_ira instead of
20922 2013-02-13 Jakub Jelinek <jakub@redhat.com>
20924 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
20926 2013-02-13 Richard Biener <rguenther@suse.de>
20929 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
20930 globals in MEM_REFs.
20932 2013-02-13 Richard Biener <rguenther@suse.de>
20934 * loop-init.c (loop_optimizer_init): Clear loop state when
20935 re-initializing preserved loops.
20936 * loop-unswitch.c (unswitch_single_loop): Return whether
20937 we unswitched the loop. Do not verify loop state here.
20938 (unswitch_loops): When we unswitched a loop discover new loops.
20940 2013-02-13 Kostya Serebryany <kcc@google.com>
20942 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
20944 * sanitizer.def: Rename __asan_init to __asan_init_v1.
20946 2013-02-12 Dodji Seketeli <dodji@redhat.com>
20948 Avoid instrumenting duplicated memory access in the same basic block
20949 * Makefile.in (asan.o): Add new dependency on hash-table.h
20950 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
20951 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
20952 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
20953 (free_mem_ref_resources, has_mem_ref_been_instrumented)
20954 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
20955 (get_mem_ref_of_assignment): New functions.
20956 (get_mem_refs_of_builtin_call): Extract from
20957 instrument_builtin_call and tweak a little bit to make it fit with
20959 (instrument_builtin_call): Use the new
20960 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
20961 of is_gimple_builtin_call.
20962 (instrument_derefs, instrument_mem_region_access): Insert the
20963 instrumented memory reference into the hash table.
20964 (maybe_instrument_assignment): Renamed instrument_assignment into
20965 this, and change it to advance the iterator when instrumentation
20966 actually happened and return true in that case. This makes it
20967 homogeneous with maybe_instrument_assignment, and thus give a
20968 chance to callers to be more 'regular'.
20969 (transform_statements): Clear the memory reference hash table
20970 whenever we enter a new BB, when we cross a function call, or when
20971 we are done transforming statements. Use
20972 maybe_instrument_assignment instead of instrumentation. No more
20973 need to special case maybe_instrument_assignment and advance the
20974 iterator after calling it; it's now handled just like
20975 maybe_instrument_call. Update comment.
20977 2013-02-13 Richard Biener <rguenther@suse.de>
20979 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
20980 Fix loop discovery code.
20982 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
20984 PR inline-asm/56148
20985 * lra-constraints.c (process_alt_operands): Match early clobber
20986 operand with itself. Check conflicts with earlyclobber only if
20987 the operand is not reloaded. Prefer to reload conflicting operand
20988 if earlyclobber and matching operands are the same.
20990 2013-02-12 Richard Biener <rguenther@suse.de>
20993 * lto-streamer-out.c (write_symbol): Do not output symbols
20994 for hard register variables.
20996 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
20999 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
21000 (umulsidi3_insn, mulsidi3_insn): New insns.
21002 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
21004 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
21005 (struct tune_params): Add vec_costs field.
21006 * config/arm/arm.c (arm_builtin_vectorization_cost)
21007 (arm_add_stmt_cost): New functions.
21008 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
21009 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
21010 (arm_default_vec_cost): New struct of type cpu_vec_costs.
21011 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
21012 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
21013 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
21014 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
21016 2013-02-12 Richard Biener <rguenther@suse.de>
21019 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
21020 decls again if possible.
21022 2013-02-12 Richard Biener <rguenther@suse.de>
21024 PR middle-end/56288
21025 * tree-ssa.c (verify_ssa_name): Fix check, move
21026 SSA_NAME_IN_FREE_LIST check up.
21028 2013-02-12 Jakub Jelinek <jakub@redhat.com>
21029 Steven Bosscher <steven@gcc.gnu.org>
21031 PR rtl-optimization/56151
21032 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
21033 equal to op0 or op1, and last_insn pattern is CODE operation
21034 with MEM dest and one of the operands matches that MEM.
21036 2013-02-11 Sriraman Tallam <tmsriramgoogle.com>
21038 * doc/extend.texi: Document Function Multiversioning and "default"
21039 parameter string to target attribute.
21040 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
21041 target attribute parameter is "default".
21042 (ix86_compare_version_priority): Remove checks for target attribute.
21043 (ix86_mangle_function_version_assembler_name): Change error to sorry.
21044 Remove check for target attribute equal to NULL. Add assert.
21045 (ix86_generate_version_dispatcher_body): Change error to sorry.
21047 2013-02-11 Iain Sandoe <iain@codesourcery.com>
21048 Jack Howarth <howarth@bromo.med.uc.edu>
21049 Patrick Marlier <patrick.marlier@gmail.com>
21052 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
21053 define ENDFILE_SPEC as TM_DESTRUCTOR.
21054 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
21056 2013-02-11 Alexander Potapenko <glider@google.com>
21057 Jack Howarth <howarth@bromo.med.uc.edu>
21058 Jakub Jelinek <jakub@redhat.com>
21061 * config/darwin.c (cdtor_record): Rename ctor_record.
21062 (sort_cdtor_records): Rename sort_ctor_records.
21063 (finalize_dtors): New routine to sort destructors by
21064 priority before use in assemble_integer.
21065 (machopic_asm_out_destructor): Use finalize_dtors if needed.
21067 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
21069 PR rtl-optimization/56275
21070 * simplify-rtx.c (avoid_constant_pool_reference): Check that
21071 offset is non-negative and less than cmode size before
21072 calling simplify_subreg.
21074 2013-02-11 Richard Biener <rguenther@suse.de>
21076 PR tree-optimization/56264
21077 * cfgloop.h (fix_loop_structure): Adjust prototype.
21078 * loop-init.c (fix_loop_structure): Return the number of
21079 newly discovered loops.
21080 * tree-cfgcleanup.c (repair_loop_structures): When new loops
21081 are discovered, do a full loop-closed SSA rewrite.
21083 2013-02-11 Richard Biener <rguenther@suse.de>
21085 PR tree-optimization/56273
21086 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
21088 (check_array_ref): Fix missing newline in dumps.
21089 (search_for_addr_array): Likewise.
21091 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
21093 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
21095 2013-02-09 Jakub Jelinek <jakub@redhat.com>
21098 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
21100 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
21102 PR rtl-optimization/56246
21103 * lra-constraints.c (simplify_operand_subreg): Try to reuse
21105 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
21106 constraints are satisfied.
21108 2013-02-08 Jeff Law <law@redhat.com>
21111 * emit-rtl.c (reg_is_parm_p): New function.
21112 * regs.h (reg_is_parm_p): New prototype.
21113 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
21114 callee-clobbered registers.
21116 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
21119 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
21120 If there is no implicit builtin declaration, just return NULL.
21122 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
21124 * config/i386/sse.md (FMAMODEM): New mode iterator.
21125 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
21126 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
21128 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
21130 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
21131 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
21132 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
21134 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
21136 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
21137 (microblaze*-*-elf): Likewise.
21138 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
21140 * config/microblaze/microblaze-c.c: Add builtin defines for
21141 _LITTLE_ENDIAN and _BIG_ENDIAN.
21142 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
21143 add to TARGET_DEFAULT flags.
21144 Expand ASM_SPEC and LINK_SPEC.
21145 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
21146 * config/microblaze/microblaze.md: Update extendsidi2 and
21147 movdi_internal instructions to use low-order / high-order reg
21149 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
21150 options and inversemask / mask of LITTLE_ENDIAN.
21151 * config/microblaze/t-microblaze: Expand multilib options to
21152 include mlittle-endian (le) and update exceptions patterns.
21154 2013-02-08 Jakub Jelinek <jakub@redhat.com>
21156 PR rtl-optimization/56195
21157 * lra-constraints.c (get_reload_reg): Don't reuse regs
21158 if they have smaller mode than requested, if they have
21159 wider mode than requested, try to return a SUBREG.
21161 PR tree-optimization/56250
21162 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
21163 if type is unsigned and code isn't MULT_EXPR.
21165 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
21167 PR tree-optimization/56064
21168 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
21169 bits according to mode.
21170 * fixed-value.h (fixed_from_double_int)
21171 (const_fixed_from_double_int): Adjust comments.
21173 2013-02-08 Richard Biener <rguenther@suse.de>
21176 * lto-streamer.h (struct data_in): Remove current_file, current_line
21177 and current_col members.
21178 * lto-streamer-out.c (lto_output_location): Stream changed bits
21179 en-block for efficiency.
21180 * lto-streamer-in.c (clear_line_info): Remove.
21181 (lto_input_location): Cache current file, line and column
21182 globally via local statics. Read changed bits en-block.
21183 (input_function): Do not call clear_line_info.
21184 (lto_read_body): Likewise.
21185 (lto_input_toplevel_asms): Likewise.
21187 2013-02-08 Michael Matz <matz@suse.de>
21189 PR tree-optimization/52448
21190 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
21191 (nt_call_phase): New static.
21192 (add_or_mark_expr): Only mark accesses with newer phase than any
21194 (nonfreeing_call_p): New.
21195 (nt_init_block): Update nt_call_phase, mark blocks as visited.
21196 (nt_fini_block): Keep blocks marked as visited.
21197 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
21199 2013-02-08 Richard Biener <rguenther@suse.de>
21201 * ira.c (ira): Free broken dominator information.
21203 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
21205 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
21207 2013-02-08 Marek Polacek <polacek@redhat.com>
21209 * cfgloop.c (verify_loop_structure): Add more checking of headers.
21211 2013-02-08 Richard Biener <rguenther@suse.de>
21213 PR middle-end/56181
21214 * cfgloop.h (flow_loops_find): Adjust.
21215 (bb_loop_header_p): Declare.
21216 * cfgloop.c (bb_loop_header_p): New function split out from ...
21217 (flow_loops_find): ... here. Adjust function signature,
21218 support incremental loop structure update.
21219 (verify_loop_structure): Cleanup. Verify a loop is a loop.
21220 * cfgloopmanip.c (fix_loop_structure): Move ...
21221 * loop-init.c (fix_loop_structure): ... here.
21222 (apply_loop_flags): Split out from ...
21223 (loop_optimizer_init): ... here.
21224 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
21225 in incremental mode, only remove dead loops here.
21227 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
21230 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
21231 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
21232 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
21233 (*round<mode>3.libgcc): New insns for fixed-modes.
21234 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
21235 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
21236 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
21237 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
21238 implementations. Define to __builtin_avr_absFX,
21239 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
21240 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
21241 __builtin_avr_countlsFX, respectively.
21242 * config/avr/avr-c.c (target.h): Include it.
21243 (enum avr_builtin_id): New enum.
21244 (avr_resolve_overloaded_builtin): New static function.
21245 (avr_register_target_pragmas): Use it to set
21246 targetm.resolve_overloaded_builtin.
21247 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
21248 tree nodes used by DEF_BUILTIN.
21249 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
21250 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
21251 <AVR_BUILTIN_xxBITS>: Same.
21253 2013-02-08 Richard Biener <rguenther@suse.de>
21255 * cfgloop.c (verify_loop_structure): Properly handle
21256 a loop exiting to another loop header.
21257 * ira-int.h (ira_loops): Remove.
21258 * ira.c (ira_loops): Remove.
21259 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
21260 (do_reload): Use loop_optimizer_finalize.
21261 * ira-build.c (create_loop_tree_nodes): Use get_loops and
21262 number_of_loops to access the loop tree.
21263 (more_one_region_p): Likewise.
21264 (finish_loop_tree_nodes): Likewise.
21265 (rebuild_regno_allocno_maps): Likewise.
21266 (mark_loops_for_removal): Likewise.
21267 (mark_all_loops_for_removal): Likewise.
21268 (remove_unnecessary_regions): Likewise.
21269 (ira_build): Likewise.
21270 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
21272 2013-02-08 Richard Biener <rguenther@suse.de>
21274 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
21275 * ipa-pure-const.c (analyze_function): Avoid calling
21276 mark_irreducible_loops twice.
21277 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
21279 2013-02-07 David S. Miller <davem@davemloft.net>
21281 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
21283 * var-tracking.c (vt_add_function_parameter): Test the presence of
21284 HAVE_window_save properly and do not remap argument registers when
21285 we have a leaf function.
21287 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
21290 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
21292 * config/i386/i386.c (ix86_print_operand): Ditto.
21294 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
21296 * lra-constraints.c (process_alt_operands): Fix recently added comment.
21298 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
21300 PR rtl-optimization/56225
21301 * lra-constraints.c (process_alt_operands): Check that reload hard
21302 reg can hold value for strict_low_part.
21304 2013-02-07 Jakub Jelinek <jakub@redhat.com>
21307 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
21308 dwarf2out_end_function.
21309 (in_first_function_p, maybe_at_text_label_p,
21310 first_loclabel_num_not_at_text_label): New variables.
21311 (dwarf2out_var_location): In the first function find out
21312 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
21313 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
21316 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
21318 PR rtl-optimization/56178
21319 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
21320 SUBREG of a register. Tidy up related block of code.
21321 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
21322 note if the source is a register or a SUBREG of a register.
21324 2013-02-07 Jakub Jelinek <jakub@redhat.com>
21327 * config/rs6000/rs6000.md (ptrm): New mode attr.
21328 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
21329 call_value_indirect_aix<pttrsize>,
21330 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
21333 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
21335 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
21336 if -bnortl. Convert to strcmp and strncmp.
21338 2013-02-07 Alan Modra <amodra@gmail.com>
21341 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
21342 addresses won't wrap when offsetting.
21343 (rs6000_secondary_reload): Provide secondary reloads needed for
21344 wrapping LO_SUM addresses.
21346 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
21348 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
21349 MACH, just __MACH__.
21351 2013-02-06 Richard Biener <rguenther@suse.de>
21353 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
21354 instead of calling fix_loop_structure.
21356 2013-02-06 Jakub Jelinek <jakub@redhat.com>
21358 PR middle-end/56217
21359 * omp-low.c (use_pointer_for_field): Return false if
21360 lower_send_shared_vars doesn't generate any copy-out code.
21362 2013-02-06 Tom de Vries <tom@codesourcery.com>
21364 PR rtl-optimization/56131
21365 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
21366 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
21367 of the label is NULL. Add comment.
21369 2013-02-05 Jakub Jelinek <jakub@redhat.com>
21371 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
21374 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
21375 (STATIC_LIBTSAN_LIBS): Likewise.
21376 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
21377 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
21378 is defined, don't add anything else beyond that.
21379 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
21380 (LINK_COMMAND_SPEC): Use them.
21382 PR tree-optimization/56205
21383 * tree-stdarg.c (check_all_va_list_escapes): Return true if
21384 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
21385 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
21387 2013-02-05 Richard Biener <rguenther@suse.de>
21389 PR tree-optimization/53342
21390 PR tree-optimization/53185
21391 * tree-vectorizer.h (vect_check_strided_load): Remove.
21392 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
21393 not disallow peeling for vectorized strided loads.
21394 (vect_check_strided_load): Make static and simplify.
21395 (vect_analyze_data_refs): Adjust.
21396 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
21397 correctly when vectorizing strided loads.
21399 2013-02-05 Richard Biener <rguenther@suse.de>
21401 * doc/install.texi: Refer to ISL, not PPL.
21403 2013-02-05 Jan Hubicka <jh@suse.cz>
21405 PR tree-optimization/55789
21406 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
21408 2013-02-05 Jan Hubicka <jh@suse.cz>
21410 PR tree-optimization/55789
21411 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
21412 the dead call anyway.
21414 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
21417 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
21419 2013-02-04 Alexander Potapenko <glider@google.com>
21420 Jack Howarth <howarth@bromo.med.uc.edu>
21421 Jakub Jelinek <jakub@redhat.com>
21424 * config/darwin.c (sort_ctor_records): Stabilized qsort
21425 on constructor priority by using original position.
21426 (finalize_ctors): New routine to sort constructors by
21427 priority before use in assemble_integer.
21428 (machopic_asm_out_constructor): Use finalize_ctors if needed.
21430 2013-02-04 Jakub Jelinek <jakub@redhat.com>
21433 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
21434 about visibility on artificial decls.
21435 * config/sol2.c (solaris_assemble_visibility): Likewise.
21437 2013-02-04 Kai Tietz <ktietz@redhat.com>
21440 * config/i386/i386.c (function_value_ms_64): Add additional valtype
21441 argument and improve checking of return-argument types for 16-byte
21443 (ix86_function_value_1): Add additional valtype argument on call
21444 of function_value_64.
21445 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
21446 handling infunction_value_64 function.
21448 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
21450 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
21452 2013-02-04 Richard Biener <rguenther@suse.de>
21454 PR tree-optimization/56188
21455 * tree-ssa-structalias.c (label_visit): Consider case with
21456 initially non-empty points-to set.
21457 (perform_var_substitution): Dump node mapping and clean up.
21459 2013-02-04 Richard Guenther <rguenther@suse.de>
21462 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
21463 node prevail as last resort.
21464 (lto_symtab_merge_decls): Remove guard on LTRANS here.
21465 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
21467 2013-02-04 Richard Biener <rguenther@suse.de>
21469 PR tree-optimization/56113
21470 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
21472 (equiv_class_lookup_or_add): ... this.
21473 (label_visit): Adjust and fix error in previous patch.
21474 (perform_var_substitution): Adjust.
21476 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
21478 * config/sh/divtab.c: Fix formatting and comments throughout the file.
21479 * config/sh/sh4-300.md: Likewise.
21480 * config/sh/sh4a.md: Likewise.
21481 * config/sh/constraints.md: Likewise.
21482 * config/sh/sh.md: Likewise.
21483 * config/sh/netbsd-elf.h: Likewise.
21484 * config/sh/predicates.md: Likewise.
21485 * config/sh/sh-protos.h: Likewise.
21486 * config/sh/ushmedia.h: Likewise.
21487 * config/sh/linux.h: Likewise.
21488 * config/sh/sh.c: Likewise.
21489 * config/sh/superh.h: Likewise.
21490 * config/sh/elf.h: Likewise.
21491 * config/sh/sh4.md: Likewise.
21492 * config/sh/sh.h: Likewise.
21494 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21496 * config/pa/constraints.md: Adjust unused letters. Change "T"
21497 constraint to match_test floating_point_store_memory_operand().
21498 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
21499 (base14_operand): New.
21500 (floating_point_store_memory_operand): New.
21501 (integer_store_memory_operand): Revise to use base14_operand and
21502 reg_plus_base_memory_operand.
21503 (move_dest_operand): Allow symbolic_memory_operands.
21504 (symbolic_memory_operand): Check for LO_SOM.
21505 (symbolic_operand): Change default case to break.
21506 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
21507 CONST_DOUBLE values to be reloaded by putting them into memory when
21508 the destination is a floating point register.
21509 (movdf): Remove code to handle CONST_DOUBLE.
21511 (reload_indf_r1): New.
21512 (reload_insf_r1): New.
21513 Consistently use "Q" and "T" constraints with integer and floating
21514 point move instructions, respectively.
21515 (movdi): Remove FAIL.
21516 Change predicate for source operand unamed DImode move from
21517 general_operand to move_src_operand.
21518 (umulsidi3): Change predicate for destination operand to
21520 Likewise for similar unamed patterns.
21521 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
21522 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
21523 (hppa_legitimize_address): Simplify mask calculation.
21524 (pa_emit_move_sequence): Revised handling of secondary reloads from
21525 REG+D addresses for floating point loads and stores. Directly handle
21526 loading CONST0_RTX (mode) to a floating point register.
21527 (pa_secondary_reload): Handle reloading DF and SFmode constant values
21528 to floating point registers. Don't restrict secondary reloads to
21529 floating point registers to integer modes. Revise some comments and
21531 (TARGET_LEGITIMATE_ADDRESS_P): Define.
21532 (pa_legitimate_address_p): New.
21533 (pa_legitimize_reload_address): New.
21534 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
21535 (STRICT_REG_OK_FOR_BASE_P): New.
21536 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
21537 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
21539 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
21540 Andrew Dixie <andrewd@gentrack.com>
21542 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
21545 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
21547 * expmed.c (extract_bit_field_1): Pass the full width of the
21548 structure to get_best_reg_extraction_insn.
21550 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
21553 * configure.ac (use_cxa_atexit): Add AIX.
21554 * configure: Regenerate.
21556 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
21558 2013-02-01 Jakub Jelinek <jakub@redhat.com>
21561 * final.c (need_profile_function): New variable.
21562 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
21563 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
21564 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
21565 notes, targetm.asm_out.function_prologue doesn't emit anything,
21566 HAVE_prologue and profiler should be emitted before prologue,
21567 set need_profile_function instead of emitting it.
21568 (final_scan_insn): If need_profile_function, emit
21569 profile_function on the first NOTE_INSN_BASIC_BLOCK or
21570 NOTE_INSN_FUNCTION_BEG note.
21572 2013-02-01 Richard Henderson <rth@redhat.com>
21574 * config/rs6000/rs6000.md (smulditi3): New.
21577 * config/alpha/alpha.md (umulditi3): New.
21579 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
21581 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
21582 (ASM_OUTPUT_ALIGNED_LOCAL): New.
21584 2013-02-01 Richard Biener <rguenther@suse.de>
21586 PR tree-optimization/56113
21587 * tree-ssa-structalias.c (label_visit): Reduce work for
21588 single-predecessor nodes.
21590 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
21592 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
21593 range isn't testing for zero.
21595 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
21597 PR middle-end/56113
21598 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
21600 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
21601 Nick Clifton <nickc@redhat.com>
21603 * config/v850/constraints.md (Q): Define as a memory constraint.
21604 * config/v850/predicates.md (label_ref_operand): New predicate.
21605 (e3v5_shift_operand): New predicate.
21606 (ior_operator): New predicate.
21607 * config/v850/t-v850: Add e3v5 multilib.
21608 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
21609 (v850_gen_movdi): Prototype.
21610 * config/v850/v850.c: Add support for e3v5 architecture.
21611 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
21613 (construct_save_jarl): Add e3v5 long JARL support.
21614 (v850_adjust_insn_length): New function. Adjust length of call
21615 insns when using e3v5 instructions.
21616 (v850_gen_movdi): New function: Generate instructions to move a
21618 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
21619 (CPP_SPEC): Define __v850e3v5__ as appropriate.
21620 (TARGET_USE_FPU): Enable for e3v5.
21621 (CONST_OK_FOR_W): New macro.
21622 (ADJUST_INSN_LENGTH): Define.
21623 * config/v850/v850.md (UNSPEC_LOOP): Define.
21624 (attr cpu): Add v850e3v5.
21625 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
21626 (movdi): New pattern.
21627 (movdi_internal): New pattern.
21628 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
21629 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
21630 (cstoresf4): Likewise.
21631 (cstoredf4): Likewise.
21632 (insv): New pattern.
21633 (rotlso3_a): New pattern.
21634 (rotlsi3_b): New pattern
21635 (rotlsi3_v850e3v5): New pattern.
21636 (doloop_begin): New pattern.
21637 (fix_loop_counter): New pattern.
21638 (doloop_end): New pattern.
21639 (branch_normal): Add e3v5 long branch support.
21640 (branch_invert): Likewise.
21641 (branch_z_normal): Likewise.
21642 (branch_z_invert): Likewise.
21643 (branch_nz_normal): Likewise.
21644 (branch_nz_invert): Likewise.
21645 (call_internal_short): Add e3v5 register-indirect JARL support.
21646 (call_internal_long): Likewise.
21647 (call_value_internal_short): Likewise.
21648 (call_value_internal_long): Likewise.
21649 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
21650 (mloop): New option.
21651 * config.gcc: Add support for configuring v840e3v5 target.
21652 * doc/invoke.texi: Document new v850 specific command line options.
21654 2013-01-31 Paul Koning <ni1d@arrl.net>
21658 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
21659 children if parent is a class.
21660 (prune_unused_types_prune): Don't add DW_AT_declaration.
21662 2013-01-31 Richard Biener <rguenther@suse.de>
21664 PR tree-optimization/56157
21665 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
21666 match up operand with SLP child.
21668 2013-01-31 Jason Merrill <jason@redhat.com>
21671 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
21672 parameters the first time.
21673 (gen_scheduled_generic_parms_dies): Check completeness here.
21675 2013-01-31 Richard Biener <rguenther@suse.de>
21677 PR middle-end/53073
21678 * common.opt (faggressive-loop-optimizations): New flag,
21679 enabled by default.
21680 * doc/invoke.texi (faggressive-loop-optimizations): Document.
21681 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
21682 infer_loop_bounds_from_undefined by it.
21684 2013-01-31 Richard Biener <rguenther@suse.de>
21686 PR tree-optimization/56150
21687 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
21688 visit virtual operands.
21689 (find_uses_to_rename_bb): Likewise.
21691 2013-01-31 Richard Biener <rguenther@suse.de>
21693 PR tree-optimization/56150
21694 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
21695 mixed store non-store stmts.
21697 2013-01-30 Jakub Jelinek <jakub@redhat.com>
21700 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
21701 LIBASAN_EARLY_SPEC is defined.
21702 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
21703 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
21705 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
21708 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
21709 invalid args instead of ICEing on it.
21710 (ix86_valid_target_attribute_tree): Return error_mark_node if
21711 ix86_valid_target_attribute_inner_p failed.
21712 (ix86_valid_target_attribute_p): Return false only if
21713 ix86_valid_target_attribute_tree returned error_mark_node. Allow
21714 target("default") attribute.
21715 (sorted_attr_string): Change argument from const char * to tree,
21716 merge in all target attribute arguments rather than just one.
21717 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
21718 instead of free. Avoid using strcat.
21719 (ix86_mangle_function_version_assembler_name): Mangle
21720 target("default") as if no target attribute is present. Adjust
21721 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
21722 instead of xmalloc and XDELETEVEC instead of free.
21723 (ix86_function_versions): Don't return true if one of the decls
21724 doesn't have target attribute. If they don't and one of the decls
21725 is DECL_FUNCTION_VERSIONED, report an error. Adjust
21726 sorted_attr_string caller. Use XDELETEVEC instead of free.
21727 (ix86_supports_function_versions): Remove.
21728 (make_name): Fix up formatting.
21729 (make_dispatcher_decl): Remove resolver_name and its initialization.
21730 Avoid leaking memory.
21731 (is_function_default_version): Return true if there is
21732 target("default") attribute rather than no target attribute at all.
21733 (make_resolver_func): Avoid leaking memory.
21734 (ix86_generate_version_dispatcher_body): Likewise.
21735 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
21736 * target.def (supports_function_versions): Remove.
21737 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
21738 * doc/tm.texi: Regenerated.
21740 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
21742 PR rtl-optimization/56144
21743 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
21744 for values with side effects.
21746 2013-01-30 Richard Biener <rguenther@suse.de>
21748 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
21749 (sparseset_pop): Likewise.
21750 * cfganal.c (compute_idf): Likewise. Increase work-stack size
21751 to be able to use quick_push in the worker loop.
21753 2013-01-30 Marek Polacek <polacek@redhat.com>
21755 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
21757 2013-01-30 Richard Biener <rguenther@suse.de>
21760 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
21762 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
21764 PR tree-optimization/56064
21765 * fixed-value.c (fixed_from_double_int): New function.
21766 * fixed-value.h (fixed_from_double_int): New prototype.
21767 (const_fixed_from_double_int): New static inline function.
21768 * fold-const.c (native_interpret_fixed): New static function.
21769 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
21770 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
21771 (native_encode_fixed): New static function.
21772 (native_encode_expr) <FIXED_CST>: Use it.
21773 (native_interpret_int): Move double_int worker code to...
21774 * double-int.c (double_int::from_buffer): ...this new static method.
21775 * double-int.h (double_int::from_buffer): Prototype it.
21777 2013-01-30 Richard Biener <rguenther@suse.de>
21779 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
21780 New pointer-map and obstack.
21781 (init_alias_vars): Allocate pointer-map and obstack.
21782 (delete_points_to_sets): Free them.
21783 (find_what_var_points_to): Cache result.
21784 (find_what_p_points_to): Adjust for changed interface of
21785 find_what_var_points_to.
21786 (compute_points_to_sets): Likewise.
21787 (ipa_pta_execute): Likewise.
21789 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
21791 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
21792 * configure: Regenerate.
21793 * config.in: Regenerate.
21794 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
21795 #nobits/#progbits if supported.
21797 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
21800 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
21801 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
21802 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
21804 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
21806 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
21807 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
21809 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
21811 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
21812 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
21814 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
21816 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
21818 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
21819 * config/arm/cortex-a7.md: New bypasses using
21820 arm_mac_accumulator_is_result.
21822 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
21824 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
21825 (cortex_a7_neon_mla): Likewise.
21826 (cortex_a7_fpfmad): New reservation.
21827 (cortex_a7_fpmacs): Use ffmas and update required units.
21828 (cortex_a7_fpmuld): Update required units and latency.
21829 (cortex_a7_fpmacd): Likewise.
21830 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
21831 (cortex_a7_neon). Likewise.
21832 (bypass) Update participating units.
21834 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
21836 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
21837 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
21839 * config/arm/vfp11.md (vfp_farith): Use ffmas.
21840 (vfp_fmul): Use ffmad.
21841 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
21842 (cortex_r4_fmacd): Use ffmad.
21843 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
21844 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
21845 (cortex_a9_fmacd): Use ffmad.
21846 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
21847 (cortex_a8_vfp_macd): Use ffmad.
21848 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
21849 (cortex_a5_fpmacd): Use ffmad.
21850 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
21851 (cortex_a15_vfp_macd): Use ffmad.
21852 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
21854 2013-01-29 Jason Merrill <jason@redhat.com>
21857 * varasm.c (default_assemble_visibility): Don't warn about
21858 visibility on artificial decls.
21860 2013-01-29 Richard Biener <rguenther@suse.de>
21862 PR tree-optimization/56113
21863 * tree-ssa-structalias.c (equiv_class_lookup): Also return
21865 (label_visit): Free duplicate bitmaps and record the leader instead.
21866 (perform_var_substitution): Adjust.
21868 2013-01-29 Richard Biener <rguenther@suse.de>
21870 PR tree-optimization/55270
21871 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
21872 the CFG, schedule loops for fixup.
21874 2013-01-29 Nick Clifton <nickc@redhat.com>
21876 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
21879 2013-01-28 Leif Ekblad <leif@rdos.net>
21881 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
21882 * config/i386/i386.h (TARGET_RDOS): New macro.
21883 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
21884 * config/i386/i386.c (ix86_option_override_internal): For 64bit
21885 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
21886 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
21887 DEFAULT_LARGE_SECTION_THRESHOLD.
21888 * config/i386/i386.md (R14_REG, R15_REG): New constants.
21889 * config/i386/rdos.h: New file.
21890 * config/i386/rdos64.h: New file.
21892 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
21895 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
21898 2013-01-28 Jakub Jelinek <jakub@redhat.com>
21900 PR rtl-optimization/56117
21901 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
21902 call cselib_lookup_from_insn on the MEM before calling
21903 add_insn_mem_dependence.
21905 2013-01-28 Richard Biener <rguenther@suse.de>
21907 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
21908 to a stmt that didn't have one.
21909 (copy_phis_for_bb): Likewise for PHI arguments.
21910 (copy_debug_stmt): Likewise for debug stmts.
21912 2013-01-28 Richard Biener <rguenther@suse.de>
21914 PR tree-optimization/56034
21915 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
21916 (partition_builtin_p): Adjust.
21917 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
21918 it is the last partition.
21919 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
21920 up the vertex for the definition.
21921 (classify_partition): Classify whether a partition is a
21922 PKIND_REDUCTION, thus has uses outside of the loop.
21923 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
21924 Merge all PKIND_REDUCTION partitions into the last partition.
21925 (tree_loop_distribution): Seed partitions from reductions as well.
21927 2013-01-28 Jakub Jelinek <jakub@redhat.com>
21929 PR tree-optimization/56125
21930 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
21931 pow(x,c) into sqrt(x) * powi(x, n/2) or
21932 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
21933 optimizing for size.
21934 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
21935 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
21938 PR tree-optimization/56094
21939 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
21940 to UNKNOWN_LOCATION while gimplifying expr.
21942 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
21945 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
21946 operand 0 in movabs insn template for -masm=intel asm alternative.
21947 (*movabs<mode>_2): Ditto for operand 1.
21949 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
21952 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
21955 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
21957 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
21960 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
21962 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
21963 Undef to avoid warning.
21965 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
21967 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
21968 * configure: Regenerate.
21970 2013-01-25 Jakub Jelinek <jakub@redhat.com>
21972 PR tree-optimization/56098
21973 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
21974 for stmts with volatile ops.
21975 (cond_store_replacement): Don't optimize if assign has volatile ops.
21976 (cond_if_else_store_replacement_1): Don't optimize if either
21977 then_assign or else_assign have volatile ops.
21978 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
21981 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
21983 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
21985 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
21987 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
21988 missing ':' in asm example.
21990 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
21992 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
21993 entries into lane and laneq entries.
21994 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
21995 Remove AdvSIMD scalar modes.
21996 (aarch64_sq<r>dmulh_laneq<mode>): New.
21997 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
21999 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
22000 builtin implementations to relfect changes in RTL in aarch64-simd.md.
22001 * config/aarch64/iterators.md (VCOND): New.
22004 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
22007 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
22008 Add NULL LIBNAME argument to existing definitions.
22009 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
22010 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
22011 * config/avr/avr.c (DEF_BUILTIN): Same.
22012 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
22013 (avr_expand_builtin): Expand to a vanilla call if a libgcc
22014 implementation is available (DECL_ASSEMBLER_NAME is set).
22015 (avr_fold_absfx): New static function.
22016 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
22017 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
22018 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
22019 AVR_BUILTIN_ABSLLK.
22020 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
22021 (abshk, absk, abslk, absllk): Provide as static inline functions.
22023 2013-01-25 Marek Polacek <polacek@redhat.com>
22025 PR tree-optimization/56035
22026 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
22028 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
22030 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
22031 (*movtf_internal_rex64): Add (!o,C) alternative
22032 (*movxf_internal_rex64): Ditto.
22033 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
22035 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
22037 * doc/invoke.texi: fix typo.
22038 * doc/objc.texi: fix typo.
22040 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
22042 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
22043 for the first two alternatives.
22045 2013-01-24 Diego Novillo <dnovillo@google.com>
22047 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
22048 (ggc-zone.o): Remove.
22049 * configure.ac: Remove option --with-gc.
22050 * configure: Re-generate.
22051 * doc/install.texi: Remove documentation for --with-gc.
22052 * gengtype.c (write_enum_defn): Remove. Update all users.
22053 (write_Types_process_field): Remove generation of gt_e_* argument.
22054 (output_type_enum): Remove. Update all users.
22055 (write_enum_defn): Remove. Update all users.
22056 (enum alloc_zone): Remove. Update all users.
22057 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
22058 * ggc-common.c (ggc_splay_alloc): Remove first argument.
22059 Update all callers.
22060 (struct ptr_data): Remove field TYPE. Update all users.
22061 (gt_pch_note_object): Remove argument TYPE. Update all users.
22062 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
22064 * ggc-none.c (ggc_alloc_typed_stat): Remove.
22065 (struct alloc_zone): Remove.
22066 (ggc_internal_alloc_zone_stat): Remove.
22067 (ggc_internal_cleared_alloc_zone_stat): Remove.
22068 * ggc-page.c (ggc_alloc_typed_stat): Remove.
22069 (ggc_pch_count_object): Remove last argument. Update all users.
22070 (ggc_pch_alloc_object): Remove last argument. Update all users.
22071 (struct alloc_zone): Remove.
22072 * ggc-zone.c: Remove.
22073 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
22074 (struct alloc_zone): Remove.
22075 (ggc_alloc_typed_stat): Remove.
22076 (ggc_alloc_typed): Remove.
22077 (ggc_splay_alloc): Remove first argument.
22078 (rtl_zone): Remove. Update all users.
22079 (tree_zone): Remove. Update all users.
22080 (tree_id_zone): Remove. Update all users.
22081 (ggc_internal_zone_alloc_stat): Remove. Update all users.
22082 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
22083 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
22084 * tree-ssanames.c: Remove references to zone allocator in comments.
22086 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
22088 * config/avr/avr.c (avr_out_fract): Make register numbers that
22089 might be outside of source operand signed.
22091 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
22093 * config/i386/constraints.md (Yf): New constraint.
22094 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
22095 of f constraint to conditionaly disable x87 register preferences.
22096 (*movdf_internal): Ditto.
22097 (*movsf_internal): Ditto.
22099 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
22101 PR inline-asm/55934
22102 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
22103 that have operands with impossible constraints.
22104 Add a FIXME for a speed-up opportunity.
22105 * lra-constraints.c (process_alt_operands): Verify that a class
22106 selected from constraints on asms is valid for the operand mode.
22107 (curr_insn_transform): Remove incorrect comment.
22109 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
22111 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
22112 TOC operand is a valid symbol ref in the constant pool.
22114 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
22116 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
22118 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
22121 * config/avr/stdfix.h: New file.
22122 * t-avr (stdfix-gcc.h): New rule to build it.
22123 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
22125 2013-01-23 Kostya Serebryany <kcc@google.com>
22127 * config/darwin.h: remove dependency on
22128 CoreFoundation (asan on Mac OS).
22130 2013-01-23 Jakub Jelinek <jakub@redhat.com>
22133 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
22134 instead of cmpdi_operand for first comparison operand.
22135 Don't assert that comparison operands aren't both constants.
22137 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
22139 * doc/install.texi (Downloading the Source): Update references to
22140 downloading separate components.
22142 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
22144 * doc/extend.texi (__int128): Improve grammar.
22146 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
22149 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
22150 alternative to (o,r).
22151 (*movdi_internal_rex64): Remove (!o,n) alternative.
22152 (DImode immediate->memory splitter): Remove.
22153 (DImode immediate->memory peephole2): Remove.
22154 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
22155 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
22156 alternative to (!o,*r).
22157 (*movtf_internal_sse): New pattern.
22158 (*movxf_internal_rex64): New pattern.
22159 (*movxf_internal): Disable for TARGET_64BIT.
22160 (*movdf_internal_rex64): Remove (!o,F) alternative.
22162 2013-01-22 Jakub Jelinek <jakub@redhat.com>
22164 PR middle-end/56074
22165 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
22166 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
22167 * tree-vect-loop-manip.c (find_loop_location): Also ignore
22168 stmt locations where LOCATION_LOCUS of the stmt location is
22169 UNKNOWN_LOCATION or BUILTINS_LOCATION.
22172 * config/i386/i386.md (UNSPEC_STOS): New.
22173 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
22174 *strsetqi_1): Add UNSPEC_STOS.
22176 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
22179 * doc/invoke.texi: Remove left over -Wsynth example.
22181 2013-01-21 Jakub Jelinek <jakub@redhat.com>
22183 PR tree-optimization/56051
22184 * fold-const.c (fold_binary_loc): Don't fold
22185 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
22186 a narrowing conversion, or widening conversion from signed
22189 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
22191 PR rtl-optimization/56023
22192 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
22193 dependent on debug instruction.
22195 2013-01-21 Martin Jambor <mjambor@suse.cz>
22197 PR middle-end/56022
22198 * function.c (allocate_struct_function): Call
22199 invoke_set_current_function_hook earlier.
22201 2013-01-21 Jakub Jelinek <jakub@redhat.com>
22203 * reload1.c (init_reload): Only initialize reload_obstack
22204 during the first call.
22206 2013-01-21 Marek Polacek <polacek@redhat.com>
22208 * cfgloop.c (verify_loop_structure): Fix up grammar.
22210 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
22212 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
22213 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
22215 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
22218 * config/arm/marvell-pj4.md: Update copyright year.
22219 Fix up use of alu to alu_reg and simple_alu_imm.
22221 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
22223 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
22225 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
22228 * lra-constraints.c (curr_insn_transform): Don't reuse original
22229 insn for secondary memory move when memory mode should be different.
22231 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
22233 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
22234 atomic_storedi_1): New patterns.
22236 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
22238 btver2 pipeline descriptions.
22239 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
22241 * config/i386/i386.md (btver2_decode): New type attributes.
22242 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
22244 * config/i386/btver2.md: New file describing btver2 pipelines.
22246 2013-01-19 Andrew Pinski <apinski@cavium.com>
22248 PR tree-optimization/52631
22249 * tree-ssa-sccvn (visit_use): Before looking up the original
22250 statement, try looking up the simplified expression.
22252 2013-01-19 Anthony Green <green@moxielogic.com>
22254 * config/moxie/moxie.c (moxie_expand_prologue): Set
22255 current_function_static_stack_size.
22257 2013-01-18 Jakub Jelinek <jakub@redhat.com>
22259 PR tree-optimization/56029
22260 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
22261 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
22263 2013-01-18 Sharad Singhai <singhai@google.com>
22265 PR tree-optimization/55995
22266 * dumpfile.c (dump_loc): Print location only if available.
22267 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
22269 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
22272 * lra-constraints.c (curr_insn_transform): Reuse original insn for
22273 secondary memory move.
22274 (inherit_reload_reg): Use rclass instead of cl for
22275 check_secondary_memory_needed_p.
22277 2013-01-18 Jakub Jelinek <jakub@redhat.com>
22279 PR middle-end/56015
22280 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
22281 the case where writing real complex part of target modifies op1.
22283 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
22285 * config/aarch64/aarch64-simd.md
22286 (aarch64_vcond_internal<mode>): Handle unordered cases.
22287 * config/aarch64/iterators.md (v_cmp_result): New.
22289 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
22290 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
22292 * config/arm/marvell-pj4.md: New file.
22293 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
22294 * config/arm/arm.md (generic_sched): Add marvell_pj4.
22295 (generic_vfp): Likewise.
22296 * config/arm/arm-cores.def: Add marvell-pj4.
22297 * config/arm/arm-tune.md: Regenerate.
22298 * config/arm/arm-tables.opt: Regenerate.
22299 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
22300 * doc/invoke.texi: Document marvell-pj4.
22302 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
22304 * config/aarch64/arm_neon.h: Map scalar types to standard types.
22306 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
22311 * var-tracking.c (negative_power_of_two_p): New.
22312 (global_get_addr_cache, local_get_addr_cache): New.
22313 (get_addr_from_global_cache, get_addr_from_local_cache): New.
22314 (vt_canonicalize_addr): Rewrite using the above. Adjust the
22316 (vt_stack_offset_p): Remove.
22317 (vt_canon_true_dep): Always canonicalize loc's address.
22318 (clobber_overlapping_mems): Make sure we have a MEM.
22319 (local_get_addr_clear_given_value): New.
22320 (val_reset): Clear local cached entries.
22321 (compute_bb_dataflow): Create and release the local cache.
22322 Disable duplicate MEMs clobbering.
22323 (emit_notes_in_bb): Clobber MEMs likewise.
22324 (vt_emit_notes): Create and release the local cache.
22325 (vt_initialize, vt_finalize): Create and release the global
22326 cache, respectively.
22327 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
22329 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
22331 PR libmudflap/53359
22332 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
22333 not found in the symtab.
22335 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
22338 PR rtl-optimization/55547
22339 PR rtl-optimization/53827
22342 * alias.c (offset_overlap_p): New, factored out of...
22343 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
22344 the conservative special case for symbolic constants. Don't
22345 adjust zero sizes on alignment.
22347 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
22349 PR rtl-optimization/52573
22350 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
22351 REG_UNUSED for the same register.
22353 2013-01-17 Richard Biener <rguenther@suse.de>
22354 Marek Polacek <polacek@redhat.com>
22356 PR rtl-optimization/55833
22357 * loop-unswitch.c (unswitch_loops): Move loop verification...
22358 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
22359 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
22360 Set it to true when we're removing a loop from hierarchy tree in
22361 an irreducible region.
22362 (fix_bb_placements): Adjust caller.
22363 (fix_loop_placements): Likewise.
22365 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
22367 * config/avr/builtins.def (DEF_BUILTIN): Factor out
22368 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
22369 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
22370 Remove ID. Adjust comments.
22371 * config/avr/avr-c.c (avr_builtin_name): Remove.
22372 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
22373 * config/avr/avr.c (avr_tolower): New static function.
22374 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
22375 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
22376 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
22379 2013-01-17 Jan Hubicka <jh@suse.cz>
22381 PR tree-optimization/55273
22382 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
22384 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
22387 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
22388 store through atomic_store<mode>_1.
22389 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
22391 2013-01-17 Martin Jambor <mjambor@suse.cz>
22393 PR tree-optimizations/55264
22394 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
22395 for virtual methods.
22396 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
22397 virtual methods before inlining is over.
22398 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
22400 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
22403 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
22405 PR rtl-optimization/56005
22406 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
22407 pending reads for prefetch.
22409 2013-01-16 Ian Bolton <ian.bolton@arm.com>
22411 * config/aarch64/aarch64.md
22412 (*cstoresi_neg_uxtw): New pattern.
22413 (*cmovsi_insn_uxtw): New pattern.
22414 (*<optab>si3_uxtw): New pattern.
22415 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
22416 (*<optab>si3_insn_uxtw): New pattern.
22417 (*bswapsi2_uxtw): New pattern.
22419 2013-01-16 Richard Biener <rguenther@suse.de>
22421 * tree-inline.c (tree_function_versioning): Remove set but
22422 never used variable.
22424 2013-01-16 Richard Biener <rguenther@suse.de>
22426 PR tree-optimization/55964
22427 * tree-flow.h (rename_variables_in_loop): Remove.
22428 (rename_variables_in_bb): Likewise.
22429 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
22430 (copy_loop_before): Adjust and delete update-ssa status.
22431 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
22432 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
22433 (rename_variables_in_loop): Remove.
22434 (slpeel_update_phis_for_duplicate_loop): Likewise.
22435 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
22436 use available cfg machinery instead of duplicating it.
22437 Update PHI nodes and perform poor-mans SSA update here.
22438 (slpeel_tree_peel_loop_to_edge): Adjust.
22440 2013-01-16 Richard Biener <rguenther@suse.de>
22442 PR tree-optimization/54767
22443 PR tree-optimization/53465
22444 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
22445 (vrp_visit_phi_node): For PHI arguments coming via backedges
22446 drop all symbolical range information.
22447 (execute_vrp): Compute backedges.
22449 2013-01-16 Richard Biener <rguenther@suse.de>
22451 * doc/install.texi: Update CLooG and ISL requirements to
22454 2013-01-16 Christian Bruel <christian.bruel@st.com>
22457 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
22458 (broken_move): Handle UNSPECV_SP_SWITCH_B.
22459 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
22461 2013-01-16 DJ Delorie <dj@redhat.com>
22463 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
22464 (UNSPECV_SP_SWITCH_E): New.
22465 (sp_switch_1): Change to an unspec.
22466 (sp_switch_2): Change to an unspec. Don't use post-inc when we
22469 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
22471 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
22472 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
22473 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
22474 (expand_mem_thread_fence): Ditto.
22475 (expand_mem_signal_fence): Ditto.
22476 (expand_atomic_load): Ditto.
22477 (expand_atomic_store): Ditto.
22479 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
22481 PR rtl-optimization/55547
22482 PR rtl-optimization/53827
22485 * alias.c (memrefs_conflict_p): Set sizes to negative after
22488 2013-01-15 Jakub Jelinek <jakub@redhat.com>
22491 * function.c (thread_prologue_and_epilogue_insns): Always
22492 add crtl->drap_reg to set_up_by_prologue.set, even if
22493 stack_realign_drap is false.
22495 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
22497 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
22498 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
22499 *call): Fix indention.
22501 2013-01-15 Tom de Vries <tom@codesourcery.com>
22504 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
22507 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
22509 PR rtl-optimization/55153
22510 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
22512 2013-01-15 Martin Jambor <mjambor@suse.cz>
22514 PR tree-optimization/55920
22515 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
22516 accesses as grp_to_be_debug_replaced.
22518 2013-01-15 Jakub Jelinek <jakub@redhat.com>
22520 PR tree-optimization/55920
22521 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
22522 there is non-useless type conversion needed from debug rhs to lhs,
22523 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
22525 2013-01-15 Joseph Myers <joseph@codesourcery.com>
22526 Mikael Pettersson <mikpe@it.uu.se>
22529 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
22531 (ASM_OUTPUT_CASE_LABEL): Remove.
22532 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
22533 * final.c (shorten_branches): Update alignment of labels before
22534 jump tables if CASE_VECTOR_SHORTEN_MODE.
22536 2013-01-15 Richard Biener <rguenther@suse.de>
22539 * system.h: Do not include gmp.h for building host tools.
22541 2013-01-15 Richard Biener <rguenther@suse.de>
22543 PR middle-end/55882
22544 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
22545 account for bitpos when computing alignment.
22547 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
22549 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
22550 (ix86_target_macros_internal): Likewise.
22552 * config/i386/i386.c (m_CORE2I7): Removed.
22553 (m_CORE_HASWELL): New macro.
22554 (m_CORE_ALL): Likewise.
22555 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
22556 (initial_ix86_arch_features): Likewise.
22557 (processor_target_table): Initializations for Core avx2.
22558 (cpu_names): New names "core-avx2".
22559 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
22560 PROCESSOR_CORE_HASWELL.
22561 (ix86_issue_rate): New case.
22562 (ia32_multipass_dfa_lookahead): Likewise.
22563 (ix86_sched_init_global): Likewise.
22565 * config/i386/i386.h (TARGET_HASWELL): New macro.
22566 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
22567 (processor_type): New PROCESSOR_HASWELL.
22569 2013-01-15 Jakub Jelinek <jakub@redhat.com>
22571 PR tree-optimization/55955
22572 * tree-vect-loop.c (vectorizable_reduction): Give up early on
22573 *SHIFT_EXPR and *ROTATE_EXPR codes.
22575 PR tree-optimization/48766
22576 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
22577 -ftrapv disable -fwrapv.
22579 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
22582 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
22583 etc. to 1 and not to __flash.
22584 Use LL suffix for __INT24_MAX__ with -mint8.
22585 Use ULL suffix for __UINT24_MAX__ with -mint8.
22587 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
22589 * config/avr/avr-arch.h
22590 (struct base_arch_s): Use typedef avr_arch_t instead.
22591 (struct arch_info_s): Use typedef avr_arch_info_t instead.
22592 (struct mcu_type_s): Use typedef avr_mcu_t instead.
22593 * config/avr/avr.c: Same.
22594 * config/avr/avr-devices.c: Same.
22595 * config/avr/driver-avr.c: Same.
22596 * config/avr/gen-avr-mmcu-texi.c: Same.
22597 * config/avr/avr-mcus.def: Adjust comment.
22599 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
22601 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
22602 * config/aarch64/iterators.md (VALLDI): New.
22604 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
22605 Andi Kleen <ak@linux.intel.com>
22608 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
22609 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
22612 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
22614 * config/avr/avr-stdint.h: Remove trailing blanks.
22615 * config/avr/avr-log.h: Same.
22616 * config/avr/avr-arch.h: Same.
22617 * config/avr/avr-devices.c: Same.
22618 * config/avr/avr-dimode.md: Same.
22619 * config/avr/predicates.md: Same.
22620 * config/avr/avr-c.c: Same. And fix typo.
22622 * config/avr/avr-protos.h: Same. And:
22623 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
22624 (init_cumulative_args): Rename to avr_init_cumulative_args.
22625 (expand_prologue): Rename to avr_expand_prologue.
22626 (expand_epilogue): Rename to avr_expand_epilogue.
22627 (adjust_insn_length): Rename to avr_adjust_insn_length.
22628 (notice_update_cc): Rename to avr_notice_update_cc.
22629 (final_prescan_insn): Rename to avr_final_prescan_insn.
22630 * config/avr/avr.c: Same.
22631 * config/avr/avr.h: Same.
22632 * config/avr/avr.md: Remove trailing blanks.
22633 (prologue): Use avr_expand_prologue.
22634 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
22636 2013-01-14 Richard Biener <rguenther@suse.de>
22638 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
22639 verify_location, collect_subblocks): New functions.
22640 (verify_gimple_in_cfg): Verify that locations only reference
22641 BLOCKs in the functions BLOCK tree.
22643 2013-01-14 Richard Biener <rguenther@suse.de>
22645 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
22647 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
22649 (insert_out_of_ssa_copy_on_edge): Likewise.
22650 (rewrite_close_phi_out_of_ssa): Likewise.
22651 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
22653 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
22654 propagated constants.
22655 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
22658 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
22660 * config/avr/avr-modes.def: Add GPL copyright notice.
22662 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
22664 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
22665 MEMMODEL_MASK to determine memory model.
22666 (atomic_store<mode>): Ditto from operands[2].
22667 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
22669 2013-01-13 Jakub Jelinek <jakub@redhat.com>
22672 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
22673 (fold_gimple_assign): Don't call unshare_expr here.
22674 (fold_ctor_reference): Call unshare_expr.
22676 2013-01-13 Terry Guo <terry.guo@arm.com>
22678 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
22679 * doc/fragments.texi: Document MULTILIB_REUSE.
22680 * gcc.c (multilib_reuse): New internal spec.
22681 (set_multilib_dir): Also search multilib from multilib_reuse.
22682 * genmultilib (tmpmultilib3): Refactor code.
22683 (tmpmultilib4): Ditto.
22684 (multilib_reuse): New multilib argument.
22686 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
22688 * Makefile.in: Update copyright.
22690 2013-01-12 Tom de Vries <tom@codesourcery.com>
22692 PR middle-end/55890
22693 * calls.c (expand_call): Check if arg_nr is valid.
22695 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
22697 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
22698 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
22699 documentation. Add missing '__' in front of
22700 __builtin_ia32_packssdw256.
22702 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
22705 * config/s390/s390.c (s390_preferred_reload_class): Do not return
22706 NO_REGS for larl operands.
22707 (s390_reload_larl_operand): Use s390_load_address instead of
22710 2013-01-11 Richard Biener <rguenther@suse.de>
22712 * tree-cfg.c (verify_node_sharing_1): Split out from ...
22713 (verify_node_sharing): ... here.
22714 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
22716 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
22718 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
22719 Substitute TREECHECKING.
22720 * configure: Regenerate.
22721 * Makefile.in (TREECHECKING): New.
22723 2013-01-11 Richard Guenther <rguenther@suse.de>
22725 PR tree-optimization/44061
22726 * tree-vrp.c (extract_range_basic): Compute zero as
22727 value-range for __builtin_constant_p of function parameters.
22729 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
22731 Update copyright years.
22733 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
22735 PR rtl-optimization/55672
22736 * lra-eliminations.c (mark_not_eliminable): Permit addition with
22737 const to be eliminable.
22739 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
22741 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
22742 * configure: Regenerate.
22744 2013-01-10 Richard Biener <rguenther@suse.de>
22746 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
22748 2013-01-10 Richard Biener <rguenther@suse.de>
22751 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
22752 locations for virtual PHI arguments.
22753 (rewrite_update_phi_arguments): Likewise.
22755 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
22757 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
22760 2013-01-10 Jakub Jelinek <jakub@redhat.com>
22762 PR tree-optimization/55921
22763 * tree-complex.c (expand_complex_asm): New function.
22764 (expand_complex_operations_1): Call it for GIMPLE_ASM.
22766 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
22769 * config/s390/s390.c (s390_symref_operand_p)
22770 (s390_loadrelative_operand_p): Merge the two functions.
22771 (s390_check_qrst_address, print_operand_address): Add parameters
22772 to s390_loadrelative_operand_p invokation.
22773 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
22774 (s390_reload_larl_operand, s390_secondary_reload): Use
22775 s390_loadrelative_operand_p instead of s390_symref_operand_p.
22776 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
22778 2013-01-09 Mike Stump <mikestump@comcast.net>
22780 * dse.c (record_store): Remove unnecessary assert.
22782 2013-01-09 Jan Hubicka <jh@suse.cz>
22784 PR tree-optimization/55569
22785 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
22786 * cfgloop.h (scale_loop_profile): Likewise.
22788 2013-01-09 Jan Hubicka <jh@suse.cz>
22791 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
22793 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
22795 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
22797 PR middle-end/55114
22798 * expr.h (maybe_emit_group_store): Declare.
22799 * expr.c (maybe_emit_group_store): New function.
22800 * builtins.c (expand_builtin_int_roundingfn): Call it.
22801 (expand_builtin_int_roundingfn_2): Likewise.
22803 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
22805 PR rtl-optimization/55829
22806 * lra-constraints.c (match_reload): Add code for absent output.
22807 (curr_insn_transform): Add code for reloads of matched inputs
22810 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
22812 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
22813 attribute of movddup insn to DF.
22814 (*vec_interleave_lowv2df): Ditto.
22815 (vec_dupv2df): Ditto.
22817 2013-01-09 Jan Hubicka <jh@suse.cz>
22819 PR tree-optimiation/55875
22820 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
22821 EVERY_ITERATION parameter.
22822 (number_of_iterations_exit): Check if exit is executed every iteration.
22823 (idx_infer_loop_bounds): Similarly here.
22824 (n_of_executions_at_most): Simplify
22825 to only test for cases where statement is dominated by the
22826 particular bound; handle correctly the "postdominance" test.
22827 (scev_probably_wraps_p): Use max loop iterations info
22828 as a global bound first.
22830 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
22831 Nick Clifton <nickc@redhat.com>
22833 * config/v850/v850.md (cbranchsf4): New pattern.
22834 (cstoresf4): New pattern.
22835 (cbranchdf4): New pattern.
22836 (cstoredf4): New pattern.
22837 (movsicc): Disallow floating point comparisons.
22838 (cmpsf_le_insn): Fix order of operators.
22839 (cmpsf_lt_insn): Likewise.
22840 (cmpsf_eq_insn): Likewise.
22841 (cmpdf_le_insn): Likewise.
22842 (cmpdf_lt_insn): Likewise.
22843 (cmpdf_eq_insn): Likewise.
22844 (cmpsf_ge_insn): Use LE comparison.
22845 (cmpdf_ge_insn): Likewise.
22846 (cmpsf_gt_insn): Use LT comparison.
22847 (cmpdf_gt_insn): Likewise.
22848 (cmpsf_ne_insn): Delete pattern.
22849 (cmpdf_ne_insn): Delete pattern.
22850 * config/v850/v850.c (v850_gen_float_compare): Use
22851 gen_cmpdf_eq_insn for NE comparison.
22852 (v850_float_z_comparison_operator)
22853 (v850_float_nz_comparison_operator): Move from here ...
22854 * config/v850/predicates.md: ... to here. Move GT and GE
22855 comparisons into v850_float_z_comparison_operator.
22856 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
22858 (v850_float_nz_comparison_operator): Likewise.
22860 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
22862 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
22863 with calls to gen_insvsi/gen_insvdi.
22865 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
22867 * config/i386/i386.c (initial_ix86_tune_features): Set up
22868 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
22870 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
22871 Jakub Jelinek <jakub@redhat.com>
22873 PR tree-optimization/48189
22874 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
22875 If nitercst is 0, don't predict the exit edge.
22877 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22879 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
22880 in asm_fprintf with reg_names.
22881 (aarch64_print_operand_address): Likewise.
22882 (aarch64_return_addr): Likewise.
22883 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
22885 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
22887 * config/pa/pa.h (VAL_U6_BITS_P): Define.
22888 (INT_U6_BITS): Likewise.
22889 * config/pa/predicates.md (uint6_operand): New predicate.
22890 (shift5_operand, shift6_operand): Likewise.
22891 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
22893 (lshrdi3): Use shift6_operand.
22894 (shrpsi4, shrpdi4): New insn patterns.
22895 (extzv): Delete expander.
22896 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
22897 predicates in unamed zero extract patterns. Tighten common constraint.
22898 (extv): Delete expander.
22899 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
22900 predicates in unamed sign extract patterns. Tighten common constraint.
22901 (insv): Delete expander.
22902 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
22903 predicates in unamed insert patterns. Tighten common constraint.
22904 Change uint32_operand predicate to uint6_operand predicate in unamed
22905 DImode pattern to insert constant values of type 1...1xxxx.
22907 2013-01-04 Jan Hubicka <jh@suse.cz>
22909 PR tree-optimization/55823
22910 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
22913 2013-01-08 Jakub Jelinek <jakub@redhat.com>
22914 Uros Bizjak <ubizjak@gmail.com>
22916 PR rtl-optimization/55845
22917 * df-problems.c (can_move_insns_across): Stop scanning at
22918 volatile_insn_p source instruction or give up if
22919 across_from .. across_to range contains any volatile_insn_p
22922 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
22924 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
22925 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
22927 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
22928 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
22930 2013-01-08 Jakub Jelinek <jakub@redhat.com>
22933 * asan.c (asan_clear_shadow): New function.
22934 (asan_emit_stack_protection): Use it.
22936 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
22938 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
22939 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
22940 with tab instead of space.
22942 2013-01-08 Nick Clifton <nickc@redhat.com>
22944 * config/rl78/rl78.c (rl78_expand_prologue): Always select
22945 register bank 0 at the start of an interrupt handler.
22946 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
22949 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
22951 * config/aarch64/aarch64-simd.md
22952 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
22953 (aarch64_simd_bsl): Likewise.
22954 (aarch64_vcond_internal<mode>): Likewise.
22955 (vcond<mode><mode>): Likewise.
22956 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
22957 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
22959 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
22961 * config/aarch64/aarch64-builtins.c
22962 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
22964 2013-01-08 Martin Jambor <mjambor@suse.cz>
22967 * tree-sra.c (analyze_access_subtree): Return true also after
22968 potentially creating a debug-only replacement.
22970 2013-01-08 Jakub Jelinek <jakub@redhat.com>
22972 PR middle-end/55890
22973 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
22975 PR tree-optimization/54120
22976 * tree-vrp.c (range_fits_type_p): Don't allow
22977 src_precision < precision from signed vr to unsigned_p
22978 if vr->min or vr->max is negative.
22979 (simplify_float_conversion_using_ranges): Test can_float_p
22980 against CODE_FOR_nothing.
22982 2013-01-08 Jakub Jelinek <jakub@redhat.com>
22983 Richard Biener <rguenther@suse.de>
22985 PR middle-end/55851
22986 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
22987 types instead of just INTEGER_TYPE types.
22989 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
22991 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
22994 2013-01-07 Steve Ellcey <sellcey@mips.com>
22997 * config/mips/mips.opt: Change mad to mmad to match documentation.
22999 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
23002 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
23003 .progmemx.data now.
23005 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
23008 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
23009 (avr_addrspace_t): Add .section_name field.
23010 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
23012 (avr_addrspace): Same. Initialize .section_name. Remove last
23013 NULL entry. Put __memx into .progmemx.data.
23014 (progmem_section_prefix): Remove.
23015 (avr_asm_init_sections): No need to initialize progmem_section.
23016 (avr_asm_named_section): Use avr_addrspace[].section_name to get
23017 section name prefix.
23018 (avr_asm_select_section): Ditto. And use get_unnamed_section to
23019 retrieve the progmem section.
23020 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
23021 boundary to run over avr_addrspace[].
23022 (avr_register_target_pragmas): Ditto.
23024 2013-01-06 Jakub Jelinek <jakub@redhat.com>
23026 * varasm.c (output_constant_def_contents): For asan_protect_global
23027 protected strings, adjust DECL_ALIGN if needed, before testing for
23029 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
23030 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
23032 (output_object_block): For asan protected decls, emit asan padding
23033 after their contents.
23034 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
23035 (asan_finish_file): Test it here instead.
23037 2013-01-07 Nick Clifton <nickc@redhat.com>
23038 Matthias Klose <doko@debian.org>
23039 Doug Kwan <dougkwan@google.com>
23040 H.J. Lu <hongjiu.lu@intel.com>
23043 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
23045 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
23047 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
23049 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
23051 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
23053 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
23056 * doc/install.texi (Cross-Compiler-Specific Options): Document
23059 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
23061 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
23062 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
23063 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
23064 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
23065 vqmovun_high_s64): Fix source operand number and update copyright.
23067 2013-01-07 Richard Biener <rguenther@suse.de>
23069 PR middle-end/55890
23070 * gimple.h (gimple_call_builtin_p): New overload.
23071 * gimple.c (validate_call): New function.
23072 (gimple_call_builtin_p): Likewise.
23073 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
23074 Use gimple_call_builtin_p.
23075 (find_func_clobbers): Likewise.
23076 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
23077 (strlen_optimize_stmt): Likewise.
23079 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
23081 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
23082 (vld1q_dup_*): Likewise.
23083 (vld1_*): Likewise.
23084 (vld1q_*): Likewise.
23085 (vld1_lane_*): Likewise.
23086 (vld1q_lane_*): Likewise.
23088 2013-01-07 Richard Biener <rguenther@suse.de>
23090 * lto-streamer.h (LTO_minor_version): Bump to 2.
23092 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
23094 * config/aarch64/aarch64-protos.h
23095 (aarch64_const_double_zero_rtx_p): Rename to...
23096 (aarch64_float_const_zero_rtx_p): ...this.
23097 (aarch64_float_const_representable_p): New.
23098 (aarch64_output_simd_mov_immediate): Likewise.
23099 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
23100 move immediate case.
23101 * config/aarch64/aarch64.c
23102 (aarch64_const_double_zero_rtx_p): Rename to...
23103 (aarch64_float_const_zero_rtx_p): ...this.
23104 (aarch64_print_operand): Allow printing of new constants.
23105 (aarch64_valid_floating_const): New.
23106 (aarch64_legitimate_constant_p): Check for valid floating-point
23108 (aarch64_simd_valid_immediate): Likewise.
23109 (aarch64_vect_float_const_representable_p): New.
23110 (aarch64_float_const_representable_p): Likewise.
23111 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
23112 (aarch64_output_simd_mov_immediate): New.
23113 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
23114 (*movdf_aarch64): Likewise.
23115 * config/aarch64/constraints.md (Ufc): New.
23116 (Y): call aarch64_float_const_zero_rtx.
23117 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
23119 2013-01-07 Richard Biener <rguenther@suse.de>
23121 PR tree-optimization/55888
23122 PR tree-optimization/55862
23123 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
23124 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
23125 not if it is contained therein.
23127 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
23129 * config/avr/t-avr: Typo.
23131 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
23134 * config/avr/t-avr: Don't automatically rebuild
23135 $(srcdir)/config/avr/t-multilib
23136 $(srcdir)/config/avr/avr-tables.opt
23137 $(srcdir)/doc/avr-mmcu.texi
23138 (avr-mcus): New phony target to build them on request.
23139 (s-avr-mlib, s-avr-mmcu-texi): Remove.
23140 * avr/avr-mcus.def: Adjust comments.
23142 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
23144 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
23146 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
23148 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
23150 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
23152 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
23154 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
23156 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
23157 to generate profiling.
23158 * config/rs6000/aix64.h (LIB_SPEC): Same.
23160 2013-01-04 Andrew Pinski <apinski@cavium.com>
23162 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
23164 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
23166 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
23168 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
23170 (ix86_expand_move): Ditto.
23171 (ix86_zero_extend_to_Pmode): Ditto.
23172 (ix86_expand_call): Ditto.
23173 (ix86_expand_special_args_builtin): Ditto.
23174 (ix86_expand_builtin): Ditto.
23176 2013-01-04 Richard Biener <rguenther@suse.de>
23178 PR tree-optimization/55862
23179 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
23180 translating them through PHI nodes.
23182 2013-01-04 Martin Jambor <mjambor@suse.cz>
23184 PR tree-optimization/55755
23185 * tree-sra.c (sra_modify_assign): Do not check that an access has no
23186 children when trying to avoid producing a VIEW_CONVERT_EXPR.
23188 2013-01-04 Marek Polacek <polacek@redhat.com>
23190 PR middle-end/55859
23191 * opts.c (default_options_optimization): Clarify error message.
23193 2013-01-04 Richard Biener <rguenther@suse.de>
23195 PR middle-end/55863
23196 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
23199 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23202 * config/pa/pa.md (movsi): Revert previous change.
23203 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
23206 2013-01-03 Richard Henderson <rth@redhat.com>
23208 * config/i386/i386.c (ix86_expand_move): Always assign to op1
23209 after eliminating TLS symbols.
23211 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
23214 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
23215 * graphite-poly.c (debug_gmp_value): Likewise.
23217 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
23220 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
23221 selected code model, define __code_mode_small__, __code_model_medium__,
23222 __code_model_large__, __code_model_32__ or __code_model_kernel__.
23223 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
23224 xchg temporary register with %k. Declare temporary register as
23226 [__x86_64__]: For medium and large code models, preserve %rbx register.
23228 2013-01-03 Richard Biener <rguenther@suse.de>
23230 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
23231 (dump_subscript): Adjust.
23232 (finalize_ddr_dependent): Do not dump redundant info.
23233 (analyze_siv_subscript): Adjust.
23234 (subscript_dependence_tester): Likewise.
23235 (compute_affine_dependence): Likewise.
23237 2013-01-03 Richard Biener <rguenther@suse.de>
23240 2013-01-03 Richard Biener <rguenther@suse.de>
23242 PR tree-optimization/55857
23243 * tree-vect-stmts.c (vectorizable_load): Do not setup
23244 re-alignment for invariant loads.
23246 2013-01-02 Richard Biener <rguenther@suse.de>
23248 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
23249 invariant load do not generate a vector load from the scalar location.
23251 2013-01-03 Richard Biener <rguenther@suse.de>
23253 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
23254 for not vectorizing.
23255 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
23256 not build INDIRECT_REFs, call get_name once only.
23257 (vect_create_data_ref_ptr): Likewise. Dump base object kind
23258 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
23260 2013-01-03 Richard Biener <rguenther@suse.de>
23262 PR tree-optimization/55857
23263 * tree-vect-stmts.c (vectorizable_load): Do not setup
23264 re-alignment for invariant loads.
23266 2013-01-03 Richard Biener <rguenther@suse.de>
23269 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
23270 prefer a built-in decl.
23272 2013-01-03 Jakub Jelinek <jakub@redhat.com>
23274 * gcc.c (process_command): Update copyright notice dates.
23275 * gcov.c (print_version): Likewise.
23276 * gcov-dump.c (print_version): Likewise.
23278 PR rtl-optimization/55838
23279 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
23280 iv0.step, iv1.step and step.
23282 2013-01-03 Jakub Jelinek <jakub@redhat.com>
23283 Marc Glisse <marc.glisse@inria.fr>
23285 PR tree-optimization/55832
23286 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
23287 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
23288 integer_{one,zero}_node.
23290 2013-01-03 Jakub Jelinek <jakub@redhat.com>
23293 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
23294 * var-tracking.c (reverse_op): Don't add reverse ops to
23295 VALUEs that have already
23296 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
23298 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
23300 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
23302 2013-01-02 Teresa Johnson <tejohnson@google.com>
23304 * dumpfile.c (dump_loc): Print filename with location.
23305 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
23306 new location_t parameter to emit complete unroll message with
23307 new dump framework.
23308 (canonicalize_loop_induction_variables): Compute loops location
23309 and pass to try_unroll_loop_completely.
23310 * loop-unroll.c (report_unroll_peel): New function.
23311 (peel_loops_completely): Use new dump format with location
23312 for main dumpfile message, and invoke report_unroll_peel on success.
23313 (decide_unrolling_and_peeling): Ditto.
23314 (decide_peel_once_rolling): Remove old dumpfile message subsumed
23315 by report_unroll_peel.
23316 (decide_peel_completely): Ditto.
23317 (decide_unroll_constant_iterations): Ditto.
23318 (decide_unroll_runtime_iterations): Ditto.
23319 (decide_peel_simple): Ditto.
23320 (decide_unroll_stupid): Ditto.
23321 * cfgloop.c (get_loop_location): New function.
23322 * cfgloop.h (get_loop_location): Declare.
23324 2013-01-02 Sriraman Tallam <tmsriram@google.com>
23326 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
23329 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23331 PR middle-end/55198
23332 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
23333 BLKmode objects when EXPAND_MEMORY is specified.
23335 2013-01-02 Sriraman Tallam <tmsriram@google.com>
23337 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
23339 (fold_builtin_cpu): Do not share cpu model decls across statements.
23341 2013-01-02 Jason Merrill <jason@redhat.com>
23344 * tree.c (build_array_type_1): Revert earlier change.
23346 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
23348 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
23350 * config/aarch64/aarch64-tune.md: Re-generate.
23352 2013-01-02 Richard Biener <rguenther@suse.de>
23354 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
23355 invariant load do not generate a vector load from the scalar location.
23357 2013-01-02 Richard Biener <rguenther@suse.de>
23360 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
23361 * configure: Regenerate.
23363 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
23365 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
23366 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
23367 (expand_builtin_int_roundingfn_2): Keep the original target around
23368 for the fallback case.
23370 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
23372 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
23373 to be clear for sign changes.
23375 2013-01-01 Jan Hubicka <jh@suse.cz>
23377 * ipa-inline-analysis.c: Fix formatting.
23379 2013-01-01 Jakub Jelinek <jakub@redhat.com>
23381 PR tree-optimization/55831
23382 * tree-vect-loop.c (get_initial_def_for_induction): Use
23383 gsi_after_labels instead of gsi_start_bb.
23385 Copyright (C) 2013 Free Software Foundation, Inc.
23387 Copying and distribution of this file, with or without modification,
23388 are permitted in any medium without royalty provided the copyright
23389 notice and this notice are preserved.