1 2013-09-26 Nick Clifton <nickc@redhat.com>
3 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
5 (msp430_print_operand_raw): Delete unused letter parameter.
6 (TARGET_PRINT_OPERAND_ADDRESS): Define.
7 (msp430_print_operand_address): New function.
8 (msp430_print_operand): Move address printing code from here to
10 * config/msp430/msp430.md (movsipsi2): Add comment in generated
12 (zero_extendpsisi2): Likewise.
13 (extendpsisi2): New pattern.
14 (andneghi3): New pattern.
16 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
18 * config/aarch64/aarch64.opt (mlra): New option.
19 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
20 (TARGET_LRA_P): Define.
22 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
24 * expr.c (expand_assignment): Remove obsolete comment.
26 2013-09-25 Jeff Law <law@redhat.com>
28 * tree-flow.h (thread_through_all_blocks): Prototype moved into
29 tree-ssa-threadupdate.h.
30 (register_jump_thread): Similarly.
31 * tree-ssa-threadupdate.h: New header file.
32 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
33 * tree-vrp.c: Likewise.
34 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
35 (thread_around_empty_blocks): Change type of path vector argument to
36 an edge,type pair from just an edge. Initialize both elements when
37 appending to a jump threading path. Tweak references to elements
39 (thread_across_edge): Similarly. Release memory for the elements
41 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
42 (dump_jump_thread_path): New function broken out from
44 (register_jump_thread): Use dump_jump_thread_path. Change type of
45 path vector entries. Search the path for NULL edges and dump
46 the path if one is found. Tweak the conversion of path to 3-edge
47 form to use the block copy type information embedded in the path.
49 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
51 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
53 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
54 Vladimir Makarov <vmakarov@redhat.com>
56 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
57 from the least significant bit.
58 (strip_address_mutations): Add bitfield operations handling.
59 (must_be_index_p): Add shifting and rotate operations handling.
60 (set_address_base): Use must_be_base_p predicate.
61 (set_address_index): Use must_be_index_p predicate.
63 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
64 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
65 Sergey Lega <sergey.s.lega@intel.com>
66 Anna Tikhonova <anna.tikhonova@intel.com>
67 Ilya Tocar <ilya.tocar@intel.com>
68 Andrey Turetskiy <andrey.turetskiy@intel.com>
69 Ilya Verbin <ilya.verbin@intel.com>
70 Kirill Yukhin <kirill.yukhin@intel.com>
71 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
73 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
75 (ix86_expand_vector_move_misalign): Support new unaligned load and
76 stores and use new names.
77 (CODE_FOR_sse2_storedqu): Rename to ...
78 (CODE_FOR_sse2_storedquv16qi): ... this.
79 (CODE_FOR_sse2_loaddqu): Rename to ...
80 (CODE_FOR_sse2_loaddquv16qi): ... this.
81 (CODE_FOR_avx_loaddqu256): Rename to ...
82 (CODE_FOR_avx_loaddquv32qi): ... this.
83 (CODE_FOR_avx_storedqu256): Rename to ...
84 (CODE_FOR_avx_storedquv32qi): ... this.
85 * config/i386/i386.md (fpint_logic): New.
86 * config/i386/sse.md (VMOVE): Extend for AVX512.
90 (VI_UNALIGNED_LOADSTORE): Ditto.
91 (sse2_avx_avx512f): Ditto.
92 (sse2_avx2): Extend for AVX512.
95 (sse): Extend for AVX512.
98 (avxsizesuffix): Ditto.
99 (sseintvecmode): Ditto.
101 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
102 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
103 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
104 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
105 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
106 (<sse2_avx_avx512f>_storedqu<mode): ... this.
107 (<sse>_movnt<mode>): Replace constraint "x" with "v".
108 (STORENT_MODE): Extend for AVX512.
109 (*absneg<mode>2): Replace constraint "x" with "v".
110 (*mul<mode>3): Ditto.
111 (*ieee_smin<mode>3): Ditto.
112 (*ieee_smax<mode>3): Ditto.
113 (avx_cmp<mode>3): Replace VF with VF_128_256.
114 (*<sse>_maskcmp<mode>3_comm): Ditto.
115 (<sse>_maskcmp<mode>3): Ditto.
116 (<sse>_andnot<mode>3): Extend for AVX512.
117 (<code><mode>3, anylogic): Replace VF with VF_128_256.
118 (<code><mode>3, fpint_logic): New.
119 (*<code><mode>3): Extend for AVX512.
120 (avx512flogicsuff): New.
121 (avx512f_<logic><mode>): Ditto.
122 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
124 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
125 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
126 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
127 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
128 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
129 (xop_vpermil2<mode>3): Ditto.
130 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
131 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
132 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
133 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
135 2013-09-25 Tom Tromey <tromey@redhat.com>
137 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
138 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
139 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
140 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
141 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
142 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
143 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
144 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
145 (GRAPHITE_HTAB_H): Remove.
147 2013-09-25 Tom Tromey <tromey@redhat.com>
149 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
151 2013-09-25 Tom Tromey <tromey@redhat.com>
153 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
155 2013-09-25 Tom Tromey <tromey@redhat.com>
157 * config/i386/t-i386 (i386.o): Remove.
158 (i386-c.o): Use COMPILE and POSTCOMPILE.
160 2013-09-25 Tom Tromey <tromey@redhat.com>
162 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
164 2013-09-25 Tom Tromey <tromey@redhat.com>
166 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
167 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
168 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
169 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
170 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
171 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
172 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
173 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
174 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
175 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
176 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
177 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
178 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
179 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
180 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
181 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
182 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
183 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
184 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
185 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
186 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
187 (tree-ssa-pre.o, tree-ssa-sccvn.o)
188 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
189 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
190 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
191 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
192 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
193 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
194 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
195 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
196 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
197 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
198 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
199 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
200 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
201 (graphite-clast-to-gimple.o, graphite-dependences.o)
202 (graphite-interchange.o, graphite-poly.o)
203 (graphite-scop-detection.o, graphite-sese-to-poly.o)
204 (graphite-optimize-isl.o, tree-vect-loop.o)
205 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
206 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
207 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
208 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
209 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
210 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
211 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
212 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
213 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
214 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
215 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
216 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
217 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
218 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
219 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
220 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
221 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
222 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
223 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
224 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
225 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
226 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
227 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
228 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
229 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
230 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
231 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
232 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
233 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
234 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
235 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
236 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
237 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
238 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
239 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
240 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
241 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
242 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
243 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
244 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
245 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
246 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
247 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
248 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
249 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
250 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
251 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
252 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
253 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
254 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
255 (gcov-dump.o): Remove.
256 (default-c.o): Use COMPILE and POSTCOMPILE.
257 (CFLAGS-gcc.o): New variable.
258 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
260 2013-09-25 Tom Tromey <tromey@redhat.com>
262 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
263 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
264 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
265 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
266 (c-family/c-pragma.o, c-family/c-pretty-print.o)
267 (c-family/c-semantics.o, c-family/c-ada-spec.o)
268 (c-family/array-notation-common.o, c-family/stub-objc.o)
269 (c-family/c-ubsan.o): Remove.
271 2013-09-25 Tom Tromey <tromey@redhat.com>
273 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
275 2013-09-25 Tom Tromey <tromey@redhat.com>
277 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
278 to add -DENABLE_SHARED_LIBGCC.
279 (gcc.o): Don't use subshell.
281 2013-09-25 Tom Tromey <tromey@redhat.com>
283 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
284 * configure.ac: Don't invoke AM_PROG_CC_C_O.
285 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
286 * configure, config.in: Rebuild.
288 2013-09-25 Tom Tromey <tromey@redhat.com>
290 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
291 (COMPILE, POSTCOMPILE): New variables.
292 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
293 (DEPFILES): New variable.
295 * configure.ac: Add checks for dependency checking.
296 * configure, aclocal.m4: Regenerate.
298 2013-09-25 Tom Tromey <tromey@redhat.com>
300 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
301 ($(ALL_HOST_OBJS)): Move order-only dependency to end
304 2013-09-25 Tom Tromey <tromey@redhat.com>
306 * Makefile.in (generated_files): Add options.h,
307 target-hooks-def.h, insn-opinit.h,
308 common/common-target-hooks-def.h, pass-instances.def,
309 c-family/c-target-hooks-def.h.
311 2013-09-25 Jeff Law <law@redhat.com>
313 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
314 than foo[foo.length () - 1] to access last member in a vec.
315 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
317 2013-09-25 Richard Biener <rguenther@suse.de>
320 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
322 2013-09-25 Jan Hubicka <jh@suse.cz>
324 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
327 2013-09-25 Marek Polacek <polacek@redhat.com>
330 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
331 when determining the type name.
333 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
335 * config/sh/sh.md: Fix formatting.
337 2013-09-24 Xinliang David Li <davidxl@google.com>
339 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
340 Check max peel iterations parameter.
341 * param.def: New parameter.
342 * doc/invoke.texi: Document New parameter.
344 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
346 * gimple-pretty-print.c: Various whitespace tweaks.
347 * tree-core.h: Likewise.
348 * tree-pretty-print.c: Likewise.
349 * tree-ssa-alias.c: Likewise.
350 * tree-ssa-copy.c: Likewise.
351 * tree-ssanames.c: Likewise.
352 * tree-ssanames.h: Likewise.
353 * tree-vrp.c: Likewise.
355 2013-09-24 Alan Modra <amodra@gmail.com>
359 * stmt.c (expand_asm_operands): Call expand_expr with
360 EXPAND_MEMORY for output operands that disallow regs. Don't
361 use EXPAND_WRITE on inout operands.
363 2013-09-24 Richard Biener <rguenther@suse.de>
366 * tree.c (reference_alias_ptr_type): Move ...
367 * alias.c (reference_alias_ptr_type): ... here and implement
368 in terms of the new reference_alias_ptr_type_1.
369 (ref_all_alias_ptr_type_p): New helper.
370 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
371 use ref_all_alias_ptr_type_p.
372 (get_deref_alias_set): Add flag_strict_aliasing check here.
373 (reference_alias_ptr_type_1): New function, split out from ...
374 (get_alias_set): ... here.
375 (alias_ptr_types_compatible_p): New function.
376 * alias.h (reference_alias_ptr_type): Declare.
377 (alias_ptr_types_compatible_p): Likewise.
378 * tree.h (reference_alias_ptr_type): Remove.
379 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
380 to compare MEM_REF alias types.
382 2013-09-24 Richard Biener <rguenther@suse.de>
384 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
386 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
388 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
389 reload helper function arrays into a single array reg_addr.
390 (reload_fpr_gpr): Likewise.
391 (reload_gpr_vsx): Likewise.
392 (reload_vsx_gpr): Likewise.
393 (struct rs6000_reg_addr): Likewise.
394 (reg_addr): Likewise.
395 (rs6000_debug_reg_global): Change rs6000_vector_reload,
396 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
397 (rs6000_init_hard_regno_mode_ok): Likewise.
398 (rs6000_secondary_reload_direct_move): Likewise.
399 (rs6000_secondary_reload): Likewise.
401 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
402 constraints: wu, ww, and wy. Repurpose wv constraint added during
403 power8 changes. Put wg constraint in alphabetical order.
405 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
406 for future work to add ISA 2.07 VSX single precision support.
407 (-mvsx-scalar-double): Change default from -1 to 1, update
408 documentation comment.
409 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
410 (-mupper-regs-df): New debug switch to control whether DF values
411 can go in the traditional Altivec registers.
412 (-mupper-regs-sf): New debug switch to control whether SF values
413 can go in the traditional Altivec registers.
415 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
417 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
418 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
419 Add new constraints, wu/ww/wy. Repurpose wv constraint.
420 (rs6000_debug_legitimate_address_p): Print if we are running
421 before, during, or after reload.
422 (rs6000_secondary_reload): Add a comment.
423 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
425 * config/rs6000/constraints.md (wa constraint): Sort w<x>
426 constraints. Update documentation string.
427 (wd constraint): Likewise.
428 (wf constraint): Likewise.
429 (wg constraint): Likewise.
430 (wn constraint): Likewise.
431 (ws constraint): Likewise.
432 (wt constraint): Likewise.
433 (wx constraint): Likewise.
434 (wz constraint): Likewise.
435 (wu constraint): New constraint for ISA 2.07 SFmode scalar
437 (ww constraint): Likewise.
438 (wy constraint): Likewise.
439 (wv constraint): Repurpose ISA 2.07 constraint that we not used in
440 the previous submissions.
441 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
443 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
445 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
446 (REG_BR_PROB): Say that the probability is stored in an int_list.
447 * reg-notes.def: Update commentary to mention INT_LIST.
448 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
450 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
451 * rtlanal.c (int_reg_note_p): New function.
452 (alloc_reg_note): Assert that the note does not have an int argument.
453 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
454 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
455 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
456 rather than an INSN_LIST. Handle INT_LIST.
457 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
458 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
459 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
460 Manipulate them as ints rather than rtxes.
461 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
462 * regmove.c (copy_src_to_dest): Likewise.
463 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
465 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
466 into the cases that need it.
467 * config/arm/arm.c (arm_unwind_emit): Likewise.
469 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
470 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
471 * loop-doloop.c (add_test, doloop_modify): Likewise.
472 * loop-unswitch.c (compare_and_jump_seq): Likewise.
473 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
474 * predict.c (combine_predictions_for_insn): Likewise.
475 * print-rtl.c (print_rtx): Handle INT_LIST.
476 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
477 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
478 * config/arm/arm.c (emit_unlikely_jump): Likewise.
479 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
480 (ix86_split_fp_branch, predict_jump): Likewise.
481 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
482 * config/sh/sh.c (expand_cbranchsi4): Likewise.
483 * config/spu/spu.c (ea_load_store_inline): Likewise.
485 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
486 value of a REG_BR_PROB note.
487 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
488 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
489 * emit-rtl.c (try_split): Likewise.
490 * predict.c (br_prob_note_reliable_p): Likewise.
491 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
492 * reorg.c (mostly_true_jump): Likewise.
493 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
494 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
495 * config/i386/i386.c (ix86_print_operand): Likewise.
496 * config/ia64/ia64.c (ia64_print_operand): Likewise.
497 * config/mmix/mmix.c (mmix_print_operand): Likewise.
498 * config/rs6000/rs6000.c (output_cbranch): Likewise.
499 * config/s390/s390.c (s390_expand_tbegin): Likewise.
500 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
501 * config/sparc/sparc.c (output_cbranch): Likewise.
502 * config/spu/spu.c (get_branch_target): Likewise.
503 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
504 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
506 2013-09-23 Jan Hubicka <jh@suse.cz>
508 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
510 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
511 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
513 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
514 (gimple_fold_call): Dump inconsistent devirtualizations; add
515 sanity check for type based devirtualizations.
516 * ipa-prop.c: Include ipa-utils.h
517 (ipa_intraprocedural_devirtualization): Add sanity check.
518 (try_make_edge_direct_virtual_call): Likewise.
520 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
522 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
523 assignment statements.
525 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
527 * gimple-pretty-print.c (dump_ssaname_info): New function.
528 (dump_gimple_phi): Call it.
529 (pp_gimple_stmt_1): Likewise.
530 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
531 (range_info_def): Declare.
532 * tree-pretty-print.c (pp_double_int): New function.
533 (dump_generic_node): Call it.
534 * tree-pretty-print.h (pp_double_int): Declare.
535 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
536 * tree-ssanames.h (range_info_def): New structure.
537 (value_range_type): Move definition here.
538 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
540 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
542 (set_range_info): New function.
543 (get_range_info): Likewise.
544 (duplicate_ssa_name_range_info): Likewise.
545 (duplicate_ssa_name_fn): Check pointer type and call
546 duplicate_ssa_name_range_info.
547 * tree-ssa-copy.c (fini_copy_prop): Likewise.
548 * tree-vrp.c (value_range_type): Remove definition, now in
550 (vrp_finalize): Call set_range_info to update value range of
552 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
553 (SSA_NAME_RANGE_INFO): New macro.
555 2013-09-23 Richard Biener <rguenther@suse.de>
557 PR tree-optimization/58464
558 * tree-ssa-pre.c (phi_trans_lookup): Remove.
559 (phi_trans_add): Change to add conditionally on being not
561 (phi_translate_1): Remove recursion detection here.
562 (phi_translate): Pre-seed the cache with NULL to catch
563 recursion here in a more generic way.
564 (bitmap_find_leader): Adjust comment.
565 (get_representative_for): Dump value-numbers.
566 (create_expression_by_pieces): Likewise.
567 (insert_into_preds_of_block): Likewise.
569 2013-09-23 Christian Bruel <christian.bruel@st.com>
572 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
573 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
575 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
578 2013-09-20 Renlin Li <renlin.li@arm.com>
580 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
581 (aarch64_expand_epilogue): Likewise.
582 (aarch64_legitimize_reload_address): Likewise.
584 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
586 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
587 NULL_TREE before pushing them onto the vector. Likewise for labels.
589 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
591 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
592 comparison operators when -fno-trapping-math is in effect.
593 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
594 comparison operators in TFmode and assert that unsupported operators
596 (ia64_print_operand): Likewise.
598 2013-09-21 Jan Hubicka <jh@suse.cz>
600 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
601 (sse_typeless_stores): Enable for core
602 (sse_load0_by_pxor): Likewise.
603 (four_jump_limit): Disable for core.
604 (pad_returns): Likewise.
605 (avoid_vector_decode): Likewise.
606 (fuse_cmp_and_branch): Enable for cores.
607 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
609 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
612 * config/pa/pa.c (pa_option_override): Disable auto increment and
613 decrement instructions until reload is completed.
615 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
616 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
617 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
619 2013-09-20 DJ Delorie <dj@redhat.com>
620 Nick Clifton <nickc@redhat.com>
622 * config/rl78/rl78.c: Various whitespace and comment tweaks.
623 (need_to_save): Save bank 0 on interrupts.
624 (characterize_address): Strip far address wrappers.
625 (rl78_as_legitimate_address): Likewise.
626 (transcode_memory_rtx): Likewise.
627 (rl78_peep_movhi_p): Disable this peephole after devirt.
628 (rl78_propogate_register_origins): Forget all origins when a
630 * config/rl78/rl78-virt.md: Various whitespace tweaks.
631 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
633 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
634 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
635 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
637 * config/rl78/constraints.md: Various whitespace and paren tweaks.
639 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
641 * config/pa/pa.md: In "scc" insn patterns, change output template to
642 handle const0_rtx in reg_or_0_operand operands.
644 2013-09-20 Martin Husemann <martin@NetBSD.org>
647 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
648 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
650 2013-09-20 Richard Biener <rguenther@suse.de>
653 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
654 remembering SSA name version and block index.
655 (new_scev_info_str): Adjust.
656 (hash_scev_info): Likewise. Also hash the block index.
657 (eq_scev_info): Adjust.
658 (find_var_scev_info): Likewise.
659 (struct instantiate_cache_entry): Remove.
660 (struct instantiate_cache_type): Use a htab to map name, block
662 (instantiate_cache_type::~instantiate_cache_type): Adjust.
663 (get_instantiated_value_entry): Likewise.
664 (hash_idx_scev_info, eq_idx_scev_info): New functions.
665 (instantiate_scev_name): Adjust.
667 2013-09-20 Jeff Law <law@redhat.com>
669 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
671 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
673 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
674 Call aarch64_simd_expand_args to update op[argc].
676 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
678 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
681 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
683 * gengtype.c (file_rules): Added rule for *.cc files.
684 (get_output_file_with_visibility): Give fatal message when no
687 2013-09-20 Renlin Li <renlin.li@arm.com>
689 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
690 (aarch64_expand_epilogue): Likewise.
691 (aarch64_legitimize_reload_address): Likewise.
693 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
696 * expr.c (expand_assignment): Remove misalignp code path.
698 2013-09-20 Marek Polacek <polacek@redhat.com>
701 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
702 TYPE_PRECISION. Add asserts.
704 2013-09-20 Richard Biener <rguenther@suse.de>
706 PR tree-optimization/58453
707 * tree-loop-distribution.c (distribute_loop): Apply the cost
708 model for -ftree-loop-distribute-patterns, too.
710 2013-09-20 Richard Biener <rguenther@suse.de>
713 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
714 make type comparison less strict.
716 2013-09-20 Alan Modra <amodra@gmail.com>
718 * configure: Regenerate.
719 * aclocal.m4: Regenerate.
721 2013-09-20 Marek Polacek <polacek@redhat.com>
724 * doc/extend.texi: Document that attribute used is meant to be used
725 on variables with static storage duration.
727 2013-09-19 Jakub Jelinek <jakub@redhat.com>
729 PR tree-optimization/58472
730 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
731 simd_lane_access set inv_p = false.
732 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
733 the simduid magic VAR_DECL.
735 2013-09-19 Jan Hubicka <jh@suse.cz>
737 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
739 2013-09-17 Jeff Law <law@redhat.com>
741 * tree-ssa-dom.c (record_temporary_equivalences): New function
742 split out of dom_opt_dom_walker::after_dom_children.
743 (dom_opt_dom_walker::thread_across_edge): Move common code
744 in here from dom_opt_dom_walker::after_dom_children.
745 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
747 2013-09-19 Jan Hubicka <jh@suse.cz>
749 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
750 (TARGET_GENERIC): Use PROCESOR_GENERIC
751 (enum processor_type): Unify generic32 and 64.
752 * i386.md (cpu): Likewise.
753 * x86-tune.def (use_leave): Enable for generic32.
754 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
755 * athlon.md: Change generic64 to generic in all occurences.
756 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
757 (ix86_target_macros_internal): Likewise.
758 * driver-i386.c (host_detect_local_cpu): Likewise.
759 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename to ..
760 (generic_memcpy, generic_memset, generic_cost): This one.
761 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
762 (m_GENERIC32, m_GENERIC64): Remove.
763 (m_GENERIC): Turn into one flag.
764 (processor_target): Unify generic tunnings.
765 (ix86_option_override_internal): Replace generic32/64 by generic.
766 (ix86_issue_rate): Likewise.
767 (ix86_adjust_cost): Likewise.
769 2013-09-19 Jan Hubicka <jh@suse.cz>
771 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
774 2013-09-19 Jakub Jelinek <jakub@redhat.com>
776 * omp-low.c (expand_omp_sections): Always pass len - 1 to
777 GOMP_sections_start, even if !exit_reachable.
779 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
781 * lra-constraints.c (need_for_all_save_p): Use macro
782 HARD_REGNO_CALL_PART_CLOBBERED.
783 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
784 set up pseudo conflict hard regs.
786 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
789 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
792 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
794 PR rtl-optimization/58438
795 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
796 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
797 from previous subpasses.
799 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
801 * arm.c (arm_get_frame_offsets): Validate architecture supports
802 LDRD/STRD before accepting the tuning preference.
803 (arm_expand_prologue): Likewise.
804 (arm_expand_epilogue): Likewise.
806 2013-09-18 Richard Biener <rguenther@suse.de>
808 PR tree-optimization/58417
809 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
810 have chrecs with symbols defined in the loop as operands.
811 (chrec_fold_multiply): Likewise.
812 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
813 parameters before folding binary operations.
814 (struct instantiate_cache_entry_hasher): Remove.
815 (struct instantiate_cache_type): Use a pointer-map.
816 (instantiate_cache_type::instantiate_cache_type): New function.
817 (instantiate_cache_type::get): Likewise.
818 (instantiate_cache_type::set): Likewise.
819 (instantiate_cache_type::~instantiate_cache_type): Adjust.
820 (get_instantiated_value_entry): Likewise.
821 (global_cache): New global.
822 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
823 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
824 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
825 (instantiate_scev_name): Adjust.
826 (instantiate_scev): Construct global instead of local cache.
827 (resolve_mixers): Likewise.
829 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
830 Paolo Carlini <paolo.carlini@oracle.com>
833 * doc/implement-cxx.texi: Fix references to the C++ standards.
835 2013-09-18 Jakub Jelinek <jakub@redhat.com>
837 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
838 * tree-vect-data-refs.c (vect_analyze_data_refs): For
839 simd_lane_access drs, update also DR_ALIGNED_TO.
841 2013-09-18 Marek Polacek <polacek@redhat.com>
844 * doc/extend.texi: Document no_sanitize_undefined attribute.
845 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
846 no_sanitize_undefined attribute.
848 2013-09-18 Nick Clifton <nickc@redhat.com>
850 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
851 (ASM_DECLARE_FUNCTION_NAME): Define.
853 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
855 * compare-elim.c (find_comparison_dom_walker): New class
856 (find_comparisons_in_bb): Rename to
857 find_comparison_dom_walker::before_dom_children
858 (find_comparisons): Adjust
859 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
861 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
862 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
863 (init_walk_dominator_tree): Remove declaration.
864 (fini_walk_dominator_tree): Remove declaration.
865 * fwprop.c (single_def_use_dom_walker): New class
866 (single_def_use_enter_block): Convert to
867 single_def_use_dom_walker::before_dom_children.
868 (single_def_use_leave_block): Convert to
869 single_def_use_dom_walker::after_dom_children.
870 (build_single_def_use_links): Adjust.
871 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
873 (find_candidates_in_block): Convert to
874 find_candidates_dom_walker::before_dom_children.
875 (execute_strength_reduction): Adjust.
876 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
877 (sese_dom_walker): New class.
878 (sese_dom_walker::sese_dom_walker): New constructor.
879 (sese_dom_walker::~sese_dom_walker): New destructor.
880 (build_sese_conditions_before): Convert to
881 sese_dom_walker::before_dom_children.
882 (build_sese_conditions_after): Convert to
883 sese_dom_walker::after_dom_children.
884 (build_poly_scop): Adjust
885 * tree-into-ssa.c (rewrite_dom_walker): New class
886 (rewrite_enter_block): Convert to
887 rewrite_dom_walker::before_dom_children.
888 (rewrite_leave_block): Convert to
889 rewrite_dom_walker::after_dom_children.
890 (rewrite_update_dom_walker): New class.
891 (rewrite_update_enter_block): Convert to
892 rewrite_update_dom_walker::before_dom_children.
893 (rewrite_update_leave_block): Convert to
894 rewrite_update_dom_walker::after_dom_children.
895 (rewrite_blocks, rewrite_into_ssa): Adjust.
896 (mark_def_dom_walker): New class.
897 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
898 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
899 (mark_def_sites_blocks): Convert to
900 mark_def_dom_walker::before_dom_children.
901 (mark_def_site_blocks): Remove.
902 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
903 (tree_ssa_dominator_optimize): Adjust.
904 (dom_thread_across_edge): Convert to method
905 dom_opt_dom_walker::thread_across_edge.
906 (dom_opt_enter_block): Convert to member function
907 dom_opt_dom_walker::before_dom_children.
908 (dom_opt_leave_block): Convert to member function
909 dom_opt_dom_walker::after_dom_children.
910 * tree-ssa-dse.c (dse_dom_walker): New class.
911 (dse_enter_block): Convert to member function
912 dse_dom_walker::before_dom_children.
913 (tree_ssa_dse): Adjust.
914 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
915 (determine_invariantness_stmt): Convert to method
916 invariantness_dom_walker::before_dom_children.
917 (determine_invariantness): Remove
918 (move_computations_dom_walker): New class.
919 (move_computations_stmt): Convert to method
920 move_computations_dom_walker::before_dom_children.
921 (move_computations, tree_ssa_lim): Adjust.
922 * tree-ssa-phiopt.c (nontrapping_dom_walker): new class
923 (nt_init_block): Make method
924 notrappping_dom_walker::before_dom_children.
925 (nt_fini_block): Make
926 method nontrapping_dom_walker::after_dom_children.
927 (get_non_trapping): Adjust.
928 * tree-ssa-pre.c (eliminate_dom_walker): New class.
929 (eliminate_bb): Make method eliminate_dom_walker::before_dom_children.
930 (eliminate_leave_block): Make method.
931 eliminate_dom_walker::after_dom_children.
933 * tree-ssa-strlen.c (strlen_dom_walker): New class.
934 (strlen_enter_block): Make method
935 strlen_dom_walker::before_dom_children.
936 (strlen_leave_block): Make
937 method strlen_dom_walker::after_dom_children.
938 (tree_ssa_strlen): Adjust.
939 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
940 (tree_ssa_uncprop): Adjust.
941 (uncprop_leave_block): Make method
942 uncprop_dom_walker::after_dom_children.
943 (uncprop_leave_block): Make method
944 uncprop_dom_walker::before_dom_children.
946 2013-09-18 Bin Cheng <bin.cheng@arm.com>
948 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn
949 before branch in same basic block. Check both src and dest of
952 2013-09-17 Nick Clifton <nickc@redhat.com>
954 * config/rl78/rl78-real.md (bf): New pattern.
956 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
957 (rl78_print_operand): Do not put a # before a %B.
958 * config/rl78/rl78.opt: Tweak doc strings.
960 2013-09-17 DJ Delorie <dj@redhat.com>
962 * config/rl78/constraints.md (Wcv): Allow up to $r31.
963 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
964 (rl78_option_override): Likewise, if -mallregs.
965 (is_virtual_register): Likewise.
966 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
967 (REGNO_OK_FOR_BASE_P): Likewise.
968 * config/rl78/rl78.opt (-mallregs): New.
970 2013-09-17 Nick Clifton <nickc@redhat.com>
972 * config/rl78/rl78.c (need_to_save): Change return type to bool.
973 For interrupt functions: save all call clobbered registers if the
974 interrupt handler is not a leaf function.
975 (rl78_expand_prologue): Always recompute the frame information.
976 For interrupt functions: only select bank 0 if one of the bank 0
977 registers is going to be psuhed.
979 2013-09-17 DJ Delorie <dj@redhat.com>
981 * config/rl78/constraints.md: For each W* constraint, rename to C*
982 and create a W* constraint that checks for an optional ES: prefix
984 * config/rl78/rl78.md (UNS_ES_ADDR): New.
985 (es_addr): New. Used to wrap far addresses.
986 * config/rl78/rl78-protos.h (rl78_es_addr): New.
988 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
989 wrapped far addresses.
990 (rl78_print_operand_1): Unwrap far addresses before processing.
991 (rl78_lo16): Wrap far addresses in unspecs.
994 (insn_ok_now): Check for not-yet-wrapped far addresses.
995 (transcode_memory_rtx): Properly re-wrap far addresses.
997 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
999 * config/sparc/t-rtems: Add leon3 multilibs.
1001 2013-09-17 Cong Hou <congh@google.com>
1003 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
1004 when checking the dot production pattern. The type of rhs operand
1005 of multiply is now checked correctly.
1007 2013-09-17 Jeff Law <law@redhat.com>
1009 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
1010 edge implied equivalences into successor phis.
1011 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
1012 here from tree-ssa-threadedge.c.
1013 (mark_threaded_blocks): When threading through a joiner, if both
1014 successors of the joiner's clone reach the same block, verify the
1015 PHI arguments are equal. If not, cancel the jump threading request.
1016 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
1017 tree-ssa-threadupdate.c
1018 (thread_across_edge): Don't check PHI argument equality when
1019 threading through joiner block here.
1021 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
1023 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
1024 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
1025 (warn_uninit, warn_uninitialized_vars, execute_early_warn_uninitialized,
1026 make_pass_early_warn_uninitialized): Move to tree-ssa-uninit.c.
1027 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c
1028 (has_undefined_value_p): New. Pass dependant parts of
1029 ssa_undefined_value_p.
1030 (uninit_undefined_value_p): Use has_undefined_value_p.
1031 (warn_uninit, warn_uninitialized_vars, execute_early_warn_uninitialized,
1032 make_pass_early_warn_uninitialized): Move from tree-ssa.c
1033 * tree-ssa.h: Adjust prototypes
1035 2013-09-17 Jan Hubicka <jh@suse.cz>
1038 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
1039 * ipa-inline.c (can_inline_edge_p): Do not downgrade
1040 FUNCTION_NOT_OPTIMIZED.
1041 * ipa-inline-analysis.c (compute_inline_parameters): Function
1042 not optimized is not inlinable unless it is alwaysinline.
1043 (inline_analyze_function): Force calls in not optimized
1044 function not inlinable.
1046 2013-09-17 Jan Hubicka <jh@suse.cz>
1049 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
1051 * ipa.c (function_and_variable_visibility): Likewise.
1052 * ipa-profile.c (ipa_profile): Likewise.
1054 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
1057 * cgraph.c (cgraph_function_body_availability): Check for ifunc
1058 attribute, and don't inline the resolver in this case.
1060 2013-09-17 Teresa Johnson <tejohnson@google.com>
1062 * coverage.c (get_coverage_counts): Add missing newline.
1064 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1066 PR tree-optimization/58088
1067 * fold-const.c (mask_with_trailing_zeros): New function.
1068 (fold_binary_loc): Make sure we don't recurse infinitely
1069 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
1070 Use mask_with_trailing_zeros where appropriate.
1072 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
1074 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
1075 of 'prev' var to get better distance estimation.
1077 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
1079 * tree-inline.h (struct copy_body_data): Add transform_parameter.
1080 * tree-inline.c (is_parameter_of): New predicate.
1081 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
1082 a parameter has been remapped.
1083 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
1084 (optimize_inline_calls): Initialize transform_parameter.
1085 (copy_gimple_seq_and_replace_locals): Likewise.
1086 (tree_function_versioning): Likewise.
1087 (maybe_inline_call_in_expr): Likewise.
1089 2013-09-17 Nick Clifton <nickc@redhat.com>
1091 * config/msp430/msp430-protos.h: Add prototypes for new functions.
1092 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
1094 (is_attr_func): New function.
1095 (msp430_is_interrupt_func): New function.
1096 (is_naked_func): New function.
1097 (is_reentrant_func): New function.
1098 (is_critical_func): New function.
1099 (msp430_start_function): Add annotations for function attributes.
1100 (msp430_attr): New function.
1101 (msp430_attribute_table): New.
1102 (msp430_function_section): New function.
1103 (TARGET_ASM_FUNCTION_SECTION): Define.
1104 (msp430_builtin): New enum.
1105 (msp430_init_builtins): New function.
1106 (msp430_builtin_devl): New function.
1107 (msp430_expand_builtin): New function.
1108 (TARGET_INIT_BUILTINS): Define.
1109 (TARGET_EXPAND_BUILTINS): Define.
1110 (TARGET_BUILTIN_DECL): Define.
1111 (msp430_expand_prologue): Add support for naked, interrupt,
1112 critical and reentrant functions.
1113 (msp430_expand_epilogue): Likewise.
1114 (msp430_print_operand): Handle 'O' character.
1115 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
1117 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
1118 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
1119 (pushm): Use a 'n' rather than an 'i' constraint.
1120 (msp_return): Add generation of the interrupt return instruction.
1121 (disable_interrupts): New pattern.
1122 (enable_interrupts): New pattern.
1123 (push_intr_state): New pattern.
1124 (pop_intr_state): New pattern.
1125 (bic_SR): New pattern.
1126 (bis_SR): New pattern.
1127 * doc/extend.texi: Document MSP430 function attributes and builtin
1130 2013-09-17 Richard Biener <rguenther@suse.de>
1132 PR tree-optimization/58432
1133 * tree-loop-distribution.c (tree_loop_distribution): Also
1134 scan PHIs for outside loop uses and seed a partition from them.
1136 2013-09-17 Bin Cheng <bin.cheng@arm.com>
1138 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
1139 (restructure_reference): Call backtrace_base_for_ref.
1141 2013-09-17 Alan Modra <amodra@gmail.com>
1144 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11 patch.
1146 2013-09-16 DJ Delorie <dj@redhat.com>
1148 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
1149 vregs location for RL78/G10.
1150 (rl78_expand_prologue): Avoid SEL on G10.
1151 (rl78_expand_epilogue): Likewise.
1152 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
1153 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
1154 __RL78_G10__ when appropriate.
1155 (ASM_SPEC): Pass -mg10 along to the assembler.
1156 * config/rl78/rl78.md (sel_rb): Disable for G10.
1157 * config/rl78/rl78.opt: Add -mg10 option.
1158 * config/rl78/t-rl78: Add -mg10 multilib.
1160 2013-09-16 Xinliang David Li <davidxl@google.com>
1162 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
1163 * omp-low.c (omp_max_vf): Ditto.
1164 (expand_omp_simd): Ditto.
1165 * tree-vectorizer.c (vectorize_loops): Ditto.
1166 (gate_vect_slp): Ditto.
1167 (gate_increase_alignment): Ditto.
1168 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
1169 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
1170 (gate_tree_vectorize): Name change.
1171 (tree_vectorize): Ditto.
1172 (pass_vectorize::gate): Call new function.
1173 (pass_vectorize::execute): Ditto.
1174 opts.c: O3 default setting change.
1175 (finish_options): Check new flag.
1176 * doc/invoke.texi: Document new flags.
1177 * common.opt: New flags.
1179 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
1181 * doc/tm.texi.in (Cond Exec Macros): Remove node.
1182 (Condition Code): Don't reference it.
1183 * doc/tm.texi: Regenerate.
1185 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
1188 * lra-constraints.c (undo_optional_reloads): Consider all optional
1189 reload even if it did not get a hard reg.
1191 2013-09-16 Teresa Johnson <tejohnson@google.com>
1193 * dumpfile.c (dump_loc): Remove newline emission.
1194 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
1195 emission to dump_printf_loc calls where missing.
1196 (vect_mark_for_runtime_alias_test): Ditto.
1197 (vect_analyze_data_ref_dependence): Ditto.
1198 (vect_analyze_data_ref_dependences): Ditto.
1199 (vect_slp_analyze_data_ref_dependence): Ditto.
1200 (vect_slp_analyze_data_ref_dependences): Ditto.
1201 (vect_compute_data_ref_alignment): Ditto.
1202 (vect_update_misalignment_for_peel): Ditto.
1203 (vect_verify_datarefs_alignment): Ditto.
1204 (vector_alignment_reachable_p): Ditto.
1205 (vect_get_data_access_cost): Ditto.
1206 (vect_enhance_data_refs_alignment): Ditto.
1207 (vect_find_same_alignment_drs): Ditto.
1208 (vect_analyze_data_refs_alignment): Ditto.
1209 (vect_analyze_group_access): Ditto.
1210 (vect_analyze_data_ref_access): Ditto.
1211 (vect_analyze_data_ref_accesses): Ditto.
1212 (vect_prune_runtime_alias_test_list): Ditto.
1213 (vect_analyze_data_refs): Ditto.
1214 (vect_create_addr_base_for_vector_ref): Ditto.
1215 (vect_create_data_ref_ptr): Ditto.
1216 (vect_grouped_store_supported): Ditto.
1217 (vect_grouped_load_supported): Ditto.
1218 * value-prof.c (check_counter): Ditto.
1219 (check_ic_target): Ditto.
1220 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
1221 (vect_recog_widen_mult_pattern): Ditto.
1222 (vect_recog_widen_sum_pattern): Ditto.
1223 (vect_recog_over_widening_pattern): Ditto.
1224 (vect_recog_widen_shift_pattern): Ditto.
1225 (vect_recog_rotate_pattern): Ditto.
1226 (vect_recog_vector_vector_shift_pattern): Ditto.
1227 (vect_recog_divmod_pattern): Ditto.
1228 (vect_recog_mixed_size_cond_pattern): Ditto.
1229 (vect_recog_bool_pattern): Ditto.
1230 (vect_pattern_recog_1): Ditto.
1231 (vect_pattern_recog): Ditto.
1232 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
1233 (vect_is_simple_iv_evolution): Ditto.
1234 (vect_analyze_scalar_cycles_1): Ditto.
1235 (vect_get_loop_niters): Ditto.
1236 (vect_analyze_loop_1): Ditto.
1237 (vect_analyze_loop_form): Ditto.
1238 (vect_analyze_loop_operations): Ditto.
1239 (vect_analyze_loop_2): Ditto.
1240 (vect_analyze_loop): Ditto.
1241 (report_vect_op): Ditto.
1242 (vect_is_slp_reduction): Ditto.
1243 (vect_is_simple_reduction_1): Ditto.
1244 (vect_get_known_peeling_cost): Ditto.
1245 (vect_estimate_min_profitable_iters): Ditto.
1246 (vect_model_reduction_cost): Ditto.
1247 (vect_model_induction_cost): Ditto.
1248 (get_initial_def_for_induction): Ditto.
1249 (vect_create_epilog_for_reduction): Ditto.
1250 (vectorizable_reduction): Ditto.
1251 (vectorizable_induction): Ditto.
1252 (vectorizable_live_operation): Ditto.
1253 (vect_loop_kill_debug_uses): Ditto.
1254 (vect_transform_loop): Ditto.
1255 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
1256 (vect_stmt_relevant_p): Ditto.
1257 (process_use): Ditto.
1258 (vect_mark_stmts_to_be_vectorized): Ditto.
1259 (vect_model_simple_cost): Ditto.
1260 (vect_model_promotion_demotion_cost): Ditto.
1261 (vect_model_store_cost): Ditto.
1262 (vect_get_store_cost): Ditto.
1263 (vect_model_load_cost): Ditto.
1264 (vect_get_load_cost): Ditto.
1265 (vect_init_vector_1): Ditto.
1266 (vect_get_vec_def_for_operand): Ditto.
1267 (vect_finish_stmt_generation): Ditto.
1268 (vectorizable_call): Ditto.
1269 (vectorizable_conversion): Ditto.
1270 (vectorizable_assignment): Ditto.
1271 (vectorizable_shift): Ditto.
1272 (vectorizable_operation): Ditto.
1273 (vectorizable_store): Ditto.
1274 (vectorizable_load): Ditto.
1275 (vectorizable_condition): Ditto.
1276 (vect_analyze_stmt): Ditto.
1277 (vect_transform_stmt): Ditto.
1278 (vect_is_simple_use): Ditto.
1279 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
1280 (vect_can_advance_ivs_p): Ditto.
1281 (vect_update_ivs_after_vectorizer): Ditto.
1282 (vect_do_peeling_for_loop_bound): Ditto.
1283 (vect_gen_niters_for_prolog_loop): Ditto.
1284 (vect_update_inits_of_drs): Ditto.
1285 (vect_create_cond_for_alias_checks): Ditto.
1286 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
1287 (vect_build_slp_tree_1): Ditto.
1288 (vect_supported_load_permutation_p): Ditto.
1289 (vect_analyze_slp_instance): Ditto.
1290 (vect_analyze_slp): Ditto.
1291 (vect_make_slp_decision): Ditto.
1292 (vect_detect_hybrid_slp): Ditto.
1293 (vect_bb_vectorization_profitable_p): Ditto.
1294 (vect_slp_analyze_bb_1): Ditto.
1295 (vect_update_slp_costs_according_to_vf): Ditto.
1296 (vect_get_mask_element): Ditto.
1297 (vect_transform_slp_perm_load): Ditto.
1298 (vect_schedule_slp_instance): Ditto.
1299 (vect_schedule_slp): Ditto.
1300 (vect_slp_transform_bb): Ditto.
1301 * profile.c (read_profile_edge_counts): Ditto.
1302 (compute_branch_probabilities): Ditto.
1303 * coverage.c (get_coverage_counts): Ditto.
1305 2013-09-16 Diego Novillo <dnovillo@google.com>
1307 * tree-core.h: Add missing comment lines from refactoring
1310 2013-09-16 Jan Hubicka <jh@suse.cz>
1312 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
1313 abstract functions; for static functions check the presence
1316 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
1318 * config/aarch64/aarch64-simd-builtins.def (fma): New.
1319 * config/aarch64/aarch64-simd.md
1320 (aarch64_mla_elt<mode>): New.
1321 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
1322 (aarch64_mls_elt<mode>): Likewise.
1323 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
1324 (aarch64_fma4_elt<mode>): Likewise.
1325 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
1326 (aarch64_fma4_elt_to_128v2df): Likewise.
1327 (aarch64_fma4_elt_to_64df): Likewise.
1328 (fnma<mode>4): Likewise.
1329 (aarch64_fnma4_elt<mode>): Likewise.
1330 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
1331 (aarch64_fnma4_elt_to_128v2df): Likewise.
1332 (aarch64_fnma4_elt_to_64df): Likewise.
1333 * config/aarch64/iterators.md (VDQSF): New.
1334 * config/aarch64/arm_neon.h
1335 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
1336 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
1338 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
1340 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
1341 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
1342 (aarch64_mul3_elt_to_128df): Likewise.
1343 (aarch64_mul3_elt_to_64v2df): Likewise.
1344 * config/aarch64/iterators.md (VEL): Also handle DFmode.
1346 (VMUL_CHANGE_NLANES) Likewise.
1349 * config/aarch64/arm_neon.h
1350 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
1352 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
1354 * config/aarch64/arm_neon.h
1355 (vcvtx_high_f32_f64): Fix parameters.
1357 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1358 Uros Bizjak <ubizjak@gmail.com>
1360 * config/alpha.c: Include tree-ssanames.h.
1362 2013-09-16 Richard Biener <rguenther@suse.de>
1364 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
1365 (dot_rdg_1): Handle control_dd.
1366 (create_edge_for_control_dependence): New function.
1367 (create_rdg_edges): Add control dependences if asked for.
1368 (build_rdg): Likewise.
1369 (generate_loops_for_partition): If there are not necessary
1370 control stmts remove all their dependencies.
1371 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
1372 (distribute_loop): Pass on control dependences.
1373 (tree_loop_distribution): Compute control dependences and remove
1374 restriction on number of loop nodes.
1376 2013-09-16 Jakub Jelinek <jakub@redhat.com>
1378 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
1381 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
1383 * cse.c (try_const_anchors): Punt on CC modes.
1385 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1387 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
1389 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
1392 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
1395 2013-09-14 Iain Sandoe <iain@codesourcery.com>
1398 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
1400 (darwin_objc1_section): Likewise.
1401 (darwin_file_end): Emit Image Info section when required.
1403 2013-09-14 Jan Hubicka <jh@suse.cz>
1405 * tree-into-ssa.c (gate_into_ssa): New.
1406 (pass_data_build_ssa): Use it.
1407 * cgraph.h (expand_thunk): Update prototype.
1408 * cgraphunit.c (analyze_function): Expand thunks early.
1409 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
1410 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
1411 set lowered flag; do not add new function.
1412 (assemble_thunks_and_aliases): Update.
1413 * tree-ssa.c (gate_init_datastructures): New gate.
1414 (pass_data_init_datastructures): Use it.
1416 2013-09-14 Iain Sandoe <iain@codesourcery.com>
1419 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
1420 xmm register set described in the psABI.
1422 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
1424 * dwarf2out.c (should_emit_struct_debug): Add check
1425 for type_decl variable is not NULL.
1427 2013-09-13 Jacek Caban <jacek@codeweavers.com>
1429 * config.gcc: Use new winnt-c.c target hooks
1430 * config/t-winnt: New file
1431 * config/winnt-c.c: New file
1432 * doc/tm.texi.in: Document new hook
1433 * doc/tm.texi: Regenerated
1435 2013-09-13 Jan Hubicka <jh@suse.cz>
1438 * ipa-inline.c (check_callers): New function.
1439 (check_caller_edge): Remove.
1440 (want_inline_function_to_all_callers_p): Also permit alises that are
1442 (inline_to_all_callers): Terminate the walk when devirtualization
1443 introduce new calls.
1445 2013-09-13 Jan Hubicka <jh@suse.cz>
1447 * ipa-inline-analysis.c (struct growth_data): Add node.
1448 (do_estimate_growth_1): Fix detection of recursion.
1450 2013-09-13 Jakub Jelinek <jakub@redhat.com>
1452 PR tree-optimization/58392
1453 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
1454 to avoid shadowing of outer loop variable. If
1455 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
1456 replace_by_duplicate_decl simduid of loops that have it set and
1457 set dest_cfun->has_simduid_loops and/or
1458 dest_cfun->has_force_vect_loops.
1459 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
1460 instead of maybe_lookup_decl.
1461 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
1462 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
1463 call. Copy over force_vect and copy and remap simduid. Set
1464 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
1465 (copy_cfg_body): Remove blocks_to_copy argument. Use
1466 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
1467 caller. Don't set cfun->has_simduid_loops and/or
1468 cfun->has_force_vect_loops here.
1469 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
1471 (expand_call_inline, tree_function_versioning): Adjust copy_body
1474 2013-09-13 Martin Jambor <mjambor@suse.cz>
1477 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
1478 the assert if the edge was a speculative one.
1480 2013-09-13 Richard Biener <rguenther@suse.de>
1482 * tree-data-ref.h (known_dependences_p): Move here ...
1483 * tree-loop-distribution.c (known_dependences_p): ... from here.
1484 (dump_rdg_component, debug_rdg_component): Remove.
1486 (generate_loops_for_partition): Use gimple_uid instead of
1487 relying on matching stmt visit order.
1488 (rdg_build_partitions): Take starting stmt vector.
1489 (ldist_gen): Merge into ...
1490 (distribute_loop): ... this function. Do not compute starting
1492 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
1494 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1496 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
1497 Set type attribute correctly. Set predicable_short_it attribute.
1498 (cmpsi_shiftsi): Remove %? from output template.
1500 2013-09-13 Richard Biener <rguenther@suse.de>
1502 * tree-loop-distribution.c (struct rdg_component,
1503 rdg_defs_used_in_other_loops_p, free_rdg_components,
1504 rdg_build_components): Remove.
1505 (stmts_from_loop): Do not record virtual PHIs.
1506 (generate_loops_for_partition): Skip virtual PHIs.
1507 (build_rdg_partition_for_component): Rename to ...
1508 (build_rdg_partition_for_vertex): ... this and adjust.
1509 (rdg_build_partitions): Take a vector of starting vertices
1510 instead of components. Remove unnecessary leftover handling.
1511 (ldist_gen): Do not build components or record other stores.
1512 (distribute_loop): Do not distribute loops containing stmts
1515 2013-09-13 Christian Bruel <christian.bruel@st.com>
1518 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
1520 2013-09-13 Kai Tietz <ktietz@redhat.com>
1522 * config.gcc: Separate cases for mingw and cygwin targets,
1523 and add 64-bit cygwin target case.
1525 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
1526 dll-export inline-functions.
1527 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
1530 2013-09-13 Jeff Law <law@redhat.com>
1534 2013-09-06 Jeff Law <law@redhat.com>
1536 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
1537 edge implied equivalences into successor phis.
1539 2013-09-12 DJ Delorie <dj@redhat.com>
1541 * config/rl78/rl78-virt.md: Change from | to \; for asm line
1544 2013-09-12 Brooks Moses <bmoses@google.com>
1547 * Makefile.in: Do not install driver binaries in $(target)/bin.
1549 2013-09-12 DJ Delorie <dj@redhat.com>
1551 * config/rl78/rl78.opt (mrelax): New.
1552 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
1553 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
1555 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
1557 (rl78_expand_epilogue): Likewise.
1559 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
1562 * lra-eliminations.c (remove_reg_equal_offset_note): New.
1563 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
1564 pointer elimination with using remove_reg_equal_offset_note.
1566 2013-09-12 DJ Delorie <dj@redhat.com>
1568 * config/msp430/: New port.
1569 * config.gcc (msp430): Added.
1570 * doc/invoke.texi: Document MSP430 options.
1571 * doc/install.texi: Document msp430-elf
1572 * doc/md.texi: Document msp430-elf
1573 * doc/contrib.texi: Document msp430-elf
1575 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
1578 2013-09-12 Martin Jambor <mjambor@suse.cz>
1581 * ipa-prop.c (remove_described_reference): Give up if the edge in the
1582 reference descriptor is NULL.
1583 (ipa_edge_removal_hook): If owning a reference descriptor, set its
1586 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
1588 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
1589 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
1590 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
1591 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
1592 make_temp_ssa_name): move to tree-ssanames.h
1593 * tree-ssa-alias.h: Move prototype.
1594 * tree-ssa.h: Include tree-ssanames.h.
1595 * tree-ssanames.c (FREE_SSANAMES): Move to here.
1596 * tree-ssanames.h: New. Move items from tree-flow*.h
1597 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
1599 2013-09-12 Richard Biener <rguenther@suse.de>
1601 PR tree-optimization/58404
1602 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
1603 propagate non-invariant addresses into dereferences wrapped
1604 in component references.
1606 2013-09-12 Richard Biener <rguenther@suse.de>
1608 PR tree-optimization/58402
1609 * passes.def: Move pass_late_warn_uninitialized later.
1611 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
1613 * tree-ssa.h: New. Move content from tree-flow.h and
1615 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
1616 Move prototypes belonging to tree-ssa.c.
1617 * tree-flow-inline.h (redirect_edge_var_map_def,
1618 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
1620 * gimple.h: Adjust prototypes.
1621 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
1623 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
1624 * tree.h: Move prototype to tree-ssa.h.
1625 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
1626 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
1627 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
1628 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
1629 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
1630 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
1631 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
1632 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
1633 graphite-blocking.c, graphite-clast-to-gimple.c,
1634 graphite-dependences.c, graphite-interchange.c,
1635 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
1636 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
1637 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
1638 ipa-reference.c, ipa-split.c, ipa-utils.c,
1639 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
1640 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
1641 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
1642 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
1643 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
1644 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
1645 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
1646 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
1647 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
1648 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
1649 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
1650 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
1651 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
1652 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
1653 value-prof.c, var-tracking.c,
1654 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
1656 2013-09-12 Richard Biener <rguenther@suse.de>
1658 PR tree-optimization/58396
1659 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
1660 (build_rdg): Take a loop-nest parameter, fix memleaks.
1661 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
1663 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
1665 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
1668 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
1670 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
1671 The FMA instruction names should have a 'v' prefix.
1673 2013-09-12 Richard Biener <rguenther@suse.de>
1675 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
1676 (dot_rdg): Use popen instead of system in optional code.
1677 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
1678 (already_processed_vertex_p): Adjust.
1679 (has_anti_or_output_dependence, predecessor_has_mem_write,
1680 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
1681 rdg_flag_uses): Remove.
1682 (rdg_flag_vertex): Simplify.
1683 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
1685 (build_rdg_partition_for_component): Process the first vertex
1686 of a component only.
1687 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
1689 2013-09-12 Alan Modra <amodra@gmail.com>
1691 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
1693 2013-09-11 DJ Delorie <dj@redhat.com>
1694 Nick Clifton <nickc@redhat.com>
1696 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
1697 (rl78_stack_based_mem): New.
1698 * config/rl78/constraints.md (Iv08): New.
1707 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
1709 (movsi): Change from expand to insn-and-split.
1710 (ashrsi3): Clobber AX.
1714 * config/rl78/rl78.md (CC_REG): Fix.
1715 (addsi3): Allow memory and immediate operands.
1716 (addsi3_internal): Split into...
1717 (addsi3_internal_virt): ...new, and ...
1718 (addsi3_internal_real): ...new.
1720 (subsi3_internal_virt): New.
1721 (subsi3_internal_real): New.
1722 (mulsi3): Add memory operand.
1723 (mulsi3_rl78): Likewise.
1724 (mulsi3_g13): Likewise.
1725 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
1726 (cbranchqi4_real): Add more constraint options.
1727 (cbranchhi4_real): Expand pattern.
1728 (cbranchhi4_real_signed): New.
1729 (cbranchhi4_real_inverted): New.
1730 (cbranchsi4_real_lt): New.
1731 (cbranchsi4_real_ge): New.
1732 (cbranchsi4_real_signed): New.
1733 (cbranchsi4_real): New.
1735 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
1737 (lshrsi3_virt): Likewise.
1738 (ashlsi3_virt): Likewise.
1739 (cbranchqi4_virt_signed): New.
1740 (cbranchhi4_virt_signed): New.
1741 (cbranchsi4_virt): New.
1742 * config/rl78/rl78.c: Whitespace fixes throughout.
1743 (move_elim_pass): New.
1744 (pass_data_rl78_move_elim): New.
1745 (pass_rl78_move_elim): New.
1746 (make_pass_rl78_move_elim): New.
1747 (rl78_devirt_info): Run devirt earlier.
1748 (rl78_move_elim_info): New.
1749 (rl78_asm_file_start): Register it.
1750 (rl78_split_movsi): New.
1751 (rl78_as_legitimate_address): Allow virtual base registers when
1753 (rl78_addr_space_convert): Remove spurious debug stuff.
1754 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
1755 (rl78_print_operand): More cases for not printing '#'.
1756 (rl78_expand_compare): Remove most of the logic.
1757 (content_memory): New.
1758 (clear_content_memory): New.
1759 (get_content_index): New.
1760 (get_content_name): New.
1761 (display_content_memory): New.
1762 (update_content): New.
1763 (record_content): New.
1764 (already_contains): New.
1765 (insn_ok_now): Re-recog insns with virtual registers.
1766 (add_postponed_content_update): New.
1767 (process_postponed_content_update): New.
1768 (gen_and_emit_move): New.
1769 (transcode_memory_rtx): Record new location content.
1770 Use gen_and_emit_move.
1771 (force_into_acc): New.
1772 (move_to_acc): Use gen_and_emit_move.
1773 (move_from_acc): Likewise.
1774 (move_acc_to_reg): Likewise.
1775 (move_to_x): Likewise.
1776 (move_to_hl): Likewise.
1777 (move_to_de): Likewise.
1778 (rl78_alloc_physical_registers_op1): Record location content.
1779 (has_constraint): New.
1780 (rl78_alloc_physical_registers_op2): Record location content.
1782 (rl78_alloc_physical_registers_ro1): Likewise.
1783 (rl78_alloc_physical_registers_cmp): Likewise.
1784 (rl78_alloc_physical_registers_umul): Likewise.
1785 (rl78_alloc_address_registers_macax): New.
1786 (rl78_alloc_physical_registers): Initialize and set location
1787 content memory as needed.
1788 (rl78_reorg): Make sure split2 is called.
1789 (rl78_rtx_costs): New.
1791 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
1793 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
1794 for (not (neg ...)) and (neg (not ...)) cases.
1796 2013-09-11 Richard Biener <rguenther@suse.de>
1799 * passes.def: Split critical edges before late uninit warning passes.
1800 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
1802 2013-09-11 Jakub Jelinek <jakub@redhat.com>
1804 PR tree-optimization/58385
1805 * fold-const.c (build_range_check): If both low and high are NULL,
1806 use omit_one_operand_loc to preserve exp side-effects.
1808 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1810 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
1812 2013-09-11 Richard Biener <rguenther@suse.de>
1814 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
1815 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
1816 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
1817 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
1818 stmts_from_loop, known_dependences_p, build_empty_rdg,
1819 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
1820 * tree-loop-distribution.c: ... here.
1821 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
1822 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
1823 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
1824 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
1825 * tree-loop-distribution.c: ... here.
1826 * tree-loop-distribution.c: Include gimple-pretty-print.h.
1827 (struct partition_s): Add loops member.
1828 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
1829 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
1830 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
1832 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1833 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1834 Sergey Lega <sergey.s.lega@intel.com>
1835 Anna Tikhonova <anna.tikhonova@intel.com>
1836 Ilya Tocar <ilya.tocar@intel.com>
1837 Andrey Turetskiy <andrey.turetskiy@intel.com>
1838 Ilya Verbin <ilya.verbin@intel.com>
1839 Kirill Yukhin <kirill.yukhin@intel.com>
1840 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1842 * config/i386/constraints.md (k): New.
1844 * config/i386/i386.c (const regclass_map): Add new mask registers.
1845 (dbx_register_map): Ditto.
1846 (dbx64_register_map): Ditto.
1847 (svr4_dbx_register_map): Ditto.
1848 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
1850 (ix86_preferred_reload_class): Disable constants for mask registers.
1851 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
1852 (ix86_hard_regno_mode_ok): Support new mask registers.
1853 (x86_order_regs_for_local_alloc): Ditto.
1854 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
1855 (FIXED_REGISTERS): Add new mask registers.
1856 (CALL_USED_REGISTERS): Ditto.
1857 (REG_ALLOC_ORDER): Ditto.
1858 (VALID_MASK_REG_MODE): New.
1859 (FIRST_MASK_REG): Ditto.
1860 (LAST_MASK_REG): Ditto.
1861 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
1862 (MAYBE_MASK_CLASS_P): New.
1863 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
1864 (REG_CLASS_CONTENTS): Ditto.
1865 (MASK_REGNO_P): New.
1866 (ANY_MASK_REG_P): Ditto.
1867 (HI_REGISTER_NAMES): Add new mask registers.
1868 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
1869 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
1871 (attribute "type"): Add mskmov, msklog.
1872 (attribute "length_immediate"): Support them.
1873 (attribute "memory"): Ditto.
1874 (attribute "prefix_0f"): Ditto.
1875 (*movhi_internal): Support new mask registers.
1876 (*movqi_internal): Ditto.
1877 (define_split): Split out clobber pattern is a logic
1878 insn on mask registers.
1879 (*k<logic><mode>): New.
1880 (*andhi_1): Extend to support mask regs.
1881 (*andqi_1): Extend to support mask regs.
1883 (define_split): Split and-not to and and not if operands
1885 (*<code><mode>_1): Separate HI mode to new pattern...
1886 (*<code>hi_1): This.
1887 (*<code>qi_1): Extend to support mask regs.
1889 (kortestzhi): Ditto.
1890 (kortestchi): Ditto.
1892 (*one_cmpl<mode>2_1): Remove HImode and handle it...
1893 (*one_cmplhi2_1): ...Here, now with mask registers support.
1894 (*one_cmplqi2_1): Support new mask registers.
1895 (HI/QImode arithmetics splitter): Don't split if mask registers
1897 (HI/QImode not splitter): Ditto.
1898 * config/i386/predicated.md (mask_reg_operand): New.
1899 (general_reg_operand): Ditto.
1901 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1903 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
1904 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
1906 2013-09-10 Jeff Law <law@redhat.com>
1908 PR tree-optimization/58380
1909 * tree-ssa-threadupdate.c (thread_block): Recognize another case
1910 of threading through a buried loop header.
1912 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
1913 return value for single successor case.
1915 2013-09-10 Jan Hubicka <jh@suse.cz>
1917 * ipa-devirt.c (ipa_devirt): Enable with LTO.
1919 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
1922 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
1923 support conditional execution.
1924 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
1926 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
1928 * lra.c (lra): Clear lra_optional_reload_pseudos before every
1930 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
1931 Check destination too to check move insn.
1932 (undo_optional_reloads): Add check that the original peudo did not
1933 changed its allocation and the optional reload was inherited on last
1934 inheritance pass. Break loop after deciding to keep optional reload.
1935 (lra_undo_inheritance): Add check that inherited pseudo still in
1938 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
1940 * config/aarch64/aarch64.md (generic_sched): New.
1941 * config/aarch64/aarch64-generic.md (load): Make conditional
1942 on generic_sched attribute.
1943 (nonload): Likewise.
1945 2013-09-10 Jan Hubicka <jh@suse.cz>
1947 * lto-cgraph.c: Include ipa-utils.h.
1948 (compute_ltrans_boundary): Also add possible targets into the boundary.
1950 2013-09-10 Jan Hubicka <jh@suse.cz>
1952 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
1953 VAR_DECL of vtable rather than full expression.
1955 2013-09-10 Jan Hubicka <jh@suse.cz>
1956 Paolo Carlini <paolo.carlini@oracle.com>
1958 * cgraphunit.c (analyze_functions): Save input_location, set it
1959 to UNKNOWN_LOCATION and restore it at the end.
1961 2013-09-10 Martin Jambor <mjambor@suse.cz>
1963 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
1964 represented by a thunk.
1966 2013-09-10 Jeff Law <law@redhat.com>
1968 PR tree-optimization/58343
1969 * tree-ssa-threadupdate.c (thread_block): Identify and disable
1970 jump threading requests through loop headers buried in the middle
1971 of a jump threading path.
1973 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
1974 in return value/type.
1976 2013-09-10 Jakub Jelinek <jakub@redhat.com>
1978 PR rtl-optimization/58365
1979 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
1980 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
1983 2013-09-10 Richard Biener <rguenther@suse.de>
1985 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
1986 * tree-data-ref.c (create_rdg_vertices): Collect all data
1987 references, signal failure to the caller, use data-ref API.
1988 (build_rdg): Compute data references only once. Maintain lifetime
1989 of data references and data dependences from within RDG.
1990 (free_rdg): Free dependence relations.
1991 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
1992 inventing extra dependences.
1993 (distribute_loop): Update for RDG API changes.
1995 2013-09-10 Kai Tietz <ktietz@redhat.com>
1997 * doc/invoke.texi (fms-extensions): Document changed
1998 behavior for ms-abi targets.
1999 * config/i386/i386.c (ix86_option_override_internal):
2000 Set default value of option -fms-extension for ms-abi targets.
2002 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
2004 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
2006 2013-09-10 Alan Modra <amodra@gmail.com>
2009 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
2011 2013-09-10 Alan Modra <amodra@gmail.com>
2013 * config/rs6000/predicates.md (add_cint_operand): New.
2014 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
2015 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
2016 using add_cint_operand.
2017 (largetoc_high_plus_aix): Likewise.
2019 2013-09-09 Jakub Jelinek <jakub@redhat.com>
2021 PR tree-optimization/58364
2022 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
2023 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
2024 the current range can't be an unconditional true or false.
2026 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
2028 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
2030 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
2032 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
2034 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
2037 * doc/invoke.texi (-Wdelete-incomplete): Document it.
2039 2013-09-09 Ian Bolton <ian.bolton@arm.com>
2041 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
2042 NO_REGS for immediate that can't be moved directly into FP_REGS.
2044 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2046 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
2047 comparison with negated operand.
2048 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
2051 2013-09-09 Richard Biener <rguenther@suse.de>
2054 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
2055 of a subloop record all its block as affecting loop-closed SSA form.
2057 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
2059 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
2060 of an rtx/bitpos pair.
2061 (store_fixed_bit_field): Update accordingly.
2063 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
2065 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
2067 * builtins.c (expand_errno_check): Likewise.
2068 * dwarf2cfi.c (init_return_column_size): Likewise.
2069 * except.c (sjlj_mark_call_sites): Likewise.
2070 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
2071 * lra-constraints.c (emit_inc): Likewise.
2072 * ree.c (combine_set_extension): Likewise.
2073 * regmove.c (fixup_match_2): Likewise.
2074 * reload1.c (inc_for_reload): Likewise.
2076 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
2078 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
2079 (force_to_mode, simplify_shift_const_1, simplify_comparison):
2080 Use gen_int_mode with the mode of the associated simplify_* call.
2081 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
2082 * expmed.c (expand_shift_1): Likewise.
2083 * function.c (instantiate_virtual_regs_in_insn): Likewise.
2084 * loop-iv.c (iv_number_of_iterations): Likewise.
2085 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
2086 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
2088 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
2090 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
2091 of the associated expand_* call.
2092 (asan_emit_stack_protection): Likewise.
2093 * builtins.c (round_trampoline_addr): Likewise.
2094 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
2095 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
2096 (emit_store_flag): Likewise.
2097 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
2098 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
2100 * function.c (instantiate_virtual_regs_in_insn): Likewise.
2101 * ifcvt.c (noce_try_store_flag_constants): Likewise.
2102 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
2103 * modulo-sched.c (generate_prolog_epilog): Likewise.
2104 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
2105 (expand_ctz, expand_ffs, expand_unop): Likewise.
2107 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
2109 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
2110 of the associated gen_rtx_* call.
2111 * caller-save.c (init_caller_save): Likewise.
2112 * combine.c (find_split_point, make_extraction): Likewise.
2113 (make_compound_operation): Likewise.
2114 * dwarf2out.c (mem_loc_descriptor): Likewise.
2115 * explow.c (plus_constant, probe_stack_range): Likewise.
2116 * expmed.c (expand_mult_const): Likewise.
2117 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
2118 * reload1.c (init_reload): Likewise.
2119 * valtrack.c (cleanup_auto_inc_dec): Likewise.
2120 * var-tracking.c (adjust_mems): Likewise.
2121 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
2122 rather than gen_rtx_fmt_ee.
2124 2013-09-09 Jan Hubicka <jh@suse.cz>
2127 * value-prof.c (gimple_ic): Copy also abnormal edges.
2129 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
2131 * asan.c (asan_shadow_cst): Use gen_int_mode.
2133 2013-09-08 Jan Hubicka <jh@suse.cz>
2135 * ipa-profile.c: Add toplevel comment.
2136 (ipa_propagate_frequency_1): Be more conservative when profile is read.
2137 (contains_hot_call_p): New function.
2138 (ipa_propagate_frequency): Set frequencies based on counts when
2140 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
2141 profile; do not tamper with profile after inlining if it is read.
2143 2013-09-08 Jan Hubicka <jh@suse.cz>
2145 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
2148 2013-09-08 Jan Hubicka <jh@suse.cz>
2150 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
2153 2013-09-08 Jeff Law <law@redhat.com>
2156 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
2159 2013-09-08 Andi Kleen <ak@linux.intel.com>
2161 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
2163 2013-09-08 Jan Hubicka <jh@suse.cz>
2165 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
2167 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
2169 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
2170 for non-debug insns.
2171 * lra.c (new_insn_reg): Take the containing insn as a parameter.
2172 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
2173 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
2176 2013-09-08 Jan Hubicka <jh@suse.cz>
2178 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
2179 targets and devirtualize to BUILT_IN_UNREACHABLE.
2180 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
2181 * ipa.c (walk_polymorphic_call_targets): New function.
2182 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
2183 functions; use the new timevar.
2184 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
2185 was removed from the program.
2186 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
2187 not consider it in the walk when its vtable is dead.
2188 (possible_polymorphic_call_targets_1): Pass anonymous flag to
2190 (devirt_variable_node_removal_hook): New function.
2191 (possible_polymorphic_call_targets): Also register
2192 devirt_variable_node_removal_hook.
2193 (ipa_devirt): Do not do non-speculative devirtualization.
2194 (gate_ipa_devirt): One execute if devirtualizing speculatively.
2196 2013-09-08 Jan Hubicka <jh@suse.cz>
2198 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
2199 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
2200 varpool_remove_variable_insertion_hook): Declare.
2201 * varpool.c (varpool_node_hook_list): New structure.
2202 (first_varpool_node_removal_hook,
2203 first_varpool_variable_insertion_hook): New variables.
2204 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
2205 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
2206 varpool_remove_variable_insertion_hook,
2207 varpool_call_variable_insertion_hooks): New functions.
2208 (varpool_remove_node): Use it.
2210 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
2213 * diagnostic.c (diagnostic_build_prefix): When s.file is
2214 "<built-in>" don't output line and column numbers.
2216 2013-09-06 Jan Hubicka <jh@suse.cz>
2218 * cgraphunit.c (expand_thunk): Get body before touching arguments.
2219 * lto-streamer-out.c: Stream thunks, too.
2220 * lto-streamer-in.c (input_function): Pop cfun here
2221 (lto_read_body): Instead of here.
2223 2013-09-06 Caroline Tice <cmtice@google.com>
2225 * doc/install.texi: Add documentation for the --enable-vtable-verify
2226 and the --disable-libvtv configure options.
2228 2013-09-06 Jeff Law <law@redhat.com>
2230 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
2231 edge implied equivalences into successor phis.
2233 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
2235 * resource.c (mark_referenced_resources): Handle COND_EXEC.
2237 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
2239 * resource.c (mark_target_live_regs): Compute resources taking
2240 into account if a call is predicated or not.
2242 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
2244 * toplev.c (output_stack_usage): Be prepared for suffixes created by
2245 the compiler in the function names.
2247 2013-09-06 Jan Hubicka <jh@suse.cz>
2250 * ipa-inline.c (has_caller_p): New function.
2251 (want_inline_function_to_all_callers_p): Use it.
2252 (sum_callers, inline_to_all_callers): Break out from ...
2253 (ipa_inline): ... here.
2255 2013-09-06 Jan Hubicka <jh@suse.cz>
2257 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
2258 only when AVX is enabled.
2260 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2262 * config/aarch64/aarch64.md
2263 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
2264 is fpsimd_<load/store>2.
2265 (load_pair<mode>): Likewise.
2266 (store_pair<mode>): Likewise.
2268 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2270 * config/arm/types.md (type): Add "mrs" type.
2271 * config/aarch64/aarch64.md
2272 (aarch64_load_tp_hard): Make type "mrs".
2274 (load_tp_hard): Make type "mrs".
2275 * config/arm/cortex-a15.md: Update with new attributes.
2276 * config/arm/cortex-a5.md: Update with new attributes.
2277 * config/arm/cortex-a53.md: Update with new attributes.
2278 * config/arm/cortex-a7.md: Update with new attributes.
2279 * config/arm/cortex-a8.md: Update with new attributes.
2280 * config/arm/cortex-a9.md: Update with new attributes.
2281 * config/arm/cortex-m4.md: Update with new attributes.
2282 * config/arm/cortex-r4.md: Update with new attributes.
2283 * config/arm/fa526.md: Update with new attributes.
2284 * config/arm/fa606te.md: Update with new attributes.
2285 * config/arm/fa626te.md: Update with new attributes.
2286 * config/arm/fa726te.md: Update with new attributes.
2288 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2290 * config/aarch64/aarch64.md
2291 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
2292 (*movtf_aarch64): Likewise.
2294 (thumb1_movdi_insn): Use "multiple" for type where more than one
2295 instruction is used for a move.
2296 (*arm32_movhf): Likewise.
2297 (*thumb_movdf_insn): Likewise.
2299 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2301 * config/arm/types.md (type): Rename fcpys to fmov.
2303 (*arm_movsi_vfp): Rename type fcpys as fmov.
2304 (*thumb2_movsi_vfp): Likewise
2305 (*movhf_vfp_neon): Likewise
2306 (*movhf_vfp): Likewise
2307 (*movsf_vfp): Likewise
2308 (*thumb2_movsf_vfp): Likewise
2309 (*movsfcc_vfp): Likewise
2310 (*thumb2_movsfcc_vfp): Likewise
2311 * config/aarch64/aarch64-simd.md
2312 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
2313 * config/aarch64/aarch64.md
2314 (*movsi_aarch64): Replace type mov_reg with fmovs.
2315 (*movdi_aarch64): Likewise
2316 (*movsf_aarch64): Likewise
2317 (*movdf_aarch64): Likewise
2319 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
2320 * config/arm/iwmmxt.md
2321 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
2322 * config/arm/arm1020e.md: Update with new attributes.
2323 * config/arm/cortex-a15-neon.md: Update with new attributes.
2324 * config/arm/cortex-a5.md: Update with new attributes.
2325 * config/arm/cortex-a53.md: Update with new attributes.
2326 * config/arm/cortex-a7.md: Update with new attributes.
2327 * config/arm/cortex-a8-neon.md: Update with new attributes.
2328 * config/arm/cortex-a9.md: Update with new attributes.
2329 * config/arm/cortex-m4-fpu.md: Update with new attributes.
2330 * config/arm/cortex-r4f.md: Update with new attributes.
2331 * config/arm/marvell-pj4.md: Update with new attributes.
2332 * config/arm/vfp11.md: Update with new attributes.
2334 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2336 * config/aarch64/aarch64.md
2337 (*madd<mode>): Fix type attribute.
2338 (*maddsi_uxtw): Likewise.
2339 (*msub<mode>): Likewise.
2340 (*msubsi_uxtw): Likewise.
2341 (<su_optab>maddsidi4): Likewise.
2342 (<su_optab>msubsidi4): Likewise.
2344 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2346 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
2347 * config/arm/arm.md (core_cycles): Remove fdiv.
2348 * config/arm/vfp.md:
2349 (*sqrtsf2_vfp): Update for attribute changes.
2350 (*sqrtdf2_vfp): Likewise.
2351 * config/aarch64/aarch64.md:
2352 (sqrt<mode>2): Update for attribute changes.
2353 * config/arm/arm1020e.md: Update with new attributes.
2354 * config/arm/cortex-a15-neon.md: Update with new attributes.
2355 * config/arm/cortex-a5.md: Update with new attributes.
2356 * config/arm/cortex-a53.md: Update with new attributes.
2357 * config/arm/cortex-a7.md: Update with new attributes.
2358 * config/arm/cortex-a8-neon.md: Update with new attributes.
2359 * config/arm/cortex-a9.md: Update with new attributes.
2360 * config/arm/cortex-m4-fpu.md: Update with new attributes.
2361 * config/arm/cortex-r4f.md: Update with new attributes.
2362 * config/arm/marvell-pj4.md: Update with new attributes.
2363 * config/arm/vfp11.md: Update with new attributes.
2365 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2367 * config/arm/types.md
2368 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
2369 * config/aarch64/aarch64.md
2370 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
2372 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
2373 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
2374 (float<GPI:mode><GPF:mode>2): Likewise.
2376 (*truncsisf2_vfp): Update with new attributes.
2377 (*truncsidf2_vfp): Likewise.
2378 (fixuns_truncsfsi2): Likewise.
2379 (fixuns_truncdfsi2): Likewise.
2380 (*floatsisf2_vfp): Likewise.
2381 (*floatsidf2_vfp): Likewise.
2382 (floatunssisf2): Likewise.
2383 (floatunssidf2): Likewise.
2384 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
2385 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
2386 * config/arm/arm1020e.md: Update with new attributes.
2387 * config/arm/cortex-a15-neon.md: Update with new attributes.
2388 * config/arm/cortex-a5.md: Update with new attributes.
2389 * config/arm/cortex-a53.md: Update with new attributes.
2390 * config/arm/cortex-a7.md: Update with new attributes.
2391 * config/arm/cortex-a8-neon.md: Update with new attributes.
2392 * config/arm/cortex-a9.md: Update with new attributes.
2393 * config/arm/cortex-m4-fpu.md: Update with new attributes.
2394 * config/arm/cortex-r4f.md: Update with new attributes.
2395 * config/arm/marvell-pj4.md: Update with new attributes.
2396 * config/arm/vfp11.md: Update with new attributes.
2398 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2400 * config/aarch64/arm_neon.h
2401 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
2402 (vqtbx<1,2,3,4><q>_s8): Likewise.
2404 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2406 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
2407 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
2408 (add<mode>3): Add type attribute.
2409 (add<mode>3): Likewise.
2410 (usadd<mode>3): Likewise.
2411 (ssadd<mode>3): Likewise.
2412 (sub<mode>3): Likewise.
2413 (sub<mode>3): Likewise.
2414 (ussub<mode>3): Likewise.
2415 (sssub<mode>3): Likewise.
2416 (ssmulsa3): Likewise.
2417 (usmulusa3): Likewise.
2418 (arm_usatsihi): Likewise.
2420 (*movdi_vfp): Add types for all instructions.
2421 (*movdi_vfp_cortexa8): Likewise.
2422 (*movhf_vfp_neon): Likewise.
2423 (*movhf_vfp): Likewise.
2424 (*movdf_vfp): Likewise.
2425 (*thumb2_movdf_vfp): Likewise.
2426 (*thumb2_movdfcc_vfp): Likewise.
2427 * config/arm/arm.md: Add type attribute to all insn patterns.
2428 (*thumb1_adddi3): Add type attribute.
2429 (*arm_adddi3): Likewise.
2430 (*adddi_sesidi_di): Likewise.
2431 (*adddi_zesidi_di): Likewise.
2432 (*thumb1_addsi3): Likewise.
2433 (addsi3_compare0): Likewise.
2434 (*addsi3_compare0_scratch): Likewise.
2435 (*compare_negsi_si): Likewise.
2436 (cmpsi2_addneg): Likewise.
2437 (*addsi3_carryin_<optab>): Likewise.
2438 (*addsi3_carryin_alt2_<optab>): Likewise.
2439 (*addsi3_carryin_clobercc_<optab>): Likewise.
2440 (*subsi3_carryin): Likewise.
2441 (*subsi3_carryin_const): Likewise.
2442 (*subsi3_carryin_compare): Likewise.
2443 (*subsi3_carryin_compare_const): Likewise.
2444 (*arm_subdi3): Likewise.
2445 (*thumb_subdi3): Likewise.
2446 (*subdi_di_zesidi): Likewise.
2447 (*subdi_di_sesidi): Likewise.
2448 (*subdi_zesidi_di): Likewise.
2449 (*subdi_sesidi_di): Likewise.
2450 (*subdi_zesidi_ze): Likewise.
2451 (thumb1_subsi3_insn): Likewise.
2452 (*arm_subsi3_insn): Likewise.
2453 (*anddi3_insn): Likewise.
2454 (*anddi_zesidi_di): Likewise.
2455 (*anddi_sesdi_di): Likewise.
2456 (*ne_zeroextracts): Likewise.
2457 (*ne_zeroextracts): Likewise.
2458 (*ite_ne_zeroextr): Likewise.
2459 (*ite_ne_zeroextr): Likewise.
2460 (*anddi_notdi_di): Likewise.
2461 (*anddi_notzesidi): Likewise.
2462 (*anddi_notsesidi): Likewise.
2463 (andsi_notsi_si): Likewise.
2464 (thumb1_bicsi3): Likewise.
2465 (*iordi3_insn): Likewise.
2466 (*iordi_zesidi_di): Likewise.
2467 (*iordi_sesidi_di): Likewise.
2468 (*thumb1_iorsi3_insn): Likewise.
2469 (*xordi3_insn): Likewise.
2470 (*xordi_zesidi_di): Likewise.
2471 (*xordi_sesidi_di): Likewise.
2472 (*arm_xorsi3): Likewise.
2473 (*andsi_iorsi3_no): Likewise.
2474 (*smax_0): Likewise.
2475 (*smax_m1): Likewise.
2476 (*arm_smax_insn): Likewise.
2477 (*smin_0): Likewise.
2478 (*arm_smin_insn): Likewise.
2479 (*arm_umaxsi3): Likewise.
2480 (*arm_uminsi3): Likewise.
2481 (*minmax_arithsi): Likewise.
2482 (*minmax_arithsi_): Likewise.
2483 (*satsi_<SAT:code>): Likewise.
2484 (arm_ashldi3_1bit): Likewise.
2485 (arm_ashrdi3_1bit): Likewise.
2486 (arm_lshrdi3_1bit): Likewise.
2487 (*arm_negdi2): Likewise.
2488 (*thumb1_negdi2): Likewise.
2489 (*arm_negsi2): Likewise.
2490 (*thumb1_negsi2): Likewise.
2491 (*negdi_extendsid): Likewise.
2492 (*negdi_zero_extend): Likewise.
2493 (*arm_abssi2): Likewise.
2494 (*thumb1_abssi2): Likewise.
2495 (*arm_neg_abssi2): Likewise.
2496 (*thumb1_neg_abss): Likewise.
2497 (one_cmpldi2): Likewise.
2498 (extend<mode>di2): Likewise.
2499 (*compareqi_eq0): Likewise.
2500 (*arm_extendhisi2addsi): Likewise.
2501 (*arm_movdi): Likewise.
2502 (*thumb1_movdi_insn): Likewise.
2503 (*arm_movt): Likewise.
2504 (*thumb1_movsi_insn): Likewise.
2505 (pic_add_dot_plus_four): Likewise.
2506 (pic_add_dot_plus_eight): Likewise.
2507 (tls_load_dot_plus_eight): Likewise.
2508 (*thumb1_movhi_insn): Likewise.
2509 (*thumb1_movsf_insn): Likewise.
2510 (*movdf_soft_insn): Likewise.
2511 (*thumb_movdf_insn): Likewise.
2512 (cbranchsi4_insn): Likewise.
2513 (cbranchsi4_scratch): Likewise.
2514 (*negated_cbranchsi4): Likewise.
2515 (*tbit_cbranch): Likewise.
2516 (*tlobits_cbranch): Likewise.
2517 (*tstsi3_cbranch): Likewise.
2518 (*cbranchne_decr1): Likewise.
2519 (*addsi3_cbranch): Likewise.
2520 (*addsi3_cbranch_scratch): Likewise.
2521 (*arm_cmpdi_insn): Likewise.
2522 (*arm_cmpdi_unsig): Likewise.
2523 (*arm_cmpdi_zero): Likewise.
2524 (*thumb_cmpdi_zero): Likewise.
2525 (*deleted_compare): Likewise.
2526 (*mov_scc): Likewise.
2527 (*mov_negscc): Likewise.
2528 (*mov_notscc): Likewise.
2529 (*cstoresi_eq0_thumb1_insn): Likewise.
2530 (cstoresi_nltu_thumb1): Likewise.
2531 (cstoresi_ltu_thu): Likewise.
2532 (thumb1_addsi3_addgeu): Likewise.
2533 (*arm_jump): Likewise.
2534 (*thumb_jump): Likewise.
2535 (*check_arch2): Likewise.
2536 (arm_casesi_internal): Likewise.
2537 (thumb1_casesi_dispatch): Likewise.
2538 (*arm_indirect_jump): Likewise.
2539 (*thumb1_indirect_jump): Likewise.
2541 (*and_scc): Likewise.
2542 (*ior_scc): Likewise.
2543 (*compare_scc): Likewise.
2544 (*cond_move): Likewise.
2545 (*cond_arith): Likewise.
2546 (*cond_sub): Likewise.
2547 (*cmp_ite0): Likewise.
2548 (*cmp_ite1): Likewise.
2549 (*cmp_and): Likewise.
2550 (*cmp_ior): Likewise.
2551 (*ior_scc_scc): Likewise.
2552 (*ior_scc_scc_cmp): Likewise.
2553 (*and_scc_scc): Likewise.
2554 (*and_scc_scc_cmp): Likewise.
2555 (*and_scc_scc_nod): Likewise.
2556 (*negscc): Likewise.
2557 (movcond_addsi): Likewise.
2558 (movcond): Likewise.
2559 (*ifcompare_plus_move): Likewise.
2560 (*if_plus_move): Likewise.
2561 (*ifcompare_move_plus): Likewise.
2562 (*if_move_plus): Likewise.
2563 (*ifcompare_arith_arith): Likewise.
2564 (*if_arith_arith): Likewise.
2565 (*ifcompare_arith_move): Likewise.
2566 (*if_arith_move): Likewise.
2567 (*ifcompare_move_arith): Likewise.
2568 (*if_move_arith): Likewise.
2569 (*ifcompare_move_not): Likewise.
2570 (*if_move_not): Likewise.
2571 (*ifcompare_not_move): Likewise.
2572 (*if_not_move): Likewise.
2573 (*ifcompare_shift_move): Likewise.
2574 (*if_shift_move): Likewise.
2575 (*ifcompare_move_shift): Likewise.
2576 (*if_move_shift): Likewise.
2577 (*ifcompare_shift_shift): Likewise.
2578 (*ifcompare_not_arith): Likewise.
2579 (*ifcompare_arith_not): Likewise.
2580 (*if_arith_not): Likewise.
2581 (*ifcompare_neg_move): Likewise.
2582 (*if_neg_move): Likewise.
2583 (*ifcompare_move_neg): Likewise.
2584 (*if_move_neg): Likewise.
2585 (prologue_thumb1_interwork): Likewise.
2586 (*cond_move_not): Likewise.
2587 (*sign_extract_onebit): Likewise.
2588 (*not_signextract_onebit): Likewise.
2589 (stack_tie): Likewise.
2590 (align_4): Likewise.
2591 (align_8): Likewise.
2592 (consttable_end): Likewise.
2593 (consttable_1): Likewise.
2594 (consttable_2): Likewise.
2595 (consttable_4): Likewise.
2596 (consttable_8): Likewise.
2597 (consttable_16): Likewise.
2598 (*thumb1_tablejump): Likewise.
2599 (prefetch): Likewise.
2600 (force_register_use): Likewise.
2601 (thumb_eh_return): Likewise.
2602 (load_tp_hard): Likewise.
2603 (load_tp_soft): Likewise.
2604 (tlscall): Likewise.
2605 (*arm_movtas_ze): Likewise.
2606 (*arm_rev): Likewise.
2607 (*arm_revsh): Likewise.
2608 (*arm_rev16): Likewise.
2609 * config/arm/thumb2.md
2610 (*thumb2_smaxsi3): Likewise.
2611 (*thumb2_sminsi3): Likewise.
2612 (*thumb32_umaxsi3): Likewise.
2613 (*thumb2_uminsi3): Likewise.
2614 (*thumb2_negdi2): Likewise.
2615 (*thumb2_abssi2): Likewise.
2616 (*thumb2_neg_abss): Likewise.
2617 (*thumb2_movsi_insn): Likewise.
2618 (tls_load_dot_plus_four): Likewise.
2619 (*thumb2_movhi_insn): Likewise.
2620 (*thumb2_mov_scc): Likewise.
2621 (*thumb2_mov_negs): Likewise.
2622 (*thumb2_mov_negs): Likewise.
2623 (*thumb2_mov_nots): Likewise.
2624 (*thumb2_mov_nots): Likewise.
2625 (*thumb2_movsicc_): Likewise.
2626 (*thumb2_movsfcc_soft_insn): Likewise.
2627 (*thumb2_indirect_jump): Likewise.
2628 (*thumb2_and_scc): Likewise.
2629 (*thumb2_ior_scc): Likewise.
2630 (*thumb2_ior_scc_strict_it): Likewise.
2631 (*thumb2_cond_move): Likewise.
2632 (*thumb2_cond_arith): Likewise.
2633 (*thumb2_cond_ari): Likewise.
2634 (*thumb2_cond_sub): Likewise.
2635 (*thumb2_negscc): Likewise.
2636 (*thumb2_movcond): Likewise.
2637 (thumb2_casesi_internal): Likewise.
2638 (thumb2_casesi_internal_pic): Likewise.
2639 (*thumb2_alusi3_short): Likewise.
2640 (*thumb2_mov<mode>_shortim): Likewise.
2641 (*thumb2_addsi_short): Likewise.
2642 (*thumb2_subsi_short): Likewise.
2643 (thumb2_addsi3_compare0): Likewise.
2644 (*thumb2_cbz): Likewise.
2645 (*thumb2_cbnz): Likewise.
2646 (*thumb2_one_cmplsi2_short): Likewise.
2647 (*thumb2_negsi2_short): Likewise.
2648 (*orsi_notsi_si): Likewise.
2649 * config/arm/arm1020e.md: Update with new attributes.
2650 * config/arm/arm1026ejs.md: Update with new attributes.
2651 * config/arm/arm1136jfs.md: Update with new attributes.
2652 * config/arm/arm926ejs.md: Update with new attributes.
2653 * config/arm/cortex-a15.md: Update with new attributes.
2654 * config/arm/cortex-a5.md: Update with new attributes.
2655 * config/arm/cortex-a53.md: Update with new attributes.
2656 * config/arm/cortex-a7.md: Update with new attributes.
2657 * config/arm/cortex-a8.md: Update with new attributes.
2658 * config/arm/cortex-a9.md: Update with new attributes.
2659 * config/arm/cortex-m4.md: Update with new attributes.
2660 * config/arm/cortex-r4.md: Update with new attributes.
2661 * config/arm/fa526.md: Update with new attributes.
2662 * config/arm/fa606te.md: Update with new attributes.
2663 * config/arm/fa626te.md: Update with new attributes.
2664 * config/arm/fa726te.md: Update with new attributes.
2666 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2668 * config/aarch64/aarch64-simd.md
2669 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
2670 <vwx> iterator to ensure correct register choice.
2671 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
2672 (aarch64_sqdmull_n<mode>): Likewise.
2673 (aarch64_sqdmull2_n<mode>_internal): Likewise.
2674 * config/aarch64/arm_neon.h
2675 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
2676 (vml<as><q>_n_<su>16): Likewise.
2677 (vml<as>l_high_lane<q>_<su>16): Likewise.
2678 (vml<as>l_high_n_<su>16): Likewise.
2679 (vml<as>l_lane<q>_<su>16): Likewise.
2680 (vml<as>l_n_<su>16): Likewise.
2681 (vmul<q>_lane<q>_<su>16): Likewise.
2682 (vmul<q>_n_<su>16): Likewise.
2683 (vmull_lane<q>_<su>16): Likewise.
2684 (vmull_n_<su>16): Likewise.
2685 (vmull_high_lane<q>_<su>16): Likewise.
2686 (vmull_high_n_<su>16): Likewise.
2687 (vqrdmulh<q>_n_s16): Likewise.
2689 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
2691 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
2692 have the correct lane parameter.
2694 2013-09-06 Richard Biener <rguenther@suse.de>
2696 * cfganal.c (control_dependences::~control_dependences):
2697 Properly free all of the vector.
2699 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
2702 * config/i386/i386.c (ix86_conditional_register_usage):
2703 Proper initialize extended SSE registers.
2705 2013-09-06 Jan Hubicka <jh@suse.cz>
2707 PR tree-optimization/58311
2708 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
2710 2013-09-06 Jan Hubicka <jh@suse.cz>
2712 * Makefile.in (tree-sra.o): Update dependencies.
2713 * tree-sra.c: Include ipa-utils.h
2714 (scan_function): Use recursive_call_p.
2715 (has_caller_p): New function.
2716 (cgraph_for_node_and_aliases): Count also callers of aliases.
2718 2013-09-06 Jan Hubicka <jh@suse.cz>
2721 * cgraph.h (symtab_semantically_equivalent_p): Declare.
2722 * tree-tailcall.c: Include ipa-utils.h.
2723 (find_tail_calls): Use it.
2724 * ipa-pure-const.c (check_call): Likewise.
2725 * ipa-utils.c (recursive_call_p): New function.
2726 * ipa-utils.h (recursive_call_p): Dclare.
2727 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
2728 (symtab_semantically_equivalent_p): New function.
2729 * Makefile.in (tree-tailcall.o): Update dependencies.
2731 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
2733 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
2736 2013-09-06 Richard Biener <rguenther@suse.de>
2738 * lto-streamer.h (lto_global_var_decls): Remove.
2739 * Makefile.in (OBJS): Remove lto-symtab.o.
2740 (lto-symtab.o): Remove.
2741 (GTFILES): Remove lto-symtab.c
2742 * lto-symtab.c: Move to lto/
2744 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2746 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
2747 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
2748 (UNSPEC_FPINT_RINT): New constant definitions.
2749 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
2750 definition with 2 attributes.
2751 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
2752 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
2755 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2757 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
2758 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
2759 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
2760 attribute to "z196_alone".
2761 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
2764 2013-09-06 Richard Biener <rguenther@suse.de>
2766 * basic-block.h (class control_dependences): New.
2767 * tree-ssa-dce.c (control_dependence_map): Remove.
2769 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
2770 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
2771 find_pdom, find_control_dependence, find_all_control_dependences):
2773 (mark_control_dependent_edges_necessary,
2774 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
2775 tree_dce_done, perform_tree_ssa_dce): Adjust.
2776 * cfganal.c (set_control_dependence_map_bit,
2777 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
2778 find_all_control_dependences): Move from tree-ssa-dce.c and
2779 implement as methods of control_dependences class.
2780 (control_dependences::control_dependences): New.
2781 (control_dependences::~control_dependences): Likewise.
2782 (control_dependences::get_edges_dependent_on): Likewise.
2783 (control_dependences::get_edge): Likewise.
2785 2013-09-04 Jan Hubicka <jh@suse.cz>
2787 * tree.c (types_same_for_odr): Drop overactive check.
2788 * ipa-devirt.c (hash_type_name): Likewise.
2790 2013-09-04 Jan Hubicka <jh@suse.cz>
2792 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
2793 (analyze_functions): ... here.
2795 2013-09-04 Jan Hubicka <jh@suse.cz>
2798 * cgraphunit.c (analyze_functions): Clear AUX fields
2799 after processing; initialize assembler name has.
2801 2013-09-05 Jeff Law <law@redhat.com>
2803 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
2804 from thread_around_empty_block. Record threading path into PATH.
2805 Recurse if threading through the initial block is successful.
2806 (thread_across_edge): Corresponding changes to slightly simplify.
2808 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
2810 * config/aarch64/aarch64.md
2811 (type): Remove frecpe, frecps, frecpx.
2812 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
2813 fix to be a TARGET_SIMD instruction.
2814 (aarch64_frecps): Remove.
2815 * config/aarch64/aarch64-simd.md
2816 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
2817 (aarch64_frecps<mode>): Handle all float/vector of float modes.
2819 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
2820 Sofiane Naci <sofiane.naci@arm.com>
2822 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
2823 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
2824 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
2825 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
2826 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
2827 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
2828 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
2829 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
2830 into "clz, "rbit". Rename "shift" to "shift_imm".
2831 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
2832 changes. Update for attribute changes all occurrences of arlo_* and
2834 * config/arm/arm-fixed.md: Update for attribute changes
2835 all occurrences of arlo_* types.
2836 * config/arm/thumb2.md: Update for attribute changes all occurrences
2838 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
2839 (cortexa7_older_only): Likewise.
2840 (cortexa7_younger): Likewise.
2841 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
2842 (1020alu_shift_op): Likewise.
2843 (1020alu_shift_reg_op): Likewise.
2844 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
2845 (alu_shift_op): Likewise.
2846 (alu_shift_reg_op): Likewise.
2847 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
2848 (11_alu_shift_op): Likewise.
2849 (11_alu_shift_reg_op): Likewise.
2850 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
2851 (9_alu_shift_reg_op): Likewise.
2852 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
2854 (cortex_a15_alu_shift): Likewise.
2855 (cortex_a15_alu_shift_reg): Likewise.
2856 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
2858 (cortex_a5_alu_shift): Likewise.
2859 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
2861 (cortex_a53_alu_shift): Likewise.
2862 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
2864 (cortex_a7_alu_reg): Likewise.
2865 (cortex_a7_alu_shift): Likewise.
2866 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
2868 (cortex_a8_alu_shift): Likewise.
2869 (cortex_a8_alu_shift_reg): Likewise.
2870 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
2871 (cortex_a9_dp_shift): Likewise.
2872 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
2874 * config/arm/cortex-r4.md
2875 (cortex_r4_alu): Update for attribute changes.
2876 (cortex_r4_mov): Likewise.
2877 (cortex_r4_alu_shift_reg): Likewise.
2878 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
2879 (526_alu_shift_op): Likewise.
2880 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
2881 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
2882 (626te_alu_shift_op): Likewise.
2883 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
2884 (726te_alu_shift_op): Likewise.
2885 (726te_alu_shift_reg_op): Likewise.
2886 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
2887 (mp626_alu_shift_op): Likewise.
2888 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
2889 (pj4_alu_conds): Likewise.
2890 (pj4_shift): Likewise.
2891 (pj4_shift_conds): Likewise.
2892 (pj4_alu_shift): Likewise.
2893 (pj4_alu_shift_conds): Likewise.
2894 * config/aarch64/aarch64.md: Update for attribute change
2895 all occurrences of arlo_* and shift* types.
2897 2013-09-05 Mike Stump <mikestump@comcast.net>
2899 * tree.h: Move documentation for tree_function_decl to tree-core.h
2900 with the declaration.
2902 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
2905 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
2906 looking for widest mode.
2908 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
2910 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
2912 2013-09-05 Richard Biener <rguenther@suse.de>
2914 PR tree-optimization/58137
2915 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
2916 Do not create vectors of pointers.
2917 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
2918 types for the components of the vector initializer.
2919 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
2920 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
2922 2013-09-05 Martin Jambor <mjambor@suse.cz>
2924 * ipa-prop.c (remove_described_reference): Accept missing references,
2925 return false if that hppens, otherwise return true.
2926 (cgraph_node_for_jfunc): New function.
2927 (try_decrement_rdesc_refcount): Likewise.
2928 (try_make_edge_direct_simple_call): Use them.
2929 (ipa_edge_removal_hook): Remove references from rdescs.
2930 (ipa_edge_duplication_hook): Clone rdescs and their references
2931 when the new edge has the same caller as the old one.
2932 * cgraph.c (cgraph_resolve_speculation): Remove speculative
2933 reference before removing any edges.
2935 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
2937 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
2939 * thumb2.md (thumb2_storewb_parisi): New pattern.
2941 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
2943 * config/aarch64/aarch64-option-extensions.def: Add
2944 AARCH64_OPT_EXTENSION of 'crc'.
2945 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
2946 (AARCH64_ISA_CRC): Ditto.
2947 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
2948 description of the CRC extension.
2950 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
2952 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
2953 * config/rs6000/linux.h: Ditto.
2954 * alpha/linux.h: Ditto.
2955 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
2956 no_c99_libc_has_function.
2957 * config/c6x/uclinux-elf.h: Ditto.
2958 * config/lm32/uclinux-elf.h: Ditto.
2959 * config/m68k/uclinux.h: Ditto.
2960 * config/moxie/uclinux.h: Ditto.
2961 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
2962 (crisv32-*-linux*, cris-*-linux*): Ditto.
2963 * config/bfin/bfin.c: Include "tm_p.h".
2965 2013-09-05 Richard Biener <rguenther@suse.de>
2967 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
2968 check for a definition without a basic-block.
2970 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
2971 Sofiane Naci <sofiane.naci@arm.com>
2973 * config/aarch64/aarch64.md
2974 (*movti_aarch64): Rename r_2_f and f_2_r.
2975 (*movsf_aarch64): Likewise.
2976 (*movdf_aarch64): Likewise.
2977 (*movtf_aarch64): Likewise.
2978 (aarch64_movdi_<mode>low): Likewise.
2979 (aarch64_movdi_<mode>high): Likewise.
2980 (aarch64_mov<mode>high_di): Likewise.
2981 (aarch64_mov<mode>low_di): Likewise.
2982 (aarch64_movtilow_tilow): Likewise.
2983 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
2984 values to config/arm/types.md
2985 (attribute "conds"): Update for attribute change.
2986 (anddi3_insn): Likewise.
2987 (iordi3_insn): Likewise.
2988 (xordi3_insn): Likewise.
2989 (one_cmpldi2): Likewise.
2990 * config/arm/types.md (type): Add Neon types.
2991 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
2992 use "type" attribute.
2993 (movmisalign<mode>_neon_store): Likewise.
2994 (movmisalign<mode>_neon_load): Likewise.
2995 (vec_set<mode>_internal): Likewise.
2996 (vec_setv2di_internal): Likewise.
2997 (vec_extract<mode>): Likewise.
2998 (vec_extractv2di): Likewise.
2999 (add<mode>3_neon): Likewise.
3000 (adddi3_neon): Likewise.
3001 (sub<mode>3_neon): Likewise.
3002 (subdi3_neon): Likewise.
3003 (mul<mode>3_neon): Likewise.
3004 (mul<mode>3add<mode>_neon): Likewise.
3005 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
3006 (fma<VCVTF:mode>4)): Likewise.
3007 (fma<VCVTF:mode>4_intrinsic): Likewise.
3008 (fmsub<VCVTF:mode>4)): Likewise.
3009 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
3010 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
3011 (ior<mode>3): Likewise.
3012 (and<mode>3): Likewise.
3013 (anddi3_neon): Likewise.
3014 (orn<mode>3_neon): Likewise.
3015 (orndi3_neon): Likewise.
3016 (bic<mode>3_neon): Likewise.
3017 (bicdi3_neon): Likewise.
3018 (xor<mode>3): Likewise.
3019 (one_cmpl<mode>2): Likewise.
3020 (abs<mode>2): Likewise.
3021 (neg<mode>2): Likewise.
3022 (umin<mode>3_neon): Likewise.
3023 (umax<mode>3_neon): Likewise.
3024 (smin<mode>3_neon): Likewise.
3025 (smax<mode>3_neon): Likewise.
3026 (vashl<mode>3): Likewise.
3027 (vashr<mode>3_imm): Likewise.
3028 (vlshr<mode>3_imm): Likewise.
3029 (ashl<mode>3_signed): Likewise.
3030 (ashl<mode>3_unsigned): Likewise.
3031 (neon_load_count): Likewise.
3032 (ashldi3_neon_noclobber): Likewise.
3033 (signed_shift_di3_neon): Likewise.
3034 (unsigned_shift_di3_neon): Likewise.
3035 (ashrdi3_neon_imm_noclobber): Likewise.
3036 (lshrdi3_neon_imm_noclobber): Likewise.
3037 (widen_ssum<mode>3): Likewise.
3038 (widen_usum<mode>3): Likewise.
3039 (quad_halves_<code>v4si): Likewise.
3040 (quad_halves_<code>v4sf): Likewise.
3041 (quad_halves_<code>v8hi): Likewise.
3042 (quad_halves_<code>v16qi): Likewise.
3043 (reduc_splus_v2di): Likewise.
3044 (neon_vpadd_internal<mode>): Likewise.
3045 (neon_vpsmin<mode>): Likewise.
3046 (neon_vpsmax<mode>): Likewise.
3047 (neon_vpumin<mode>): Likewise.
3048 (neon_vpumax<mode>): Likewise.
3049 (ss_add<mode>_neon): Likewise.
3050 (us_add<mode>_neon): Likewise.
3051 (ss_sub<mode>_neon): Likewise.
3052 (us_sub<mode>_neon): Likewise.
3053 (neon_vadd<mode>_unspec): Likewise.
3054 (neon_vaddl<mode>): Likewise.
3055 (neon_vaddw<mode>): Likewise.
3056 (neon_vhadd<mode>): Likewise.
3057 (neon_vqadd<mode>): Likewise.
3058 (neon_vaddhn<mode>): Likewise.
3059 (neon_vmul<mode>): Likewise.
3060 (neon_vmla<mode>): Likewise.
3061 (neon_vmlal<mode>): Likewise.
3062 (neon_vmls<mode>): Likewise.
3063 (neon_vmlsl<mode>): Likewise.
3064 (neon_vqdmulh<mode>): Likewise.
3065 (neon_vqdmlal<mode>): Likewise.
3066 (neon_vqdmlsl<mode>): Likewise.
3067 (neon_vmull<mode>): Likewise.
3068 (neon_vqdmull<mode>): Likewise.
3069 (neon_vsub<mode>_unspec): Likewise.
3070 (neon_vsubl<mode>): Likewise.
3071 (neon_vsubw<mode>): Likewise.
3072 (neon_vqsub<mode>): Likewise.
3073 (neon_vhsub<mode>): Likewise.
3074 (neon_vsubhn<mode>): Likewise.
3075 (neon_vceq<mode>): Likewise.
3076 (neon_vcge<mode>): Likewise.
3077 (neon_vcgeu<mode>): Likewise.
3078 (neon_vcgt<mode>): Likewise.
3079 (neon_vcgtu<mode>): Likewise.
3080 (neon_vcle<mode>): Likewise.
3081 (neon_vclt<mode>): Likewise.
3082 (neon_vcage<mode>): Likewise.
3083 (neon_vcagt<mode>): Likewise.
3084 (neon_vtst<mode>): Likewise.
3085 (neon_vabd<mode>): Likewise.
3086 (neon_vabdl<mode>): Likewise.
3087 (neon_vaba<mode>): Likewise.
3088 (neon_vabal<mode>): Likewise.
3089 (neon_vmax<mode>): Likewise.
3090 (neon_vmin<mode>): Likewise.
3091 (neon_vpaddl<mode>): Likewise.
3092 (neon_vpadal<mode>): Likewise.
3093 (neon_vpmax<mode>): Likewise.
3094 (neon_vpmin<mode>): Likewise.
3095 (neon_vrecps<mode>): Likewise.
3096 (neon_vrsqrts<mode>): Likewise.
3097 (neon_vqabs<mode>): Likewise.
3098 (neon_vqneg<mode>): Likewise.
3099 (neon_vcls<mode>): Likewise.
3100 (clz<mode>2): Likewise.
3101 (popcount<mode>2): Likewise.
3102 (neon_vrecpe): Likewise.
3103 (neon_vrsqrte): Likewise.
3104 (neon_vget_lane<mode>_sext_internal): Likewise.
3105 (neon_vget_lane<mode>_zext_internal): Likewise.
3106 (neon_vdup_n<mode>): Likewise.
3107 (neon_vdup_nv2di): Likewise.
3108 (neon_vdpu_lane<mode>_internal): Likewise.
3109 (neon_vswp<mode>): Likewise.
3110 (float<mode><V_cvtto>2): Likewise.
3111 (floatuns<mode><V_cvtto>2): Likewise.
3112 (fix_trunc<mode><V_cvtto>)2): Likewise
3113 (fixuns_trunc<mode><V_cvtto)2): Likewise.
3114 (neon_vcvt<mode>): Likewise.
3115 (neon_vcvtv4sfv4hf): Likewise.
3116 (neon_vcvtv4hfv4sf): Likewise.
3117 (neon_vcvt_n<mode>): Likewise.
3118 (neon_vmovn<mode>): Likewise.
3119 (neon_vqmovn<mode>): Likewise.
3120 (neon_vqmovun<mode>): Likewise.
3121 (neon_vmovl<mode>): Likewise.
3122 (neon_vmul_lane<mode>): Likewise.
3123 (neon_vmull_lane<mode>): Likewise.
3124 (neon_vqdmull_lane<mode>): Likewise.
3125 (neon_vqdmulh_lane<mode>): Likewise.
3126 (neon_vmla_lane<mode>): Likewise.
3127 (neon_vmlal_lane<mode>): Likewise.
3128 (neon_vqdmlal_lane<mode>): Likewise.
3129 (neon_vmls_lane<mode>): Likewise.
3130 (neon_vmlsl_lane<mode>): Likewise.
3131 (neon_vqdmlsl_lane<mode>): Likewise.
3132 (neon_vext<mode>): Likewise.
3133 (neon_vrev64<mode>): Likewise.
3134 (neon_vrev32<mode>): Likewise.
3135 (neon_vrev16<mode>): Likewise.
3136 (neon_vbsl<mode>_internal): Likewise.
3137 (neon_vshl<mode>): Likewise.
3138 (neon_vqshl<mode>): Likewise.
3139 (neon_vshr_n<mode>): Likewise.
3140 (neon_vshrn_n<mode>): Likewise.
3141 (neon_vqshrn_n<mode>): Likewise.
3142 (neon_vqshrun_n<mode>): Likewise.
3143 (neon_vshl_n<mode>): Likewise.
3144 (neon_vqshl_n<mode>): Likewise.
3145 (neon_vqshlu_n<mode>): Likewise.
3146 (neon_vshll_n<mode>): Likewise.
3147 (neon_vsra_n<mode>): Likewise.
3148 (neon_vsri_n<mode>): Likewise.
3149 (neon_vsli_n<mode>): Likewise.
3150 (neon_vtbl1v8qi): Likewise.
3151 (neon_vtbl2v8qi): Likewise.
3152 (neon_vtbl3v8qi): Likewise.
3153 (neon_vtbl4v8qi): Likewise.
3154 (neon_vtbx1v8qi): Likewise.
3155 (neon_vtbx2v8qi): Likewise.
3156 (neon_vtbx3v8qi): Likewise.
3157 (neon_vtbx4v8qi): Likewise.
3158 (neon_vtrn<mode>_internal): Likewise.
3159 (neon_vzip<mode>_internal): Likewise.
3160 (neon_vuzp<mode>_internal): Likewise.
3161 (neon_vld1<mode>): Likewise.
3162 (neon_vld1_lane<mode>): Likewise.
3163 (neon_vld1_dup<mode>): Likewise.
3164 (neon_vld1_dupv2di): Likewise.
3165 (neon_vst1<mode>): Likewise.
3166 (neon_vst1_lane<mode>): Likewise.
3167 (neon_vld2<mode>): Likewise.
3168 (neon_vld2_lane<mode>): Likewise.
3169 (neon_vld2_dup<mode>): Likewise.
3170 (neon_vst2<mode>): Likewise.
3171 (neon_vst2_lane<mode>): Likewise.
3172 (neon_vld3<mode>): Likewise.
3173 (neon_vld3qa<mode>): Likewise.
3174 (neon_vld3qb<mode>): Likewise.
3175 (neon_vld3_lane<mode>): Likewise.
3176 (neon_vld3_dup<mode>): Likewise.
3177 (neon_vst3<mode>): Likewise.
3178 (neon_vst3qa<mode>): Likewise.
3179 (neon_vst3qb<mode>): Likewise.
3180 (neon_vst3_lane<mode>): Likewise.
3181 (neon_vld4<mode>): Likewise.
3182 (neon_vld4qa<mode>): Likewise.
3183 (neon_vld4qb<mode>): Likewise.
3184 (neon_vld4_lane<mode>): Likewise.
3185 (neon_vld4_dup<mode>): Likewise.
3186 (neon_vst4<mode>): Likewise.
3187 (neon_vst4qa<mode>): Likewise.
3188 (neon_vst4qb<mode>): Likewise.
3189 (neon_vst4_lane<mode>): Likewise.
3190 (neon_vec_unpack<US>_lo_<mode>): Likewise.
3191 (neon_vec_unpack<US>_hi_<mode>): Likewise.
3192 (neon_vec_<US>mult_lo_<mode>): Likewise.
3193 (neon_vec_<US>mult_hi_<mode>): Likewise.
3194 (neon_vec_<US>shiftl_<mode>): Likewise.
3195 (neon_unpack<US>_<mode>): Likewise.
3196 (neon_vec_<US>mult_<mode>): Likewise.
3197 (vec_pack_trunc_<mode>): Likewise.
3198 (neon_vec_pack_trunk_<mode>): Likewise.
3199 (neon_vabd<mode>_2): Likewise.
3200 (neon_vabd<mode>_3): Likewise.
3201 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
3202 (thumb2_movsi_vfp): Likewise.
3203 (movdi_vfp): Likewise.
3204 (movdi_vfp_cortexa8): Likewise.
3205 (movhf_vfp_neon): Likewise.
3206 (movhf_vfp): Likewiwse.
3207 (movsf_vfp): Likewiwse.
3208 (thumb2_movsf_vfp): Likewiwse.
3209 (movdf_vfp): Likewise.
3210 (thumb2_movdf_vfp): Likewise.
3211 (movsfcc_vfp): Likewise.
3212 (thumb2_movsfcc_vfp): Likewise.
3213 (movdfcc_vfp): Likewise.
3214 (thumb2_movdfcc_vfp): Likewise.
3215 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
3216 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
3217 (v10_v2c): Likewise.
3218 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
3220 (cortex_a15_neon_int_2): Likewise.
3221 (cortex_a15_neon_int_3): Likewise.
3222 (cortex_a15_neon_int_4): Likewise.
3223 (cortex_a15_neon_int_5): Likewise.
3224 (cortex_a15_neon_vqneg_vqabs): Likewise.
3225 (cortex_a15_neon_vmov): Likewise.
3226 (cortex_a15_neon_vaba): Likewise.
3227 (cortex_a15_neon_vaba_qqq): Likewise.
3228 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3229 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
3230 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
3231 scalar_64_32_long_scalar): Likewise.
3232 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3233 (cortex_a15_neon_mla_qqq_8_16): Likewise.
3234 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
3235 lotype_qdd_64_32_long): Likewise.
3236 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
3237 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
3238 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
3239 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
3240 (cortex_a15_neon_shift_1): Likewise.
3241 (cortex_a15_neon_shift_2): Likewise.
3242 (cortex_a15_neon_shift_3): Likewise.
3243 (cortex_a15_neon_vshl_ddd): Likewise.
3244 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
3245 (cortex_a15_neon_vsra_vrsra): Likewise.
3246 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
3247 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
3248 (cortex_a15_neon_fp_vmul_ddd): Likewise.
3249 (cortex_a15_neon_fp_vmul_qqd): Likewise.
3250 (cortex_a15_neon_fp_vmla_ddd): Likewise.
3251 (cortex_a15_neon_fp_vmla_qqq): Likewise.
3252 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
3253 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
3254 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
3255 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
3256 (cortex_a15_neon_bp_simple): Likewise.
3257 (cortex_a15_neon_bp_2cycle): Likewise.
3258 (cortex_a15_neon_bp_3cycle): Likewise.
3259 (cortex_a15_neon_vld1_1_2_regs): Likewise.
3260 (cortex_a15_neon_vld1_3_4_regs): Likewise.
3261 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
3262 (cortex_a15_neon_vld2_4_regs): Likewise.
3263 (cortex_a15_neon_vld3_vld4): Likewise.
3264 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
3265 (cortex_a15_neon_vst1_3_4_regs): Likewise.
3266 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
3267 (cortex_a15_neon_vst3_vst4): Likewise.
3268 (cortex_a15_neon_vld1_vld2_lane): Likewise.
3269 (cortex_a15_neon_vld3_vld4_lane" 10
3270 (cortex_a15_neon_vst1_vst2_lane): Likewise.
3271 (cortex_a15_neon_vst3_vst4_lane): Likewise.
3272 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
3273 (cortex_a15_neon_ldm_2): Likewise.0
3274 (cortex_a15_neon_stm_2): Likewise.
3275 (cortex_a15_neon_mcr): Likewise.
3276 (cortex_a15_neon_mcr_2_mcrr): Likewise.
3277 (cortex_a15_neon_mrc): Likewise.
3278 (cortex_a15_neon_mrrc): Likewise.
3279 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
3281 (cortex_a15_alu_shift): Likewise.
3282 (cortex_a15_alu_shift_reg): Likewise.
3283 (cortex_a15_mult32): Likewise.
3284 (cortex_a15_mult64): Likewise.
3285 (cortex_a15_block): Likewise.
3286 (cortex_a15_branch): Likewise.
3287 (cortex_a15_load1): Likewise.
3288 (cortex_a15_load3): Likewise.
3289 (cortex_a15_store1): Likewise.
3290 (cortex_a15_store3): Likewise.
3291 (cortex_a15_call): Likewise.
3292 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
3293 (cortex_a5_f2r): Likewise.
3294 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
3296 (cortex_a53_f2r): Likewise.
3297 * config/arm/cortex-a7.md
3298 (cortex_a7_branch): Update for attribute change.
3299 (cortex_a7_call): Likewise.
3300 (cortex_a7_alu_imm): Likewise.
3301 (cortex_a7_alu_reg): Likewise.
3302 (cortex_a7_alu_shift): Likewise.
3303 (cortex_a7_mul): Likewise.
3304 (cortex_a7_load1): Likewise.
3305 (cortex_a7_store1): Likewise.
3306 (cortex_a7_load2): Likewise.
3307 (cortex_a7_store2): Likewise.
3308 (cortex_a7_load3): Likewise.
3309 (cortex_a7_store3): Likewise.
3310 (cortex_a7_load4): Likewise.
3311 (cortex_a7_store4): Likewise.
3312 (cortex_a7_fpalu): Likewise.
3313 (cortex_a7_fconst): Likewise.
3314 (cortex_a7_fpmuls): Likewise.
3315 (cortex_a7_neon_mul): Likewise.
3316 (cortex_a7_fpmacs): Likewise.
3317 (cortex_a7_neon_mla: Likewise.
3318 (cortex_a7_fpmuld: Likewise.
3319 (cortex_a7_fpmacd: Likewise.
3320 (cortex_a7_fpfmad: Likewise.
3321 (cortex_a7_fdivs: Likewise.
3322 (cortex_a7_fdivd: Likewise.
3323 (cortex_a7_r2f: Likewise.
3324 (cortex_a7_f2r: Likewise.
3325 (cortex_a7_f_flags: Likewise.
3326 (cortex_a7_f_loads: Likewise.
3327 (cortex_a7_f_loadd: Likewise.
3328 (cortex_a7_f_stores: Likewise.
3329 (cortex_a7_f_stored: Likewise.
3330 (cortex_a7_neon): Likewise.
3331 * config/arm/cortex-a8-neon.md
3332 (cortex_a8_neon_mrc): Update for attribute change.
3333 (cortex_a8_neon_mrrc): Likewise.
3334 (cortex_a8_neon_int_1): Likewise.
3335 (cortex_a8_neon_int_2): Likewise.
3336 (cortex_a8_neon_int_3): Likewise.
3337 (cortex_a8_neon_int_4): Likewise.
3338 (cortex_a8_neon_int_5): Likewise.
3339 (cortex_a8_neon_vqneg_vqabs): Likewise.
3340 (cortex_a8_neon_vmov): Likewise.
3341 (cortex_a8_neon_vaba): Likewise.
3342 (cortex_a8_neon_vaba_qqq): Likewise.
3343 (cortex_a8_neon_vsma): Likewise.
3344 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3345 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
3346 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
3347 long_scalar): Likewise.
3348 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3349 (cortex_a8_neon_mla_qqq_8_16): Likewise.
3350 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
3351 long_scalar_qdd_64_32_long): Likewise.
3352 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
3353 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
3354 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
3355 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
3356 (cortex_a8_neon_shift_1): Likewise.
3357 (cortex_a8_neon_shift_2): Likewise.
3358 (cortex_a8_neon_shift_3): Likewise.
3359 (cortex_a8_neon_vshl_ddd): Likewise.
3360 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
3361 (cortex_a8_neon_vsra_vrsra): Likewise.
3362 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
3363 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
3364 (cortex_a8_neon_fp_vsum): Likewise.
3365 (cortex_a8_neon_fp_vmul_ddd): Likewise.
3366 (cortex_a8_neon_fp_vmul_qqd): Likewise.
3367 (cortex_a8_neon_fp_vmla_ddd): Likewise.
3368 (cortex_a8_neon_fp_vmla_qqq): Likewise.
3369 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
3370 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
3371 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
3372 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
3373 (cortex_a8_neon_bp_simple): Likewise.
3374 (cortex_a8_neon_bp_2cycle): Likewise.
3375 (cortex_a8_neon_bp_3cycle): Likewise.
3376 (cortex_a8_neon_ldr): Likewise.
3377 (cortex_a8_neon_str): Likewise.
3378 (cortex_a8_neon_vld1_1_2_regs): Likewise.
3379 (cortex_a8_neon_vld1_3_4_regs): Likewise.
3380 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
3381 (cortex_a8_neon_vld2_4_regs): Likewise.
3382 (cortex_a8_neon_vld3_vld4): Likewise.
3383 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
3384 (cortex_a8_neon_vst1_3_4_regs): Likewise.
3385 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
3386 (cortex_a8_neon_vst3_vst4): Likewise.
3387 (cortex_a8_neon_vld1_vld2_lane): Likewise.
3388 (cortex_a8_neon_vld3_vld4_lane): Likewise.
3389 (cortex_a8_neon_vst1_vst2_lane): Likewise.
3390 (cortex_a8_neon_vst3_vst4_lane): Likewise.
3391 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
3392 (cortex_a8_neon_mcr): Likewise.
3393 (cortex_a8_neon_mcr_2_mcrr): Likewise.
3394 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
3395 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
3397 (ca9_neon_mrrc): Likewise.
3398 (cortex_a9_neon_int_1): Likewise.
3399 (cortex_a9_neon_int_2): Likewise.
3400 (cortex_a9_neon_int_3): Likewise.
3401 (cortex_a9_neon_int_4): Likewise.
3402 (cortex_a9_neon_int_5): Likewise.
3403 (cortex_a9_neon_vqneg_vqabs): Likewise.
3404 (cortex_a9_neon_vmov): Likewise.
3405 (cortex_a9_neon_vaba): Likewise.
3406 (cortex_a9_neon_vaba_qqq): Likewise.
3407 (cortex_a9_neon_vsma): Likewise.
3408 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3409 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
3410 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
3411 long_scalar): Likewise.
3412 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3413 (cortex_a9_neon_mla_qqq_8_16): Likewise.
3414 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
3415 long_scalar_qdd_64_32_long): Likewise.
3416 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
3417 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
3418 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
3419 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
3420 (cortex_a9_neon_shift_1): Likewise.
3421 (cortex_a9_neon_shift_2): Likewise.
3422 (cortex_a9_neon_shift_3): Likewise.
3423 (cortex_a9_neon_vshl_ddd): Likewise.
3424 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
3425 (cortex_a9_neon_vsra_vrsra): Likewise.
3426 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
3427 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
3428 (cortex_a9_neon_fp_vsum): Likewise.
3429 (cortex_a9_neon_fp_vmul_ddd): Likewise.
3430 (cortex_a9_neon_fp_vmul_qqd): Likewise.
3431 (cortex_a9_neon_fp_vmla_ddd): Likewise.
3432 (cortex_a9_neon_fp_vmla_qqq): Likewise.
3433 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
3434 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
3435 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
3436 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
3437 (cortex_a9_neon_bp_simple): Likewise.
3438 (cortex_a9_neon_bp_2cycle): Likewise.
3439 (cortex_a9_neon_bp_3cycle): Likewise.
3440 (cortex_a9_neon_ldr): Likewise.
3441 (cortex_a9_neon_str): Likewise.
3442 (cortex_a9_neon_vld1_1_2_regs): Likewise.
3443 (cortex_a9_neon_vld1_3_4_regs): Likewise.
3444 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
3445 (cortex_a9_neon_vld2_4_regs): Likewise.
3446 (cortex_a9_neon_vld3_vld4): Likewise.
3447 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
3448 (cortex_a9_neon_vst1_3_4_regs): Likewise.
3449 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
3450 (cortex_a9_neon_vst3_vst4): Likewise.
3451 (cortex_a9_neon_vld1_vld2_lane): Likewise.
3452 (cortex_a9_neon_vld3_vld4_lane): Likewise.
3453 (cortex_a9_neon_vst1_vst2_lane): Likewise.
3454 (cortex_a9_neon_vst3_vst4_lane): Likewise.
3455 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
3456 (cortex_a9_neon_mcr): Likewise.
3457 (cortex_a9_neon_mcr_2_mcrr): Likewise.
3458 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
3459 (cortex_a9_fps): Likewise.
3460 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
3462 (cortex_m4_fmuls): Likewise.
3463 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
3465 (cortex_r4_mrc): Likewise.
3466 * config/arm/iterators.md: Update comment referring to neon_type.
3467 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
3468 (iwmmxt_movsi_insn): Likewise.
3469 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
3471 (pj4_core_to_vfp): Likewise.
3472 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
3474 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
3475 (vfp_fstore): Likewise.
3476 * doc/md.texi: Change references to neon_type to refer to type.
3478 2013-09-04 Dodji Seketeli <dodji@redhat.com>
3480 * tree.h (DECL_BUILT_IN): Fix typo in comment.
3482 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
3484 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
3487 2013-09-04 Easwaran Raman <eraman@google.com>
3490 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
3491 (build_and_add_sum): Use it.
3492 (appears_later_in_bb): Simplify code.
3494 2013-09-04 Teresa Johnson <tejohnson@google.com>
3496 * dumpfile.c (dump_finish): Don't close stderr/stdout.
3498 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
3500 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
3502 2013-09-04 Jan Hubicka <jh@suse.cz>
3504 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
3505 * ipa-devirt.c: Include diganostic.h
3506 (odr_type_d): Add types and types_set.
3507 (hash_type_name): Work for types with vtables during LTO.
3508 (odr_hasher::remove): Fix comment; destroy types_set.
3509 (add_type_duplicate): New function,
3510 (get_odr_type): Use it.
3511 (dump_type_inheritance_graph): Dump type duplicates.
3512 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
3514 * tree.c (types_same_for_odr): Give exact answers on types with
3517 2013-09-04 Dodji Seketeli <dodji@redhat.com>
3519 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
3520 explaining their differences.
3522 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
3524 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
3526 2013-09-03 Jeff Law <law@redhat.com>
3528 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
3529 when not threading through a joiner block. Pass joiner/no joiner
3530 state to register_jump_thread.
3531 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
3532 state from argument rather than implying on path length.
3533 Dump the entire jump thread path into debugging dump.
3534 * tree-flow.h (register_jump_thread): Update prototype.
3536 2013-08-29 Xinliang David Li <davidxl@google.com>
3538 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
3539 Remove a trivial gcc_assert.
3541 2013-08-29 Xinliang David Li <davidxl@google.com>
3543 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
3544 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
3545 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
3546 Delay base decl alignment adjustment.
3547 * tree-vectorizer.c (vect_destroy_datarefs): New function.
3548 * tree-vectorizer.h: New data structure.
3549 (set_dr_misalignment): New function.
3550 (dr_misalignment): Ditto.
3551 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
3552 (vectorizable_load): Ditto.
3553 (ensure_base_align): New function.
3554 (vectorize_loops): Add dbg_cnt support.
3555 (execute_vect_slp): Ditto.
3556 * dbgcnt.def: New debug counter.
3557 * Makefile: New dependency.
3559 2013-09-03 Meador Inge <meadori@codesourcery.com>
3563 2013-08-30 Meador Inge <meadori@codesourcery.com>
3565 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
3567 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
3569 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
3570 function descriptor.
3572 2013-09-03 Richard Biener <rguenther@suse.de>
3574 * tree-affine.c (add_elt_to_tree): Fix association issue,
3575 avoid useless converts and make sure to always return a
3576 properly typed result.
3578 2013-09-03 Richard Biener <rguenther@suse.de>
3581 * fold-const.c (negate_expr_p): Fix division case.
3582 (negate_expr): Likewise.
3584 2013-09-03 Richard Biener <rguenther@suse.de>
3587 * tree-streamer-out.c: Include tm.h.
3588 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
3590 2013-09-03 Jan Hubicka <jh@suse.cz>
3592 * tree-profile.c (tree_profiling): Cleanup CFG when done.
3594 2013-09-03 Alan Modra <amodra@gmail.com>
3596 * config.gcc (powerpc*-*-linux*): Add support for little-endian
3597 multilibs to big-endian target and vice versa.
3598 * config/rs6000/t-linux64: Use := assignment on all vars.
3599 (MULTILIB_EXTRA_OPTS): Remove fPIC.
3600 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
3601 * config/rs6000/t-linux64le: New file.
3602 * config/rs6000/t-linux64bele: New file.
3603 * config/rs6000/t-linux64lebe: New file.
3605 2013-09-02 Jan Hubicka <jh@suse.cz>
3607 * ipa-inline-transform.c (inline_transform): Do not
3608 optimize_inline_calls when not optimizing.
3610 2013-09-02 Jan Hubicka <jh@suse.cz>
3612 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
3613 duplicated nodes for assembler names.
3614 * symtab.c (symtab_unregister_node): Do not attempt to unlink
3615 hard registers from assembler name hash.
3617 2013-09-02 Jan Hubicka <jh@suse.cz>
3619 * ipa-split.c (execute_split_functions): Split externally visible
3620 functions called once.
3622 2013-09-02 Martin Jambor <mjambor@suse.cz>
3625 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
3626 linked list. When finding the correct duplicate, also consider also
3627 the caller in additon to its inlined_to node.
3629 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
3631 * config/aarch64/aarch64-simd-builtins.def
3632 (dup_lane_scalar): Remove.
3633 * config/aarch64/aarch64-simd.md
3634 (aarch64_simd_dup): Add 'w->w' alternative.
3635 (aarch64_dup_lane<mode>): Allow for VALL.
3636 (aarch64_dup_lane_scalar<mode>): Remove.
3637 (aarch64_dup_lane_<vswap_width_name><mode>): New.
3638 (aarch64_get_lane_signed<mode>): Add w->w altenative.
3639 (aarch64_get_lane_unsigned<mode>): Likewise.
3640 (aarch64_get_lane<mode>): Likewise.
3641 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
3642 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
3643 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
3644 (VCON): Change container of V2SF.
3645 (vswap_width_name): Likewise.
3646 * config/aarch64/arm_neon.h
3647 (__aarch64_vdup_lane_any): New.
3648 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
3649 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
3650 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
3652 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
3655 * expr.c (emit_move_complex): Do not move complex FP values as parts if
3656 the source or the destination is a single hard register.
3658 2013-09-02 Richard Biener <rguenther@suse.de>
3661 * tree-scalar-evolution.c (instantiate_scev_name): Allow
3664 2013-09-02 Richard Biener <rguenther@suse.de>
3666 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
3667 arithmetic to sizetype.
3669 2013-09-02 Bin Cheng <bin.cheng@arm.com>
3671 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
3672 Find auto-increment use both before and after candidate.
3674 2013-09-02 Marek Polacek <polacek@redhat.com>
3676 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
3678 2013-09-01 Jan Hubicka <jh@suse.cz>
3680 * Makefile.in: Add ipa-profile.o
3681 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
3682 * cgraph.c (struct cgraph_propagate_frequency_data,
3683 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
3684 ipa-profile.c; replace cgraph_ by ipa_ prefix.
3685 * cgraph.h (cgraph_propagate_frequency): Remove.
3686 * ipa-inline-analysis.c: Include ipa-utils.h;
3687 drop duplicated cfgloop.h.
3688 (inline_update_callee_summaries): Update.
3689 * ipa-profile.c: New file.
3690 * ipa-utils.h (ipa_propagate_frequency): Declare.
3691 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
3692 data-streamer.h, value-prof.h.
3693 (symtab_remove_unreachable_nodes): Update profile.
3694 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
3695 account_time_size, cmp_counts, dump_histogram,
3696 ipa_profile_generate_summary, ipa_profile_write_summary,
3697 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
3698 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
3699 Move to ipa-profile.c.
3701 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
3703 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
3705 2013-09-01 Jan Hubicka <jh@suse.cz>
3707 * common.opt (fdevirtualize-speculatively): New function.
3708 * invoke.texi (fdevirtualize-speculatively): Document.
3709 * ipa-devirt.c: Include ipa-inline.h
3710 (likely_target_p): New function.
3711 (ipa_devirt): New function.
3712 (gate_ipa_devirt): New function.
3713 (pass_data_ipa_devirt): New static var.
3714 (pass_ipa_devirt): Likewise.
3715 (make_pass_ipa_devirt): New function.
3716 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
3717 (common_handle_option): Disable devirtualization when
3718 value range profiling is available.
3719 * passes.def (pass_ipa_devirt): Add.
3720 * timever.def (TV_IPA_DEVIRT): New timevar.
3721 * tree-pass.h (make_pass_ipa_devirt):
3723 2013-09-01 Iain Sandoe <iain@codesourcery.com>
3725 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
3726 include sanitize(undefined).
3728 2013-08-31 Diego Novillo <dnovillo@google.com>
3730 * Makefile.in (TREE_CORE_H): Define.
3732 (GTFILES): Add tree-core.h.
3733 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
3735 * tree-core.h: New file.
3736 Move all data structures, enum, typedefs, global
3737 declarations and constants from ...
3740 2013-08-31 Jan Hubicka <jh@suse.cz>
3742 * bulitins.c (expand_builtin): Do not early exit for gcov
3743 instrumented functions.
3745 2013-08-31 Marek Polacek <polacek@redhat.com>
3747 * ubsan.c: Include tm_p.h.
3749 2013-08-31 Jan Hubicka <jh@suse.cz>
3751 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
3754 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
3755 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
3756 * tree-cfg.c (verify_gimple_label): ... here.
3757 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
3758 (ipa_merge_profiles): New function.
3759 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
3760 (lto_input_function_body): Likewise.
3761 * ipa-utils.h (ipa_merge_profiles): Declare.
3762 * lto-streamer.h (lto_input_function_body): Update prototype.
3763 (emit_label_in_global_context_p): Remove.
3764 * lto-symtab.c: Include ipa-utils.h
3765 (lto_cgraph_replace_node): Use ipa_merge_profiles.
3767 2013-08-31 Jan Hubicka <jh@suse.cz>
3769 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
3771 2013-08-31 Jan Hubicka <jh@suse.cz>
3773 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
3775 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
3777 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
3778 "cmp" RTX before signed_comparison_operator check to account
3781 2013-08-30 Jan Hubicka <jh@suse.cz>
3783 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
3784 (detect_type_change_1): Rename to ...
3785 (detect_type_change): ... this one; early return on non-polymorphic
3787 (detect_type_change_ssa): Add comp_type parameter; update
3788 use of detect_type_change.
3789 (compute_complex_assign_jump_func): Add param_type parameter;
3790 update use of detect_type_change_ssa.
3791 (compute_complex_ancestor_jump_func): Likewise.
3792 (ipa_get_callee_param_type): New function.
3793 (ipa_compute_jump_functions_for_edge): Compute parameter type;
3794 update calls to the jump function computation functions.
3796 2013-08-30 Teresa Johnson <tejohnson@google.com>
3797 Steven Bosscher <steven@gcc.gnu.org>
3799 * cfgrtl.c (fixup_new_cold_bb): New routine.
3800 (commit_edge_insertions): Invoke fixup_partitions.
3801 (find_partition_fixes): New routine.
3802 (fixup_partitions): Ditto.
3803 (verify_hot_cold_block_grouping): Update comments.
3804 (rtl_verify_edges): Invoke find_partition_fixes.
3805 (rtl_verify_bb_pointers): Update comments.
3806 (rtl_verify_bb_layout): Ditto.
3807 * basic-block.h (probably_never_executed_edge_p): Declare.
3808 (fixup_partitions): Ditto.
3809 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
3810 * bb-reorder.c (sanitize_hot_paths): New function.
3811 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
3813 * predict.c (probably_never_executed_edge_p): New routine.
3814 * cfg.c (check_bb_profile): Add partition insanity warnings.
3816 2013-08-30 Meador Inge <meadori@codesourcery.com>
3818 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
3820 2013-08-30 Marek Polacek <polacek@redhat.com>
3822 * Makefile.in (ubsan.o): Add.
3823 (c-family/c-ubsan.o): Add.
3824 (builtins.o): Add ubsan.h dependency.
3825 * ubsan.h: New file.
3826 * ubsan.c: New file.
3827 * common.opt: Add -fsanitize=undefined option.
3828 (flag_sanitize): Add variable.
3829 (fsanitize=): Add option. Add Driver.
3830 (fsanitize=thread): Remove option.
3831 (fsanitize=address): Likewise.
3832 (static-libubsan): New option.
3833 * doc/invoke.texi: Document the new flag and -static-libubsan.
3834 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
3835 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
3836 * builtin-attrs.def (ATTR_COLD): Define.
3837 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
3838 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
3839 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
3840 * flag-types.h (sanitize_code): New enum.
3841 * opts.c (common_handle_option): Parse command line arguments
3842 of -fsanitize=. Add -fsanitize=unreachable option.
3843 * varasm.c (get_variable_section): Adjust.
3844 (assemble_noswitch_variable): Likewise.
3845 (assemble_variable): Likewise.
3846 (output_constant_def_contents): Likewise.
3847 (categorize_decl_for_section): Likewise.
3848 (place_block_symbol): Likewise.
3849 (output_object_block): Likewise.
3850 * builtins.def: Likewise.
3851 * toplev.c (compile_file): Likewise.
3852 (process_options): Likewise.
3853 * cppbuiltin.c: Likewise.
3854 * tsan.c (tsan_pass): Likewise.
3855 (tsan_gate): Likewise.
3856 (tsan_gate_O0): Likewise.
3857 * cfgexpand.c (partition_stack_vars): Likewise.
3858 (expand_stack_vars): Likewise.
3859 (defer_stack_allocation): Likewise.
3860 (expand_used_vars): Likewise.
3861 * cfgcleanup.c (old_insns_match_p): Likewise.
3862 * asan.c (asan_finish_file): Likewise.
3863 (asan_instrument): Likewise.
3864 (gate_asan): Likewise.
3865 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
3866 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
3867 (asan_global_struct): Use pointer_sized_int_node instead
3868 calling build_nonstandard_integer_type.
3869 (initialize_sanitizer_builtins): Likewise.
3870 (asan_finish_file): Likewise.
3871 * gcc.c: Document %{%:function(args):X}.
3872 (static_spec_functions): Add sanitize.
3873 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
3874 store funcval != NULL there.
3875 (do_spec_1): Adjust handle_spec_function caller.
3876 (handle_braces): Allow %:function(args) as condition.
3877 (sanitize_spec_function): New function.
3878 (ADD_STATIC_LIBUBSAN_LIBS): Define.
3879 (LIBUBSAN_SPEC): Likewise.
3880 (LIBUBSAN_EARLY_SPEC): Likewise.
3881 (SANITIZER_SPEC): Handle libubsan.
3882 (SANITIZER_EARLY_SPEC): Likewise.
3883 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
3884 instead of fsanitize=address.
3885 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
3886 instead of fsanitize=address*.
3887 * builtins.c: Include ubsan.h.
3888 (fold_builtin_0): Instrument __builtin_unreachable.
3889 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
3890 instead of flag_asan.
3891 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
3892 (pointer_sized_int_node): Define.
3893 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
3895 2013-08-30 Mike Stump <mikestump@comcast.net>
3897 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
3900 2013-08-29 Jan Hubicka <jh@suse.cz>
3902 * cgraph.c (cgraph_function_body_availability): Handle weakref
3904 * passes.def: Remove pass_fixup_cfg.
3905 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
3906 track when we need to remove functions.
3907 (gate_ipa_inline): Execute inlining always; add comment why.
3908 (pass_data_ipa_inline): Remove TODO_remove_functions.
3909 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
3910 do not produce summaries.
3911 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
3912 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
3913 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
3914 constant pool and vtable.
3916 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
3918 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
3919 New arm_neon.h's internal macros to specify 64-bit constants.
3920 Avoid using stdint.h's macros.
3922 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
3924 * recog.c (verify_changes): Verify that changes[i].old is non-zero
3925 before applying REG_P.
3927 2013-08-30 Jakub Jelinek <jakub@redhat.com>
3929 PR tree-optimization/58277
3930 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
3931 after seeing too many stmts with vdef in between dombb and current
3932 bb, invalidate everything.
3934 2013-08-30 Richard Biener <rguenther@suse.de>
3936 * fold-const.c (fold_single_bit_test): Fix overflow test.
3938 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
3940 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
3941 and which can live in a register, always retrieve the value on entry.
3942 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
3943 passed by invisible reference specially.
3944 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
3945 (vt_add_function_parameter): Correctly deal with a parameter passed by
3946 invisible reference.
3948 2013-08-30 Jan Hubicka <jh@suse.cz>
3950 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
3952 2013-08-30 Richard Biener <rguenther@suse.de>
3954 PR tree-optimization/58228
3955 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
3956 allow invariant loads in nested loop vectorization.
3958 2013-08-30 Richard Biener <rguenther@suse.de>
3960 PR tree-optimization/58223
3961 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
3962 (has_anti_or_output_dependence): ... this and adjust to also
3963 look for output dependences.
3964 (mark_nodes_having_upstream_mem_writes): Adjust.
3965 (rdg_flag_uses): Likewise.
3967 2013-08-30 Richard Biener <rguenther@suse.de>
3969 PR tree-optimization/58010
3970 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
3971 assert that we have a loop-closed PHI.
3973 2013-08-29 Jan Hubicka <jh@suse.cz>
3975 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
3976 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
3977 * lto-section-in.c (lto_free_function_in_decl_state): New function.
3978 (lto_free_function_in_decl_state_for_node): New function.
3980 2013-08-29 Xinliang David Li <davidxl@google.com>
3982 * loop-unroll.c (report_unroll_peel): Minor message change.
3983 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
3984 Emit alignment peeling message with default -fopt-info.
3985 (vect_loop_versioning): Emit loop version info message.
3986 * tree-vectorizer.c (vectorize_loops): Minor message change.
3987 (execute_vect_slp): Ditto.
3989 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
3991 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
3992 of the clone from the DECL_NAME of the original function.
3994 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
3996 * passes.c (register_pass): Add overload.
3997 * tree-pass.h (register_pass): Forward declare it. Add comment.
3999 2013-08-29 Jan Hubicka <jh@suse.cz>
4001 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
4002 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
4003 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
4004 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
4005 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
4006 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
4007 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
4008 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
4010 2013-08-29 Teresa Johnson <tejohnson@google.com>
4012 * dumpfile.c (dump_loc): Output column number.
4013 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
4014 * doc/invoke.texi: Document optall -fopt-info flag.
4015 * profile.c (read_profile_edge_counts): Use new dump framework.
4016 (compute_branch_probabilities): Ditto.
4017 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
4018 when pass not in any opt group.
4019 * pass_manager.h (pass_manager::get_pass_profile): New method.
4020 * value-prof.c (check_counter): Use new dump framework.
4021 (check_ic_target): Ditto.
4022 * coverage.c (get_coverage_counts): Ditto.
4023 (coverage_init): Setup new dump framework.
4025 2013-08-29 Richard Biener <rguenther@suse.de>
4027 PR tree-optimization/58246
4028 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
4029 handle the dominance check inside a basic-block.
4031 2013-08-29 Richard Biener <rguenther@suse.de>
4034 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
4035 of default defs that appear in abnormal PHI nodes.
4037 2013-08-29 Richard Biener <rguenther@suse.de>
4039 PR tree-optimization/57685
4040 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
4041 single-use operands to avoid exponential complexity.
4043 2013-08-28 Dehao Chen <dehao@google.com>
4045 * ipa-inline.c (edge_badness): Fix integer underflow.
4047 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
4049 * gtm-builtins.def (_ITM_free): Declare leaf.
4051 2013-08-28 Jakub Jelinek <jakub@redhat.com>
4054 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
4055 (*tls_local_dynamic_base_64_largepic): Likewise.
4056 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
4057 Remove predicate from call operand.
4058 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
4059 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
4061 2013-08-28 Jeff Law <law@redhat.com>
4063 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
4064 checks for the number of predecessors and successors allowed.
4065 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
4066 which require copying a joiner block if there is a request which
4067 is a subpath that requires no joiner block copying.
4069 2013-08-28 Jan Hubicka <jh@suse.cz>
4071 * lto-streamer-out.c (DFS_write_tree_body): Drop
4072 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
4073 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
4074 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
4075 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
4076 Do not read DECL_ERROR_ISSUED.
4077 (unpack_ts_decl_with_vis_value_fields): Do not read
4079 (lto_input_ts_binfo_tree_pointers): Do not read
4080 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
4081 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
4082 write DECL_ERROR_ISSUED..
4083 (pack_ts_decl_with_vis_value_fields): Do not write
4085 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
4086 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
4087 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
4088 * tree.h (tree_decl_common): Update comment.
4089 (DECL_ERROR_ISSUED): Remove.
4091 2013-08-28 Jakub Jelinek <jakub@redhat.com>
4094 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
4096 2013-08-28 Jan Hubicka <jh@suse.cz>
4098 * builtins.def (free): Declare leaf.
4100 2013-08-27 David Malcolm <dmalcolm@redhat.com>
4103 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
4104 * configure: Regenerate.
4106 2013-08-27 Martin Jambor <mjambor@suse.cz>
4108 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
4109 (ipa_ancestor_jf_data): Likewise.
4110 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
4111 (ipa_get_jf_pass_through_type_preserved): New function.
4112 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
4113 (ipa_get_jf_ancestor_type_preserved): New function.
4114 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
4115 (ipa_get_jf_ancestor_result): Likewise.
4116 (propagate_vals_accross_pass_through): Use
4117 ipa_get_jf_pass_through_result to do all the value mappings.
4118 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
4119 type_preserved flag.
4120 (ipa_set_jf_cst_copy): New function.
4121 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
4122 (ipa_set_jf_arith_pass_through): Likewise.
4123 (ipa_set_ancestor_jf): Likewise.
4124 (compute_complex_assign_jump_func): Set type_preserved instead of
4126 (ipa_compute_jump_functions_for_edge): Likewise.
4127 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
4128 (update_jump_functions_after_inlining): Update type_preserved.
4129 Explicitely create jump functions when combining one with pass_through.
4130 (ipa_write_jump_function): Stream the type_preserved flags.
4131 (ipa_read_jump_function): Likewise.
4133 2013-08-27 Jakub Jelinek <jakub@redhat.com>
4134 Aldy Hernandez <aldyh@redhat.com>
4136 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
4137 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
4138 * function.h (struct function): Add has_force_vect_loops and
4140 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
4141 * gimple.c (gimple_build_omp_critical): Add KIND argument and
4143 * gimple.def: Update CLAUSES comments.
4144 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
4145 (gimple_build_omp_for): Add argument to prototype.
4146 (gimple_omp_for_kind): New.
4147 (gimple_omp_for_set_kind): New.
4148 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
4149 GOVD_DATA_SHARE_CLASS.
4150 (enum omp_region_type): Add ORT_SIMD.
4151 (gimple_add_tmp_var): Handle ORT_SIMD.
4152 (gimplify_var_or_parm_decl): Same.
4153 (is_gimple_stmt): Same.
4154 (omp_firstprivatize_variable): Same.
4155 (omp_add_variable): Only use splay_tree_insert if lookup failed.
4156 (omp_notice_variable): Handle ORT_SIMD.
4157 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
4158 (omp_check_private): Handle ORT_SIMD.
4159 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
4161 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
4162 Handle OMP_CLAUSE_LASTPRIVATE.
4163 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
4165 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
4166 (gimplify_expr): Handle OMP_SIMD.
4167 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
4168 (expand_GOMP_SIMD_VF): New.
4169 (expand_GOMP_SIMD_LAST_LANE): New.
4170 * internal-fn.def (GOMP_SIMD_LANE): New.
4171 (GOMP_SIMD_VF): New.
4172 (GOMP_SIMD_LAST_LANE): New.
4173 * omp-low.c: Include target.h.
4174 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
4176 (check_omp_nesting_restrictions): Same.
4178 (lower_rec_simd_input_clauses): New.
4179 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
4181 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
4182 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
4183 (expand_omp_build_assign): New.
4184 (expand_omp_for_init_counts): New.
4185 (expand_omp_for_init_vars): New.
4186 (extract_omp_for_update_vars): New.
4187 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
4188 and rewrite accordingly.
4189 (expand_omp_simd): New.
4190 (expand_omp_for): Use expand_omp_simd.
4191 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
4192 (lower_omp_for): Do not lower the body.
4193 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
4195 * tree-flow.h (find_omp_clause): Remove prototype.
4196 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
4197 forcing vectorization of the loop, or if flag_tree_vectorize.
4198 (gate_tree_if_conversion): Similarly.
4199 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
4200 gimple_build_omp_for.
4201 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
4202 * tree-parloops (create_parallel_loop): Pass kind argument to
4203 gimple_build_omp_for.
4204 * tree-pretty-print.c (dump_omp_clause): Add cases for
4205 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
4206 OMP_CLAUSE__SIMDUID_.
4207 (dump_generic_node): Handle OMP_SIMD.
4208 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
4209 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
4210 unroll OMP_SIMD loops here.
4211 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
4212 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
4214 (vect_analyze_data_refs): Handle simd loops.
4215 * tree-vect-loop.c (vectorizable_live_operation): Handle
4217 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
4218 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
4219 (vectorizable_load): Same.
4220 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
4221 (struct simduid_to_vf): New.
4222 (simduid_to_vf::hash): New.
4223 (simduid_to-vf::equal): New.
4224 (struct simd_array_to_simduid): New.
4225 (simd_array_to_simduid::hash): New.
4226 (simd_array_to_simduid::equal): New.
4227 (adjust_simduid_builtins): New.
4228 (struct note_simd_array_uses_struct): New.
4229 (note_simd_array_uses_cb): New.
4230 (note_simd_array_uses): New.
4231 (vectorize_loops): Handle simd hints and adjust simd builtins
4233 * tree-vectorizer.h (struct _stmt_vec_info): Add
4234 simd_lane_access_p field.
4235 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
4236 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
4237 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
4238 (omp_clause_code_name): Same.
4239 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
4240 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
4241 * tree.def (OMP_SIMD): New entry.
4242 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
4243 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
4244 (OMP_CLAUSE_DECL): Adjust range for new clauses.
4245 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
4246 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
4247 (OMP_CLAUSE_LINEAR_STEP): New.
4248 (OMP_CLAUSE_SAFELEN_EXPR): New.
4249 (OMP_CLAUSE__SIMDUID__DECL): New.
4250 (find_omp_clause): New prototype.
4252 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
4254 * config/i386/driver-i386.c (host_detect_local_cpu): Update
4255 Haswell processor detection.
4257 2013-08-27 Christian Widmer <shadow@umbrox.de>
4260 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
4261 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
4262 AVX2 capable processors.
4264 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
4266 * config/aarch64/arm_neon.h: Replace all inline asm implementations
4267 of vget_low_* with implementations in terms of other intrinsics.
4269 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
4272 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
4273 values to -1, 0 and 1.
4275 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
4277 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
4278 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
4279 (<optab><mode>3_insn): Remove.
4280 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
4281 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
4282 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
4283 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
4284 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
4285 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
4286 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
4287 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
4288 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
4289 (ror<mode>3_insn): Likewise.
4290 * config/aarch64/predicates.md (aarch64_simd_register): New.
4292 2013-08-27 Richard Biener <rguenther@suse.de>
4294 PR tree-optimization/57521
4295 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
4296 one edge is non-critical.
4297 (find_phi_replacement_condition): Make sure to use a non-critical
4298 edge. Cleanup and remove old bug workarounds.
4299 (bb_postdominates_preds): Remove.
4300 (if_convertible_loop_p_1): Do not compute post-dominators.
4301 (combine_blocks): Do not free post-dominators.
4302 (main_tree_if_conversion): Likewise.
4303 (pass_data_if_conversion): Add TODO_verify_ssa.
4305 2013-08-27 DJ Delorie <dj@redhat.com>
4307 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
4309 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
4311 * function.c (assign_parm_find_data_types): Set passed_mode and
4312 nominal_mode to the TYPE_MODE of nominal_type for the built
4313 pointer type in case of the struct-pass-by-reference.
4315 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
4317 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
4318 (UINT16_TYPE): Change default to "unsigned int".
4320 * config/avr/avr.opt (mfract-convert-truncate): New option.
4321 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
4322 is set, round negative fractional integers according to n1169
4323 when converting to integer types.
4325 2013-08-26 Jan Hubicka <jh@suse.cz>
4327 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
4328 methods can not be called indirectly when their address is not taken.
4330 2013-08-26 Jan Hubicka <jh@suse.cz>
4332 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
4335 2013-08-26 Jan Hubicka <jh@suse.cz>
4337 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
4340 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
4342 * reload.c (find_valid_class): Allow classes that do not include
4343 FIRST_PSEUDO_REGISTER - 1.
4345 2013-08-26 Jan Hubicka <jh@suse.cz>
4347 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
4348 fix edge count/frequency when speculation failed; fix type check
4349 for the direct call.
4351 2013-08-26 Jan Hubicka <jh@suse.cz>
4353 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
4355 2013-08-26 Jan Hubicka <jh@suse.cz>
4357 * ipa-inline-transform.c (inline_transform): Be ready for basic block
4358 to be changed by edge redirection.
4360 2013-08-26 Jan Hubicka <jh@suse.cz>
4362 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
4363 formating; add sanity check.
4364 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
4365 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
4366 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
4367 (resolve_noninline_speculation): Update callee keys, too.
4369 2013-08-26 Jan Hubicka <jh@suse.cz>
4371 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
4372 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
4374 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
4376 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
4379 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
4381 * config/i386/i386.c (ix86_debug_options): Remove prototype.
4382 (x86_64_elf_select_section): Ditto.
4383 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
4385 (ix86_pass_by_reference): Ditto.
4386 (output_set_got): Ditto.
4387 (ix86_unary_operator_ok): Ditto.
4388 (ix86_expand_builtin): Ditto.
4390 2013-08-23 Jan Hubicka <jh@suse.cz>
4392 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
4394 2013-08-23 Jan Hubicka <jh@suse.cz>
4396 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
4397 (tree_decl_with_vis): Add FINAL field.
4399 2013-08-23 Jeff Law <law@redhat.com>
4401 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
4402 the debugging dump when the expression is fully redundant.
4404 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
4406 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
4407 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
4408 * pretty-print.c (pp_formatted_text_data): Likewise.
4409 (pp_write_text_to_stream): Likewise.
4410 (pp_write_text_as_dot_label_to_stream): Likewise.
4411 (pp_append_r): Likewise.
4412 (pp_format): Likewise.
4413 (pp_flush): Likewise.
4414 (pp_clear_output_area): Likewise.
4415 (pp_append_text): Likewise.
4416 (pp_formatted_text): Likewise.
4417 (pp_remaining_character_count_for_line): Likewise.
4418 (pp_newline): Likewise.
4419 (pp_character): Likewise.
4420 (output_buffer::~output_buffer): Define.
4421 (pretty_printer::~pretty_printer): Destruct output buffer.
4422 * pretty-print.h (output_buffer::~output_buffer): Declare.
4423 (pretty_printer::~pretty_printer): Declare virtual.
4425 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
4428 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
4429 HOST_WIDE_INT_M1U): New macros.
4430 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
4431 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
4432 unsigned -1 for lshift.
4433 * cse.c (cse_insn): Likewise.
4434 * double-int.c (rshift_double, lshift_double): Likewise.
4435 * builtins.c (fold_builtin_bitop): Likewise.
4436 * combine.c (force_to_mode): Likewise.
4437 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
4438 * simplify-rtx.c (simplify_const_unary_operation,
4439 simplify_const_binary_operation): Likewise.
4440 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
4441 check_va_list_escapes): Likewise.
4442 * rtlanal.c (nonzero_bits1): Likewise.
4443 * expmed.c (expand_smod_pow2): Likewise.
4444 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
4446 2013-08-23 Jan Hubicka <jh@suse.cz>
4448 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
4449 as having address taken.
4451 2013-08-23 Jan Hubicka <jh@suse.cz>
4453 * ipa-utils.h (method_class_type): Declare.
4454 * ipa-devirt.c (method_class_type): Export.
4456 * cgraphunit.c (analyze_functions): Do basic devirtualization;
4457 do not walk base classes of anonymous types.
4459 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
4461 PR rtl-optimization/58220
4463 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
4464 handle SEQUENCE insns properly.
4466 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
4468 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
4470 (pp_newline_and_indent): Likewise.
4471 (pp_separate_with): Likewise.
4472 * pretty-print.c (pp_newline_and_flush): Define.
4473 (pp_newline_and_indent): Likewise.
4474 (pp_separate_with): Likewise.
4476 2013-08-23 Jakub Jelinek <jakub@redhat.com>
4479 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
4480 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
4482 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
4484 * config/i386/predicates.md (ext_sse_reg_operand): New.
4485 * config/i386/i386.md (*movti_internal): Use
4486 predicate to determine if EVEX is needed.
4487 (*movsi_internal): Ditto.
4488 (*movdf_internal): Ditto.
4489 (*movsf_internal): Ditto.
4490 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
4492 2013-08-23 Jakub Jelinek <jakub@redhat.com>
4494 PR tree-optimization/58209
4495 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
4496 (find_tail_calls): Give up for pointer result types if m is non-NULL.
4497 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
4498 emit POINTER_PLUS_EXPR.
4499 (create_tailcall_accumulator): For pointer result type accumulate in
4502 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
4504 * configure.ac: Add backslashes missing from the last change.
4505 * configure: Regenerate.
4507 2013-08-22 Jan Hubicka <jh@susue.cz>
4509 * ipa.c (function_and_variable_visibility): First remember function
4510 was global and then make it local.
4512 2013-08-22 Julian Brown <julian@codesourcery.com>
4514 * configure.ac: Add aarch64 to list of arches which use "nop" in
4516 * configure: Regenerate.
4518 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4520 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
4521 gnu_libc_has_function.
4522 * config/s390/tpf.h: Likewise.
4524 2013-08-22 Jan Hubicka <jh@susue.cz>
4526 * timevar.c (validate_phases): Add cast.
4528 2013-08-22 Jan Hubicka <jh@susue.cz>
4530 * timevar.c (validate_phases): Use size_t for memory.
4531 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
4533 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
4535 * pretty-print.h (output_buffer::output_buffer): Declare.
4536 (pretty_printer::pretty_printer): Likewise.
4537 (pp_construct): Remove.
4538 * pretty-print.c (output_buffer::output_buffer): Define.
4539 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
4540 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
4541 (print_gimple_expr): Likewise.
4542 (print_gimple_seq): Likewise.
4543 (gimple_dump_bb): Likewise.
4544 * sched-vis.c (dump_value_slim): Likewise.
4545 (dump_insn_slim): Likewise.
4546 (dump_rtl_slim): Likewise.
4547 (str_pattern_slim): Likewise.
4548 * tree-mudflap.c (mf_varname_tree): Likewise.
4549 * graph.c (print_graph_cfg): Likewise.
4550 (start_graph_dump): Likewise.
4551 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
4553 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
4554 pretty printer initialization.
4555 * coretypes.h (diagnostic_context): Remove superflous type alias
4557 (pretty_printer): Likewise. Declare directly as a class.
4558 (pretty_print_info): Remove declaration as class.
4559 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
4560 and pp_clear_output_area.
4561 (asan_add_global): Likewise.
4563 2013-08-22 Jan Hubicka <jh@suse.cz>
4565 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
4566 * ipa-utils.h (update_type_inheritance_graph): Declare.
4567 (possible_polymorphic_call_target_p): Declare.
4568 (possible_polymorphic_call_target_p): New.
4569 * ipa-devirt.c: Update toplevel comments.
4570 (cached_polymorphic_call_targets): Move up.
4571 (odr_type_d): Move ID down.
4572 (polymorphic_type_binfo_p): Update comment.
4573 (odr_hasher::remove): Likewise;
4574 (get_odr_type): Set anonymous_namespace.
4575 (dump_odr_type): Dump it.
4576 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
4577 (maybe_record_node): Record node in cached_polymorphic_call_targets.
4578 (record_binfo): Add comment.
4579 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
4581 (devirt_node_removal_hook): Do not iCE when cache is freed.
4582 (possible_polymorphic_call_target_p): New predicate.
4583 (update_type_inheritance_graph): New function.
4585 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
4586 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4587 Sergey Lega <sergey.s.lega@intel.com>
4588 Anna Tikhonova <anna.tikhonova@intel.com>
4589 Ilya Tocar <ilya.tocar@intel.com>
4590 Andrey Turetskiy <andrey.turetskiy@intel.com>
4591 Ilya Verbin <ilya.verbin@intel.com>
4592 Kirill Yukhin <kirill.yukhin@intel.com>
4593 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4595 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
4596 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
4597 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
4598 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
4599 (OPTION_MASK_ISA_AVX2_UNSET): Update.
4600 (OPTION_MASK_ISA_AVX512F_UNSET): New.
4601 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
4602 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
4603 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
4604 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
4605 OPT_mavx512pf, OPT_mavx512er cases.
4606 * config/i386/constraints.md (v): New constraint.
4607 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
4608 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
4609 (bit_AVX512CD): New.
4610 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
4611 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
4612 * config/i386/i386-c.c (ix86_target_macros_internal):
4613 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
4615 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
4616 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
4617 * config/i386/i386.c (regclass_map, dbx_register_map)
4618 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
4619 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
4620 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
4622 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
4623 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
4624 -mavx512pf options. Fix formatting.
4625 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
4626 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
4627 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
4628 -mavx512cd, -mavx512pf options.
4629 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
4630 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
4631 (ix86_preferred_output_reload_class): Replace SSE_REGS with
4633 (ix86_hard_regno_mode_ok): Support 512-bit registers.
4634 (ix86_set_reg_reg_cost): Ditto.
4635 (x86_order_regs_for_local_alloc): Ditto.
4636 (MAX_VECT_LEN): Extend to 64-byte.
4637 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
4638 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
4639 (TARGET_AVX512ER, TARGET_AVX512CD): New.
4640 (BIGGEST_ALIGNMENT): Extend to 512-bits.
4641 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
4642 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
4643 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
4644 (SSE_REG_MODE_P): Support new modes.
4645 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
4646 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
4647 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
4648 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
4649 (REG_CLASS_CONTENTS): Add new registers.
4650 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
4651 (EXT_REX_SSE_REGNO_P): New.
4652 (HI_REGISTER_NAMES): Add new registers.
4653 * config/i386/i386.md: Define constants for new registers.
4654 (mode): Add new 512-bit modes.
4655 (prefix): Support evex prefix.
4656 (isa): Support avx512f, noavx512f, fma_avx512f.
4657 (ssemodesuffix): Add new 512-bit modes.
4659 (*movxi_internal_avx512f): Ditto.
4660 (*movdi_internal): Replace constraint "x" with the new constraint "v".
4662 (*movsi_internal): Likewise.
4663 (*movdf_internal): Likewise.
4664 (*movsf_internal): Likewise.
4665 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
4666 (<code><mode>3): Likewise.
4667 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
4669 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
4670 with the new constraint "v".
4671 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
4673 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
4674 with the new constraint "v".
4675 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
4676 (<sse2>_storedqu<avxsizesuffix>): Likewise.
4677 (*<plusminus_insn><mode>3): Likewise.
4678 (<sse>_vm<plusminus_insn><mode>3): Likewise.
4679 (*mul<mode>3): Likewise.
4680 (<sse>_vmmul<mode>3): Likewise.
4681 (<sse>_div<mode>3): Likewise.
4682 (<sse>_vmdiv<mode>3): Likewise.
4683 (<sse>_sqrt<mode>2): Likewise.
4684 (<sse>_vmsqrt<mode>2): Likewise.
4685 (*<code><mode>3_finite): Likewise.
4686 (*<code><mode>3) <smaxmin>: Likewise.
4687 (<sse>_vm<code><mode>3): Likewise.
4688 (*<code><mode>3) <any_logic>: Likewise.
4689 (*fma_fmadd_<mode>): Likewise.
4690 (*fma_fmsub_<mode>): Likewise.
4691 (*fma_fnmadd_<mode>): Likewise.
4692 (*fma_fnmsub_<mode>): Likewise.
4693 (*fma_fmaddsub_<mode>): Likewise.
4694 (*fma_fmsubadd_<mode>): Likewise.
4695 (*fmai_fmadd_<mode>): Likewise.
4696 (*fmai_fmsub_<mode>): Likewise.
4697 (*fmai_fnmadd_<mode>): Likewise.
4698 (*fmai_fnmsub_<mode>): Likewise.
4699 (sse_cvtsi2ss): Likewise.
4700 (sse_cvtsi2ssq): Likewise.
4701 (sse_cvtss2si): Likewise.
4702 (sse_cvtss2si_2): Likewise.
4703 (sse_cvtss2siq): Likewise.
4704 (sse_cvtss2siq_2): Likewise.
4705 (sse_cvttss2si): Likewise.
4706 (sse_cvtss2siq_2): Likewise.
4707 (float<sseintvecmodelower><mode>2): Likewise.
4708 (sse2_cvtsd2si_2): Likewise.
4709 (sse2_cvtsd2siq_2): Likewise.
4710 (*<plusminus_insn><mode>3): Likewise.
4711 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
4712 (*<sse4_1_avx2>_mul<mode>3): Likewise.
4713 (ashr<mode>3): Likewise.
4714 (<shift_insn><mode>3): Likewise.
4715 (avx2_<code><mode>3): Likewise.
4716 (*avx2_<code><mode>3): Likewise.
4717 (*andnot<mode>3): Likewise.
4718 (*<code><mode>3) <any_logic>: Likewise.
4719 (abs<mode>2): Likewise.
4720 (avx2_permvar<mode>): Likewise.
4721 (avx2_perm<mode>_1): Likewise.
4722 (*avx_vpermilp<mode>): Likewise.
4723 (avx_vpermilvar<mode>3): Likewise.
4724 (avx2_ashrv<mode>): Likewise.
4725 (avx2_<shift_insn>v<mode>): Likewise.
4726 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
4728 * doc/rtl.texi: Document XImode.
4730 2013-08-21 Jeff Law <law@redhat.com>
4732 * tree-flow.h (register_jump_thread): Pass vector of edges
4733 instead of each important edge.
4734 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
4735 thread path into a vector and pass that to register_jump_thread.
4736 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
4737 passed in edge vector to the current 3-edge form.
4740 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
4742 * dce.c (fini_dce): Call df_analyze again just in case
4743 delete_unmarked_insns removed anything.
4745 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
4747 * reload.h (struct reg_equivs): Rename to ..
4748 (struct reg_equivs_s): .. this.
4750 2013-08-20 Martin Liska <marxin.liska@gmail.com>
4752 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
4754 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4756 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
4758 2013-08-21 Jeff Law <law@redhat.com>
4760 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
4761 simplify assignments too. If the RHS collapses to a singleton
4762 range, then return the value for the range.
4764 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
4766 * config/i386/sse.md (V16): Rename to...
4768 (mov<mode>): Update iterator name.
4769 (*mov<mode>_internal): Ditto.
4770 (push<mode>1): Ditto.
4771 (movmisalign<mode>): Ditto.
4773 2013-08-20 Jan Hubicka <jh@suse.cz>
4776 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
4777 entry for direct edges.
4778 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
4780 2013-08-20 David Malcolm <dmalcolm@redhat.com>
4782 Revert my last two changes, r201865 and r201864:
4785 2013-08-20 David Malcolm <dmalcolm@redhat.com>
4787 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
4788 instances can own GC refs.
4790 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
4791 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
4792 (gcc::context::gt_pch_nx): Likewise.
4793 (gcc::context::gt_pch_nx): Likewise.
4794 * ggc.h (gt_ggc_mx <T>): New.
4795 (gt_pch_nx_with_op <T>): New.
4796 (gt_pch_nx <T>): New.
4797 * passes.c (opt_pass::gt_ggc_mx): New.
4798 (opt_pass::gt_pch_nx): New.
4799 (opt_pass::gt_pch_nx_with_op): New.
4800 (pass_manager::gt_ggc_mx): New.
4801 (pass_manager::gt_pch_nx): New.
4802 (pass_manager::gt_pch_nx_with_op): New.
4803 (pass_manager::operator new): Use
4804 ggc_internal_cleared_alloc_stat rather than xcalloc.
4805 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
4806 (pass_manager::gt_ggc_mx): New.
4807 (pass_manager::gt_pch_nx): New.
4808 (pass_manager::gt_pch_nx_with_op): New.
4809 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
4810 (opt_pass::operator new): New.
4811 (opt_pass::gt_ggc_mx): New.
4812 (opt_pass::gt_pch_nx): New.
4813 (opt_pass::gt_pch_nx_with_op): New.
4816 2013-08-20 David Malcolm <dmalcolm@redhat.com>
4818 * Makefile.in (GTFILES): Add context.h.
4819 * context.c (gcc::context::operator new): New.
4820 (gcc::context::gt_ggc_mx): New.
4821 (gcc::context::gt_pch_nx): New.
4822 (gcc::context::gt_pch_nx): New.
4823 * context.h (gcc::context): Add GTY((user)) marking.
4824 (gcc::context::operator new): New.
4825 (gcc::context::gt_ggc_mx): New.
4826 (gcc::context::gt_pch_nx): New.
4827 (gcc::context::gt_pch_nx): New.
4828 (g): Add GTY marking.
4829 (gt_ggc_mx (gcc::context *)): New.
4830 (gt_pch_nx (gcc::context *)): New.
4831 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
4832 void *cookie)): New.
4833 * gengtype.c (open_base_files) <ifiles>: Add context.h.
4835 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
4837 * dce.c (fini_dce): Call df_analyze again just in case
4838 delete_unmarked_insns removed anything.
4840 2013-08-20 Teresa Johnson <tejohnson@google.com>
4842 PR rtl-optimizations/57451
4843 * final.c (reemit_insn_block_notes): Prevent lexical blocks
4844 from crossing split section boundaries.
4846 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
4848 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
4849 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
4850 with MULTLIB_DEFAULTS.
4852 2013-08-20 Nick Clifton <nickc@redhat.com>
4854 * target.def (narrow_volatile_bitfield): Note that the default
4855 value is false, not !TARGET_STRICT_ALIGN.
4856 * doc/tm.texi: Regenerate.
4858 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
4860 Fix LIB_SPEC for systems without libpthread.
4862 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
4863 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
4865 * config/i386/linux-common.h: Likewise.
4866 * config/mips/linux-common.h: Likewise.
4868 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
4870 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
4873 2013-08-20 David Malcolm <dmalcolm@redhat.com>
4875 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
4876 instances can own GC refs.
4878 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
4879 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
4880 (gcc::context::gt_pch_nx): Likewise.
4881 (gcc::context::gt_pch_nx): Likewise.
4882 * ggc.h (gt_ggc_mx <T>): New.
4883 (gt_pch_nx_with_op <T>): New.
4884 (gt_pch_nx <T>): New.
4885 * passes.c (opt_pass::gt_ggc_mx): New.
4886 (opt_pass::gt_pch_nx): New.
4887 (opt_pass::gt_pch_nx_with_op): New.
4888 (pass_manager::gt_ggc_mx): New.
4889 (pass_manager::gt_pch_nx): New.
4890 (pass_manager::gt_pch_nx_with_op): New.
4891 (pass_manager::operator new): Use
4892 ggc_internal_cleared_alloc_stat rather than xcalloc.
4893 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
4894 (pass_manager::gt_ggc_mx): New.
4895 (pass_manager::gt_pch_nx): New.
4896 (pass_manager::gt_pch_nx_with_op): New.
4897 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
4898 (opt_pass::operator new): New.
4899 (opt_pass::gt_ggc_mx): New.
4900 (opt_pass::gt_pch_nx): New.
4901 (opt_pass::gt_pch_nx_with_op): New.
4903 2013-08-20 David Malcolm <dmalcolm@redhat.com>
4905 * Makefile.in (GTFILES): Add context.h.
4906 * context.c (gcc::context::operator new): New.
4907 (gcc::context::gt_ggc_mx): New.
4908 (gcc::context::gt_pch_nx): New.
4909 (gcc::context::gt_pch_nx): New.
4910 * context.h (gcc::context): Add GTY((user)) marking.
4911 (gcc::context::operator new): New.
4912 (gcc::context::gt_ggc_mx): New.
4913 (gcc::context::gt_pch_nx): New.
4914 (gcc::context::gt_pch_nx): New.
4915 (g): Add GTY marking.
4916 (gt_ggc_mx (gcc::context *)): New.
4917 (gt_pch_nx (gcc::context *)): New.
4918 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
4919 void *cookie)): New.
4920 * gengtype.c (open_base_files) <ifiles>: Add context.h.
4922 2013-08-20 Alan Modra <amodra@gmail.com>
4925 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
4926 (rs6000_emit_epilogue): Likewise.
4928 2013-08-19 Dehao Chen <dehao@google.com>
4930 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
4932 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
4933 Jakub Jelinek <jakub@redhat.com>
4935 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
4936 (BUILT_IN_FABSD64): Likewise.
4937 (BUILT_IN_FABSD128): Likewise.
4938 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
4939 (fold_builtin_1): Likewise.
4940 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
4941 destination and source operands.
4942 (*abstd2_fpr): Likewise.
4943 (*nabstd2_fpr): Likewise.
4945 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
4947 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
4950 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
4952 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
4955 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
4957 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
4958 * builtins.c (default_libc_has_function): New.
4959 (gnu_libc_has_function): Ditto.
4960 (no_c99_libc_has_function): Ditto.
4961 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
4962 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
4963 (fold_builtin_sincos): Likewise.
4964 (fold_builtin_cexp): Likewise.
4965 * builtins.def (DEF_C94_BUILTIN): Likewise.
4966 (DEF_C99_BUILTIN): Likewise.
4967 (DEF_C99_C90RES_BUILTIN): Likewise.
4968 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
4969 definitions to using this define.
4970 * config/darwin-protos.h (darwin_libc_has_function): New.
4971 * config/darwin.c (darwin_libc_has_function): Ditto.
4972 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
4973 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
4974 * config/darwin.h: Ditto.
4975 * config/elfos.h: Ditto.
4976 * config/freebsd.h: Ditto.
4977 * config/i386/cygming.h: Ditto.
4978 * config/i386/djgpp.h: Ditto.
4979 * config/i386/i386-interix.h: Ditto.
4980 * config/microblaze/microblaze.h: Ditto.
4981 * config/mmix/mmix.h: Ditto.
4982 * config/gnu-user.h: Ditto.
4983 * config/ia64/hpux.h: Ditto.
4984 * config/pa/pa-hpux.h: Ditto.
4985 * config/pdp11/pdp11.h: Ditto.
4986 * config/picochip/picochip.h: Ditto.
4987 * config/linux.h: Ditto.
4988 * config/netbsd.h: Ditto.
4989 * config/openbsd.h: Ditto.
4990 * config/rs6000/aix43.h: Ditto.
4991 * config/rs6000/aix51.h: Ditto.
4992 * config/rs6000/aix52.h: Ditto.
4993 * config/rs6000/aix53.h: Ditto.
4994 * config/rs6000/aix61.h: Ditto.
4995 * config/rs6000/darwin.h: Ditto.
4996 * config/rs6000/linux.h: Ditto.
4997 * config/rs6000/linux64.h: Ditto.
4998 * config/s390/tpf.h: Ditto.
4999 * config/sol2-10.h: Ditto.
5000 * config/sol2.h: Ditto.
5001 * config/vms/vms.h: Ditto.
5002 * config/vxworks.h: Ditto.
5003 * config/linux-android.c (linux_android_libc_has_function):
5004 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
5005 * config/linux-protos.h (linux_android_libc_has_function):
5007 * config/i386/i386.c (ix86_libc_has_function): New.
5008 * config/i386/i386-protos.h
5009 (ix86_libc_has_function): New declaration.
5010 * config/i386/i386.md
5011 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
5012 ("isinf<mode>2): Likewise.
5013 * convert.c (convert_to_integer): Using new target hook
5014 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
5015 TARGET_C99_FUNCTIONS.
5016 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
5017 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
5018 * coretypes.h (function_class): New enum for different
5019 classes of functions.
5020 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
5021 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
5022 (TARGET_HAS_SINCOS): Likewise.
5023 (TARGET_LIBC_HAS_FUNCTION): New.
5024 * doc/tm.texi: Regenerated.
5025 * targhooks.h (default_libc_has_function): New declaration.
5026 (no_c99_libc_has_function): Ditto.
5027 (gnu_libc_has_function): Ditto.
5028 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
5029 and TARGET_HAS_SINCOS.
5031 2013-08-18 Jan Hubicka <jh@suse.cz>
5033 * Makeifle-in (ipa-devirt.o): New.
5034 (GTFILES): Add ipa-utils.h and ipa-devirt.c
5035 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
5036 (analyze_functions): Look into possible targets of polymorphic call.
5037 * dumpfile.c (dump_files): Add type-inheritance dump.
5038 * dumpfile.h (TDI_inheritance): New.
5039 * ipa-devirt.c: New file.
5040 * ipa-utils.h (odr_type_d): Forward declare.
5041 (odr_type): New type.
5042 (build_type_inheritance_graph): Declare.
5043 (possible_polymorphic_call_targets): Declare and introduce inline
5044 variant when only edge is pased.
5045 (dump_possible_polymorphic_call_targets): Likewise.
5046 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
5047 * tree.c (type_in_anonymous_namespace_p): Break out from ...
5048 (types_same_for_odr): ... here.
5049 * tree.h (type_in_anonymous_namespace_p): Declare.
5051 2013-08-18 Jakub Jelinek <jakub@redhat.com>
5053 PR tree-optimization/58006
5054 * tree-parloops.c (take_address_of): Don't ICE if get_name
5056 (eliminate_local_variables_stmt): Remove clobber stmts.
5058 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
5060 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
5061 error message is issued for an alias to undefined symbol.
5063 2013-08-18 Jan Hubicka <jh@suse.cz>
5065 * cgraph.c (cgraph_create_indirect_edge): Discover
5066 polymorphic calls and record basic info into indirect_info.
5067 * gimple-fold.c (gimple_fold_call): When doing BINFO based
5068 devirtualization, ignore objc function calls.
5069 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
5070 call with no parm index info.
5071 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
5072 * tree.c (virtual_method_call_p): New function.
5073 * tree.h (virtual_method_call_p): Declare.
5075 2013-08-16 Jan Hubicka <jh@suse.cz>
5078 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
5080 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
5082 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
5085 2013-08-16 David Malcolm <dmalcolm@redhat.com>
5087 * gengtype.c (type_for_name): Add special-case support for
5088 locating types within the "gcc::" namespace.
5089 (open_base_files): Emit a "using namespace gcc" directive.
5091 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
5094 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
5095 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
5097 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
5098 array instead of each individual operand as a separate argument.
5099 (emit_fusion_gpr_load): Likewise.
5100 (expand_fusion_gpr_load): Add new function declaration.
5102 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
5103 signature to have the operands passed as an array, instead of as
5104 separate arguments. Allow ZERO_EXTEND to be in the memory
5105 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
5106 depend on the register live/dead flags when peepholes are run.
5107 (expand_fusion_gpr_load): New function to be called from the
5108 peephole2 pass, to change the register that addis sets to be the
5110 (emit_fusion_gpr_load): Change the calling signature to have the
5111 operands passed as an array, instead of as separate arguments.
5112 Allow ZERO_EXTEND to be in the memory address, and also
5113 SIGN_EXTEND if -mpower8-fusion-sign.
5115 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
5117 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
5118 adjust the register addis loads up in the peephole2 pass. Do not
5119 depend on the register live/dead state when the peephole pass is done.
5121 2013-08-16 David Malcolm <dmalcolm@redhat.com>
5123 * gengtype.c (create_user_defined_type): Ensure that the kind
5124 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
5125 declaration is seen before the GTY((user)) marking.
5127 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
5130 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
5132 2013-08-16 Jan Hubicka <jh@suse.cz>
5134 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
5135 arugment expected_type.
5136 (gimple_fold_call): Use it.
5137 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
5138 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
5139 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
5140 (try_make_edge_direct_virtual_call): Likewise.
5141 * tree.c (obj_type_ref_class): New.
5142 * tree.h (obj_type_ref_class): Use it.
5144 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
5146 * sched-vis.c (rtl_slim_pp_initialized): Remove.
5147 (rtl_slim_pp): Likewise.
5148 (init_rtl_slim_pretty_print): Likewise.
5149 (dump_value_slim): Don't call it. Use local pretty printer.
5150 (dump_insn_slim): Likewise.
5151 (dump_rtl_slim): Likewise.
5152 (str_pattern_slim): Likewise.
5153 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
5156 2013-08-16 Jakub Jelinek <jakub@redhat.com>
5158 PR tree-optimization/58164
5159 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
5160 walk gimple_goto_dest of GIMPLE_GOTO.
5162 PR tree-optimization/58165
5163 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
5164 bi_call must be the last stmt in a bb, don't split_block, instead
5165 use fallthru edge from it and give up if there is none.
5166 Release conds vector when returning early.
5168 2013-08-14 Xinliang David Li <davidxl@google.com>
5170 * config/i386/i386.c (ix86_option_override_internal):
5171 Remove unused variable and field.
5173 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5176 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
5177 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
5178 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
5179 and Linux, correct BLKmode alignment when 128-bit alignment is
5180 required and compatibility flag is not set.
5181 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
5182 for zero-size arguments when compatibility flag is not set.
5184 2013-08-14 Jakub Jelinek <jakub@redhat.com>
5186 PR tree-optimization/58145
5187 * tree-sra.c (build_ref_for_offset): If prev_base has
5188 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
5190 2013-08-14 Xinliang David Li <davidxl@google.com>
5192 * config/i386/i386.c (ix86_option_override_internal):
5193 Fix uninitialized variable error.
5195 2013-08-14 Xinliang David Li <davidxl@google.com>
5197 * config/i386/i386.opt: Define two new options.
5198 * config/i386/x86-tune.def: Add arch selector field in macros.
5199 * config/i386/i386.h: Adjust macro definition.
5200 * config/i386/i386.c (ix86_option_override_internal):
5202 (parse_mtune_ctrl_str): New function.
5203 (set_ix86_tune_features): New function.
5204 (ix86_function_specific_restore): Call the new helper function.
5206 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
5208 PR rtl-optimization/57662
5209 * sel-sched.c (code_motion_process_successors): When the current insn
5210 is removed after the recursive traversal, break from the loop.
5211 Add comments and debug printouts.
5213 2013-08-14 Jakub Jelinek <jakub@redhat.com>
5214 Alexandre Oliva <aoliva@redhat.com>
5217 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
5218 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
5219 there also UNSPEC_PLTOFF.
5221 2013-08-14 Marek Polacek <polacek@redhat.com>
5223 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
5224 NUM_CONDITIONS bit positions.
5226 2013-08-13 Cary Coutant <ccoutant@google.com>
5228 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
5229 (attr_checksum): Hash vector contents instead of pointer.
5230 (attr_checksum_ordered): Likewise.
5232 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
5234 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
5235 when Pmode != word_mode. Add length_address attribute.
5236 (sse3_monitor_<mode>): Merge from sse3_monitor and
5237 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
5238 Pmode != word_mode. Update insn length attribute.
5239 * config/i386/i386.c (ix86_option_override_internal): Update
5240 ix86_gen_monitor selection for merged sse3_monitor insn.
5242 2013-08-13 Julian Brown <julian@codesourcery.com>
5244 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
5245 perform invalid legitimization on greater-than-word-size modes for
5248 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
5250 * ira.c (setup_class_translate_array): Use aclass instead of cl
5251 for classes not fully covered by allocno classes.
5253 2013-08-13 Jakub Jelinek <jakub@redhat.com>
5255 PR tree-optimization/57661
5256 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
5257 * tree-inline.c (tree_function_versioning): Initialize it.
5258 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
5259 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
5260 that is not being copied.
5263 * asan.c (instrument_strlen_call): Fix typo in comment.
5264 Use char * type even for the lhs of POINTER_PLUS_EXPR.
5266 2013-08-13 Steve Ellcey <sellcey@mips.com>
5268 * config/mips/mips.md (prefetch): Use lw instead of ld on
5269 loongson in 32bit mode.
5271 2013-08-13 Nick Clifton <nickc@redhat.com>
5273 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
5275 2013-08-13 Jan Hubicka <jh@suse.cz>
5277 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
5278 introduced edge; fix typo in sanity check.
5279 (cgraph_resolve_speculation): Export; improve diagnostic.
5280 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
5281 speculation at type mismatch.
5282 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
5283 (cgraph_resolve_speculation): Declare.
5284 (symtab_can_be_discarded): New function.
5285 * value-prof.c (gimple_ic_transform): Remove actual transform code.
5286 * ipa-inline-transform.c (speculation_removed): New global var.
5287 (clone_inlined_nodes): See if speculation can be removed.
5288 (inline_call): If speculations was removed, we growths may not match.
5289 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
5290 (speculation_useful_p): New function.
5291 (resolve_noninline_speculation): New function.
5292 (inline_small_functions): Resolve useless speculations.
5293 * ipa-inline.h (speculation_useful_p): Declare
5294 * ipa.c (can_replace_by_local_alias): Simplify.
5295 (ipa_profile): Produce speculative calls in non-lto, too;
5296 add simple cost model; produce local aliases.
5298 2013-08-13 David Malcolm <dmalcolm@redhat.com>
5300 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
5303 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
5305 * config/i386/i386.c (ix86_function_versions): Use error + inform.
5307 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
5309 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
5310 iterator instead of X87MODEF.
5312 2013-08-12 Perez Read <netfirewall@gmail.com>
5315 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
5316 operand 0 for intel asm alternative.
5317 (*movabs<mode>_2): Ditto for operand 1.
5319 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
5321 * config/aarch64/arm_none.h
5322 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
5324 2013-08-12 Nick Clifton <nickc@redhat.com>
5326 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
5328 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
5330 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
5331 expand for QI/HImode operand to produce more effictive code for
5332 unsigned char(short) --> float(double) conversion.
5334 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
5336 * doc/invoke.texi: Mention that -ftls-model does not force the final
5339 2013-08-12 Marek Polacek <polacek@redhat.com>
5340 Marc Glisse <marc.glisse@inria.fr>
5342 PR tree-optimization/57980
5343 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
5344 when creating -1 constant.
5346 2013-08-10 Jan Hubicka <jh@suse.cz>
5348 Workaround binutils PR14342.
5349 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
5350 (gimple_init_edge_profiler): Likewise.
5351 (gimple_gen_ic_func_profiler): Likewise.
5353 2013-08-09 Jan Hubicka <jh@suse.cz>
5355 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
5357 2013-08-09 Xinliang David Li <davidxl@google.com>
5359 * config/i386/stringop.def: New file.
5360 * config/i386/stringop.opt: New file.
5361 * config/i386/i386-opts.h: Include stringopt.def.
5362 * config/i386/i386.opt: Include stringopt.opt.
5363 * config/i386/i386.c (ix86_option_override_internal):
5364 Override default size based stringop inline strategies with options.
5365 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
5368 2013-08-09 Jan Hubicka <jh@suse.cz>
5370 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
5372 2013-08-09 Jan Hubicka <jh@suse.cz>
5374 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
5376 (dump_cgraph_node): Dump profile-id.
5377 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
5378 and common_target_probability.
5379 * lto-cgraph.c (lto_output_edge): Stream common targets.
5380 (lto_output_node): Stream profile ids.
5381 (input_node): Stream profile ids.
5382 (input_edge): Stream common targets.
5383 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
5384 * ipa.c: Include value-prof.h
5385 (ipa_profile_generate_summary): Turn indirect call statement histograms
5386 into common targets.
5387 (ipa_profile): Turn common targets into speculative edges.
5389 2013-08-09 Jan Hubicka <jh@suse.cz>
5391 * cgraph.h (cgraph_node): Add profile_id.
5392 * value-prof.c (cgraph_node_map): Turn into pointer_map.
5393 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
5394 (del_node_map): Update.
5395 (find_func_by_funcdef_no): Replace by ...
5396 (find_func_by_profile_id): ... this one.
5397 (gimple_ic_transform): Do not remove useful histograms when
5398 speculation is not done; dump info when indirect call removal
5400 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
5401 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
5402 (__gcov_indirect_call_profiler_v2): .. this one.
5403 * profile.h (init_node_map): Update.
5404 * coverage.c (coverage_compute_profile_id): New function.
5405 * coverage.h (coverage_compute_profile_id): Declare.
5406 * tree-profile.c (init_ic_make_global_vars): Make
5407 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
5408 (gimple_init_edge_profiler): Update prototype of
5409 __gcov_indirect_call_profiler.
5410 (gimple_gen_ic_func_profiler): Simplify.
5411 (tree_profiling): Use init_node_map
5413 2013-08-09 Jan Hubicka <jh@suse.cz>
5415 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
5416 non-speculative refs.
5417 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
5418 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
5419 (cgraph_set_call_stmt): Likewise.
5420 (cgraph_create_edge_1): Fix release checking compilatoin;
5422 (cgraph_free_edge): Free indirect info.
5423 (cgraph_turn_edge_to_speculative): New function.
5424 (cgraph_speculative_call_info): New function.
5425 (cgraph_make_edge_direct): Return direct edge; handle speculation.
5426 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
5427 (dump_cgraph_node): Dump speculation.
5428 (verify_edge_count_and_frequency): Accept speculative edges.
5429 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
5430 (verify_cgraph_node): Handle speculation.
5431 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
5432 (cgraph_set_call_stmt): Update prototype.
5433 (cgraph_make_edge_direct): Update prototype.
5434 (cgraph_speculative_call_info): Declare.
5435 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
5436 to change; update call of ipa_find_references.
5437 * ipa-ref.c (ipa_record_reference): Fix return value; clear
5438 lto_stmt_uid and speculative flags.
5439 (ipa_dump_references): Dump speculation.
5440 (ipa_clone_references): Clone speculative flag.
5441 (ipa_clone_referring): Likewise.
5442 (ipa_clone_ref): New function.
5443 (ipa_find_reference): Look into lto_stmt_uids
5444 (ipa_clear_stmts_in_references): Do not clear speculative calls.
5445 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
5446 (ipa_find_reference): Update declaration.
5447 (ipa_clone_ref): Declare.
5448 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
5449 stream speculative flag.
5450 (lto_output_ref): Stream statements uids and speculation.
5451 (input_ref): Likewise.
5452 (input_edge): Stream speuclation.
5453 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
5454 (cgraph_set_call_stmt_including_clones): Handle speculation.
5455 * ipa-inline.c (heap_edge_removal_hook): New function.
5456 (inline_small_functions): Register it.
5457 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
5458 also initialize refs.
5459 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
5461 (try_make_edge_direct_simple_call): Likewise.
5462 (try_make_edge_direct_simple_call): Likewise.
5463 (update_indirect_edges_after_inlining): Likewise.
5464 (remove_described_reference): Look proper lto_stmt_uid.
5465 (propagate_controlled_uses): Likewise.
5466 (propagate_controlled_uses): Liekwise.
5467 * tree-inline.c (copy_bb): Copy speculative edges.
5468 (redirect_all_calls): New function.
5469 (copy_cfg_body): Do redirection after loop info is updated.
5470 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
5472 2013-08-09 Jan Hubicka <jh@suse.cz>
5474 * lto-streamer-out.c (output_function): Renumber PHIs.
5475 * lto-streamer-in.c (input_function): Likewise.
5477 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
5479 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
5480 (get_lane_unsigned): Likewise.
5481 (dup_lane_scalar): Likewise.
5482 (get_lane): enable for VALL.
5483 * config/aarch64/aarch64-simd.md
5484 (aarch64_dup_lane_scalar<mode>): Remove.
5485 (aarch64_get_lane_signed<mode>): Likewise.
5486 (aarch64_get_lane_unsigned<mode>): Likewise.
5487 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
5488 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
5489 (aarch64_get_lane<mode>): Enable for all vector modes.
5490 (aarch64_get_lanedi): Remove misleading constraints.
5491 * config/aarch64/arm_neon.h
5492 (__aarch64_vget_lane_any): Define.
5493 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
5494 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
5495 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
5496 * config/aarch64/iterators.md (VDQQH): New.
5500 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
5502 * configure.ac: Add GAS check for LEON instructions on SPARC.
5503 * configure: Regenerate.
5504 * config.in: Likewise.
5505 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
5507 * config/sparc/sparc.opt (LEON, LEON3): New masks.
5508 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
5510 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
5511 (AS_LEON_FLAG): New macro.
5512 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
5513 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
5514 Deal with LEON and LEON3 for the memory model.
5515 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
5516 (atomic_compare_and_swap<mode>_1): Likewise.
5517 (*atomic_compare_and_swap<mode>_1): Likewise.
5519 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
5521 * config/arm/neon.md (vcond): Fix floating-point vector
5522 comparisons against 0.
5524 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
5526 * lra-constraints.c (emit_spill_move): Remove assert.
5527 (process_alt_operands): Add more debugging
5528 output. Increase reject for spilling into memory. Decrease
5529 reject for reloading scratch.
5530 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
5532 2013-08-08 Steve Ellcey <sellcey@mips.com>
5534 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
5535 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
5536 micromips incompatible. Add nan2008.
5537 (MULTILIB_DIRNAMES): Add nan2008.
5538 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
5539 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
5540 and micromips incompatible. Add nan2008.
5541 (MULTILIB_DIRNAMES): Add nan2008.
5542 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
5544 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
5546 PR rtl-optimization/58079
5547 * combine.c (combine_simplify_rtx): Avoid using SUBST if
5548 simplify_comparison has widened a comparison with an integer.
5550 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5552 * config/arm/neon.md (movmisalign<mode>): Disable when we
5553 don't allow unaligned accesses.
5554 (*movmisalign<mode>_neon_store): Likewise.
5555 (*movmisalign<mode>_neon_load): Likewise.
5556 (*movmisalign<mode>_neon_store): Likewise.
5557 (*movmisalign<mode>_neon_load): Likewise.
5559 2013-08-08 Jan Hubicka <jh@suse.cz>
5561 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
5562 (make_pass_rebuild_cgraph_edges): Also clear references.
5563 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
5564 * ipa-inline-transform.c (inline_transform): Remove all references
5566 * cgraphunit.c (expand_function): Remove all references after
5568 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
5569 (ipa_find_reference): Rewrite to iterator.
5570 (remove_stmt_references): Likewise.
5571 (ipa_clear_stmts_in_references): New function.
5572 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
5573 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
5575 * ipa-split.c (split_function): Remove references in split function.
5577 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
5580 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
5581 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
5583 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
5586 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
5587 suggested mode for the assignment isn't compatible with the
5590 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
5593 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
5595 2013-08-07 Xinliang David Li <davidxl@google.com>
5597 * config/i386/i386.opt: New option -mtune-ctrl=.
5598 * config/i386/x86-tune.def: New file.
5599 * config/i386/i386.h: include x86-tune.def.
5600 * config/i386/i386.c (ix86_option_override_internal):
5601 Parsing -mtune-ctrl= option and set tune features.
5603 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
5606 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
5607 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
5610 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
5612 * rtl.h (update_alignments): Declare.
5613 * final.c (grow_label_align): New function extracted from...
5614 (shorten_branches): ...here. Call it.
5615 (update_alignments): New function.
5616 * reorg.c (sibling_labels): New variable.
5617 (get_label_before): Add SIBLING parameter. If it is non-zero, push
5618 the new label along with it onto the sibling_labels vector.
5619 (fill_simple_delay_slots): Adjust call to get_label_before.
5620 (fill_slots_from_thread): Likewise.
5621 (relax_delay_slots): Likewise.
5622 (make_return_insns): Likewise.
5623 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
5625 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
5627 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
5628 document its semantics.
5629 (diagnostic_report_diagnostic): Adjust accordingly.
5631 2013-08-07 David Malcolm <dmalcolm@redhat.com>
5633 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
5634 (sparc_option_override): ...and port to new C++ pass API.
5635 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
5637 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
5639 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
5641 2013-08-06 Caroline Tice <cmtice@google.com>
5643 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
5644 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
5645 * tree-pass.h: Add pass_vtable_verify.
5646 * varasm.c (assemble_variable): Add code to properly set the comdat
5647 section and name for the .vtable_map_vars section.
5648 (assemble_vtyv_preinit_initializer): New function.
5649 (default_sectin_type_flags): Make sure .vtable_map_vars section has
5651 * output.h: Add function decl for assemble_vtv_preinit_initializer.
5652 * vtable-verify.c: New file.
5653 * vtable-verify.h: New file.
5654 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
5655 initialiation levels.
5656 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
5657 * passes.def: Insert pass_vtable_verify.
5658 * aclocal.m4: Reorder includes.
5659 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
5660 -fvtv-counts options.
5661 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
5662 as appropriate, if -fvtable-verify=... is used.
5663 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
5664 -fvtable-verify=... is used.
5665 * Makefile.in (OBJS): Add vtable-verify.o to list.
5666 (vtable-verify.o): Add new build rule.
5667 (GTFILES): Add vtable-verify.c to list.
5668 * common.opt (fvtable-verify=): New flag.
5669 (vtv_priority): Values for fvtable-verify= flag.
5670 (fvtv-counts): New flag.
5671 (fvtv-debug): New flag.
5672 * tree.h (save_vtable_map_decl): New extern function decl.
5674 2013-08-07 David Malcolm <dmalcolm@redhat.com>
5676 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
5677 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
5678 (pass_data_rl78_devirt): ...new pass_data instance and...
5679 (make_pass_rl78_devirt): ...new function.
5680 (rl78_asm_file_start): Port pass registration to new C++ API.
5682 2013-08-07 David Malcolm <dmalcolm@redhat.com>
5684 * coretypes.h (rtl_opt_pass): Add.
5685 (gcc::context): Add.
5686 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
5687 (epiphany_init): Port to new C++ pass API.
5688 (epiphany_optimize_mode_switching): Likewise.
5689 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
5690 (pass_manager::get_pass_mode_switching): New.
5691 (pass_manager::get_pass_peephole2): New.
5692 * mode-switching.c (pass_mode_switching): Add clone method.
5693 * recog.c (pass_peephole2): Add clone method.
5694 (pass_split_all_insns): Add clone method.
5696 2013-08-06 David Malcolm <dmalcolm@redhat.com>
5698 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
5699 (mips_option_override): ...here, porting to new C++ API for passes.
5701 2013-08-06 Jan Hubicka <jh@suse.cz>
5703 * cgraph.c (cgraph_get_body): New function based on lto.c
5705 * cgraph.h (cgraph_get_body): Declare.
5706 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
5708 * cgraphunit.c (expand_function): Get body prior expanding.
5709 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
5710 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
5712 * passes.c (do_per_function_toporder): Get body.
5713 * tree-inline.c (expand_call_inline): Get body prior inlining it.
5714 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
5716 2013-08-06 Martin Jambor <mjambor@suse.cz>
5719 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
5720 is not re-finalized. Rename second parameter to no_collect.
5722 2013-08-06 Martin Jambor <mjambor@suse.cz>
5725 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
5726 MEM_REF has proper alignment information.
5728 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
5731 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
5733 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
5734 rtx (*) (rtx, ...) with insn_gen_fn.
5735 * genoutput.c (output_insn_data): Cast gen_? function pointers to
5736 insn_gen_fn::stored_funcptr. Add initializer braces.
5738 2013-08-05 David Malcolm <dmalcolm@redhat.com>
5740 Rewrite how instances of passes are cloned to remove assumptions
5741 about their sizes (thus allowing pass subclasses to have
5742 additional data fields, albeit non-GC-managed ones at this point).
5744 * passes.c (make_pass_instance): Now that passes have clone
5745 methods, rewrite this function to eliminate XNEW and memcpy
5746 calls that used hardcoded sizes. Since this function no longer
5747 creates pass instances, rename it to...
5748 (add_pass_instance): ...this. Document the old way that passes were
5749 numbered and flagged, and rework this function to continue using it.
5750 (next_pass_1): Add an initial_pass argument for use by
5752 (position_pass): When adding multiple instances of a pass, use
5753 the pass's clone method, rather than relying on the XNEW/memcpy
5754 within the former make_pass_instance (now add_pass_instance).
5755 (pass_manager::pass_manager): When invoking next_pass_1, also supply
5756 the initial instance of the current pass within the pass manager.
5758 2013-08-05 David Malcolm <dmalcolm@redhat.com>
5760 This is the automated part of the conversion of passes from C
5761 structs to C++ classes.
5763 Patch autogenerated by refactor_passes.py from
5764 https://github.com/davidmalcolm/gcc-refactoring-scripts
5765 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
5767 * asan.c (pass_asan): Convert from a global struct to a subclass of
5768 gimple_opt_pass along with...
5769 (pass_data_asan): ...new pass_data instance and...
5770 (make_pass_asan): ...new function.
5771 (pass_asan_O0): Convert from a global struct to a subclass of
5772 gimple_opt_pass along with...
5773 (pass_data_asan_O0): ...new pass_data instance and...
5774 (make_pass_asan_O0): ...new function.
5775 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
5776 subclass of rtl_opt_pass along with...
5777 (pass_data_inc_dec): ...new pass_data instance and...
5778 (make_pass_inc_dec): ...new function.
5779 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
5780 a subclass of rtl_opt_pass along with...
5781 (pass_data_reorder_blocks): ...new pass_data instance and...
5782 (make_pass_reorder_blocks): ...new function.
5783 (pass_duplicate_computed_gotos): Convert from a global struct to a
5784 subclass of rtl_opt_pass along with...
5785 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
5786 (make_pass_duplicate_computed_gotos): ...new function.
5787 (pass_partition_blocks): Convert from a global struct to a subclass of
5788 rtl_opt_pass along with...
5789 (pass_data_partition_blocks): ...new pass_data instance and...
5790 (make_pass_partition_blocks): ...new function.
5791 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
5792 struct to a subclass of rtl_opt_pass along with...
5793 (pass_data_branch_target_load_optimize1): ...new pass_data instance
5795 (make_pass_branch_target_load_optimize1): ...new function.
5796 (pass_branch_target_load_optimize2): Convert from a global struct to a
5797 subclass of rtl_opt_pass along with...
5798 (pass_data_branch_target_load_optimize2): ...new pass_data instance
5800 (make_pass_branch_target_load_optimize2): ...new function.
5801 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
5802 of rtl_opt_pass along with...
5803 (pass_data_jump): ...new pass_data instance and...
5804 (make_pass_jump): ...new function.
5805 (pass_jump2): Convert from a global struct to a subclass of
5806 rtl_opt_pass along with...
5807 (pass_data_jump2): ...new pass_data instance and...
5808 (make_pass_jump2): ...new function.
5809 * cfgexpand.c (pass_expand): Convert from a global struct to a
5810 subclass of rtl_opt_pass along with...
5811 (pass_data_expand): ...new pass_data instance and...
5812 (make_pass_expand): ...new function.
5813 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
5814 of rtl_opt_pass along with...
5815 (pass_data_free_cfg): ...new pass_data instance and...
5816 (make_pass_free_cfg): ...new function.
5817 (pass_into_cfg_layout_mode): Convert from a global struct to a
5818 subclass of rtl_opt_pass along with...
5819 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
5820 (make_pass_into_cfg_layout_mode): ...new function.
5821 (pass_outof_cfg_layout_mode): Convert from a global struct to a
5822 subclass of rtl_opt_pass along with...
5823 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
5824 (make_pass_outof_cfg_layout_mode): ...new function.
5825 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
5826 struct to a subclass of gimple_opt_pass along with...
5827 (pass_data_build_cgraph_edges): ...new pass_data instance and...
5828 (make_pass_build_cgraph_edges): ...new function.
5829 (pass_rebuild_cgraph_edges): Convert from a global struct to a
5830 subclass of gimple_opt_pass along with...
5831 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
5832 (make_pass_rebuild_cgraph_edges): ...new function.
5833 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
5834 subclass of gimple_opt_pass along with...
5835 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
5837 (make_pass_remove_cgraph_callee_edges): ...new function.
5838 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
5839 struct to a subclass of rtl_opt_pass along with...
5840 (pass_data_stack_adjustments): ...new pass_data instance and...
5841 (make_pass_stack_adjustments): ...new function.
5842 * combine.c (pass_combine): Convert from a global struct to a subclass
5843 of rtl_opt_pass along with...
5844 (pass_data_combine): ...new pass_data instance and...
5845 (make_pass_combine): ...new function.
5846 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
5847 global struct to a subclass of rtl_opt_pass along with...
5848 (pass_data_compare_elim_after_reload): ...new pass_data instance
5850 (make_pass_compare_elim_after_reload): ...new function.
5851 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
5852 of rtl_opt_pass along with...
5853 (pass_data_rtl_cprop): ...new pass_data instance and...
5854 (make_pass_rtl_cprop): ...new function.
5855 * cse.c (pass_cse): Convert from a global struct to a subclass of
5856 rtl_opt_pass along with...
5857 (pass_data_cse): ...new pass_data instance and...
5858 (make_pass_cse): ...new function.
5859 (pass_cse2): Convert from a global struct to a subclass of
5860 rtl_opt_pass along with...
5861 (pass_data_cse2): ...new pass_data instance and...
5862 (make_pass_cse2): ...new function.
5863 (pass_cse_after_global_opts): Convert from a global struct to a
5864 subclass of rtl_opt_pass along with...
5865 (pass_data_cse_after_global_opts): ...new pass_data instance and...
5866 (make_pass_cse_after_global_opts): ...new function.
5867 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
5868 of rtl_opt_pass along with...
5869 (pass_data_ud_rtl_dce): ...new pass_data instance and...
5870 (make_pass_ud_rtl_dce): ...new function.
5871 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
5872 rtl_opt_pass along with...
5873 (pass_data_fast_rtl_dce): ...new pass_data instance and...
5874 (make_pass_fast_rtl_dce): ...new function.
5875 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
5876 a subclass of rtl_opt_pass along with...
5877 (pass_data_df_initialize_opt): ...new pass_data instance and...
5878 (make_pass_df_initialize_opt): ...new function.
5879 (pass_df_initialize_no_opt): Convert from a global struct to a
5880 subclass of rtl_opt_pass along with...
5881 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
5882 (make_pass_df_initialize_no_opt): ...new function.
5883 (pass_df_finish): Convert from a global struct to a subclass of
5884 rtl_opt_pass along with...
5885 (pass_data_df_finish): ...new pass_data instance and...
5886 (make_pass_df_finish): ...new function.
5887 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
5888 rtl_opt_pass along with...
5889 (pass_data_rtl_dse1): ...new pass_data instance and...
5890 (make_pass_rtl_dse1): ...new function.
5891 (pass_rtl_dse2): Convert from a global struct to a subclass of
5892 rtl_opt_pass along with...
5893 (pass_data_rtl_dse2): ...new pass_data instance and...
5894 (make_pass_rtl_dse2): ...new function.
5895 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
5896 subclass of rtl_opt_pass along with...
5897 (pass_data_dwarf2_frame): ...new pass_data instance and...
5898 (make_pass_dwarf2_frame): ...new function.
5899 * except.c (pass_set_nothrow_function_flags): Convert from a global
5900 struct to a subclass of rtl_opt_pass along with...
5901 (pass_data_set_nothrow_function_flags): ...new pass_data instance
5903 (make_pass_set_nothrow_function_flags): ...new function.
5904 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
5905 subclass of rtl_opt_pass along with...
5906 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
5908 (make_pass_convert_to_eh_region_ranges): ...new function.
5909 * final.c (pass_compute_alignments): Convert from a global struct to a
5910 subclass of rtl_opt_pass along with...
5911 (pass_data_compute_alignments): ...new pass_data instance and...
5912 (make_pass_compute_alignments): ...new function.
5913 (pass_final): Convert from a global struct to a subclass of
5914 rtl_opt_pass along with...
5915 (pass_data_final): ...new pass_data instance and...
5916 (make_pass_final): ...new function.
5917 (pass_shorten_branches): Convert from a global struct to a subclass of
5918 rtl_opt_pass along with...
5919 (pass_data_shorten_branches): ...new pass_data instance and...
5920 (make_pass_shorten_branches): ...new function.
5921 (pass_clean_state): Convert from a global struct to a subclass of
5922 rtl_opt_pass along with...
5923 (pass_data_clean_state): ...new pass_data instance and...
5924 (make_pass_clean_state): ...new function.
5925 * function.c (pass_instantiate_virtual_regs): Convert from a global
5926 struct to a subclass of rtl_opt_pass along with...
5927 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
5928 (make_pass_instantiate_virtual_regs): ...new function.
5929 (pass_leaf_regs): Convert from a global struct to a subclass of
5930 rtl_opt_pass along with...
5931 (pass_data_leaf_regs): ...new pass_data instance and...
5932 (make_pass_leaf_regs): ...new function.
5933 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
5934 subclass of rtl_opt_pass along with...
5935 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
5937 (make_pass_thread_prologue_and_epilogue): ...new function.
5938 (pass_match_asm_constraints): Convert from a global struct to a
5939 subclass of rtl_opt_pass along with...
5940 (pass_data_match_asm_constraints): ...new pass_data instance and...
5941 (make_pass_match_asm_constraints): ...new function.
5942 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
5943 subclass of rtl_opt_pass along with...
5944 (pass_data_rtl_fwprop): ...new pass_data instance and...
5945 (make_pass_rtl_fwprop): ...new function.
5946 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
5947 rtl_opt_pass along with...
5948 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
5949 (make_pass_rtl_fwprop_addr): ...new function.
5950 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
5951 rtl_opt_pass along with...
5952 (pass_data_rtl_pre): ...new pass_data instance and...
5953 (make_pass_rtl_pre): ...new function.
5954 (pass_rtl_hoist): Convert from a global struct to a subclass of
5955 rtl_opt_pass along with...
5956 (pass_data_rtl_hoist): ...new pass_data instance and...
5957 (make_pass_rtl_hoist): ...new function.
5958 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
5959 subclass of gimple_opt_pass along with...
5960 (pass_data_lower_cf): ...new pass_data instance and...
5961 (make_pass_lower_cf): ...new function.
5962 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
5963 from a global struct to a subclass of gimple_opt_pass along with...
5964 (pass_data_strength_reduction): ...new pass_data instance and...
5965 (make_pass_strength_reduction): ...new function.
5966 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
5967 of rtl_opt_pass along with...
5968 (pass_data_rtl_ifcvt): ...new pass_data instance and...
5969 (make_pass_rtl_ifcvt): ...new function.
5970 (pass_if_after_combine): Convert from a global struct to a subclass of
5971 rtl_opt_pass along with...
5972 (pass_data_if_after_combine): ...new pass_data instance and...
5973 (make_pass_if_after_combine): ...new function.
5974 (pass_if_after_reload): Convert from a global struct to a subclass of
5975 rtl_opt_pass along with...
5976 (pass_data_if_after_reload): ...new pass_data instance and...
5977 (make_pass_if_after_reload): ...new function.
5978 * init-regs.c (pass_initialize_regs): Convert from a global struct to
5979 a subclass of rtl_opt_pass along with...
5980 (pass_data_initialize_regs): ...new pass_data instance and...
5981 (make_pass_initialize_regs): ...new function.
5982 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
5983 of ipa_opt_pass_d along with...
5984 (pass_data_ipa_cp): ...new pass_data instance and...
5985 (make_pass_ipa_cp): ...new function.
5986 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
5987 global struct to a subclass of gimple_opt_pass along with...
5988 (pass_data_inline_parameters): ...new pass_data instance and...
5989 (make_pass_inline_parameters): ...new function.
5990 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
5991 subclass of gimple_opt_pass along with...
5992 (pass_data_early_inline): ...new pass_data instance and...
5993 (make_pass_early_inline): ...new function.
5994 (pass_ipa_inline): Convert from a global struct to a subclass of
5995 ipa_opt_pass_d along with...
5996 (pass_data_ipa_inline): ...new pass_data instance and...
5997 (make_pass_ipa_inline): ...new function.
5998 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
5999 struct to a subclass of gimple_opt_pass along with...
6000 (pass_data_local_pure_const): ...new pass_data instance and...
6001 (make_pass_local_pure_const): ...new function.
6002 (pass_ipa_pure_const): Convert from a global struct to a subclass of
6003 ipa_opt_pass_d along with...
6004 (pass_data_ipa_pure_const): ...new pass_data instance and...
6005 (make_pass_ipa_pure_const): ...new function.
6006 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
6007 to a subclass of ipa_opt_pass_d along with...
6008 (pass_data_ipa_reference): ...new pass_data instance and...
6009 (make_pass_ipa_reference): ...new function.
6010 * ipa-split.c (pass_split_functions): Convert from a global struct to
6011 a subclass of gimple_opt_pass along with...
6012 (pass_data_split_functions): ...new pass_data instance and...
6013 (make_pass_split_functions): ...new function.
6014 (pass_feedback_split_functions): Convert from a global struct to a
6015 subclass of gimple_opt_pass along with...
6016 (pass_data_feedback_split_functions): ...new pass_data instance and...
6017 (make_pass_feedback_split_functions): ...new function.
6018 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
6019 global struct to a subclass of simple_ipa_opt_pass along with...
6020 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
6022 (make_pass_ipa_function_and_variable_visibility): ...new function.
6023 (pass_ipa_free_inline_summary): Convert from a global struct to a
6024 subclass of simple_ipa_opt_pass along with...
6025 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
6026 (make_pass_ipa_free_inline_summary): ...new function.
6027 (pass_ipa_whole_program_visibility): Convert from a global struct to a
6028 subclass of ipa_opt_pass_d along with...
6029 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
6031 (make_pass_ipa_whole_program_visibility): ...new function.
6032 (pass_ipa_profile): Convert from a global struct to a subclass of
6033 ipa_opt_pass_d along with...
6034 (pass_data_ipa_profile): ...new pass_data instance and...
6035 (make_pass_ipa_profile): ...new function.
6036 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
6037 ipa_opt_pass_d along with...
6038 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
6039 (make_pass_ipa_cdtor_merge): ...new function.
6040 * ira.c (pass_ira): Convert from a global struct to a subclass of
6041 rtl_opt_pass along with...
6042 (pass_data_ira): ...new pass_data instance and...
6043 (make_pass_ira): ...new function.
6044 (pass_reload): Convert from a global struct to a subclass of
6045 rtl_opt_pass along with...
6046 (pass_data_reload): ...new pass_data instance and...
6047 (make_pass_reload): ...new function.
6048 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
6049 subclass of rtl_opt_pass along with...
6050 (pass_data_cleanup_barriers): ...new pass_data instance and...
6051 (make_pass_cleanup_barriers): ...new function.
6052 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
6053 of rtl_opt_pass along with...
6054 (pass_data_loop2): ...new pass_data instance and...
6055 (make_pass_loop2): ...new function.
6056 (pass_rtl_loop_init): Convert from a global struct to a subclass of
6057 rtl_opt_pass along with...
6058 (pass_data_rtl_loop_init): ...new pass_data instance and...
6059 (make_pass_rtl_loop_init): ...new function.
6060 (pass_rtl_loop_done): Convert from a global struct to a subclass of
6061 rtl_opt_pass along with...
6062 (pass_data_rtl_loop_done): ...new pass_data instance and...
6063 (make_pass_rtl_loop_done): ...new function.
6064 (pass_rtl_move_loop_invariants): Convert from a global struct to a
6065 subclass of rtl_opt_pass along with...
6066 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
6067 (make_pass_rtl_move_loop_invariants): ...new function.
6068 (pass_rtl_unswitch): Convert from a global struct to a subclass of
6069 rtl_opt_pass along with...
6070 (pass_data_rtl_unswitch): ...new pass_data instance and...
6071 (make_pass_rtl_unswitch): ...new function.
6072 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
6073 subclass of rtl_opt_pass along with...
6074 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
6076 (make_pass_rtl_unroll_and_peel_loops): ...new function.
6077 (pass_rtl_doloop): Convert from a global struct to a subclass of
6078 rtl_opt_pass along with...
6079 (pass_data_rtl_doloop): ...new pass_data instance and...
6080 (make_pass_rtl_doloop): ...new function.
6081 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
6082 a subclass of rtl_opt_pass along with...
6083 (pass_data_lower_subreg): ...new pass_data instance and...
6084 (make_pass_lower_subreg): ...new function.
6085 (pass_lower_subreg2): Convert from a global struct to a subclass of
6086 rtl_opt_pass along with...
6087 (pass_data_lower_subreg2): ...new pass_data instance and...
6088 (make_pass_lower_subreg2): ...new function.
6089 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
6090 struct to a subclass of ipa_opt_pass_d along with...
6091 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
6092 (make_pass_ipa_lto_gimple_out): ...new function.
6093 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
6094 of ipa_opt_pass_d along with...
6095 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
6096 (make_pass_ipa_lto_finish_out): ...new function.
6097 * mode-switching.c (pass_mode_switching): Convert from a global struct
6098 to a subclass of rtl_opt_pass along with...
6099 (pass_data_mode_switching): ...new pass_data instance and...
6100 (make_pass_mode_switching): ...new function.
6101 * modulo-sched.c (pass_sms): Convert from a global struct to a
6102 subclass of rtl_opt_pass along with...
6103 (pass_data_sms): ...new pass_data instance and...
6104 (make_pass_sms): ...new function.
6105 * omp-low.c (pass_expand_omp): Convert from a global struct to a
6106 subclass of gimple_opt_pass along with...
6107 (pass_data_expand_omp): ...new pass_data instance and...
6108 (make_pass_expand_omp): ...new function.
6109 (pass_lower_omp): Convert from a global struct to a subclass of
6110 gimple_opt_pass along with...
6111 (pass_data_lower_omp): ...new pass_data instance and...
6112 (make_pass_lower_omp): ...new function.
6113 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
6114 of gimple_opt_pass along with...
6115 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
6116 (make_pass_diagnose_omp_blocks): ...new function.
6117 * passes.c (pass_early_local_passes): Convert from a global struct to
6118 a subclass of simple_ipa_opt_pass along with...
6119 (pass_data_early_local_passes): ...new pass_data instance and...
6120 (make_pass_early_local_passes): ...new function.
6121 (pass_all_early_optimizations): Convert from a global struct to a
6122 subclass of gimple_opt_pass along with...
6123 (pass_data_all_early_optimizations): ...new pass_data instance and...
6124 (make_pass_all_early_optimizations): ...new function.
6125 (pass_all_optimizations): Convert from a global struct to a subclass
6126 of gimple_opt_pass along with...
6127 (pass_data_all_optimizations): ...new pass_data instance and...
6128 (make_pass_all_optimizations): ...new function.
6129 (pass_all_optimizations_g): Convert from a global struct to a subclass
6130 of gimple_opt_pass along with...
6131 (pass_data_all_optimizations_g): ...new pass_data instance and...
6132 (make_pass_all_optimizations_g): ...new function.
6133 (pass_rest_of_compilation): Convert from a global struct to a subclass
6134 of rtl_opt_pass along with...
6135 (pass_data_rest_of_compilation): ...new pass_data instance and...
6136 (make_pass_rest_of_compilation): ...new function.
6137 (pass_postreload): Convert from a global struct to a subclass of
6138 rtl_opt_pass along with...
6139 (pass_data_postreload): ...new pass_data instance and...
6140 (make_pass_postreload): ...new function.
6141 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
6142 subclass of rtl_opt_pass along with...
6143 (pass_data_gcse2): ...new pass_data instance and...
6144 (make_pass_gcse2): ...new function.
6145 * postreload.c (pass_postreload_cse): Convert from a global struct to
6146 a subclass of rtl_opt_pass along with...
6147 (pass_data_postreload_cse): ...new pass_data instance and...
6148 (make_pass_postreload_cse): ...new function.
6149 * predict.c (pass_profile): Convert from a global struct to a subclass
6150 of gimple_opt_pass along with...
6151 (pass_data_profile): ...new pass_data instance and...
6152 (make_pass_profile): ...new function.
6153 (pass_strip_predict_hints): Convert from a global struct to a subclass
6154 of gimple_opt_pass along with...
6155 (pass_data_strip_predict_hints): ...new pass_data instance and...
6156 (make_pass_strip_predict_hints): ...new function.
6157 * recog.c (pass_peephole2): Convert from a global struct to a subclass
6158 of rtl_opt_pass along with...
6159 (pass_data_peephole2): ...new pass_data instance and...
6160 (make_pass_peephole2): ...new function.
6161 (pass_split_all_insns): Convert from a global struct to a subclass of
6162 rtl_opt_pass along with...
6163 (pass_data_split_all_insns): ...new pass_data instance and...
6164 (make_pass_split_all_insns): ...new function.
6165 (pass_split_after_reload): Convert from a global struct to a subclass
6166 of rtl_opt_pass along with...
6167 (pass_data_split_after_reload): ...new pass_data instance and...
6168 (make_pass_split_after_reload): ...new function.
6169 (pass_split_before_regstack): Convert from a global struct to a
6170 subclass of rtl_opt_pass along with...
6171 (pass_data_split_before_regstack): ...new pass_data instance and...
6172 (make_pass_split_before_regstack): ...new function.
6173 (pass_split_before_sched2): Convert from a global struct to a subclass
6174 of rtl_opt_pass along with...
6175 (pass_data_split_before_sched2): ...new pass_data instance and...
6176 (make_pass_split_before_sched2): ...new function.
6177 (pass_split_for_shorten_branches): Convert from a global struct to a
6178 subclass of rtl_opt_pass along with...
6179 (pass_data_split_for_shorten_branches): ...new pass_data instance
6181 (make_pass_split_for_shorten_branches): ...new function.
6182 * ree.c (pass_ree): Convert from a global struct to a subclass of
6183 rtl_opt_pass along with...
6184 (pass_data_ree): ...new pass_data instance and...
6185 (make_pass_ree): ...new function.
6186 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
6187 subclass of rtl_opt_pass along with...
6188 (pass_data_stack_regs): ...new pass_data instance and...
6189 (make_pass_stack_regs): ...new function.
6190 (pass_stack_regs_run): Convert from a global struct to a subclass of
6191 rtl_opt_pass along with...
6192 (pass_data_stack_regs_run): ...new pass_data instance and...
6193 (make_pass_stack_regs_run): ...new function.
6194 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
6195 subclass of rtl_opt_pass along with...
6196 (pass_data_cprop_hardreg): ...new pass_data instance and...
6197 (make_pass_cprop_hardreg): ...new function.
6198 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
6199 subclass of rtl_opt_pass along with...
6200 (pass_data_reginfo_init): ...new pass_data instance and...
6201 (make_pass_reginfo_init): ...new function.
6202 * regmove.c (pass_regmove): Convert from a global struct to a subclass
6203 of rtl_opt_pass along with...
6204 (pass_data_regmove): ...new pass_data instance and...
6205 (make_pass_regmove): ...new function.
6206 * regrename.c (pass_regrename): Convert from a global struct to a
6207 subclass of rtl_opt_pass along with...
6208 (pass_data_regrename): ...new pass_data instance and...
6209 (make_pass_regrename): ...new function.
6210 * reorg.c (pass_delay_slots): Convert from a global struct to a
6211 subclass of rtl_opt_pass along with...
6212 (pass_data_delay_slots): ...new pass_data instance and...
6213 (make_pass_delay_slots): ...new function.
6214 (pass_machine_reorg): Convert from a global struct to a subclass of
6215 rtl_opt_pass along with...
6216 (pass_data_machine_reorg): ...new pass_data instance and...
6217 (make_pass_machine_reorg): ...new function.
6218 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
6219 of rtl_opt_pass along with...
6220 (pass_data_sched): ...new pass_data instance and...
6221 (make_pass_sched): ...new function.
6222 (pass_sched2): Convert from a global struct to a subclass of
6223 rtl_opt_pass along with...
6224 (pass_data_sched2): ...new pass_data instance and...
6225 (make_pass_sched2): ...new function.
6226 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
6227 to a subclass of rtl_opt_pass along with...
6228 (pass_data_stack_ptr_mod): ...new pass_data instance and...
6229 (make_pass_stack_ptr_mod): ...new function.
6230 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
6231 to a subclass of rtl_opt_pass along with...
6232 (pass_data_rtl_store_motion): ...new pass_data instance and...
6233 (make_pass_rtl_store_motion): ...new function.
6234 * tracer.c (pass_tracer): Convert from a global struct to a subclass
6235 of gimple_opt_pass along with...
6236 (pass_data_tracer): ...new pass_data instance and...
6237 (make_pass_tracer): ...new function.
6238 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
6239 to a subclass of gimple_opt_pass along with...
6240 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
6241 (make_pass_diagnose_tm_blocks): ...new function.
6242 (pass_lower_tm): Convert from a global struct to a subclass of
6243 gimple_opt_pass along with...
6244 (pass_data_lower_tm): ...new pass_data instance and...
6245 (make_pass_lower_tm): ...new function.
6246 (pass_tm_init): Convert from a global struct to a subclass of
6247 gimple_opt_pass along with...
6248 (pass_data_tm_init): ...new pass_data instance and...
6249 (make_pass_tm_init): ...new function.
6250 (pass_tm_mark): Convert from a global struct to a subclass of
6251 gimple_opt_pass along with...
6252 (pass_data_tm_mark): ...new pass_data instance and...
6253 (make_pass_tm_mark): ...new function.
6254 (pass_tm_edges): Convert from a global struct to a subclass of
6255 gimple_opt_pass along with...
6256 (pass_data_tm_edges): ...new pass_data instance and...
6257 (make_pass_tm_edges): ...new function.
6258 (pass_tm_memopt): Convert from a global struct to a subclass of
6259 gimple_opt_pass along with...
6260 (pass_data_tm_memopt): ...new pass_data instance and...
6261 (make_pass_tm_memopt): ...new function.
6262 (pass_ipa_tm): Convert from a global struct to a subclass of
6263 simple_ipa_opt_pass along with...
6264 (pass_data_ipa_tm): ...new pass_data instance and...
6265 (make_pass_ipa_tm): ...new function.
6266 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
6267 subclass of gimple_opt_pass along with...
6268 (pass_data_call_cdce): ...new pass_data instance and...
6269 (make_pass_call_cdce): ...new function.
6270 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
6271 subclass of gimple_opt_pass along with...
6272 (pass_data_build_cfg): ...new pass_data instance and...
6273 (make_pass_build_cfg): ...new function.
6274 (pass_split_crit_edges): Convert from a global struct to a subclass of
6275 gimple_opt_pass along with...
6276 (pass_data_split_crit_edges): ...new pass_data instance and...
6277 (make_pass_split_crit_edges): ...new function.
6278 (pass_warn_function_return): Convert from a global struct to a
6279 subclass of gimple_opt_pass along with...
6280 (pass_data_warn_function_return): ...new pass_data instance and...
6281 (make_pass_warn_function_return): ...new function.
6282 (pass_warn_function_noreturn): Convert from a global struct to a
6283 subclass of gimple_opt_pass along with...
6284 (pass_data_warn_function_noreturn): ...new pass_data instance and...
6285 (make_pass_warn_function_noreturn): ...new function.
6286 (pass_warn_unused_result): Convert from a global struct to a subclass
6287 of gimple_opt_pass along with...
6288 (pass_data_warn_unused_result): ...new pass_data instance and...
6289 (make_pass_warn_unused_result): ...new function.
6290 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
6291 a subclass of gimple_opt_pass along with...
6292 (pass_data_merge_phi): ...new pass_data instance and...
6293 (make_pass_merge_phi): ...new function.
6294 * tree-complex.c (pass_lower_complex): Convert from a global struct to
6295 a subclass of gimple_opt_pass along with...
6296 (pass_data_lower_complex): ...new pass_data instance and...
6297 (make_pass_lower_complex): ...new function.
6298 (pass_lower_complex_O0): Convert from a global struct to a subclass of
6299 gimple_opt_pass along with...
6300 (pass_data_lower_complex_O0): ...new pass_data instance and...
6301 (make_pass_lower_complex_O0): ...new function.
6302 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
6303 subclass of gimple_opt_pass along with...
6304 (pass_data_lower_eh): ...new pass_data instance and...
6305 (make_pass_lower_eh): ...new function.
6306 (pass_refactor_eh): Convert from a global struct to a subclass of
6307 gimple_opt_pass along with...
6308 (pass_data_refactor_eh): ...new pass_data instance and...
6309 (make_pass_refactor_eh): ...new function.
6310 (pass_lower_resx): Convert from a global struct to a subclass of
6311 gimple_opt_pass along with...
6312 (pass_data_lower_resx): ...new pass_data instance and...
6313 (make_pass_lower_resx): ...new function.
6314 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
6315 of gimple_opt_pass along with...
6316 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
6317 (make_pass_lower_eh_dispatch): ...new function.
6318 (pass_cleanup_eh): Convert from a global struct to a subclass of
6319 gimple_opt_pass along with...
6320 (pass_data_cleanup_eh): ...new pass_data instance and...
6321 (make_pass_cleanup_eh): ...new function.
6322 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
6323 to a subclass of simple_ipa_opt_pass along with...
6324 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
6325 (make_pass_ipa_lower_emutls): ...new function.
6326 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
6327 a subclass of gimple_opt_pass along with...
6328 (pass_data_if_conversion): ...new pass_data instance and...
6329 (make_pass_if_conversion): ...new function.
6330 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
6331 subclass of gimple_opt_pass along with...
6332 (pass_data_build_ssa): ...new pass_data instance and...
6333 (make_pass_build_ssa): ...new function.
6334 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
6335 global struct to a subclass of gimple_opt_pass along with...
6336 (pass_data_loop_distribution): ...new pass_data instance and...
6337 (make_pass_loop_distribution): ...new function.
6338 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
6339 subclass of gimple_opt_pass along with...
6340 (pass_data_mudflap_1): ...new pass_data instance and...
6341 (make_pass_mudflap_1): ...new function.
6342 (pass_mudflap_2): Convert from a global struct to a subclass of
6343 gimple_opt_pass along with...
6344 (pass_data_mudflap_2): ...new pass_data instance and...
6345 (make_pass_mudflap_2): ...new function.
6346 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
6347 subclass of gimple_opt_pass along with...
6348 (pass_data_mudflap_1): ...new pass_data instance and...
6349 (make_pass_mudflap_1): ...new function.
6350 (pass_mudflap_2): Convert from a global struct to a subclass of
6351 gimple_opt_pass along with...
6352 (pass_data_mudflap_2): ...new pass_data instance and...
6353 (make_pass_mudflap_2): ...new function.
6354 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
6355 gimple_opt_pass along with...
6356 (pass_data_nrv): ...new pass_data instance and...
6357 (make_pass_nrv): ...new function.
6358 (pass_return_slot): Convert from a global struct to a subclass of
6359 gimple_opt_pass along with...
6360 (pass_data_return_slot): ...new pass_data instance and...
6361 (make_pass_return_slot): ...new function.
6362 * tree-object-size.c (pass_object_sizes): Convert from a global struct
6363 to a subclass of gimple_opt_pass along with...
6364 (pass_data_object_sizes): ...new pass_data instance and...
6365 (make_pass_object_sizes): ...new function.
6366 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
6367 global struct to a subclass of gimple_opt_pass along with...
6368 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
6370 (make_pass_cleanup_cfg_post_optimizing): ...new function.
6371 (pass_fixup_cfg): Convert from a global struct to a subclass of
6372 gimple_opt_pass along with...
6373 (pass_data_fixup_cfg): ...new pass_data instance and...
6374 (make_pass_fixup_cfg): ...new function.
6375 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
6376 (make_pass_mudflap_1): ...new function.
6377 (pass_mudflap_2): Replace declaration with that of...
6378 (make_pass_mudflap_2): ...new function.
6379 (pass_asan): Replace declaration with that of...
6380 (make_pass_asan): ...new function.
6381 (pass_asan_O0): Replace declaration with that of...
6382 (make_pass_asan_O0): ...new function.
6383 (pass_tsan): Replace declaration with that of...
6384 (make_pass_tsan): ...new function.
6385 (pass_tsan_O0): Replace declaration with that of...
6386 (make_pass_tsan_O0): ...new function.
6387 (pass_lower_cf): Replace declaration with that of...
6388 (make_pass_lower_cf): ...new function.
6389 (pass_refactor_eh): Replace declaration with that of...
6390 (make_pass_refactor_eh): ...new function.
6391 (pass_lower_eh): Replace declaration with that of...
6392 (make_pass_lower_eh): ...new function.
6393 (pass_lower_eh_dispatch): Replace declaration with that of...
6394 (make_pass_lower_eh_dispatch): ...new function.
6395 (pass_lower_resx): Replace declaration with that of...
6396 (make_pass_lower_resx): ...new function.
6397 (pass_build_cfg): Replace declaration with that of...
6398 (make_pass_build_cfg): ...new function.
6399 (pass_early_tree_profile): Replace declaration with that of...
6400 (make_pass_early_tree_profile): ...new function.
6401 (pass_cleanup_eh): Replace declaration with that of...
6402 (make_pass_cleanup_eh): ...new function.
6403 (pass_sra): Replace declaration with that of...
6404 (make_pass_sra): ...new function.
6405 (pass_sra_early): Replace declaration with that of...
6406 (make_pass_sra_early): ...new function.
6407 (pass_early_ipa_sra): Replace declaration with that of...
6408 (make_pass_early_ipa_sra): ...new function.
6409 (pass_tail_recursion): Replace declaration with that of...
6410 (make_pass_tail_recursion): ...new function.
6411 (pass_tail_calls): Replace declaration with that of...
6412 (make_pass_tail_calls): ...new function.
6413 (pass_tree_loop): Replace declaration with that of...
6414 (make_pass_tree_loop): ...new function.
6415 (pass_tree_loop_init): Replace declaration with that of...
6416 (make_pass_tree_loop_init): ...new function.
6417 (pass_lim): Replace declaration with that of...
6418 (make_pass_lim): ...new function.
6419 (pass_tree_unswitch): Replace declaration with that of...
6420 (make_pass_tree_unswitch): ...new function.
6421 (pass_predcom): Replace declaration with that of...
6422 (make_pass_predcom): ...new function.
6423 (pass_iv_canon): Replace declaration with that of...
6424 (make_pass_iv_canon): ...new function.
6425 (pass_scev_cprop): Replace declaration with that of...
6426 (make_pass_scev_cprop): ...new function.
6427 (pass_empty_loop): Replace declaration with that of...
6428 (make_pass_empty_loop): ...new function.
6429 (pass_record_bounds): Replace declaration with that of...
6430 (make_pass_record_bounds): ...new function.
6431 (pass_graphite): Replace declaration with that of...
6432 (make_pass_graphite): ...new function.
6433 (pass_graphite_transforms): Replace declaration with that of...
6434 (make_pass_graphite_transforms): ...new function.
6435 (pass_if_conversion): Replace declaration with that of...
6436 (make_pass_if_conversion): ...new function.
6437 (pass_loop_distribution): Replace declaration with that of...
6438 (make_pass_loop_distribution): ...new function.
6439 (pass_vectorize): Replace declaration with that of...
6440 (make_pass_vectorize): ...new function.
6441 (pass_slp_vectorize): Replace declaration with that of...
6442 (make_pass_slp_vectorize): ...new function.
6443 (pass_complete_unroll): Replace declaration with that of...
6444 (make_pass_complete_unroll): ...new function.
6445 (pass_complete_unrolli): Replace declaration with that of...
6446 (make_pass_complete_unrolli): ...new function.
6447 (pass_parallelize_loops): Replace declaration with that of...
6448 (make_pass_parallelize_loops): ...new function.
6449 (pass_loop_prefetch): Replace declaration with that of...
6450 (make_pass_loop_prefetch): ...new function.
6451 (pass_iv_optimize): Replace declaration with that of...
6452 (make_pass_iv_optimize): ...new function.
6453 (pass_tree_loop_done): Replace declaration with that of...
6454 (make_pass_tree_loop_done): ...new function.
6455 (pass_ch): Replace declaration with that of...
6456 (make_pass_ch): ...new function.
6457 (pass_ccp): Replace declaration with that of...
6458 (make_pass_ccp): ...new function.
6459 (pass_phi_only_cprop): Replace declaration with that of...
6460 (make_pass_phi_only_cprop): ...new function.
6461 (pass_build_ssa): Replace declaration with that of...
6462 (make_pass_build_ssa): ...new function.
6463 (pass_build_alias): Replace declaration with that of...
6464 (make_pass_build_alias): ...new function.
6465 (pass_build_ealias): Replace declaration with that of...
6466 (make_pass_build_ealias): ...new function.
6467 (pass_dominator): Replace declaration with that of...
6468 (make_pass_dominator): ...new function.
6469 (pass_dce): Replace declaration with that of...
6470 (make_pass_dce): ...new function.
6471 (pass_dce_loop): Replace declaration with that of...
6472 (make_pass_dce_loop): ...new function.
6473 (pass_cd_dce): Replace declaration with that of...
6474 (make_pass_cd_dce): ...new function.
6475 (pass_call_cdce): Replace declaration with that of...
6476 (make_pass_call_cdce): ...new function.
6477 (pass_merge_phi): Replace declaration with that of...
6478 (make_pass_merge_phi): ...new function.
6479 (pass_split_crit_edges): Replace declaration with that of...
6480 (make_pass_split_crit_edges): ...new function.
6481 (pass_pre): Replace declaration with that of...
6482 (make_pass_pre): ...new function.
6483 (pass_profile): Replace declaration with that of...
6484 (make_pass_profile): ...new function.
6485 (pass_strip_predict_hints): Replace declaration with that of...
6486 (make_pass_strip_predict_hints): ...new function.
6487 (pass_lower_complex_O0): Replace declaration with that of...
6488 (make_pass_lower_complex_O0): ...new function.
6489 (pass_lower_complex): Replace declaration with that of...
6490 (make_pass_lower_complex): ...new function.
6491 (pass_lower_vector): Replace declaration with that of...
6492 (make_pass_lower_vector): ...new function.
6493 (pass_lower_vector_ssa): Replace declaration with that of...
6494 (make_pass_lower_vector_ssa): ...new function.
6495 (pass_lower_omp): Replace declaration with that of...
6496 (make_pass_lower_omp): ...new function.
6497 (pass_diagnose_omp_blocks): Replace declaration with that of...
6498 (make_pass_diagnose_omp_blocks): ...new function.
6499 (pass_expand_omp): Replace declaration with that of...
6500 (make_pass_expand_omp): ...new function.
6501 (pass_expand_omp_ssa): Replace declaration with that of...
6502 (make_pass_expand_omp_ssa): ...new function.
6503 (pass_object_sizes): Replace declaration with that of...
6504 (make_pass_object_sizes): ...new function.
6505 (pass_strlen): Replace declaration with that of...
6506 (make_pass_strlen): ...new function.
6507 (pass_fold_builtins): Replace declaration with that of...
6508 (make_pass_fold_builtins): ...new function.
6509 (pass_stdarg): Replace declaration with that of...
6510 (make_pass_stdarg): ...new function.
6511 (pass_early_warn_uninitialized): Replace declaration with that of...
6512 (make_pass_early_warn_uninitialized): ...new function.
6513 (pass_late_warn_uninitialized): Replace declaration with that of...
6514 (make_pass_late_warn_uninitialized): ...new function.
6515 (pass_cse_reciprocals): Replace declaration with that of...
6516 (make_pass_cse_reciprocals): ...new function.
6517 (pass_cse_sincos): Replace declaration with that of...
6518 (make_pass_cse_sincos): ...new function.
6519 (pass_optimize_bswap): Replace declaration with that of...
6520 (make_pass_optimize_bswap): ...new function.
6521 (pass_optimize_widening_mul): Replace declaration with that of...
6522 (make_pass_optimize_widening_mul): ...new function.
6523 (pass_warn_function_return): Replace declaration with that of...
6524 (make_pass_warn_function_return): ...new function.
6525 (pass_warn_function_noreturn): Replace declaration with that of...
6526 (make_pass_warn_function_noreturn): ...new function.
6527 (pass_cselim): Replace declaration with that of...
6528 (make_pass_cselim): ...new function.
6529 (pass_phiopt): Replace declaration with that of...
6530 (make_pass_phiopt): ...new function.
6531 (pass_forwprop): Replace declaration with that of...
6532 (make_pass_forwprop): ...new function.
6533 (pass_phiprop): Replace declaration with that of...
6534 (make_pass_phiprop): ...new function.
6535 (pass_tree_ifcombine): Replace declaration with that of...
6536 (make_pass_tree_ifcombine): ...new function.
6537 (pass_dse): Replace declaration with that of...
6538 (make_pass_dse): ...new function.
6539 (pass_nrv): Replace declaration with that of...
6540 (make_pass_nrv): ...new function.
6541 (pass_rename_ssa_copies): Replace declaration with that of...
6542 (make_pass_rename_ssa_copies): ...new function.
6543 (pass_sink_code): Replace declaration with that of...
6544 (make_pass_sink_code): ...new function.
6545 (pass_fre): Replace declaration with that of...
6546 (make_pass_fre): ...new function.
6547 (pass_check_data_deps): Replace declaration with that of...
6548 (make_pass_check_data_deps): ...new function.
6549 (pass_copy_prop): Replace declaration with that of...
6550 (make_pass_copy_prop): ...new function.
6551 (pass_vrp): Replace declaration with that of...
6552 (make_pass_vrp): ...new function.
6553 (pass_uncprop): Replace declaration with that of...
6554 (make_pass_uncprop): ...new function.
6555 (pass_return_slot): Replace declaration with that of...
6556 (make_pass_return_slot): ...new function.
6557 (pass_reassoc): Replace declaration with that of...
6558 (make_pass_reassoc): ...new function.
6559 (pass_rebuild_cgraph_edges): Replace declaration with that of...
6560 (make_pass_rebuild_cgraph_edges): ...new function.
6561 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
6562 (make_pass_remove_cgraph_callee_edges): ...new function.
6563 (pass_build_cgraph_edges): Replace declaration with that of...
6564 (make_pass_build_cgraph_edges): ...new function.
6565 (pass_local_pure_const): Replace declaration with that of...
6566 (make_pass_local_pure_const): ...new function.
6567 (pass_tracer): Replace declaration with that of...
6568 (make_pass_tracer): ...new function.
6569 (pass_warn_unused_result): Replace declaration with that of...
6570 (make_pass_warn_unused_result): ...new function.
6571 (pass_diagnose_tm_blocks): Replace declaration with that of...
6572 (make_pass_diagnose_tm_blocks): ...new function.
6573 (pass_lower_tm): Replace declaration with that of...
6574 (make_pass_lower_tm): ...new function.
6575 (pass_tm_init): Replace declaration with that of...
6576 (make_pass_tm_init): ...new function.
6577 (pass_tm_mark): Replace declaration with that of...
6578 (make_pass_tm_mark): ...new function.
6579 (pass_tm_memopt): Replace declaration with that of...
6580 (make_pass_tm_memopt): ...new function.
6581 (pass_tm_edges): Replace declaration with that of...
6582 (make_pass_tm_edges): ...new function.
6583 (pass_split_functions): Replace declaration with that of...
6584 (make_pass_split_functions): ...new function.
6585 (pass_feedback_split_functions): Replace declaration with that of...
6586 (make_pass_feedback_split_functions): ...new function.
6587 (pass_strength_reduction): Replace declaration with that of...
6588 (make_pass_strength_reduction): ...new function.
6589 (pass_ipa_lower_emutls): Replace declaration with that of...
6590 (make_pass_ipa_lower_emutls): ...new function.
6591 (pass_ipa_function_and_variable_visibility): Replace declaration with
6593 (make_pass_ipa_function_and_variable_visibility): ...new function.
6594 (pass_ipa_tree_profile): Replace declaration with that of...
6595 (make_pass_ipa_tree_profile): ...new function.
6596 (pass_early_local_passes): Replace declaration with that of...
6597 (make_pass_early_local_passes): ...new function.
6598 (pass_ipa_whole_program_visibility): Replace declaration with that
6600 (make_pass_ipa_whole_program_visibility): ...new function.
6601 (pass_ipa_lto_gimple_out): Replace declaration with that of...
6602 (make_pass_ipa_lto_gimple_out): ...new function.
6603 (pass_ipa_increase_alignment): Replace declaration with that of...
6604 (make_pass_ipa_increase_alignment): ...new function.
6605 (pass_ipa_inline): Replace declaration with that of...
6606 (make_pass_ipa_inline): ...new function.
6607 (pass_ipa_free_lang_data): Replace declaration with that of...
6608 (make_pass_ipa_free_lang_data): ...new function.
6609 (pass_ipa_free_inline_summary): Replace declaration with that of...
6610 (make_pass_ipa_free_inline_summary): ...new function.
6611 (pass_ipa_cp): Replace declaration with that of...
6612 (make_pass_ipa_cp): ...new function.
6613 (pass_ipa_reference): Replace declaration with that of...
6614 (make_pass_ipa_reference): ...new function.
6615 (pass_ipa_pure_const): Replace declaration with that of...
6616 (make_pass_ipa_pure_const): ...new function.
6617 (pass_ipa_pta): Replace declaration with that of...
6618 (make_pass_ipa_pta): ...new function.
6619 (pass_ipa_lto_finish_out): Replace declaration with that of...
6620 (make_pass_ipa_lto_finish_out): ...new function.
6621 (pass_ipa_tm): Replace declaration with that of...
6622 (make_pass_ipa_tm): ...new function.
6623 (pass_ipa_profile): Replace declaration with that of...
6624 (make_pass_ipa_profile): ...new function.
6625 (pass_ipa_cdtor_merge): Replace declaration with that of...
6626 (make_pass_ipa_cdtor_merge): ...new function.
6627 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
6629 (make_pass_cleanup_cfg_post_optimizing): ...new function.
6630 (pass_init_datastructures): Replace declaration with that of...
6631 (make_pass_init_datastructures): ...new function.
6632 (pass_fixup_cfg): Replace declaration with that of...
6633 (make_pass_fixup_cfg): ...new function.
6634 (pass_expand): Replace declaration with that of...
6635 (make_pass_expand): ...new function.
6636 (pass_instantiate_virtual_regs): Replace declaration with that of...
6637 (make_pass_instantiate_virtual_regs): ...new function.
6638 (pass_rtl_fwprop): Replace declaration with that of...
6639 (make_pass_rtl_fwprop): ...new function.
6640 (pass_rtl_fwprop_addr): Replace declaration with that of...
6641 (make_pass_rtl_fwprop_addr): ...new function.
6642 (pass_jump): Replace declaration with that of...
6643 (make_pass_jump): ...new function.
6644 (pass_jump2): Replace declaration with that of...
6645 (make_pass_jump2): ...new function.
6646 (pass_lower_subreg): Replace declaration with that of...
6647 (make_pass_lower_subreg): ...new function.
6648 (pass_cse): Replace declaration with that of...
6649 (make_pass_cse): ...new function.
6650 (pass_fast_rtl_dce): Replace declaration with that of...
6651 (make_pass_fast_rtl_dce): ...new function.
6652 (pass_ud_rtl_dce): Replace declaration with that of...
6653 (make_pass_ud_rtl_dce): ...new function.
6654 (pass_rtl_dce): Replace declaration with that of...
6655 (make_pass_rtl_dce): ...new function.
6656 (pass_rtl_dse1): Replace declaration with that of...
6657 (make_pass_rtl_dse1): ...new function.
6658 (pass_rtl_dse2): Replace declaration with that of...
6659 (make_pass_rtl_dse2): ...new function.
6660 (pass_rtl_dse3): Replace declaration with that of...
6661 (make_pass_rtl_dse3): ...new function.
6662 (pass_rtl_cprop): Replace declaration with that of...
6663 (make_pass_rtl_cprop): ...new function.
6664 (pass_rtl_pre): Replace declaration with that of...
6665 (make_pass_rtl_pre): ...new function.
6666 (pass_rtl_hoist): Replace declaration with that of...
6667 (make_pass_rtl_hoist): ...new function.
6668 (pass_rtl_store_motion): Replace declaration with that of...
6669 (make_pass_rtl_store_motion): ...new function.
6670 (pass_cse_after_global_opts): Replace declaration with that of...
6671 (make_pass_cse_after_global_opts): ...new function.
6672 (pass_rtl_ifcvt): Replace declaration with that of...
6673 (make_pass_rtl_ifcvt): ...new function.
6674 (pass_into_cfg_layout_mode): Replace declaration with that of...
6675 (make_pass_into_cfg_layout_mode): ...new function.
6676 (pass_outof_cfg_layout_mode): Replace declaration with that of...
6677 (make_pass_outof_cfg_layout_mode): ...new function.
6678 (pass_loop2): Replace declaration with that of...
6679 (make_pass_loop2): ...new function.
6680 (pass_rtl_loop_init): Replace declaration with that of...
6681 (make_pass_rtl_loop_init): ...new function.
6682 (pass_rtl_move_loop_invariants): Replace declaration with that of...
6683 (make_pass_rtl_move_loop_invariants): ...new function.
6684 (pass_rtl_unswitch): Replace declaration with that of...
6685 (make_pass_rtl_unswitch): ...new function.
6686 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
6687 (make_pass_rtl_unroll_and_peel_loops): ...new function.
6688 (pass_rtl_doloop): Replace declaration with that of...
6689 (make_pass_rtl_doloop): ...new function.
6690 (pass_rtl_loop_done): Replace declaration with that of...
6691 (make_pass_rtl_loop_done): ...new function.
6692 (pass_web): Replace declaration with that of...
6693 (make_pass_web): ...new function.
6694 (pass_cse2): Replace declaration with that of...
6695 (make_pass_cse2): ...new function.
6696 (pass_df_initialize_opt): Replace declaration with that of...
6697 (make_pass_df_initialize_opt): ...new function.
6698 (pass_df_initialize_no_opt): Replace declaration with that of...
6699 (make_pass_df_initialize_no_opt): ...new function.
6700 (pass_reginfo_init): Replace declaration with that of...
6701 (make_pass_reginfo_init): ...new function.
6702 (pass_inc_dec): Replace declaration with that of...
6703 (make_pass_inc_dec): ...new function.
6704 (pass_stack_ptr_mod): Replace declaration with that of...
6705 (make_pass_stack_ptr_mod): ...new function.
6706 (pass_initialize_regs): Replace declaration with that of...
6707 (make_pass_initialize_regs): ...new function.
6708 (pass_combine): Replace declaration with that of...
6709 (make_pass_combine): ...new function.
6710 (pass_if_after_combine): Replace declaration with that of...
6711 (make_pass_if_after_combine): ...new function.
6712 (pass_ree): Replace declaration with that of...
6713 (make_pass_ree): ...new function.
6714 (pass_partition_blocks): Replace declaration with that of...
6715 (make_pass_partition_blocks): ...new function.
6716 (pass_match_asm_constraints): Replace declaration with that of...
6717 (make_pass_match_asm_constraints): ...new function.
6718 (pass_regmove): Replace declaration with that of...
6719 (make_pass_regmove): ...new function.
6720 (pass_split_all_insns): Replace declaration with that of...
6721 (make_pass_split_all_insns): ...new function.
6722 (pass_fast_rtl_byte_dce): Replace declaration with that of...
6723 (make_pass_fast_rtl_byte_dce): ...new function.
6724 (pass_lower_subreg2): Replace declaration with that of...
6725 (make_pass_lower_subreg2): ...new function.
6726 (pass_mode_switching): Replace declaration with that of...
6727 (make_pass_mode_switching): ...new function.
6728 (pass_sms): Replace declaration with that of...
6729 (make_pass_sms): ...new function.
6730 (pass_sched): Replace declaration with that of...
6731 (make_pass_sched): ...new function.
6732 (pass_ira): Replace declaration with that of...
6733 (make_pass_ira): ...new function.
6734 (pass_reload): Replace declaration with that of...
6735 (make_pass_reload): ...new function.
6736 (pass_clean_state): Replace declaration with that of...
6737 (make_pass_clean_state): ...new function.
6738 (pass_branch_prob): Replace declaration with that of...
6739 (make_pass_branch_prob): ...new function.
6740 (pass_value_profile_transformations): Replace declaration with that
6742 (make_pass_value_profile_transformations): ...new function.
6743 (pass_postreload_cse): Replace declaration with that of...
6744 (make_pass_postreload_cse): ...new function.
6745 (pass_gcse2): Replace declaration with that of...
6746 (make_pass_gcse2): ...new function.
6747 (pass_split_after_reload): Replace declaration with that of...
6748 (make_pass_split_after_reload): ...new function.
6749 (pass_branch_target_load_optimize1): Replace declaration with that
6751 (make_pass_branch_target_load_optimize1): ...new function.
6752 (pass_thread_prologue_and_epilogue): Replace declaration with that
6754 (make_pass_thread_prologue_and_epilogue): ...new function.
6755 (pass_stack_adjustments): Replace declaration with that of...
6756 (make_pass_stack_adjustments): ...new function.
6757 (pass_peephole2): Replace declaration with that of...
6758 (make_pass_peephole2): ...new function.
6759 (pass_if_after_reload): Replace declaration with that of...
6760 (make_pass_if_after_reload): ...new function.
6761 (pass_regrename): Replace declaration with that of...
6762 (make_pass_regrename): ...new function.
6763 (pass_cprop_hardreg): Replace declaration with that of...
6764 (make_pass_cprop_hardreg): ...new function.
6765 (pass_reorder_blocks): Replace declaration with that of...
6766 (make_pass_reorder_blocks): ...new function.
6767 (pass_branch_target_load_optimize2): Replace declaration with that
6769 (make_pass_branch_target_load_optimize2): ...new function.
6770 (pass_leaf_regs): Replace declaration with that of...
6771 (make_pass_leaf_regs): ...new function.
6772 (pass_split_before_sched2): Replace declaration with that of...
6773 (make_pass_split_before_sched2): ...new function.
6774 (pass_compare_elim_after_reload): Replace declaration with that of...
6775 (make_pass_compare_elim_after_reload): ...new function.
6776 (pass_sched2): Replace declaration with that of...
6777 (make_pass_sched2): ...new function.
6778 (pass_stack_regs): Replace declaration with that of...
6779 (make_pass_stack_regs): ...new function.
6780 (pass_stack_regs_run): Replace declaration with that of...
6781 (make_pass_stack_regs_run): ...new function.
6782 (pass_df_finish): Replace declaration with that of...
6783 (make_pass_df_finish): ...new function.
6784 (pass_compute_alignments): Replace declaration with that of...
6785 (make_pass_compute_alignments): ...new function.
6786 (pass_duplicate_computed_gotos): Replace declaration with that of...
6787 (make_pass_duplicate_computed_gotos): ...new function.
6788 (pass_variable_tracking): Replace declaration with that of...
6789 (make_pass_variable_tracking): ...new function.
6790 (pass_free_cfg): Replace declaration with that of...
6791 (make_pass_free_cfg): ...new function.
6792 (pass_machine_reorg): Replace declaration with that of...
6793 (make_pass_machine_reorg): ...new function.
6794 (pass_cleanup_barriers): Replace declaration with that of...
6795 (make_pass_cleanup_barriers): ...new function.
6796 (pass_delay_slots): Replace declaration with that of...
6797 (make_pass_delay_slots): ...new function.
6798 (pass_split_for_shorten_branches): Replace declaration with that of...
6799 (make_pass_split_for_shorten_branches): ...new function.
6800 (pass_split_before_regstack): Replace declaration with that of...
6801 (make_pass_split_before_regstack): ...new function.
6802 (pass_convert_to_eh_region_ranges): Replace declaration with that
6804 (make_pass_convert_to_eh_region_ranges): ...new function.
6805 (pass_shorten_branches): Replace declaration with that of...
6806 (make_pass_shorten_branches): ...new function.
6807 (pass_set_nothrow_function_flags): Replace declaration with that of...
6808 (make_pass_set_nothrow_function_flags): ...new function.
6809 (pass_dwarf2_frame): Replace declaration with that of...
6810 (make_pass_dwarf2_frame): ...new function.
6811 (pass_final): Replace declaration with that of...
6812 (make_pass_final): ...new function.
6813 (pass_rtl_seqabstr): Replace declaration with that of...
6814 (make_pass_rtl_seqabstr): ...new function.
6815 (pass_release_ssa_names): Replace declaration with that of...
6816 (make_pass_release_ssa_names): ...new function.
6817 (pass_early_inline): Replace declaration with that of...
6818 (make_pass_early_inline): ...new function.
6819 (pass_inline_parameters): Replace declaration with that of...
6820 (make_pass_inline_parameters): ...new function.
6821 (pass_update_address_taken): Replace declaration with that of...
6822 (make_pass_update_address_taken): ...new function.
6823 (pass_convert_switch): Replace declaration with that of...
6824 (make_pass_convert_switch): ...new function.
6825 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
6826 to a subclass of simple_ipa_opt_pass along with...
6827 (pass_data_ipa_tree_profile): ...new pass_data instance and...
6828 (make_pass_ipa_tree_profile): ...new function.
6829 * tree-sra.c (pass_sra_early): Convert from a global struct to a
6830 subclass of gimple_opt_pass along with...
6831 (pass_data_sra_early): ...new pass_data instance and...
6832 (make_pass_sra_early): ...new function.
6833 (pass_sra): Convert from a global struct to a subclass of
6834 gimple_opt_pass along with...
6835 (pass_data_sra): ...new pass_data instance and...
6836 (make_pass_sra): ...new function.
6837 (pass_early_ipa_sra): Convert from a global struct to a subclass of
6838 gimple_opt_pass along with...
6839 (pass_data_early_ipa_sra): ...new pass_data instance and...
6840 (make_pass_early_ipa_sra): ...new function.
6841 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
6842 subclass of gimple_opt_pass along with...
6843 (pass_data_ccp): ...new pass_data instance and...
6844 (make_pass_ccp): ...new function.
6845 (pass_fold_builtins): Convert from a global struct to a subclass of
6846 gimple_opt_pass along with...
6847 (pass_data_fold_builtins): ...new pass_data instance and...
6848 (make_pass_fold_builtins): ...new function.
6849 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
6850 subclass of gimple_opt_pass along with...
6851 (pass_data_copy_prop): ...new pass_data instance and...
6852 (make_pass_copy_prop): ...new function.
6853 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
6854 global struct to a subclass of gimple_opt_pass along with...
6855 (pass_data_rename_ssa_copies): ...new pass_data instance and...
6856 (make_pass_rename_ssa_copies): ...new function.
6857 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
6858 subclass of gimple_opt_pass along with...
6859 (pass_data_dce): ...new pass_data instance and...
6860 (make_pass_dce): ...new function.
6861 (pass_dce_loop): Convert from a global struct to a subclass of
6862 gimple_opt_pass along with...
6863 (pass_data_dce_loop): ...new pass_data instance and...
6864 (make_pass_dce_loop): ...new function.
6865 (pass_cd_dce): Convert from a global struct to a subclass of
6866 gimple_opt_pass along with...
6867 (pass_data_cd_dce): ...new pass_data instance and...
6868 (make_pass_cd_dce): ...new function.
6869 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
6870 subclass of gimple_opt_pass along with...
6871 (pass_data_dominator): ...new pass_data instance and...
6872 (make_pass_dominator): ...new function.
6873 (pass_phi_only_cprop): Convert from a global struct to a subclass of
6874 gimple_opt_pass along with...
6875 (pass_data_phi_only_cprop): ...new pass_data instance and...
6876 (make_pass_phi_only_cprop): ...new function.
6877 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
6878 subclass of gimple_opt_pass along with...
6879 (pass_data_dse): ...new pass_data instance and...
6880 (make_pass_dse): ...new function.
6881 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
6882 a subclass of gimple_opt_pass along with...
6883 (pass_data_forwprop): ...new pass_data instance and...
6884 (make_pass_forwprop): ...new function.
6885 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
6886 struct to a subclass of gimple_opt_pass along with...
6887 (pass_data_tree_ifcombine): ...new pass_data instance and...
6888 (make_pass_tree_ifcombine): ...new function.
6889 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
6890 subclass of gimple_opt_pass along with...
6891 (pass_data_ch): ...new pass_data instance and...
6892 (make_pass_ch): ...new function.
6893 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
6894 subclass of gimple_opt_pass along with...
6895 (pass_data_tree_loop): ...new pass_data instance and...
6896 (make_pass_tree_loop): ...new function.
6897 (pass_tree_loop_init): Convert from a global struct to a subclass of
6898 gimple_opt_pass along with...
6899 (pass_data_tree_loop_init): ...new pass_data instance and...
6900 (make_pass_tree_loop_init): ...new function.
6901 (pass_lim): Convert from a global struct to a subclass of
6902 gimple_opt_pass along with...
6903 (pass_data_lim): ...new pass_data instance and...
6904 (make_pass_lim): ...new function.
6905 (pass_tree_unswitch): Convert from a global struct to a subclass of
6906 gimple_opt_pass along with...
6907 (pass_data_tree_unswitch): ...new pass_data instance and...
6908 (make_pass_tree_unswitch): ...new function.
6909 (pass_predcom): Convert from a global struct to a subclass of
6910 gimple_opt_pass along with...
6911 (pass_data_predcom): ...new pass_data instance and...
6912 (make_pass_predcom): ...new function.
6913 (pass_vectorize): Convert from a global struct to a subclass of
6914 gimple_opt_pass along with...
6915 (pass_data_vectorize): ...new pass_data instance and...
6916 (make_pass_vectorize): ...new function.
6917 (pass_graphite): Convert from a global struct to a subclass of
6918 gimple_opt_pass along with...
6919 (pass_data_graphite): ...new pass_data instance and...
6920 (make_pass_graphite): ...new function.
6921 (pass_graphite_transforms): Convert from a global struct to a subclass
6922 of gimple_opt_pass along with...
6923 (pass_data_graphite_transforms): ...new pass_data instance and...
6924 (make_pass_graphite_transforms): ...new function.
6925 (pass_check_data_deps): Convert from a global struct to a subclass of
6926 gimple_opt_pass along with...
6927 (pass_data_check_data_deps): ...new pass_data instance and...
6928 (make_pass_check_data_deps): ...new function.
6929 (pass_iv_canon): Convert from a global struct to a subclass of
6930 gimple_opt_pass along with...
6931 (pass_data_iv_canon): ...new pass_data instance and...
6932 (make_pass_iv_canon): ...new function.
6933 (pass_scev_cprop): Convert from a global struct to a subclass of
6934 gimple_opt_pass along with...
6935 (pass_data_scev_cprop): ...new pass_data instance and...
6936 (make_pass_scev_cprop): ...new function.
6937 (pass_record_bounds): Convert from a global struct to a subclass of
6938 gimple_opt_pass along with...
6939 (pass_data_record_bounds): ...new pass_data instance and...
6940 (make_pass_record_bounds): ...new function.
6941 (pass_complete_unroll): Convert from a global struct to a subclass of
6942 gimple_opt_pass along with...
6943 (pass_data_complete_unroll): ...new pass_data instance and...
6944 (make_pass_complete_unroll): ...new function.
6945 (pass_complete_unrolli): Convert from a global struct to a subclass of
6946 gimple_opt_pass along with...
6947 (pass_data_complete_unrolli): ...new pass_data instance and...
6948 (make_pass_complete_unrolli): ...new function.
6949 (pass_parallelize_loops): Convert from a global struct to a subclass
6950 of gimple_opt_pass along with...
6951 (pass_data_parallelize_loops): ...new pass_data instance and...
6952 (make_pass_parallelize_loops): ...new function.
6953 (pass_loop_prefetch): Convert from a global struct to a subclass of
6954 gimple_opt_pass along with...
6955 (pass_data_loop_prefetch): ...new pass_data instance and...
6956 (make_pass_loop_prefetch): ...new function.
6957 (pass_iv_optimize): Convert from a global struct to a subclass of
6958 gimple_opt_pass along with...
6959 (pass_data_iv_optimize): ...new pass_data instance and...
6960 (make_pass_iv_optimize): ...new function.
6961 (pass_tree_loop_done): Convert from a global struct to a subclass of
6962 gimple_opt_pass along with...
6963 (pass_data_tree_loop_done): ...new pass_data instance and...
6964 (make_pass_tree_loop_done): ...new function.
6965 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
6966 struct to a subclass of gimple_opt_pass along with...
6967 (pass_data_cse_reciprocals): ...new pass_data instance and...
6968 (make_pass_cse_reciprocals): ...new function.
6969 (pass_cse_sincos): Convert from a global struct to a subclass of
6970 gimple_opt_pass along with...
6971 (pass_data_cse_sincos): ...new pass_data instance and...
6972 (make_pass_cse_sincos): ...new function.
6973 (pass_optimize_bswap): Convert from a global struct to a subclass of
6974 gimple_opt_pass along with...
6975 (pass_data_optimize_bswap): ...new pass_data instance and...
6976 (make_pass_optimize_bswap): ...new function.
6977 (pass_optimize_widening_mul): Convert from a global struct to a
6978 subclass of gimple_opt_pass along with...
6979 (pass_data_optimize_widening_mul): ...new pass_data instance and...
6980 (make_pass_optimize_widening_mul): ...new function.
6981 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
6982 subclass of gimple_opt_pass along with...
6983 (pass_data_phiopt): ...new pass_data instance and...
6984 (make_pass_phiopt): ...new function.
6985 (pass_cselim): Convert from a global struct to a subclass of
6986 gimple_opt_pass along with...
6987 (pass_data_cselim): ...new pass_data instance and...
6988 (make_pass_cselim): ...new function.
6989 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
6990 subclass of gimple_opt_pass along with...
6991 (pass_data_phiprop): ...new pass_data instance and...
6992 (make_pass_phiprop): ...new function.
6993 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
6994 subclass of gimple_opt_pass along with...
6995 (pass_data_pre): ...new pass_data instance and...
6996 (make_pass_pre): ...new function.
6997 (pass_fre): Convert from a global struct to a subclass of
6998 gimple_opt_pass along with...
6999 (pass_data_fre): ...new pass_data instance and...
7000 (make_pass_fre): ...new function.
7001 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
7002 subclass of gimple_opt_pass along with...
7003 (pass_data_reassoc): ...new pass_data instance and...
7004 (make_pass_reassoc): ...new function.
7005 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
7006 subclass of gimple_opt_pass along with...
7007 (pass_data_sink_code): ...new pass_data instance and...
7008 (make_pass_sink_code): ...new function.
7009 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
7010 subclass of gimple_opt_pass along with...
7011 (pass_data_strlen): ...new pass_data instance and...
7012 (make_pass_strlen): ...new function.
7013 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
7014 struct to a subclass of gimple_opt_pass along with...
7015 (pass_data_build_alias): ...new pass_data instance and...
7016 (make_pass_build_alias): ...new function.
7017 (pass_build_ealias): Convert from a global struct to a subclass of
7018 gimple_opt_pass along with...
7019 (pass_data_build_ealias): ...new pass_data instance and...
7020 (make_pass_build_ealias): ...new function.
7021 (pass_ipa_pta): Convert from a global struct to a subclass of
7022 simple_ipa_opt_pass along with...
7023 (pass_data_ipa_pta): ...new pass_data instance and...
7024 (make_pass_ipa_pta): ...new function.
7025 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
7026 subclass of gimple_opt_pass along with...
7027 (pass_data_uncprop): ...new pass_data instance and...
7028 (make_pass_uncprop): ...new function.
7029 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
7030 global struct to a subclass of gimple_opt_pass along with...
7031 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
7032 (make_pass_late_warn_uninitialized): ...new function.
7033 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
7034 to a subclass of gimple_opt_pass along with...
7035 (pass_data_init_datastructures): ...new pass_data instance and...
7036 (make_pass_init_datastructures): ...new function.
7037 (pass_early_warn_uninitialized): Convert from a global struct to a
7038 subclass of gimple_opt_pass along with...
7039 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
7040 (make_pass_early_warn_uninitialized): ...new function.
7041 (pass_update_address_taken): Convert from a global struct to a
7042 subclass of gimple_opt_pass along with...
7043 (pass_data_update_address_taken): ...new pass_data instance and...
7044 (make_pass_update_address_taken): ...new function.
7045 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
7046 struct to a subclass of gimple_opt_pass along with...
7047 (pass_data_release_ssa_names): ...new pass_data instance and...
7048 (make_pass_release_ssa_names): ...new function.
7049 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
7050 subclass of gimple_opt_pass along with...
7051 (pass_data_stdarg): ...new pass_data instance and...
7052 (make_pass_stdarg): ...new function.
7053 * tree-switch-conversion.c (pass_convert_switch): Convert from a
7054 global struct to a subclass of gimple_opt_pass along with...
7055 (pass_data_convert_switch): ...new pass_data instance and...
7056 (make_pass_convert_switch): ...new function.
7057 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
7058 to a subclass of gimple_opt_pass along with...
7059 (pass_data_tail_recursion): ...new pass_data instance and...
7060 (make_pass_tail_recursion): ...new function.
7061 (pass_tail_calls): Convert from a global struct to a subclass of
7062 gimple_opt_pass along with...
7063 (pass_data_tail_calls): ...new pass_data instance and...
7064 (make_pass_tail_calls): ...new function.
7065 * tree-vect-generic.c (pass_lower_vector): Convert from a global
7066 struct to a subclass of gimple_opt_pass along with...
7067 (pass_data_lower_vector): ...new pass_data instance and...
7068 (make_pass_lower_vector): ...new function.
7069 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
7070 gimple_opt_pass along with...
7071 (pass_data_lower_vector_ssa): ...new pass_data instance and...
7072 (make_pass_lower_vector_ssa): ...new function.
7073 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
7074 to a subclass of gimple_opt_pass along with...
7075 (pass_data_slp_vectorize): ...new pass_data instance and...
7076 (make_pass_slp_vectorize): ...new function.
7077 (pass_ipa_increase_alignment): Convert from a global struct to a
7078 subclass of simple_ipa_opt_pass along with...
7079 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
7080 (make_pass_ipa_increase_alignment): ...new function.
7081 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
7082 gimple_opt_pass along with...
7083 (pass_data_vrp): ...new pass_data instance and...
7084 (make_pass_vrp): ...new function.
7085 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
7086 subclass of simple_ipa_opt_pass along with...
7087 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
7088 (make_pass_ipa_free_lang_data): ...new function.
7089 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
7090 gimple_opt_pass along with...
7091 (pass_data_tsan): ...new pass_data instance and...
7092 (make_pass_tsan): ...new function.
7093 (pass_tsan_O0): Convert from a global struct to a subclass of
7094 gimple_opt_pass along with...
7095 (pass_data_tsan_O0): ...new pass_data instance and...
7096 (make_pass_tsan_O0): ...new function.
7097 * var-tracking.c (pass_variable_tracking): Convert from a global
7098 struct to a subclass of rtl_opt_pass along with...
7099 (pass_data_variable_tracking): ...new pass_data instance and...
7100 (make_pass_variable_tracking): ...new function.
7101 * web.c (pass_web): Convert from a global struct to a subclass of
7102 rtl_opt_pass along with...
7103 (pass_data_web): ...new pass_data instance and...
7104 (make_pass_web): ...new function.
7105 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
7106 declaration with that of...
7107 (make_pass_mode_switch_use): ...new function.
7108 (pass_resolve_sw_modes): Replace declaration with that of...
7109 (make_pass_resolve_sw_modes): ...new function.
7110 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
7111 from a global struct to a subclass of rtl_opt_pass along with...
7112 (pass_data_mode_switch_use): ...new pass_data instance and...
7113 (make_pass_mode_switch_use): ...new function.
7114 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
7115 from a global struct to a subclass of rtl_opt_pass along with...
7116 (pass_data_resolve_sw_modes): ...new pass_data instance and...
7117 (make_pass_resolve_sw_modes): ...new function.
7118 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
7119 struct to a subclass of rtl_opt_pass along with...
7120 (pass_data_insert_vzeroupper): ...new pass_data instance and...
7121 (make_pass_insert_vzeroupper): ...new function.
7122 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
7123 global struct to a subclass of rtl_opt_pass along with...
7124 (pass_data_work_around_errata): ...new pass_data instance and...
7125 (make_pass_work_around_errata): ...new function.
7126 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
7127 struct to a subclass of rtl_opt_pass along with...
7128 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
7129 (make_pass_mips_machine_reorg2): ...new function.
7131 2013-08-05 David Malcolm <dmalcolm@redhat.com>
7133 * passes.c (pass_manager::operator new): New.
7135 2013-08-05 David Malcolm <dmalcolm@redhat.com>
7137 Handwritten part of conversion of passes to C++ classes.
7139 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
7140 (toplev.o): Add dep on PASS_MANAGER_H.
7141 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
7142 of early local pases to reflect this moving from a global to a
7143 member of gcc::pass_manager.
7144 (cgraph_add_new_function): Likewise.
7145 * lto-cgraph.c (lto_output_node): Update for conversion of
7146 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
7147 * passes.c (opt_pass::clone): New.
7148 (opt_pass::gate): New.
7149 (opt_pass::execute): New.
7150 (opt_pass::opt_pass): New.
7151 (pass_manager::execute_early_local_passes): New.
7152 (pass_manager::execute_pass_mode_switching): new.
7153 (finish_optimization_passes): Convert to...
7154 (pass_manager::finish_optimization_passes): ...this.
7155 (finish_optimization_passes): Update for conversion of passes to
7157 (register_dump_files_1): Use has_gate since we cannot portably
7158 check a vtable entry against NULL.
7159 (dump_one_pass): Likewise.
7160 (ipa_write_summaries_2): Likewise.
7161 (ipa_write_optimization_summaries_1): Likewise.
7162 (ipa_read_summaries_1): Likewise.
7163 (ipa_read_optimization_summaries_1): Likewise.
7164 (execute_ipa_stmt_fixups): Likewise.
7165 (pass_manager::pass_manager): Rewrite pass-creation, invoking
7166 pass-creation functions rather than wiring up globals, and
7167 storing the results in fields of pass_manager generated using
7169 (pass_manager::dump_profile_report): Update for conversion of
7170 passes to C++ classes.
7171 (pass_manager::execute_ipa_summary_passes): Likewise.
7172 (execute_one_ipa_transform_pass): Likewise.
7173 (execute_one_pass): Use has_gate and has_execute since we cannot
7174 portably check a vtable entry against NULL.
7175 * pass_manager.h (pass_manager::finish_optimization_passes): New.
7176 (pass_manager): Use pass-instances.def to add fields for the
7177 various pass instances.
7178 * toplev.c (finalize): Update for move of
7179 finish_optimization_passes to a method of gcc::pass_manager.
7180 * toplev.h (finish_optimization_passes): Move to method of class
7182 * tree-pass.h (struct pass_data): New.
7183 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
7184 (opt_pass::gate): Convert to virtual function.
7185 (opt_pass::~opt_pass): New.
7186 (opt_pass::clone): New.
7187 (opt_pass::execute): Convert to virtual function.
7188 (opt_pass::opt_pass): New.
7189 (opt_pass::ctxt_): new.
7190 (gimple_opt_pass): Convert to subclass of opt_pass.
7191 (gimple_opt_pass::gimple_opt_pass): New.
7192 (rtl_opt_pass): Convert to subclass of opt_pass.
7193 (rtl_opt_pass::rtl_opt_pass): New.
7194 (ipa_opt_pass_d): Convert to subclass of opt_pass.
7195 (ipa_opt_pass_d::ipa_opt_pass_d): New.
7196 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
7197 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
7198 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
7199 invocation of pass_mode_switching to reflect this moving from a
7200 global to a member of gcc::pass_manager.
7201 (ix86_option_override): Rework how pass_insert_vzeroupper is
7202 added to the pass_manager to reflect autogenerated changes.
7203 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
7205 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
7207 PR rtl-optimization/57708
7208 * recog.c (peep2_find_free_register): Validate all regs in a
7211 2013-08-05 Jan Hubicka <jh@suse.cz>
7214 * cgraph.c (verify_cgraph_node): Accept local flags from other
7216 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
7217 (function_and_variable_visibility): Likewise.
7218 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
7220 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
7222 * graph.c (init_graph_slim_pretty_print): Remove.
7223 (print_graph_cfg): Do not call it. Use local pretty printer.
7224 (start_graph_dump): Likewise.
7226 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
7228 * gimple-pretty-print.c (buffer): Remove.
7229 (initialized): Likewise.
7230 (maybe_init_pretty_print): Likewise.
7231 (print_gimple_stmt): Do not call it. Use non-static local
7232 pretty_printer variable.
7233 (print_gimple_expr): Likewise.
7234 (print_gimple_seq): Likewise.
7235 (gimple_dump_bb): Likewise.
7237 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
7239 * asan.c (asan_pp): Remove.
7240 (asan_pp_initialized): Likewise.
7241 (asan_pp_initialize): Likewise.
7242 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
7243 (asan_emit_stack_protection): Tidy. Use local pretty printer.
7244 (asan_add_global): Likewise.
7246 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
7248 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
7249 * diagnostic.h (diagnostic_flush_buffer): Adjust.
7250 * pretty-print.c (pp_formatted_text_data): Likewise.
7251 (pp_indent): Rename from pp_base_indent.
7252 (pp_format): Rename from pp_base_format.
7253 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
7254 (pp_format_verbatim): Rename from pp_base_format_verbatim.
7255 (pp_flush): Rename from pp_base_flush.
7256 (pp_set_line_maximum_length): Rename from
7257 pp_base_set_line_maximum_length.
7258 (pp_clear_output_area): Rename from pp_base_clear_output_area.
7259 (pp_set_prefix): Rename from pp_base_set_prefix.
7260 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
7261 (pp_emit_prefix): Rename from pp_base_emit_prefix.
7262 (pp_append_text): Rename from pp_base_append_text.
7263 (pp_formatted_text): Rename from pp_base_formatted_text.
7264 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
7265 (pp_remaining_character_count_for_line): Rename from
7266 pp_base_remaining_character_count_for_line.
7267 (pp_newline): Rename from pp_base_newline.
7268 (pp_character): Rename from pp_base_character.
7269 (pp_string): Rename from pp_base_string.
7270 (pp_maybe_space): Rename from pp_base_maybe_space.
7271 * asan.c (asan_pp_string): Adjust.
7272 (asan_emit_stack_protection): Likewise.
7273 (asan_add_global): Likewise.
7274 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
7275 * tree-mudflap.c (mf_varname_tree): Likewise.
7276 * tree-pretty-print.c (pp_tree_identifier): Rename from
7277 pp_base_tree_identifier.
7278 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
7279 Declare as function.
7281 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
7283 * pretty-print.h (pp_bar_bar): New.
7284 (pp_ampersand_ampersand): Likewise.
7285 (pp_less_equal): Likewise.
7286 (pp_greater_equal): Likewise.
7287 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
7288 printer functions instead of pp_string or operators and punctuators.
7289 (dump_gimple_call): Likewise.
7290 (dump_gimple_omp_for): Likewise.
7291 (dump_gimple_transaction): Likewise.
7292 (dump_gimple_phi): Likewise.
7293 (pp_gimple_stmt_1): Likewise.
7294 * sched-vis.c (print_insn): Likewise.
7295 * tree-mudflap.c (mf_varname_tree): Likewise.
7296 * tree-pretty-print.c (dump_block_node): Likewise.
7297 (dump_generic_node): Likewise.
7299 2013-08-02 Jan Hubicka <jh@suse.cz>
7301 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
7303 * lto-streamer-out.c (tree_is_indexable): Results decls and
7304 parm decls are not indexable.
7305 (DFS_write_tree_body): Do not follow args and results.
7306 (hash_tree): Likewise.
7307 (output_functions): Rearrange so struct function is needed
7308 only when real body is output; be able to also ouptut abstract
7309 functions; output DECL_ARGUMENTS and DECL_RESULT.
7310 (lto_output): When not in WPA, ale store abstract functions.
7311 (write_symbol): Do not care about RESULT_DECL.
7312 (output_symbol_p): Handle correctly sbtract decls.
7313 * lto-streamer-in.c (input_function): Rearrange so struct
7314 function can be NULL at entry; allow streaming of
7315 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
7316 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
7317 sanity check during LTO.
7318 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
7319 RESULT_DECl and DECL_ARGUMENTS.
7320 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
7323 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
7325 * pretty-print.h (pp_underscore): New.
7327 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
7328 printer functions instead of pp_character.
7329 (dump_binary_rhs): Likewise.
7330 (dump_ternary_rhs): Likewise.
7331 (dump_gimple_call_args): Likewise.
7332 (pp_points_to_solution): Likewise.
7333 (dump_gimple_call): Likewise.
7334 (dump_gimple_switch): Likewise.
7335 (dump_gimple_cond): Likewise.
7336 (dump_gimple_bind): Likewise.
7337 (dump_gimple_try): Likewise.
7338 (dump_gimple_omp_for): Likewise.
7339 (dump_gimple_omp_continue): Likewise.
7340 (dump_gimple_omp_single): Likewise.
7341 (dump_gimple_omp_sections): Likewise.
7342 (dump_gimple_omp_block): Likewise.
7343 (dump_gimple_omp_critical): Likewise.
7344 (dump_gimple_transaction): Likewise.
7345 (dump_gimple_asm): Likewise.
7346 (dump_gimple_phi): Likewise.
7347 (dump_gimple_omp_parallel): Likewise.
7348 (dump_gimple_omp_task): Likewise.
7349 (dump_gimple_omp_atomic_load): Likewise.
7350 (dump_gimple_omp_atomic_store): Likewise.
7351 (dump_gimple_mem_ops): Likewise.
7352 (pp_gimple_stmt_1): Likewise.
7353 (pp_cfg_jump): Likewise.
7354 (dump_implicit_edges): Likewise.
7355 (gimple_dump_bb_for_graph): Likewise.
7356 * graph.c (draw_cfg_node): Likewise.
7357 * langhooks.c (lhd_print_error_function): Likewise.
7358 * sched-vis.c (print_exp): Likewise.
7359 (print_value): Likewise.
7360 (print_pattern): Likewise.
7361 (print_insn): Likewise.
7362 (rtl_dump_bb_for_graph): Likewise.
7363 * tree-pretty-print.c (dump_function_declaration): Likewise.
7364 (dump_array_domain): Likewise.
7365 (dump_omp_clause): Likewise.
7366 (dump_location): Likewise.
7367 (dump_generic_node): Likewise.
7368 (print_struct_decl): Likewise.
7369 * diagnostic.c (diagnostic_show_locus): Use pp_space.
7371 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
7373 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
7374 candidate table when replacing a candidate statement.
7375 (replace_rhs_if_not_dup): Likewise.
7376 (replace_one_candidate): Likewise.
7378 2013-08-02 Jan Hubicka <jh@suse.cz>
7379 Martin Liska <marxin.liska@gmail.com>
7381 * cgraphunit.c (add_new_function): Fix logic when adding from
7383 (assemble_thunk): Rename to ...
7384 (expand_thunk); .. this one; export; get it working with
7385 general functions; make produced gimple valid.
7386 * cgraph.h (expand_thunk): Declare.
7388 2013-08-02 Jan Hubicka <jh@suse.cz>
7390 * ipa-cp.c (gather_context_independent_values): Use
7391 ipa_get_param_move_cost.
7392 (get_replacement_map): Remove PARAM; move parameter folding
7394 (create_specialized_node): Update.
7395 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
7396 assert that we have gimple body; update move_cost.
7397 (count_formal_params): Assert that we have gimple body.
7398 (ipa_dump_param): New function.
7399 (ipa_alloc_node_params): Break out from ...
7400 (ipa_initialize_node_params): ... here.
7401 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
7402 (ipa_write_node_info): Stream move costs.
7403 (ipa_read_node_info): Read move costs.
7404 (ipa_update_after_lto_read): Do not recompute node params.
7405 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
7406 (ipa_get_param): Check we are not in WPA.
7407 (ipa_get_param_move_cost): New.
7408 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
7409 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
7410 parm numbers to be present.
7412 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
7414 PR rtl-optimization/58048
7415 * lra-constraints.c (process_alt_operands): Don't check asm
7416 operand on register.
7418 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
7420 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
7421 the implied StoreLoad barrier for atomic operations if before.
7423 2013-08-02 Jan Hubicka <jh@suse.cz>
7424 Martin Liska <marxin.liska@gmail.com>
7426 * cgraph.c (cgraph_function_body_availability): Do not check
7428 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
7429 symtab_node_availability): Declare.
7430 * ipa.c (can_replace_by_local_alias): New.
7431 (function_and_variable_visibility): Use it.
7432 * symtab.c (symtab_for_node_and_aliases,
7433 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
7435 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
7437 PR rtl-optimization/57963
7438 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
7439 (lra_constraints): Use them.
7441 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
7443 * config/arm/types.md (define_attr "type"): Add "load_acq"
7445 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
7447 (cortex_a53_store1): Likewise.
7449 2013-08-01 Jan Hubicka <jh@suse.cz>
7451 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
7452 partitions are not needed.
7454 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
7456 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
7457 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
7458 MAYBE_NON_Q_CLASS_P where appropriate.
7460 2013-08-01 Jan Hubicka <jh@suse.cz>
7462 * cgraph.h (release_function_body): Declare.
7463 * tree.c (free_lang_data_in_decl): Free, parameters and return values
7464 of unused delcarations.
7466 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7468 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
7469 RTL form when subtracting a constant.
7471 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7473 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
7474 Generate canonical plus rtx with negated immediate instead of minus
7476 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
7478 2013-08-01 Jan Hubicka <jh@suse.cz>
7480 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
7481 (cgraph_release_function_body): Likewise.
7482 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
7483 * cgraph.h (cgrpah_node): Rename abstract_and_needed
7484 to used_as_abstract_origin.
7485 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
7486 symbols used as abstract origins.
7487 * cgraphunit.c (analyze_functions): Update.
7488 * ipa.c (symtab_remove_unreachable_nodes): Recompute
7489 used_as_abstract_origin.
7490 * tree-inline.c (tree_function_versioning): Update
7491 used_as_abstract_origin; be ready for DECL_RESULT and
7492 DECL_ARGUMENTS to be NULL.
7494 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
7495 for abstract functions.
7496 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
7499 2013-08-01 Jan Hubicka <jh@suse.cz>
7501 * profile.c (compute_value_histograms): Fix thinko.
7503 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
7505 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
7506 aarch-common-protos.h to extra_headers.
7507 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
7508 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
7509 * config/aarch64/t-aarch64 (aarch-common.o): Define.
7511 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
7513 * config/aarch64/aarch64.md (define_attr "type"): Delete.
7514 Include "../arm/types.md". Define "type" attribute for all patterns.
7515 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
7518 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
7520 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
7521 to support power8 load fusion.
7522 (fusion_gpr_mem_load): Likewise.
7524 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
7526 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
7527 declarations for power8 load fusion.
7528 (emit_fusion_gpr_load): Likewise.
7530 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
7531 tuning for power8, turn on fusion mode by default. Turn on sign
7532 extending fusion mode if normal fusion mode is on, and we are at
7534 (fusion_gpr_load_p): New function, return true if we can fuse an
7535 addis instruction with a dependent load to a GPR.
7536 (emit_fusion_gpr_load): Emit the instructions for power8 load
7539 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
7540 (VSX load fusion peepholes): New peepholes to fuse together an
7541 addi instruction with a VSX load instruction.
7543 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
7544 peepholes to fuse an addis instruction with a load to a GPR base
7545 register. If we are supporting sign extending fusions, convert
7546 sign extending loads to zero extending loads and add an explicit
7549 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
7551 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
7552 aarch-common-protos.h to extra_headers.
7553 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
7554 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
7555 (arm_early_store_addr_dep): Likewise.
7556 (arm_no_early_alu_shift_dep): Likewise.
7557 (arm_no_early_alu_shift_value_dep): Likewise.
7558 (arm_no_early_mul_dep): Likewise.
7559 (arm_no_early_store_addr_dep): Likewise.
7560 (arm_mac_accumulator_is_mul_result): Likewise.
7561 (arm_mac_accumulator_is_result): Likewise.
7562 * config/arm/aarch-common.c: ... here. New file.
7563 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
7565 (arm_early_store_addr_dep): Likewise.
7566 (arm_no_early_alu_shift_dep): Likewise.
7567 (arm_no_early_alu_shift_value_dep): Likewise.
7568 (arm_no_early_mul_dep): Likewise.
7569 (arm_no_early_store_addr_dep): Likewise.
7570 (arm_mac_accumulator_is_mul_result): Likewise.
7571 (arm_mac_accumulator_is_result): Likewise.
7572 * config/arm/aarch-common-protos.h: ... here. New file.
7573 * config/arm/t-arm (aarch-common.o): Define.
7575 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
7577 * config/arm/arm.md: Include new file "types.md".
7578 (define_attr "type"): Move from here to ...
7579 (define_attr "mul32"): Likewise.
7580 (define_attr "mul64"): Likewise.
7581 * config/arm/types.md: ... here. New file.
7583 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
7585 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
7586 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
7588 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7590 * gen-pass-instances.awk: Fix offset of substr().
7592 2013-07-31 David Malcolm <dmalcolm@redhat.com>
7594 * Makefile.in (pass-instances.def): New.
7595 (passes.o): Replace dependency on passes.def with one on
7598 * gen-pass-instances.awk: New.
7600 * passes.c (pass_manager::pass_manager): Use pass-instances.def
7601 rather than passes.def, updating local definition of NEXT_PASS
7602 macro to add an extra NUM parameter (currently unused).
7604 2013-07-30 David Malcolm <dmalcolm@redhat.com>
7606 * Makefile.in (PASS_MANAGER_H): New.
7607 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
7608 (passes.o): Likewise.
7609 (statistics.o): Likewise.
7610 (cgraphunit.o): Likewise.
7611 (context.o): Depend on PASS_MANAGER_H.
7613 * pass_manager.h: New.
7615 * cgraphunit.c (cgraph_add_new_function): Update for moves
7616 of globals to fields of pass_manager.
7617 (analyze_function): Likewise.
7618 (expand_function): Likewise.
7619 (ipa_passes): Likewise.
7620 (compile): Likewise.
7622 * context.c (context::context): New.
7623 * context.h (context::context): New.
7624 (context::get_passes): New.
7625 (context::passes_): New.
7627 * lto-cgraph.c (input_node): Update for moves of globals to
7628 fields of pass_manager.
7630 * passes.c (all_passes): Remove, in favor of a field of the
7631 same name within the new class pass_manager.
7632 (all_small_ipa_passes): Likewise.
7633 (all_lowering_passes): Likewise.
7634 (all_regular_ipa_passes): Likewise.
7635 (all_late_ipa_passes): Likewise.
7636 (all_lto_gen_passes): Likewise.
7637 (passes_by_id): Likewise.
7638 (passes_by_id_size): Likewise.
7639 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
7640 the new class pass_manager.
7641 (set_pass_for_id): Convert to...
7642 (pass_manager::set_pass_for_id): ...method.
7643 (get_pass_for_id): Convert to...
7644 (pass_manager::get_pass_for_id): ...method.
7645 (register_one_dump_file): Move body of implementation into...
7646 (pass_manager::register_one_dump_file): ...here.
7647 (register_dump_files_1): Convert to...
7648 (pass_manager::register_dump_files_1): ...method.
7649 (register_dump_files): Convert to...
7650 (pass_manager::register_dump_files): ...method.
7651 (create_pass_tab): Update for moves of globals to fields of
7653 (dump_passes): Move body of implementation into...
7654 (pass_manager::dump_passes): ...here.
7655 (register_pass): Move body of implementation into...
7656 (pass_manager::register_pass): ...here.
7657 (init_optimization_passes): Convert into...
7658 (pass_manager::pass_manager): ...constructor for new
7659 pass_manager class, and initialize the pass_lists array.
7660 (check_profile_consistency): Update for moves of globals to
7661 fields of pass_manager.
7662 (dump_profile_report): Move body of implementation into...
7663 (pass_manager::dump_profile_report): ...here.
7664 (ipa_write_summaries_1): Update for moves of pass lists from
7665 being globals to fields of pass_manager.
7666 (ipa_write_optimization_summaries): Likewise.
7667 (ipa_read_summaries): Likewise.
7668 (ipa_read_optimization_summaries): Likewise.
7669 (execute_all_ipa_stmt_fixups): Likewise.
7671 * statistics.c (statistics_fini): Update for moves of globals to
7672 fields of pass_manager.
7674 * toplev.c (general_init): Replace call to
7675 init_optimization_passes with construction of the pass_manager
7678 * tree-pass.h (all_passes): Remove, in favor of a field of the
7679 same name within the new class pass_manager.
7680 (all_small_ipa_passes): Likewise.
7681 (all_lowering_passes): Likewise.
7682 (all_regular_ipa_passes): Likewise.
7683 (all_lto_gen_passes): Likewise.
7684 (all_late_ipa_passes): Likewise.
7685 (passes_by_id): Likewise.
7686 (passes_by_id_size): Likewise.
7687 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
7688 the new class pass_manager.
7689 (get_pass_for_id): Remove.
7691 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
7693 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
7696 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
7698 * arm.md (mulhi3): New expand pattern.
7700 2013-07-30 Jan Hubicka <jh@suse.cz>
7701 Martin Liska <marxin.liska@gmail.com>
7703 * profile.c (compute_value_histograms): Do not ICE when
7704 there is mismatch only on some counters.
7706 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
7708 PR rtl-optimization/57637
7709 * function.c (move_insn_for_shrink_wrap): Also check the
7710 GEN set of the LIVE problem for the liveness analysis
7711 if it exists, otherwise give up.
7713 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
7715 PR tree-optimization/57993
7716 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
7717 replaced statement in the candidate table.
7718 (phi_add_costs): Return infinite cost when the hidden basis does
7719 not dominate all phis on which the candidate is dependent.
7720 (replace_one_candidate): Record replaced statement in the
7723 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
7725 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
7726 (ashlv2si3): New expander.
7727 (*ashlv2si3_i): New define_insn_and_split.
7728 * predicates.md (float_operation): Allow patterns with three
7731 PR rtl-optimization/58021
7732 * mode-switching.c (create_pre_exit): Always split off preceding
7733 insns if we are not at the basic block head.
7735 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
7737 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
7738 (UCLIBC_DYNAMIC_LINKER): New macro.
7739 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
7741 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
7742 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
7744 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
7745 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
7746 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
7747 for SF and DF modes. Use ieee_quad_format for TF mode.
7748 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
7749 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
7750 (mips_option_override): Handle `-mnan=legacy'.
7751 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
7752 `-mabs=2008' and `-mnan=2008'.
7753 (OPTION_DEFAULT_SPECS): Add "nan" default.
7754 (ASM_SPEC): Handle `-mnan='.
7755 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
7756 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
7757 comment accordingly.
7758 (neg<mode>2): Likewise.
7759 * config/mips/mips.opt (mabs, mnan): New options.
7760 * doc/install.texi (Configuration): Document `--with-nan=' option.
7761 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
7763 (MIPS Options): Document them.
7764 * config.gcc <mips*-*-*>: Handle `--with-nan='.
7765 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
7766 * configure: Regenerate.
7767 * config.in: Regenerate.
7769 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
7771 * config/i386/i386.md (float post-reload splitters): Do not check
7772 for subregs of SSE registers.
7774 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
7775 H.J. Lu <hongjiu.lu@intel.com>
7779 * config/i386/i386.md (post-reload splitter
7780 to avoid partial SSE reg dependency stalls): New pattern.
7782 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
7784 * config/s390/s390.md ("movcc"): Swap load and store instructions.
7786 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
7788 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
7789 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
7791 2013-07-26 Cary Coutant <ccoutant@google.com>
7793 * dwarf2out.c (die_checksum_ordered): Don't include template
7794 instantiations in signature.
7795 (is_template_parameter): New function.
7796 (is_template_instantiation): New function.
7797 (generate_skeleton_bottom_up): Don't include template instantiations
7799 (generate_skeleton): Likewise.
7800 (break_out_comdat_types): Move recursive call to break out nested
7802 (prune_unused_types_mark_generic_parms_dies): Call
7803 is_template_parameter.
7805 2013-07-26 Ian Bolton <ian.bolton@arm.com>
7807 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
7808 uses vector registers.
7809 * config/aarch64/iterators.md: Add attributes rtn and vas.
7811 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7812 Richard Earnshaw <richard.earnshaw@arm.com>
7814 * combine.c (simplify_comparison): Re-canonicalize operands
7816 * config/arm/arm.md (movcond_addsi): New splitter.
7818 2013-07-25 Sterling Augustine <saugustine@google.com>
7820 * dwarf2out.c (size_of_pubnames): Move code to...
7821 (include_pubname_in_output): ...here. New.
7822 (want_pubnames): Rearrange.
7823 (output_pubnames): Call include_pubname_in_output. Move assertion.
7825 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
7827 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
7829 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
7832 * doc/extend.texi: Remove obsolete builtins. Fix
7833 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
7835 2013-07-25 Jan Hubicka <jh@suse.cz>
7837 * cgraph.c (release_function_body): Break out from ...
7838 (cgraph_release_function_body): ... this one; also release DECL_RESULT
7840 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
7841 old_tree in the map.
7842 (create_specialized_node): Update.
7843 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
7845 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
7846 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
7847 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
7848 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
7851 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7853 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
7854 addsi3_carryin_alt2_<optab>): Correct output template.
7856 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7858 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
7859 Adjust for arm_restrict_it.
7860 Remove trailing whitespace.
7862 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
7864 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
7865 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
7867 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
7869 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
7871 PR rtl-optimization/57960
7872 * lra-constraints.c (process_alt_operands): Use the right mode
7873 when checking strict_low.
7875 2013-07-25 Jan Hubicka <jh@suse.cz>
7877 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
7878 * cgraph.c (cgraph_remove_node): Do not release function body
7879 when in cgraph streaming.
7880 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
7881 in other partitions are not considered reachable; fix handling of
7884 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7886 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
7888 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7893 * config/arm/arm.md ("*sibcall_insn): Replace use of
7894 Ss with US. Adjust output for v5 and v4t.
7895 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
7897 * config/arm/constraints.md ("Ss"): Rename to US.
7899 2013-07-25 Terry Guo <terry.guo@arm.com>
7901 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
7902 shift_add/shift_sub0/shift_sub1 RTXs.
7904 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
7905 Anton Blanchard <anton@au1.ibm.com>
7907 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
7908 (altivec_vpks<VI_char>ss): Likewise.
7909 (altivec_vpks<VI_char>us): Likewise.
7910 (altivec_vpku<VI_char>us): Likewise.
7911 (altivec_vpku<VI_char>um): Likewise.
7913 2013-07-24 David Malcolm <dmalcolm@redhat.com>
7915 Introduce context class.
7917 * Makefile.in (CONTEXT_H): New.
7918 (OBJS): Add context.o.
7919 (toplev.o): Add CONTEXT_H to dependencies.
7922 * toplev.c (general_init): Create the singleton gcc::context instance.
7928 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
7930 PR rtl-optimization/57968
7931 * mode-switching.c (create_pre_exit): Allow instructions that
7932 don't set a return register to need a non-exit mode.
7934 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
7935 Anton Blanchard <anton@au1.ibm.com>
7937 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
7938 operands to vperm for little endian.
7939 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
7940 of lvsl to create the control mask for a vperm for little endian.
7942 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7943 Anton Blanchard <anton@au1.ibm.com>
7945 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
7946 two operands for little-endian.
7948 2013-07-23 Steve Ellcey <sellcey@mips.com>
7950 * config/mips/mips.c (mips_case_values_threshold): New.
7951 (TARGET_CASE_VALUES_THRESHOLD): Define.
7953 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7954 Anton Blanchard <anton@au1.ibm.com>
7956 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
7957 selection of field for vector splat in little endian mode.
7959 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
7961 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
7962 expanders to rs6000.md.
7963 (ior<mode>3): Likewise.
7964 (and<mode>3): Likewise.
7965 (one_cmpl<mode>2): Likewise.
7966 (nor<mode>3): Likewise.
7967 (andc<mode>3): Likewise.
7968 (eqv<mode>3): Likewise.
7969 (nand<mode>3): Likewise.
7970 (orc<mode>3): Likewise.
7972 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
7975 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
7976 to split multi-word logical operations.
7977 (rs6000_split_logical_di): Likewise.
7978 (rs6000_split_logical): Likewise.
7980 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
7981 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
7982 and allow TImode operations in 32-bit.
7983 (vsx_and<mode>3_64bit): Likewise.
7984 (vsx_ior<mode>3_32bit): Likewise.
7985 (vsx_ior<mode>3_64bit): Likewise.
7986 (vsx_xor<mode>3_32bit): Likewise.
7987 (vsx_xor<mode>3_64bit): Likewise.
7988 (vsx_one_cmpl<mode>2_32bit): Likewise.
7989 (vsx_one_cmpl<mode>2_64bit): Likewise.
7990 (vsx_nor<mode>3_32bit): Likewise.
7991 (vsx_nor<mode>3_64bit): Likewise.
7992 (vsx_andc<mode>3_32bit): Likewise.
7993 (vsx_andc<mode>3_64bit): Likewise.
7994 (vsx_eqv<mode>3_32bit): Likewise.
7995 (vsx_eqv<mode>3_64bit): Likewise.
7996 (vsx_nand<mode>3_32bit): Likewise.
7997 (vsx_nand<mode>3_64bit): Likewise.
7998 (vsx_orc<mode>3_32bit): Likewise.
7999 (vsx_orc<mode>3_64bit): Likewise.
8001 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
8002 logical types in GPRs.
8004 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
8005 logical insns to rs6000.md, and allow TImode operations in
8007 (altivec_ior<mode>3): Likewise.
8008 (altivec_xor<mode>3): Likewise.
8009 (altivec_one_cmpl<mode>2): Likewise.
8010 (altivec_nor<mode>3): Likewise.
8011 (altivec_andc<mode>3): Likewise.
8013 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
8014 attributes for moving the 128-bit logical operations into
8016 (BOOL_REGS_OUTPUT): Likewise.
8017 (BOOL_REGS_OP1): Likewise.
8018 (BOOL_REGS_OP2): Likewise.
8019 (BOOL_REGS_UNARY): Likewise.
8020 (BOOL_REGS_AND_CR0): Likewise.
8021 (one_cmpl<mode>2): Add support for DI logical operations on
8022 32-bit, splitting the operations to 32-bit.
8026 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
8027 changes to combine the 32/64-bit code, allow logical operations on
8028 TI mode in 32-bit, and to use similar match_operator patterns like
8029 scalar mode uses. Combine the Altivec and VSX code for logical
8030 operations, and move it here.
8031 (ior<mode>3, 128-bit types): Likewise.
8032 (xor<mode>3, 128-bit types): Likewise.
8033 (one_cmpl<mode>3, 128-bit types): Likewise.
8034 (nor<mode>3, 128-bit types): Likewise.
8035 (andc<mode>3, 128-bit types): Likewise.
8036 (eqv<mode>3, 128-bit types): Likewise.
8037 (nand<mode>3, 128-bit types): Likewise.
8038 (orc<mode>3, 128-bit types): Likewise.
8039 (and<mode>3_internal): Likewise.
8040 (bool<mode>3_internal): Likewise.
8041 (boolc<mode>3_internal1): Likewise.
8042 (boolc<mode>3_internal2): Likewise.
8043 (boolcc<mode>3_internal1): Likewise.
8044 (boolcc<mode>3_internal2): Likewise.
8045 (eqv<mode>3_internal1): Likewise.
8046 (eqv<mode>3_internal2): Likewise.
8047 (one_cmpl1<mode>3_internal): Likewise.
8049 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
8051 * config/microblaze/microblaze.c (microblaze_expand_prologue):
8052 Rename flag_stack_usage to flag_stack_usage_info.
8054 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
8056 * config/microblaze/sync.md: New file.
8057 * config/microblaze/microblaze.md: Include sync.md
8058 * config/microblaze/microblaze.c: Add print_operand 'y'.
8059 * config/microblaze/constraints.md: Add memory_contraint
8060 'Q' which is a single register.
8062 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
8064 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
8066 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
8068 * reload.c (find_reloads): Exit loop once we find this operand
8069 cannot be reloaded somehow for this alternative.
8071 * reload.c (find_reloads): Exit loop once we find a hard register.
8073 * rtlanal.c (computed_jump_p): Exit loop once we find label
8076 * i386.c (ix86_pad_returns): Exit loop after setting replace.
8078 * cfgloopmanip.c (remove_path): Exit loop after setting
8081 * gensupport.c (subst_dup): Avoid loop if code is not
8082 MATCH_DUP nor MATCH_OP_DUP.
8084 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
8086 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
8088 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
8090 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
8091 true for SP_REGNUM if mode == ptr_mode.
8092 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
8093 with value R0_REGNUM + 31.
8095 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
8097 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
8098 pad pointer-typed argument downward.
8100 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
8102 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
8103 and __ILP32__ when the ILP32 model is in use.
8105 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
8107 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
8108 (aarch64_load_symref_appropriately): In the case of
8109 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
8110 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
8111 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
8112 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
8113 if 'mode' doesn't equal to 'ptr_mode'.
8114 (aarch64_output_mi_thunk): Add an assertion on the alignment of
8115 'vcall_offset'; change to call aarch64_emit_move differently depending
8116 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
8117 to calculate the upper bound of 'vcall_offset'.
8118 (aarch64_cannot_force_const_mem): Change to also return true if
8120 (aarch64_legitimize_reload_address): In the case of large
8121 displacements, add new local variable 'xmode' and an assertion
8122 based on it; change to use 'xmode' to generate the new rtx and
8124 (aarch64_asm_trampoline_template): Change to generate the template
8125 differently depending on TARGET_ILP32 or not; change to use
8126 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
8127 (aarch64_trampoline_size): Removed.
8128 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
8129 and replace immediate literals with it. Change to use 'ptr_mode'
8130 instead of 'DImode' and call convert_memory_address if the mode
8131 of 'fnaddr' doesn't equal to 'ptr_mode'.
8132 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
8134 (aarch64_elf_asm_destructor): Likewise.
8135 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
8136 on TARGET_ILP32 instead of aarch64_trampoline_size.
8137 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
8138 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
8139 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
8140 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
8141 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
8142 (storewb_pair<GPI:mode>_<P:mode>): ... this.
8143 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
8144 depending on the value of 'mode'.
8145 (add_losym_<mode>): New.
8146 (ldr_got_small_<mode>): New, based on ldr_got_small.
8147 (ldr_got_small): Remove.
8148 (ldr_got_small_sidi): New.
8149 * config/aarch64/iterators.md (P): New.
8150 (PTR): Change to 'ptr_mode' in the condition.
8152 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
8154 * config.gcc (aarch64*-*-*): Support --with-abi.
8155 (aarch64*-*-elf): Support --with-multilib-list.
8156 (aarch64*-*-linux*): Likewise.
8157 (supported_defaults): Add abi to aarch64*-*-*.
8158 * configure.ac: Mention AArch64 for --with-multilib-list.
8159 * configure: Re-generated.
8160 * config/aarch64/biarchilp32.h: New file.
8161 * config/aarch64/biarchlp64.h: New file.
8162 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
8164 (MULTILIB_DEFAULTS): Ditto.
8165 (DRIVER_SELF_SPECS): Ditto.
8166 (ASM_SPEC): Update to also substitute -mabi.
8167 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
8168 file whose name depends on -mabi= and -mbig-endian.
8169 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
8171 (POINTER_SIZE): New define.
8172 (POINTERS_EXTEND_UNSIGNED): Ditto.
8173 (enum aarch64_abi_type): New enumeration tag.
8174 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
8175 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
8176 (TARGET_ILP32): New define.
8177 * config/aarch64/aarch64.opt (mabi): New.
8179 (ilp32, lp64): New values for -mabi.
8180 * config/aarch64/t-aarch64 (comma): New define.
8181 (MULTILIB_OPTIONS): Ditto.
8182 (MULTILIB_DIRNAMES): Ditto.
8183 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
8184 * doc/invoke.texi: Document -mabi for AArch64.
8186 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
8188 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
8190 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
8191 Anton Blanchard <anton@au1.ibm.com>
8193 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
8194 endianness when selecting field to splat.
8196 2013-07-22 Eric Christopher <echristo@gmail.com>
8198 * dwarf2out.c (die_odr_checksum): New function to use
8199 CHECKSUM_ macros and ULEB128 for DIE tag.
8200 (generate_type_signature): Use.
8202 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
8204 * config.gcc (sparc*-*-*): Accept leon3 processor.
8205 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
8206 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
8207 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
8208 * config/sparc/sparc.opt (enum processor_type): Add leon3.
8209 (mfix-ut699): Adjust comment.
8210 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
8211 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
8212 (CPP_CPU_SPEC): Likewise.
8213 (ASM_CPU_SPEC): Likewise.
8214 * config/sparc/sparc.c (leon3_cost): New constant.
8215 (sparc_option_override): Add leon3 support.
8216 (mem_ref): New function.
8217 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
8218 (sparc_do_work_around_errata): Look into the instruction in the delay
8219 slot and adjust accordingly. Add fix for the data cache nullify issues
8220 of the UT699. Change insertion position for the NOP.
8221 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
8222 (leon3_load): New reservation.
8223 (leon_store): Bump latency to 2.
8224 (grfpu): New automaton.
8225 (grfpu_alu): New unit.
8226 (grfpu_ds): Likewise.
8227 (leon_fp_alu): Adjust.
8228 (leon_fp_mult): Delete.
8229 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
8230 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
8231 * config/sparc/sparc.md (cpu): Add leon3.
8232 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
8234 (atomic_test_and_set): Likewise.
8237 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
8239 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
8240 default for R5900 targets.
8241 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
8242 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
8243 * config/mips/mips.c (mips_option_override): Report an error for
8244 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
8245 for -march=r5900 -mhard-float.
8247 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
8249 * df-problems.c (can_move_insns_across): Exit loop once we
8250 find a non-fixed, non-global register.
8252 * ipa-pure-const.c (propagate_nothrow): Exit loop after
8255 * omega.c (omega_eliminate_red): Break after setting red_found.
8256 (omega_problem_has_red_equations): Similarly after setting found.
8257 (omega_query_variable): Similarly after setting coupled.
8259 2013-07-22 Marek Polacek <polacek@redhat.com>
8261 * gimplify.c: Don't include gimple.h twice.
8263 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8265 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
8266 instead of TARGET_THUMB1.
8267 (Pz): New constraint.
8268 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
8270 (compare_negsi_si): Likewise.
8271 (compare_addsi2_op0): Likewise.
8272 (compare_addsi2_op1): Likewise.
8273 (addsi3_carryin_<optab>): Likewise.
8274 (addsi3_carryin_alt2_<optab>): Likewise.
8275 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
8276 for arm_restrict_it.
8277 (subsi3_carryin): Likewise.
8278 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
8279 (minmax_arithsi): Disable for arm_restrict_it.
8280 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
8281 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
8282 (satsi_<SAT:code>_shift): Likewise.
8283 (arm_shiftsi3): Add alternative for 16-bit encoding.
8284 (arm32_movhf): Disable for arm_restrict_it.
8285 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
8286 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
8288 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
8290 * config/arm/arm.md (attribute "insn"): Delete.
8291 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
8292 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
8293 (not_shiftsi): Update for attribute change.
8294 (not_shiftsi_compare0): Likewise.
8295 (not_shiftsi_compare0_scratch): Likewise.
8296 (arm_one_cmplsi2): Likewise.
8297 (thumb1_one_cmplsi2): Likewise.
8298 (notsi_compare0): Likewise.
8299 (notsi_compare0_scratch): Likewise.
8300 (thumb1_movdi_insn): Likewise.
8301 (arm_movsi_insn): Likewise.
8302 (movhi_insn_arch4): Likewise.
8303 (movhi_bytes): Likewise.
8304 (arm_movqi_insn): Likewise.
8305 (thumb1_movqi_insn): Likewise.
8306 (arm32_movhf): Likewise.
8307 (thumb1_movhf): Likewise.
8308 (arm_movsf_soft_insn): Likewise.
8309 (thumb1_movsf_insn): Likewise.
8310 (thumb_movdf_insn): Likewise.
8311 (movsicc_insn): Likewise.
8312 (movsfcc_soft_insn): Likewise.
8313 (and_scc): Likewise.
8314 (cond_move): Likewise.
8315 (if_move_not): Likewise.
8316 (if_not_move): Likewise.
8317 (if_shift_move): Likewise.
8318 (if_move_shift): Likewise.
8319 (if_shift_shift): Likewise.
8320 (if_not_arith): Likewise.
8321 (if_arith_not): Likewise.
8322 (cond_move_not): Likewise.
8323 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
8324 (neon_mov<mode>): Likewise.
8325 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
8326 (thumb2_movsi_vfp): Likewise.
8327 (movsf_vfp): Likewise.
8328 (thumb2_movsf_vfp): Likewise.
8329 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
8331 (cortexa7_older_only): Likewise.
8332 (cortexa7_younger): Likewise.
8333 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
8334 (1020alu_shift_op): Likewise.
8335 (1020alu_shift_reg_op): Likewise.
8336 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
8337 (alu_shift_op): Likewise.
8338 (alu_shift_reg_op): Likewise.
8339 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
8340 (11_alu_shift_op): Likewise.
8341 (11_alu_shift_reg_op): Likewise.
8342 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
8343 (9_alu_shift_reg_op): Likewise.
8344 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
8346 (cortex_a15_alu_shift): Likewise.
8347 (cortex_a15_alu_shift_reg): Likewise.
8348 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
8349 (cortex_a5_alu_shift): Likewise.
8350 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
8352 (cortex_a53_alu_shift): Likewise.
8353 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
8355 (cortex_a7_alu_reg): Likewise.
8356 (cortex_a7_alu_shift): Likewise.
8357 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
8358 (cortex_a8_alu_shift): Likewise.
8359 (cortex_a8_alu_shift_reg): Likewise.
8360 (cortex_a8_mov): Likewise.
8361 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
8362 (cortex_a9_dp_shift): Likewise.
8363 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
8364 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
8365 (cortex_r4_mov): Likewise.
8366 (cortex_r4_alu_shift): Likewise.
8367 (cortex_r4_alu_shift_reg): Likewise.
8368 * config/arm/fa526.md (526_alu_op): Update for attribute change.
8369 (526_alu_shift_op): Likewise.
8370 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
8371 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
8372 (626te_alu_shift_op): Likewise.
8373 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
8374 (726te_alu_op): Likewise.
8375 (726te_alu_shift_op): Likewise.
8376 (726te_alu_shift_reg_op): Likewise.
8377 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
8378 (mp626_alu_shift_op): Likewise.
8379 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
8380 (pj4_alu_e1_conds): Likewise.
8381 (pj4_alu): Likewise.
8382 (pj4_alu_conds): Likewise.
8383 (pj4_shift): Likewise.
8384 (pj4_shift_conds): Likewise.
8385 (pj4_alu_shift): Likewise.
8386 (pj4_alu_shift_conds): Likewise.
8388 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8390 * config/arm/predicates.md (shiftable_operator_strict_it):
8392 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
8393 Disable cond_exec version for arm_restrict_it.
8394 (thumb2_smaxsi3): Convert to generate cond_exec.
8395 (thumb2_sminsi3): Likewise.
8396 (thumb32_umaxsi3): Likewise.
8397 (thumb2_uminsi3): Likewise.
8398 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
8399 (thumb2_neg_abssi2): Likewise.
8400 (thumb2_mov_scc): Add alternative for 16-bit encoding.
8401 (thumb2_movsicc_insn): Adjust alternatives.
8402 (thumb2_mov_negscc): Disable for arm_restrict_it.
8403 (thumb2_mov_negscc_strict_it): New pattern.
8404 (thumb2_mov_notscc_strict_it): New pattern.
8405 (thumb2_mov_notscc): Disable for arm_restrict_it.
8406 (thumb2_ior_scc): Likewise.
8407 (thumb2_ior_scc_strict_it): New pattern.
8408 (thumb2_cond_move): Adjust for arm_restrict_it.
8409 (thumb2_cond_arith): Disable for arm_restrict_it.
8410 (thumb2_cond_arith_strict_it): New pattern.
8411 (thumb2_cond_sub): Adjust for arm_restrict_it.
8412 (thumb2_movcond): Likewise.
8413 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
8414 (thumb2_zero_extendhisi2_v6): Likewise.
8415 (thumb2_zero_extendqisi2_v6): Likewise.
8416 (orsi_notsi_si): Likewise.
8417 (orsi_not_shiftsi_si): Likewise.
8419 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
8421 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
8422 instruction sequence is 1 byte shorter.
8424 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
8426 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
8427 it is not needed after split.
8429 2013-07-20 Iain Sandoe <iain@codesourcery.com>
8432 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
8433 second label for nonlocal goto receivers. Don't output pic base labels
8434 unless we're producing PIC; mark that action unreachable().
8435 (ix86_save_reg): If the function contains a nonlocal label, save the
8437 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
8438 * config/darwin.c (emitted_pic_label_num): New GTY.
8439 (update_pic_label_number_if_needed): New.
8440 (machopic_output_function_base_name): Adjust for nonlocal receiver
8442 (machopic_should_output_picbase_label): New.
8443 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
8444 (nonlocal_goto_receiver): New insn and split.
8446 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
8448 * config/aarch64/aarch64-builtins.c
8449 (aarch64_fold_builtin): Fold abs in all modes.
8450 * config/aarch64/aarch64-simd-builtins.def
8451 (abs): Enable for all modes.
8452 * config/aarch64/arm_neon.h
8453 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
8454 (vabs_f64): Add missing intrinsic.
8456 2013-07-19 Ian Bolton <ian.bolton@arm.com>
8458 * config/aarch64/arm_neon.h (vabs_s64): New function
8460 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
8463 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
8464 * config/avr/avr.md (adjust_len): Add `round'.
8465 * config/avr/avr-protos.h (avr_out_round): New prototype.
8466 (avr_out_plus): Add `out_label' argument.
8467 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
8468 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
8469 Handle the case where `insn' is just a pattern.
8470 (avr_out_bitop): Handle the case where `insn' is just a pattern.
8471 (avr_out_round): New function.
8472 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
8474 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
8476 * config/microblaze/microblaze.c (microblaze_expand_prologue):
8477 Add check for flag_stack_usage to handle -fstack-usage support
8479 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
8481 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
8482 interaction for new Power8 flags and VSX.
8484 2013-07-18 Sriraman Tallam <tmsriram@google.com>
8487 * tree-inline.c (expand_call_inline): Emit errors during
8488 early_inlining only if optimization is not turned on.
8490 2013-07-18 David Malcolm <dmalcolm@redhat.com>
8494 * passes.c (init_optimization_passes): Move the construction of
8495 the pass hierarchy into a new passes.def file.
8497 * Makefile.in (passes.o): Add dependency on passes.def.
8499 2013-07-18 David Malcolm <dmalcolm@redhat.com>
8501 * passes.c (init_optimization_passes): Introduce macros for
8502 constructing the tree of passes (INSERT_PASSES_AFTER,
8503 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
8504 TERMINATE_PASS_LIST).
8506 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
8507 Wei Mi <wmi@google.com>
8509 PR rtl-optimization/57878
8510 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
8512 (reload_pseudo_compare_func): Check nregs first for reload
8515 2013-07-18 David Malcolm <dmalcolm@redhat.com>
8517 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
8519 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
8521 * read-rtl.c (validate_const_int): Once an invalid character is
8522 seen, quit the loop.
8524 * gengtype.c (write_roots): Similarly once we find the "deletable"
8525 or "if_marked" option.
8527 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
8529 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
8530 "xtab" and "sat". Move value "clz" from here to ...
8531 (attriubte "type"): ... here.
8532 (satsi_<SAT:code>): Delete "insn" attribute.
8533 (satsi_<SAT:code>_shift): Likewise.
8534 (arm_zero_extendqisi2addsi): Likewise.
8535 (arm_extendqisi2addsi): Likewise.
8536 (clzsi2): Update for attribute changes.
8537 (rbitsi2): Likewise.
8538 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
8540 (arm_usatsihi): Likewise.
8541 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
8543 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
8545 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
8546 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
8547 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
8548 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
8549 in alphabetical order.
8550 (attribute "core_cycles"): Update for attribute changes.
8551 (arm_addsi3): Likewise.
8552 (addsi3_compare0): Likewise.
8553 (addsi3_compare0_scratch): Likewise.
8554 (addsi3_compare_op1): Likewise.
8555 (addsi3_compare_op2): Likewise.
8556 (compare_addsi2_op0): Likewise.
8557 (compare_addsi2_op1): Likewise.
8558 (addsi3_carryin_shift_<optab>): Likewise.
8559 (subsi3_carryin_shift): Likewise.
8560 (rsbsi3_carryin_shift): Likewise.
8561 (arm_subsi3_insn): Likewise.
8562 (subsi3_compare0): Likewise.
8563 (subsi3_compare): Likewise.
8564 (arm_andsi3_insn): Likewise.
8565 (thumb1_andsi3_insn): Likewise.
8566 (andsi3_compare0): Likewise.
8567 (andsi3_compare0_scratch): Likewise.
8568 (zeroextractsi_compare0_scratch
8569 (andsi_not_shiftsi_si): Likewise.
8570 (iorsi3_insn): Likewise.
8571 (iorsi3_compare0): Likewise.
8572 (iorsi3_compare0_scratch): Likewise.
8573 (arm_xorsi3): Likewise.
8574 (thumb1_xorsi3_insn): Likewise.
8575 (xorsi3_compare0): Likewise.
8576 (xorsi3_compare0_scratch): Likewise.
8577 (satsi_<SAT:code>_shift): Likewise.
8579 (arm_shiftsi3): Likewise.
8580 (shiftsi3_compare0): Likewise.
8581 (not_shiftsi): Likewise.
8582 (not_shiftsi_compare0): Likewise.
8583 (not_shiftsi_compare0_scratch): Likewise.
8584 (arm_one_cmplsi2): Likewise.
8585 (thumb_one_complsi2): Likewise.
8586 (notsi_compare0): Likewise.
8587 (notsi_compare0_scratch): Likewise.
8588 (thumb1_zero_extendhisi2): Likewise.
8589 (arm_zero_extendhisi2): Likewise.
8590 (arm_zero_extendhisi2_v6): Likewise.
8591 (arm_zero_extendhisi2addsi): Likewise.
8592 (thumb1_zero_extendqisi2): Likewise.
8593 (thumb1_zero_extendqisi2_v6): Likewise.
8594 (arm_zero_extendqisi2): Likewise.
8595 (arm_zero_extendqisi2_v6): Likewise.
8596 (arm_zero_extendqisi2addsi): Likewise.
8597 (thumb1_extendhisi2): Likewise.
8598 (arm_extendhisi2): Likewise.
8599 (arm_extendhisi2_v6): Likewise.
8600 (arm_extendqisi): Likewise.
8601 (arm_extendqisi_v6): Likewise.
8602 (arm_extendqisi2addsi): Likewise.
8603 (thumb1_extendqisi2): Likewise.
8604 (thumb1_movdi_insn): Likewise.
8605 (arm_movsi_insn): Likewise.
8606 (movsi_compare0): Likewise.
8607 (movhi_insn_arch4): Likewise.
8608 (movhi_bytes): Likewise.
8609 (arm_movqi_insn): Likewise.
8610 (thumb1_movqi_insn): Likewise.
8611 (arm32_movhf): Likewise.
8612 (thumb1_movhf): Likewise.
8613 (arm_movsf_soft_insn): Likewise.
8614 (thumb1_movsf_insn): Likewise.
8615 (movdf_soft_insn): Likewise.
8616 (thumb_movdf_insn): Likewise.
8617 (arm_cmpsi_insn): Likewise.
8618 (cmpsi_shiftsi): Likewise.
8619 (cmpsi_shiftsi_swp): Likewise.
8620 (arm_cmpsi_negshiftsi_si): Likewise.
8621 (movsicc_insn): Likewise.
8622 (movsfcc_soft_insn): Likewise.
8623 (arith_shiftsi): Likewise.
8624 (arith_shiftsi_compare0
8625 (arith_shiftsi_compare0_scratch
8626 (sub_shiftsi): Likewise.
8627 (sub_shiftsi_compare0
8628 (sub_shiftsi_compare0_scratch
8629 (and_scc): Likewise.
8630 (cond_move): Likewise.
8631 (if_plus_move): Likewise.
8632 (if_move_plus): Likewise.
8633 (if_move_not): Likewise.
8634 (if_not_move): Likewise.
8635 (if_shift_move): Likewise.
8636 (if_move_shift): Likewise.
8637 (if_shift_shift): Likewise.
8638 (if_not_arith): Likewise.
8639 (if_arith_not): Likewise.
8640 (cond_move_not): Likewise.
8641 (thumb1_ashlsi3): Set type attribute.
8642 (thumb1_ashrsi3): Likewise.
8643 (thumb1_lshrsi3): Likewise.
8644 (thumb1_rotrsi3): Likewise.
8645 (shiftsi3_compare0_scratch): Likewise.
8646 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
8647 (neon_mov<mode>): Likewise.
8648 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
8650 (thumb2_movsi_insn): Likewise.
8651 (thumb2_cmpsi_neg_shiftsi): Likewise.
8652 (thumb2_extendqisi_v6): Likewise.
8653 (thumb2_zero_extendhisi2_v6): Likewise.
8654 (thumb2_zero_extendqisi2_v6): Likewise.
8655 (thumb2_shiftsi3_short): Likewise.
8656 (thumb2_addsi3_compare0_scratch): Likewise.
8657 (orsi_not_shiftsi_si): Likewise.
8658 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
8659 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
8661 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
8662 (1020alu_shift_op): Likewise.
8663 (1020alu_shift_reg_op): Likewise.
8664 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
8665 (alu_shift_op): Likewise.
8666 (alu_shift_reg_op): Likewise.
8667 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
8668 (11_alu_shift_op): Likewise.
8669 (11_alu_shift_reg_op): Likewise.
8670 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
8671 (9_alu_shift_reg_op): Likewise.
8672 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
8674 (cortex_a15_alu_shift): Likewise.
8675 (cortex_a15_alu_shift_reg): Likewise.
8676 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
8678 (cortex_a5_alu_shift): Likewise.
8679 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
8681 (cortex_a53_alu_shift): Likewise.
8682 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
8684 (cortex_a7_alu_reg): Likewise.
8685 (cortex_a7_alu_shift): Likewise.
8686 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
8688 (cortex_a8_alu_shift): Likewise.
8689 (cortex_a8_alu_shift_reg): Likewise.
8690 (cortex_a8_mov): Likewise.
8691 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
8692 (cortex_a9_dp_shift): Likewise.
8693 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
8695 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
8697 (cortex_r4_mov): Likewise.
8698 (cortex_r4_alu_shift): Likewise.
8699 (cortex_r4_alu_shift_reg): Likewise.
8700 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
8701 (526_alu_shift_op): Likewise.
8702 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
8703 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
8704 (626te_alu_shift_op): Likewise.
8705 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
8706 (726te_alu_op): Likewise.
8707 (726te_alu_shift_op): Likewise.
8708 (726te_alu_shift_reg_op): Likewise.
8709 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
8710 (mp626_alu_shift_op): Likewise.
8711 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
8712 (pj4_alu_e1_conds): Likewise.
8713 (pj4_alu): Likewise.
8714 (pj4_alu_conds): Likewise.
8715 (pj4_shift): Likewise.
8716 (pj4_shift_conds): Likewise.
8717 (pj4_alu_shift): Likewise.
8718 (pj4_alu_shift_conds): Likewise.
8719 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
8721 (cortexa7_older_only): Likewise.
8722 (cortexa7_younger): Likewise.
8724 2013-07-18 David Malcolm <dmalcolm@redhat.com>
8726 * ipa-pure-const.c (generate_summary): Rename to...
8727 (pure_const_generate_summary): ... this.
8729 2013-07-17 Iain Sandoe <iain@codesourcery.com>
8731 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
8733 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
8736 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
8739 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8741 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
8742 enabled without -march=zEC12.
8743 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
8746 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
8748 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
8749 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
8750 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
8751 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
8752 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
8753 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
8754 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
8756 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
8757 and ISA_HAS_NMADD3_NMSUB3.
8758 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
8759 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
8760 (nmsub4<mode>, nmsub3<mode>): Likewise.
8761 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
8763 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
8765 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
8766 TARGET_MIPS5400 checking.
8768 2013-07-16 Jakub Jelinek <jakub@redhat.com>
8769 Peter Bergner <bergner@vnet.ibm.com>
8771 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
8772 registers in the comment.
8773 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
8774 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
8775 rather than FIRST_PSEUDO_REGISTERS.
8777 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
8779 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
8780 enable extra ISA flags with TARGET_HTM.
8782 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
8784 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
8787 2013-07-15 Cong Hou <congh@google.com>
8789 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
8790 in compare function for sorting.
8792 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
8794 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
8795 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
8796 * config/rs6000/rs6000.opt: Add -mhtm option.
8797 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
8798 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
8799 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
8800 __HTM__ if the HTM instructions are available.
8801 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
8802 htm_spr_reg_operand): New define_predicates.
8803 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
8804 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
8806 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
8807 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
8808 HTM builtin functions.
8809 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
8810 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
8811 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
8812 (rs6000_builtin_mask_calculate): Likewise.
8813 (rs6000_option_override_internal): Likewise.
8814 (bdesc_htm): Add new HTM builtin support.
8815 (htm_spr_num): New function.
8816 (htm_spr_regno): Likewise.
8817 (rs6000_htm_spr_icode): Likewise.
8818 (htm_expand_builtin): Likewise.
8819 (htm_init_builtins): Likewise.
8820 (rs6000_expand_builtin): Add support for HTM builtin functions.
8821 (rs6000_init_builtins): Likewise.
8822 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
8824 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
8825 (TARGET_HTM, MASK_HTM): Define macros.
8826 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
8827 (FIXED_REGISTERS): Likewise.
8828 (CALL_USED_REGISTERS): Likewise.
8829 (CALL_REALLY_USED_REGISTERS): Likewise.
8830 (REG_ALLOC_ORDER): Likewise.
8831 (enum reg_class): Likewise.
8832 (REG_CLASS_NAMES): Likewise.
8833 (REG_CLASS_CONTENTS): Likewise.
8834 (REGISTER_NAMES): Likewise.
8835 (ADDITIONAL_REGISTER_NAMES): Likewise.
8836 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
8837 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
8838 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
8839 * config/rs6000/htm.md: New file.
8840 * config/rs6000/htmintrin.h: New file.
8841 * config/rs6000/htmxlintrin.h: New file.
8843 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
8845 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
8846 Define SYMBOL_TINY_GOT, update comment.
8847 * config/aarch64/aarch64.c
8848 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
8849 (aarch64_expand_mov_immediate): Likewise.
8850 (aarch64_print_operand): Likewise.
8851 (aarch64_classify_symbol): Likewise.
8852 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
8853 (ldr_got_tiny): Define.
8855 2013-07-13 Tobias Grosser <tobias@grosser.es>
8857 PR tree-optimization/54094
8858 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
8859 scheduling dimension for the parallelism check from the polyhedral
8860 information in the AST.
8861 * graphite-dependences.c (carries_deps): Do not assume the schedule is
8864 2013-07-13 Jason Merrill <jason@redhat.com>
8866 * print-tree.c (debug_vec_tree): Use debug_raw.
8867 (debug_raw (vec<tree, va_gc> &)): New.
8868 (debug_raw (vec<tree, va_gc> *)): New.
8869 * tree.h: Declare them.
8871 2013-07-13 Bin Cheng <bin.cheng@arm.com>
8873 * ifcvt.c (ifcvt_after_combine): New static variable.
8874 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
8876 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
8877 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
8878 rest_of_handle_if_after_reload): Pass new argument for if_convert.
8880 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
8882 * config/mips/mips.c (mips_expand_call): Remove empty statement.
8884 2013-07-12 Michael Matz <matz@suse.de>
8887 * convert.c (convert_to_real): Reject non-float inner types.
8889 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
8891 * config/aarch64/aarch64-protos.h
8892 (aarch64_simd_immediate_valid_for_move): Remove.
8893 * config/aarch64/aarch64.c (simd_immediate_info): New member.
8894 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
8896 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
8898 2013-07-11 Steve Ellcey <sellcey@mips.com>
8900 * config/mips/mips.c (mips_conditional_register_usage): Do not
8901 use t[0-7] registers in MIPS16 mode when optimizing for size.
8903 2013-07-11 Sriraman Tallam <tmsriram@google.com>
8905 * config/i386/i386.c (dispatch_function_versions): Fix array
8906 indexing of function_version_info to match actual_versions.
8908 2013-07-11 Teresa Johnson <tejohnson@google.com>
8910 * vec.h (struct va_gc): Move release out-of-line.
8911 (va_gc::release): Call ggc_free on released vec.
8913 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
8915 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
8916 Require GOT register as additional operand in UNSPEC.
8917 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
8918 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
8919 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
8920 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
8921 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
8922 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
8923 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
8925 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
8928 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
8929 name seen by assembler/linker rather if available.
8931 2013-07-11 Andreas Schwab <schwab@suse.de>
8933 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
8935 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
8937 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
8939 2013-07-10 Joseph Myers <joseph@codesourcery.com>
8941 * doc/tm.texi.in: Move hook documentation to ....
8942 * target.def: ... here.
8944 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
8946 * doc/tm.texi: Regenerate.
8948 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
8951 * doc/invoke.texi: Document -Wconditionally-supported.
8953 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
8956 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
8959 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
8962 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
8963 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
8964 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
8965 Remove duplicate devices.
8966 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
8967 * config/avr/t-multilib: Regenerate.
8968 * config/avr/avr-tables.opt: Regenerate.
8969 * doc/avr-mmcu.texi: Regenerate.
8971 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
8974 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
8976 2013-07-10 Graham Stott <graham.stott@btinternet.com>
8978 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
8979 the cost of MULT when optimizing for size.
8981 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8983 * config/cr16/cr16-protos.h: Don't include target.h.
8985 2013-07-09 Joseph Myers <joseph@codesourcery.com>
8987 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
8988 adjust register size for TDmode and TFmode for VSX registers.
8990 2013-07-08 Kai Tietz <ktietz@redhat.com>
8993 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
8994 hook_bool_const_tree_true.
8996 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8998 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
8999 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
9000 * config/s390/s390.md: Define FPR*_REGNUM constants.
9001 Fix FPR2_REGNUM constant (18 -> 17).
9002 ("*trunc<BFP:mode><DFP_ALL:mode>2")
9003 ("*trunc<DFP_ALL:mode><BFP:mode>2")
9004 ("trunc<BFP:mode><DFP_ALL:mode>2")
9005 ("trunc<DFP_ALL:mode><BFP:mode>2")
9006 ("*extend<BFP:mode><DFP_ALL:mode>2")
9007 ("*extend<DFP_ALL:mode><BFP:mode>2")
9008 ("extend<BFP:mode><DFP_ALL:mode>2")
9009 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
9012 2013-07-08 Graham Stott <graham.stott@btinternet.com>
9014 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
9016 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9018 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
9019 and cfun_fpr_bit_p to cfun_fpr_save_p.
9020 (s390_frame_area, s390_register_info, s390_frame_info)
9021 (s390_emit_prologue, s390_emit_epilogue)
9022 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
9024 * config/s390/s390.h: Define *_REGNUM macros for floating point
9027 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
9029 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
9031 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
9033 PR rtl-optimization/57786
9034 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
9035 and break out of the loop when it is set to false.
9037 2013-07-08 Jakub Jelinek <jakub@redhat.com>
9040 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
9041 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
9042 (const_int 63)) 0)).
9043 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
9044 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
9045 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
9047 PR rtl-optimization/57829
9048 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
9049 mask bits outside of mode are just sign-extension from mode to HWI.
9051 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
9053 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
9054 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
9055 adjust_address instead of change_address to keep info about alignment.
9056 (emit_strmov): Remove.
9057 (emit_memmov): New function.
9058 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
9059 (expand_movmem_epilogue): Likewise and return updated rtx for
9061 (expand_constant_movmem_prologue): Likewise and return updated rtx for
9062 destination and source.
9063 (decide_alignment): Refactor, handle vector_loop.
9064 (ix86_expand_movmem): Likewise.
9065 (ix86_expand_setmem): Likewise.
9066 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
9068 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
9070 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
9071 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
9073 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
9075 * config/i386/sse.md (sse_movlhps): Change alternative 3
9076 of operand 2 to "m".
9078 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
9081 * config/i386/sse.md (iptr): New mode attribute.
9082 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
9083 (<sse>_vm<plusminus_insn><mode>3): Ditto.
9084 (<sse>_vmmul<mode>3): Ditto.
9085 (<sse>_vmdiv<mode>3): Ditto.
9086 (sse_vmrcpv4sf2): Ditto.
9087 (<sse>_vmsqrt<mode>2): Ditto.
9088 (sse_vmrsqrtv4sf2): Ditto.
9089 (<sse>_vm<code><mode>3): Ditto.
9090 (avx_vmcmp<mode>3): Ditto.
9091 (<sse>_vmmaskcmp<mode>3): Ditto.
9092 (<sse>_comi): Ditto.
9093 (<sse>_ucomi): Ditto.
9094 (*xop_vmfrcz_<mode>): Ditto.
9095 (*fmai_fmadd_<mode>): Ditto.
9096 (*fmai_fmsub_<mode>): Ditto.
9097 (*fmai_fnmadd_<mode>): Ditto.
9098 (*fmai_fnmsub_<mode>): Ditto.
9099 (*fma4i_vmfmadd_<mode>): Ditto.
9100 (*fma4i_vmfmsub_<mode>): Ditto.
9101 (*fma4i_vmfnmadd_<mode>): Ditto.
9102 (*fma4i_vmfnmsub_<mode>): Ditto.
9103 (*xop_vmfrcz_<mode>): Ditto.
9104 (sse_cvtps2pi): Ditto.
9105 (sse_cvttps2pi): Ditto.
9106 (sse_cvtss2si): Ditto.
9107 (sse_cvtss2si_2): Ditto.
9108 (sse_cvtss2siq_2): Ditto.
9109 (sse_cvttss2si): Ditto.
9110 (sse_cvttss2siq): Ditto.
9111 (sse_cvtsd2si): Ditto.
9112 (sse_cvtsd2si_2): Ditto.
9113 (sse_cvtsd2siq_2): Ditto.
9114 (sse_cvttsd2si): Ditto.
9115 (sse_cvttsd2siq): Ditto.
9116 (sse_cvtsd2ss): Ditto.
9117 (sse_cvtss2sd): Ditto.
9118 (avx2_pbroadcast<mode>): Ditto.
9119 (avx2_pbroadcast<mode>_1): Ditto.
9120 (*avx_vperm_broadcast_v4sf): Ditto.
9122 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
9123 (sse_movlhps): Ditto.
9124 (sse_storehps): Ditto.
9125 (sse_loadhps): Ditto.
9126 (sse_storelps): Ditto.
9127 (sse_loadlps): Ditto.
9128 (*vec_concatv4sf): Ditto.
9129 (*vec_interleave_highv2df): Ditto.
9130 (*vec_interleave_lowv2df): Ditto.
9131 (*vec_extractv2df_1_sse): Ditto.
9132 (*vec_extractv2df_0_sse): Ditto.
9133 (sse2_storelpd): Ditto.
9134 (sse2_loadlpd): Ditto.
9135 (sse2_movsd): Ditto.
9136 (*vec_concatv4si): Ditto.
9137 (vec_concatv2di): Ditto.
9139 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
9140 for Intel asm dialect.
9141 (mmx_punpcklwd): Ditto.
9142 (mmx_punpckldq): Ditto.
9144 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
9145 for intel assembler dialect.
9147 2013-07-06 Jakub Jelinek <jakub@redhat.com>
9150 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
9151 for BUILT_IN_C{LZ,LRSB}*.
9152 * tree.h (CASE_INT_FN): Add FN##IMAX case.
9153 * tree-vrp.c (extract_range_basic): Handle
9154 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
9155 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
9156 fall thru to code calling set_value*.
9157 * builtins.c (expand_builtin): Remove *IMAX cases.
9158 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
9159 if width is bigger than 2*HWI.
9161 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
9163 PR rtl-optimization/55342
9164 * lra-int.h (lra_subreg_reload_pseudos): New.
9165 * lra.c: Add undoing optional reloads to the block diagram.
9166 (lra_subreg_reload_pseudos): New.
9167 (lra_optional_reload_pseudos): Change comments.
9168 (lra): Init and clear lra_subreg_reload_pseudos. Clear
9169 lra_optional_reload_pseudos after undo transformations.
9170 * lra-assigns.c (pseudo_prefix_title): New.
9171 (lra_setup_reg_renumber): Use it.
9172 (spill_for): Ditto. Check subreg reload pseudos too.
9173 (assign_by_spills): Consider subreg reload pseudos too.
9174 * lra-constraints.c (simplify_operand_subreg): Use
9175 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
9176 (curr_insn_transform): Recognize and do optional reloads.
9177 (undo_optional_reloads): New.
9178 (lra_undo_inheritance): Call undo_optional_reloads.
9180 2013-07-05 Thomas Quinot <quinot@adacore.com>
9182 * tree-complex.c (expand_complex_operations_1): Fix typo.
9184 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
9186 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
9187 (tune_params): New member 'const vec_costs'.
9188 * config/aarch64/aarch64.c (generic_vector_cost): New.
9189 (generic_tunings): New member 'generic_vector_cost'.
9190 (aarch64_builtin_vectorization_cost): New.
9191 (aarch64_add_stmt_cost): New.
9192 (TARGET_VECTORIZE_ADD_STMT_COST): New.
9193 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
9195 2013-07-03 Jakub Jelinek <jakub@redhat.com>
9198 * config/i386/predicates.md (vsib_address_operand): Disallow
9199 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
9201 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
9204 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
9205 expand_builtin_setjmp_receiver.
9206 (expand_label): Adjust, call expand_builtin_setjmp_receiver
9207 with NULL for the label parameter.
9208 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
9209 the frame-pointer. Adjust comments.
9210 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
9211 only if LABEL is non-NULL.
9213 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
9215 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
9216 (ARM_ABI_AAPCS64): Ditto.
9218 (ARM_DEFAULT_ABI): Ditto.
9220 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
9222 * config/aarch64/aarch64-builtins.c
9223 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
9224 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
9226 * config/aarch64/aarch64-simd.md
9227 (aarch64_ld1<VALL:mode>): New.
9228 (aarch64_st1<VALL:mode>): Likewise.
9229 * config/aarch64/arm_neon.h
9230 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
9232 2013-07-02 Sriraman Tallam <tmsriram@google.com>
9234 * config/i386/i386.c (gate_insert_vzeroupper): Check if
9236 (ix86_option_override_internal):Turn on all -mavx target flags by
9237 default as they are dependent on AVX anyway.
9239 2013-07-02 Cary Coutant <ccoutant@google.com>
9241 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
9243 (loc_checksum_ordered): Likewise.
9244 (hash_loc_operands): Remove inline keyword.
9246 2013-07-02 Jakub Jelinek <jakub@redhat.com>
9248 PR tree-optimization/57741
9249 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
9250 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
9251 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
9252 Allow REAL_CST step_exprs if flag_associative_math.
9253 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
9255 2013-07-02 Ian Bolton <ian.bolton@arm.com>
9257 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
9259 2013-07-02 Ian Bolton <ian.bolton@arm.com>
9261 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
9263 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9265 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
9267 (iorsi3_insn): Likewise.
9268 (arm_xorsi3): Likewise.
9270 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
9272 * arm.md (attribute "wtype"): Delete. Move attribute values from here
9274 (attribute "type"): ... here, and prefix with "wmmx_".
9275 (attribute "core_cycles"): Update for attribute changes.
9276 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
9277 (tbcstv4hi): Likewise.
9278 (tbcstv2si): Likewise.
9279 (iwmmxt_iordi3): Likewise.
9280 (iwmmxt_xordi3): Likewise.
9281 (iwmmxt_anddi3): Likewise.
9282 (iwmmxt_nanddi3): Likewise.
9283 (iwmmxt_arm_movdi): Likewise.
9284 (iwmmxt_movsi_insn): Likewise.
9285 (mov<mode>_internal): Likewise.
9286 (and<mode>3_iwmmxt): Likewise.
9287 (ior<mode>3_iwmmxt): Likewise.
9288 (xor<mode>3_iwmmxt): Likewise.
9289 (add<mode>3_iwmmxt): Likewise.
9290 (ssaddv8qi3): Likewise.
9291 (ssaddv4hi3): Likewise.
9292 (ssaddv2si3): Likewise.
9293 (usaddv8qi3): Likewise.
9294 (usaddv4hi3): Likewise.
9295 (usaddv2si3): Likewise.
9296 (sub<mode>3_iwmmxt): Likewise.
9297 (sssubv8qi3): Likewise.
9298 (sssubv4hi3): Likewise.
9299 (sssubv2si3): Likewise.
9300 (ussubv8qi3): Likewise.
9301 (ussubv4hi3): Likewise.
9302 (ussubv2si3): Likewise.
9303 (mulv4hi3_iwmmxt): Likewise.
9304 (smulv4hi3_highpart): Likewise.
9305 (umulv4hi3_highpart): Likewise.
9306 (iwmmxt_wmacs): Likewise.
9307 (iwmmxt_wmacsz): Likewise.
9308 (iwmmxt_wmacu): Likewise.
9309 (iwmmxt_wmacuz): Likewise.
9310 (iwmmxt_clrdi): Likewise.
9311 (iwmmxt_clrv8qi): Likewise.
9312 (iwmmxt_clr4hi): Likewise.
9313 (iwmmxt_clr2si): Likewise.
9314 (iwmmxt_uavgrndv8qi3): Likewise.
9315 (iwmmxt_uavgrndv4hi3): Likewise.
9316 (iwmmxt_uavgv8qi3): Likewise.
9317 (iwmmxt_uavgv4hi3): Likewise.
9318 (iwmmxt_tinsrb): Likewise.
9319 (iwmmxt_tinsrh): Likewise.
9320 (iwmmxt_tinsrw): Likewise.
9321 (iwmmxt_textrmub): Likewise.
9322 (iwmmxt_textrmsb): Likewise.
9323 (iwmmxt_textrmuh): Likewise.
9324 (iwmmxt_textrmsh): Likewise.
9325 (iwmmxt_textrmw): Likewise.
9326 (iwmxxt_wshufh): Likewise.
9327 (eqv8qi3): Likewise.
9328 (eqv4hi3): Likewise.
9329 (eqv2si3): Likewise.
9330 (gtuv8qi3): Likewise.
9331 (gtuv4hi3): Likewise.
9332 (gtuv2si3): Likewise.
9333 (gtv8qi3): Likewise.
9334 (gtv4hi3): Likewise.
9335 (gtv2si3): Likewise.
9336 (smax<mode>3_iwmmxt): Likewise.
9337 (umax<mode>3_iwmmxt): Likewise.
9338 (smin<mode>3_iwmmxt): Likewise.
9339 (umin<mode>3_iwmmxt): Likewise.
9340 (iwmmxt_wpackhss): Likewise.
9341 (iwmmxt_wpackwss): Likewise.
9342 (iwmmxt_wpackdss): Likewise.
9343 (iwmmxt_wpackhus): Likewise.
9344 (iwmmxt_wpackwus): Likewise.
9345 (iwmmxt_wpackdus): Likewise.
9346 (iwmmxt_wunpckihb): Likewise.
9347 (iwmmxt_wunpckihh): Likewise.
9348 (iwmmxt_wunpckihw): Likewise.
9349 (iwmmxt_wunpckilb): Likewise.
9350 (iwmmxt_wunpckilh): Likewise.
9351 (iwmmxt_wunpckilw): Likewise.
9352 (iwmmxt_wunpckehub): Likewise.
9353 (iwmmxt_wunpckehuh): Likewise.
9354 (iwmmxt_wunpckehuw): Likewise.
9355 (iwmmxt_wunpckehsb): Likewise.
9356 (iwmmxt_wunpckehsh): Likewise.
9357 (iwmmxt_wunpckehsw): Likewise.
9358 (iwmmxt_wunpckelub): Likewise.
9359 (iwmmxt_wunpckeluh): Likewise.
9360 (iwmmxt_wunpckeluw): Likewise.
9361 (iwmmxt_wunpckelsb): Likewise.
9362 (iwmmxt_wunpckelsh): Likewise.
9363 (iwmmxt_wunpckelsw): Likewise.
9364 (ror<mode>3): Likewise.
9365 (ashr<mode>3_iwmmxt): Likewise.
9366 (lshr<mode>3_iwmmxt): Likewise.
9367 (ashl<mode>3_iwmmxt): Likewise.
9368 (ror<mode>3_di): Likewise.
9369 (ashr<mode>3_di): Likewise.
9370 (lshr<mode>3_di): Likewise.
9371 (ashl<mode>3_di): Likewise.
9372 (iwmmxt_wmadds): Likewise.
9373 (iwmmxt_wmaddu): Likewise.
9374 (iwmmxt_tmia): Likewise.
9375 (iwmmxt_tmiaph): Likewise.
9376 (iwmmxt_tmiabb): Likewise.
9377 (iwmmxt_tmiatb): Likewise.
9378 (iwmmxt_tmiabt): Likewise.
9379 (iwmmxt_tmiatt): Likewise.
9380 (iwmmxt_tmovmskb): Likewise.
9381 (iwmmxt_tmovmskh): Likewise.
9382 (iwmmxt_tmovmskw): Likewise.
9383 (iwmmxt_waccb): Likewise.
9384 (iwmmxt_wacch): Likewise.
9385 (iwmmxt_waccw): Likewise.
9386 (iwmmxt_waligni): Likewise.
9387 (iwmmxt_walignr): Likewise.
9388 (iwmmxt_walignr0): Likewise.
9389 (iwmmxt_walignr1): Likewise.
9390 (iwmmxt_walignr2): Likewise.
9391 (iwmmxt_walignr3): Likewise.
9392 (iwmmxt_wsadb): Likewise.
9393 (iwmmxt_wsadh): Likewise.
9394 (iwmmxt_wsadbz): Likewise.
9395 (iwmmxt_wsadhz): Likewise.
9396 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
9397 (iwmmxt_wabsdiffb): Likewise.
9398 (iwmmxt_wabsdiffh): Likewise.
9399 (iwmmxt_wabsdiffw): Likewise.
9400 (iwmmxt_waddsubhx): Likewise
9401 (iwmmxt_wsubaddhx): Likewise.
9402 (addc<mode>3): Likewise.
9403 (iwmmxt_avg4): Likewise.
9404 (iwmmxt_avg4r): Likewise.
9405 (iwmmxt_wmaddsx): Likewise.
9406 (iwmmxt_wmaddux): Likewise.
9407 (iwmmxt_wmaddsn): Likewise.
9408 (iwmmxt_wmaddun): Likewise.
9409 (iwmmxt_wmulwsm): Likewise.
9410 (iwmmxt_wmulwum): Likewise.
9411 (iwmmxt_wmulsmr): Likewise.
9412 (iwmmxt_wmulumr): Likewise.
9413 (iwmmxt_wmulwsmr): Likewise.
9414 (iwmmxt_wmulwumr): Likewise.
9415 (iwmmxt_wmulwl): Likewise.
9416 (iwmmxt_wqmulm): Likewise.
9417 (iwmmxt_wqmulwm): Likewise.
9418 (iwmmxt_wqmulmr): Likewise.
9419 (iwmmxt_wqmulwmr): Likewise.
9420 (iwmmxt_waddbhusm): Likewise.
9421 (iwmmxt_waddbhusl): Likewise.
9422 (iwmmxt_wqmiabb): Likewise.
9423 (iwmmxt_wqmiabt): Likewise.
9424 (iwmmxt_wqmiatb): Likewise.
9425 (iwmmxt_wqmiatt): Likewise.
9426 (iwmmxt_wqmiabbn): Likewise.
9427 (iwmmxt_wqmiabtn): Likewise.
9428 (iwmmxt_wqmiatbn): Likewise.
9429 (iwmmxt_wqmiattn): Likewise.
9430 (iwmmxt_wmiabb): Likewise.
9431 (iwmmxt_wmiabt): Likewise.
9432 (iwmmxt_wmiatb): Likewise.
9433 (iwmmxt_wmiatt): Likewise.
9434 (iwmmxt_wmiabbn): Likewise.
9435 (iwmmxt_wmiabtn): Likewise.
9436 (iwmmxt_wmiatbn): Likewise.
9437 (iwmmxt_wmiattn): Likewise.
9438 (iwmmxt_wmiawbb): Likewise.
9439 (iwmmxt_wmiawbt): Likewise.
9440 (iwmmxt_wmiawtb): Likewise.
9441 (iwmmxt_wmiawtt): Likewise.
9442 (iwmmxt_wmiawbbn): Likewise.
9443 (iwmmxt_wmiawbtn): Likewise.
9444 (iwmmxt_wmiawtbn): Likewise.
9445 (iwmmxt_wmiawttn): Likewise.
9446 (iwmmxt_wmerge): Likewise.
9447 (iwmmxt_tandc<mode>3): Likewise.
9448 (iwmmxt_torc<mode>3): Likewise.
9449 (iwmmxt_torvsc<mode>3): Likewise.
9450 (iwmmxt_textrc<mode>3): Likewise.
9451 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
9452 (wmmxt_pack): Likewise.
9453 (wmmxt_mult_c1): Likewise.
9454 (wmmxt_mult_c2): Likewise.
9455 (wmmxt_alu_c1): Likewise.
9456 (wmmxt_alu_c2): Likewise.
9457 (wmmxt_alu_c3): Likewise.
9458 (wmmxt_transfer_c1): Likewise.
9459 (wmmxt_transfer_c2): Likewise.
9460 (wmmxt_transfer_c3): Likewise.
9461 (marvell_f_iwmmxt_wstr): Likewise.
9462 (marvell_f_iwmmxt_wldr): Likewise.
9464 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
9466 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
9468 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
9471 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
9472 * lra-constraints.c (need_for_split_p): Check call used hard regs
9473 living through calls.
9475 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
9476 call used regs for call insn.
9478 2013-06-28 Jakub Jelinek <jakub@redhat.com>
9481 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
9482 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
9484 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
9486 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
9488 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
9490 * lra-constraints.c (need_for_split_p): Check call used hard regs
9491 living through calls.
9493 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
9496 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
9497 to tie with any other modes. Eliminate Altivec vector mode tests,
9498 since these are a subset of ALTIVEC or VSX vector modes. Simplify
9499 code, to return 0 if testing MODE2 for a condition, if we've
9500 already tested MODE1 for the same condition.
9502 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
9504 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
9507 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
9509 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
9510 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
9512 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
9514 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
9516 (aarch64_symbolic_constant_p): Remove.
9517 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
9518 static. Fix line length and white space.
9519 (aarch64_symbolic_constant_p): Remove.
9520 * config/aarch64/predicates.md (aarch64_valid_symref):
9521 Use aarch64_classify_symbol_expression.
9523 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9525 * config/arm/constraints.md (Ts): New constraint.
9526 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
9528 (compare_scc): Use "Ts" constraint for operand 0.
9529 (ior_scc_scc): Likewise.
9530 (and_scc_scc): Likewise.
9531 (and_scc_scc_nodom): Likewise.
9532 (ior_scc_scc_cmp): Likewise for operand 7.
9533 (and_scc_scc_cmp): Likewise.
9534 * config/arm/thumb2.md (thumb2_movsi_insn):
9535 Add alternatives for 16-bit encodings.
9536 (thumb2_movhi_insn): Likewise.
9537 (thumb2_movsicc_insn): Likewise.
9538 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
9539 (thumb2_negscc): Use "Ts" constraint.
9540 Move mvn instruction outside cond_exec block.
9541 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
9542 for 16-bit encodings.
9544 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9546 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
9548 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
9549 (mulsi3subsi): Likewise.
9550 (mulsidi3adddi): Likewise.
9551 (mulsidi3_v6): Likewise.
9552 (umulsidi3_v6): Likewise.
9553 (umulsidi3adddi_v6): Likewise.
9554 (smulsi3_highpart_v6): Likewise.
9555 (umulsi3_highpart_v6): Likewise.
9556 (mulhisi3tb): Likewise.
9557 (mulhisi3bt): Likewise.
9558 (mulhisi3tt): Likewise.
9559 (maddhisi4): Likewise.
9560 (maddhisi4tb): Likewise.
9561 (maddhisi4tt): Likewise.
9562 (maddhidi4): Likewise.
9563 (maddhidi4tb): Likewise.
9564 (maddhidi4tt): Likewise.
9565 (zeroextractsi_compare0_scratch): Likewise.
9566 (insv_zero): Likewise.
9567 (insv_t2): Likewise.
9568 (anddi_notzesidi_di): Likewise.
9569 (anddi_notsesidi_di): Likewise.
9570 (andsi_notsi_si): Likewise.
9571 (iordi_zesidi_di): Likewise.
9572 (xordi_zesidi_di): Likewise.
9573 (andsi_iorsi3_notsi): Likewise.
9575 (smax_m1): Likewise.
9577 (not_shiftsi): Likewise.
9578 (unaligned_loadsi): Likewise.
9579 (unaligned_loadhis): Likewise.
9580 (unaligned_loadhiu): Likewise.
9581 (unaligned_storesi): Likewise.
9582 (unaligned_storehi): Likewise.
9583 (extv_reg): Likewise.
9584 (extzv_t2): Likewise.
9586 (udivsi3): Likewise.
9587 (arm_zero_extendhisi2addsi): Likewise.
9588 (arm_zero_extendqisi2addsi): Likewise.
9589 (compareqi_eq0): Likewise.
9590 (arm_extendhisi2_v6): Likewise.
9591 (arm_extendqisi2addsi): Likewise.
9592 (arm_movt): Likewise.
9593 (thumb2_ldrd): Likewise.
9594 (thumb2_ldrd_base): Likewise.
9595 (thumb2_ldrd_base_neg): Likewise.
9596 (thumb2_strd): Likewise.
9597 (thumb2_strd_base): Likewise.
9598 (thumb2_strd_base_neg): Likewise.
9599 (arm_negsi2): Add alternative for 16-bit encoding.
9600 (arm_one_cmplsi2): Likewise.
9602 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9604 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
9605 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
9606 (movdfcc): Likewise.
9607 * config/arm/vfp.md (*thumb2_movsf_vfp):
9608 Disable predication for arm_restrict_it.
9609 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
9610 (*thumb2_movdfcc_vfp): Likewise.
9611 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
9612 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
9613 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
9614 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
9615 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
9616 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
9617 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
9618 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
9619 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
9620 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
9621 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
9622 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
9623 Disable predication for arm_restrict_it.
9625 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
9627 * config/i386/bmiintrin.h (_bextr_u32): New.
9628 (_bextr_u64): Ditto.
9630 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
9632 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
9633 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
9634 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
9635 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
9636 * config/mips/n32-elf.h: ...this new file.
9638 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
9641 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
9642 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
9644 2013-06-27 Catherine Moore <clm@codesourcery.com>
9646 * config/mips/mips-tables.opt: Regenerate.
9647 * config/mips/mips-cpus.def: Add m14ke and m14kec.
9648 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
9649 * doc/invoke.texi: Add -m14kc.
9651 2013-06-27 Jakub Jelinek <jakub@redhat.com>
9654 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
9655 constraints of operand 1 and 2.
9658 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
9659 to match RTL canonicalization. Swap predicates and
9660 constraints of operand 1 and 2.
9662 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
9664 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
9665 Process OP_INOUT regs for splitting too.
9667 2013-06-27 Jakub Jelinek <jakub@redhat.com>
9669 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
9670 decl before the loop, initialize to NULL.
9671 (vectorizable_load): Initialize ptr_incr to NULL.
9673 2013-06-27 Martin Jambor <mjambor@suse.cz>
9676 * ipa-ref.h (ipa_maybe_record_reference): Declare.
9677 * ipa-ref.c (ipa_maybe_record_reference): New function.
9678 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
9679 * ipa-cp.c (create_specialized_node): Record potential references from
9681 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
9683 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
9685 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
9686 parameter 'mode' of type 'enum machine_mode mode'; change to pass
9687 'mode' to force_reg.
9688 (aarch64_add_offset): Update calls to aarch64_force_temporary.
9689 (aarch64_expand_mov_immediate): Likewise.
9691 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
9693 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
9694 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
9696 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9698 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
9699 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
9700 (struct machine_function): Add tbegin_p.
9701 (s390_canonicalize_comparison): Fold CC mode compares to
9702 conditional jump if possible.
9703 (s390_emit_jump): Return the emitted jump.
9704 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
9705 Handle CCRAWmode compares.
9706 (s390_option_override): Default to -mhtm if available.
9707 (s390_reg_clobbered_rtx): Handle floating point regs as well.
9708 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
9709 FPRs instead of df_regs_ever_live_p.
9710 (s390_optimize_nonescaping_tx): New function.
9711 (s390_init_frame_layout): Extend clobbered_regs array to cover
9713 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
9714 (s390_expand_tbegin): New function.
9715 (enum s390_builtin): New enum definition.
9716 (code_for_builtin): New array definition.
9717 (s390_init_builtins): New function.
9718 (s390_expand_builtin): New function.
9719 (TARGET_INIT_BUILTINS): Define.
9720 (TARGET_EXPAND_BUILTIN): Define.
9721 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
9722 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
9723 (s390_alc_comparison): Likewise.
9724 * config/s390/s390-modes.def: Add CCRAWmode.
9725 * config/s390/s390.h (processor_flags): Add PF_TX.
9726 (TARGET_CPU_HTM): Define macro.
9727 (TARGET_HTM): Define macro.
9728 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
9729 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
9730 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
9731 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
9732 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
9733 (TBEGIN_MASK, TBEGINC_MASK): New constants.
9734 ("*cc_to_int"): Move up.
9735 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
9736 constants other than 0.
9737 ("*ccraw_to_int"): New insn and splitter definition.
9738 ("tbegin", "tbegin_nofloat", "tbegin_retry")
9739 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
9740 ("tx_assist"): New expander.
9741 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
9742 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
9743 * config/s390/s390.opt: Add -mhtm option.
9744 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
9745 * config/s390/htmxlintrin.h: New file.
9746 * config/s390/htmintrin.h: New file.
9747 * config/s390/s390intrin.h: New file.
9748 * doc/extend.texi: Document htm builtins.
9749 * config.gcc: Add the new header files to extra_headers.
9751 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
9753 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
9754 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
9756 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
9757 Pat Haugen <pthaugen@us.ibm.com>
9758 Peter Bergner <bergner@vnet.ibm.com>
9760 * config/rs6000/power8.md: New.
9761 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
9762 setting for power8 entry.
9763 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
9764 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
9765 test for Power4/Power5 only.
9766 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
9768 (force_new_group): Adjust comment.
9769 * config/rs6000/rs6000.md: Include power8.md.
9771 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
9773 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
9774 * config/arm/arm-protos.h (arm_max_conditional_execute): New
9776 (tune_params): Update comment.
9777 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
9778 (arm_max_conditional_execute): New function.
9779 (thumb2_final_prescan_insn): Use max_insn_skipped and
9780 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
9782 2013-06-25 Jakub Jelinek <jakub@redhat.com>
9784 PR tree-optimization/57705
9785 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
9786 SSA_NAME step, provided that it is not defined inside the loop.
9787 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
9788 (get_initial_def_for_induction): Handle SSA_NAME IV step.
9790 2013-06-25 Martin Jambor <mjambor@suse.cz>
9793 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
9794 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
9796 (ipa_note_param_call): Initialize member_ptr flag.
9797 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
9798 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
9799 (ipa_write_indirect_edge_info): Stream member_ptr flag.
9800 (ipa_read_indirect_edge_info): Likewise.
9802 2013-06-25 Richard Biener <rguenther@suse.de>
9805 * passes.c (init_optimization_passes): Move pass_fold_builtins
9806 and pass_dce earlier with -Og.
9808 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
9810 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
9811 <BIT_FIELD_REF>: Remove trailing TAB.
9812 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
9815 2013-06-24 Martin Jambor <mjambor@suse.cz>
9817 PR tree-optimization/57358
9818 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
9819 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
9820 (ipa_analyze_params_uses): Generate pessimistic info when true.
9822 2013-06-24 Martin Jambor <mjambor@suse.cz>
9824 PR tree-optimization/57539
9825 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
9826 global.inlined_to of the new node to it. All callers changed.
9827 * ipa-inline-transform.c (clone_inlined_nodes): New variable
9828 inlining_into, pass it to cgraph_clone_node.
9829 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
9830 ipa_free_edge_args_substructures.
9831 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
9832 rdesc linked list. Do not assert rdesc edges have inlined caller.
9833 Assert we have found an rdesc in the rdesc list.
9835 2013-06-24 Richard Biener <rguenther@suse.de>
9837 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
9838 (pointer_set_lookup): Declare.
9839 (class pointer_map): New template class implementing a
9840 generic pointer to T map.
9841 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
9842 pointer_map<T>::contains, pointer_map<T>::insert,
9843 pointer_map<T>::traverse): New functions.
9844 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
9845 (pointer_set_lookup): New function.
9846 (pointer_set_contains): Use pointer_set_lookup.
9847 (pointer_set_insert): Likewise.
9848 (insert_aux): Remove.
9849 (struct pointer_map_t): Embed a pointer_set_t.
9850 (pointer_map_create): Adjust.
9851 (pointer_map_destroy): Likewise.
9852 (pointer_map_contains): Likewise.
9853 (pointer_map_insert): Likewise.
9854 (pointer_map_traverse): Likewise.
9855 * tree-streamer.h (struct streamer_tree_cache_d): Use a
9856 pointer_map<unsigned> instead of a pointer_map_t.
9857 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
9858 (streamer_tree_cache_lookup): Likewise.
9859 (streamer_tree_cache_create): Likewise.
9860 (streamer_tree_cache_delete): Likewise.
9861 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
9862 pointer_map<unsigned> instead of a pointer_map_t.
9863 (lto_init_tree_ref_encoder): Adjust.
9864 (lto_destroy_tree_ref_encoder): Likewise.
9865 * lto-section-out.c (lto_output_decl_index): Likewise.
9866 (lto_record_function_out_decl_state): Likewise.
9867 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
9869 2013-06-24 Richard Biener <rguenther@suse.de>
9871 PR tree-optimization/57488
9872 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
9874 2013-06-24 Alan Modra <amodra@gmail.com>
9876 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
9877 (gen_easy_altivec_constant): Likewise.
9878 * config/rs6000/predicates.md (easy_vector_constant_add_self,
9879 easy_vector_constant_msb): Likewise.
9881 2013-06-23 Jakub Jelinek <jakub@redhat.com>
9884 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
9885 add missing return true.
9887 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
9890 * config/sh/predicates.md (general_extend_operand): Invoke
9891 general_movsrc_operand for memory operands.
9892 (general_movsrc_operand): Allow reg+reg addressing, do not use
9893 general_operand for memory operands.
9895 2013-06-23 Sriraman Tallam <tmsriram@google.com>
9897 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
9898 when current target options does not apply.
9899 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
9900 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
9901 * config/i386/bmiintrin.h: Pass appropriate target
9902 attributes to header.
9903 * config/i386/mmintrin.h: Ditto.
9904 * config/i386/nmmintrin.h: Ditto.
9905 * config/i386/avx2intrin.h: Ditto.
9906 * config/i386/fxsrintrin.h: Ditto.
9907 * config/i386/tbmintrin.h: Ditto.
9908 * config/i386/xsaveintrin.h: Ditto.
9909 * config/i386/f16cintrin.h: Ditto.
9910 * config/i386/xtestintrin.h: Ditto.
9911 * config/i386/xsaveoptintrin.h: Ditto.
9912 * config/i386/bmi2intrin.h: Ditto.
9913 * config/i386/lzcntintrin.h: Ditto.
9914 * config/i386/smmintrin.h: Ditto.
9915 * config/i386/wmmintrin.h: Ditto.
9916 * config/i386/x86intrin.h: Remove all header include guards.
9917 * config/i386/prfchwintrin.h: Ditto.
9918 * config/i386/pmmintrin.h: Ditto.
9919 * config/i386/tmmintrin.h: Ditto.
9920 * config/i386/xmmintrin.h: Ditto.
9921 * config/i386/popcntintrin.h: Ditto.
9922 * config/i386/rdseedintrin.h: Ditto.
9923 * config/i386/ammintrin.h: Ditto.
9924 * config/i386/emmintrin.h: Ditto.
9925 * config/i386/immintrin.h: Remove all header include guards.
9926 * config/i386/fma4intrin.h: Ditto.
9927 * config/i386/lwpintrin.h: Ditto.
9928 * config/i386/xopintrin.h: Ditto.
9929 * config/i386/ia32intrin.h: Ditto.
9930 * config/i386/avxintrin.h: Ditto.
9931 * config/i386/rtmintrin.h: Ditto.
9932 * config/i386/fmaintrin.h: Ditto.
9933 * config/i386/mm3dnow.h: Ditto.
9935 2013-06-22 Sriraman Tallam <tmsriram@google.com>
9937 * common/config/i386/i386-common.c: Handle LZCNT.
9939 2013-06-22 Andi Kleen <ak@linux.intel.com>
9941 * doc/extend.texi: Use __atomic_store_n instead of
9942 __atomic_store in HLE example.
9944 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
9946 * config/sh/sh.c: Remove <cstdlib> workaround.
9948 2013-06-21 Andi Kleen <ak@linux.intel.com>
9950 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
9952 2013-06-21 Andi Kleen <ak@linux.intel.com>
9954 * doc/extend.texi: Document that __atomic_clear and
9955 __atomic_test_and_set should only be used with bool.
9957 2013-06-20 Jan Hubicka <jh@suse.cz>
9959 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
9961 * tree.c (decls_same_for_odr): New function.
9962 (same_for_edr): New function.
9963 (types_same_for_odr): New function.
9964 (get_binfo_at_offset): Use it.
9965 * tree.h (types_same_for_odr): Declare.
9967 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
9968 Jason Merrill <jason@redhat.com>
9970 * system.h: Include <cstdlib> as well as <stdlib.h>.
9972 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
9975 * config/i386/i386.c (construct_container): Report error if
9976 long double is used with disabled x87 float returns.
9978 2013-06-20 Jan Hubicka <jh@suse.cz>
9980 * lto-cgraph.c (input_symtab): Do not set cgraph state.
9982 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
9984 PR rtl-optimization/57425
9985 PR rtl-optimization/57569
9986 * alias.c (write_dependence_p): Remove parameters mem_mode and
9987 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
9988 Changed all callers.
9989 (canon_anti_dependence): Get comments and semantics in sync.
9990 Add parameter mem_canonicalized. Changed all callers.
9991 * rtl.h (canon_anti_dependence): Update prototype.
9993 2013-06-20 Richard Biener <rguenther@suse.de>
9995 * data-streamer-in.c (streamer_read_uhwi): Optimize single
9996 byte case, inline streamer_read_uchar and defer section
9999 2013-06-20 Richard Biener <rguenther@suse.de>
10001 PR tree-optimization/57584
10002 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
10003 SSA names into the expanded expression that take part in
10004 abnormal coalescing.
10006 2013-06-19 Sharad Singhai <singhai@google.com>
10008 * gcov.c (print_usage): Handle new option.
10009 (process_args): Ditto.
10010 (get_gcov_intermediate_filename): New function.
10011 (output_intermediate_file): New function.
10012 (output_gcov_file): New function
10013 (generate_results): Handle new option.
10014 (release_function): Relase demangled name.
10015 (read_graph_file): Handle demangled name.
10016 (output_lines): Ditto.
10017 * doc/gcov.texi: Document gcov intermediate format.
10019 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
10022 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
10023 (lra_emit_add): Use the functions. Add comment about Y as an
10026 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
10029 * collect2.c (collect_atexit): New.
10030 (collect_exit): Delete.
10031 (main): Register collect_atexit with atexit.
10032 (collect_wait): Change collect_exit to exit.
10034 * collect2.h (collect_exit): Delete.
10035 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
10037 2013-06-19 Wei Mi <wmi@google.com>
10039 PR rtl-optimization/57518
10040 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
10041 if regno is used in paradoxical subreg.
10042 (update_equiv_regs): Check pdx_subregs[regno] before
10043 set a reg to be equivalent with a mem.
10045 2013-06-19 Matthias Klose <doko@ubuntu.com>
10048 * file-find.h (find_a_file): Add a mode parameter.
10049 * file-find.c (find_a_file): Likewise.
10050 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
10051 with X_OK for the executables.
10052 * collect2.c (main): Call find_a_file with X_OK.
10054 2013-06-19 Steve Ellcey <sellcey@mips.com>
10057 * config/mips/mips.md (casesi_internal_mips16_<mode>):
10058 Use NEXT_INSN instead of next_real_insn.
10060 2013-06-19 Jan Hubicka <jh@suse.cz>
10062 * cgraph.h (const_value_known_p): Replace by ...
10063 (ctor_for_folding): .. this one.
10064 * cgraphunit.c (process_function_and_variable_attributes): Use it.
10065 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
10066 * expr.c (expand_expr_real_1): Likewise.
10067 (string_constant): Likewise.
10068 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
10069 * ipa.c (process_references): Likewise.
10070 (symtab_remove_unreachable_nodes): Likewise.
10071 * ipa-inline-analysis.c (param_change_prob): Likewise.
10072 * gimple-fold.c (canonicalize_constructor_val): Likewise.
10073 (get_base_constructor): Likwise.
10074 * varpool.c (varpool_remove_node): Likewise.
10075 (varpool_remove_initializer): LIkewise.
10076 (dump_varpool_node): LIkwise.
10077 (const_value_known_p): Rewrite to ...
10078 (ctor_for_folding): ... this one.
10080 2013-06-19 Jakub Jelinek <jakub@redhat.com>
10083 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
10084 PERSONALITY in $PATH derived prefixes.
10086 2013-06-19 Jeff Law <law@redhat.com>
10088 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
10091 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
10092 (simplify_bitwise_binary): Use it to simpify certain binary ops on
10095 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
10097 * config/arm/vfp.md: Move VFP instruction classification documentation
10099 * config/arm/arm.md: ... here. Update instruction classification
10102 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
10104 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
10105 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
10106 pattern. Use more efficient sequences on ARMv5 and Thumb2.
10108 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
10111 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
10112 with NEXT_INSN. Use tablejump_p to check for jump table data
10115 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
10118 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
10119 that now in C++ the value is correct per the C++ standards.
10121 2013-06-19 Richard Biener <rguenther@suse.de>
10123 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
10124 for global context.
10126 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10129 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10132 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
10133 with next_active_insn.
10135 2013-06-18 Sriraman Tallam <tmsriram@google.com>
10137 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
10138 functions are inlined during failures to flag an error.
10139 * tree-inline.c (expand_call_inline): Allow the error to be flagged
10140 in early inline pass.
10142 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
10144 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
10147 2013-06-18 Julian Brown <julian@codesourcery.com>
10149 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
10150 Permit virtual register pre-reload if !strict.
10151 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
10153 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
10155 * config/arm/neon.md (movmisalign<mode>): Use
10156 neon_perm_struct_or_reg_operand instead of
10157 neon_struct_or_register_operand.
10158 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
10159 neon_permissive_struct_operand instead of neon_struct_operand.
10160 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
10161 neon_vector_mem_operand.
10162 * config/arm/predicates.md (neon_struct_operand): Adjust call to
10163 neon_vector_mem_operand.
10164 (neon_permissive_struct_operand): New.
10165 (neon_struct_or_register_operand): Rename to...
10166 (neon_perm_struct_or_reg_operand): This. Adjust call to
10167 neon_vector_mem_operand.
10169 2013-06-18 Richard Biener <rguenther@suse.de>
10171 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
10172 * lto-streamer.h: Include pointer-set.h.
10173 (struct lto_decl_slot): Remove.
10174 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
10175 Remove next_index entry.
10176 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
10177 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
10178 (lto_init_tree_ref_encoder): Adjust.
10179 (lto_destroy_tree_ref_encoder): Likewise.
10180 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
10181 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
10182 (lto_output_decl_index): Adjust.
10183 (lto_new_out_decl_state): Likewise.
10184 (lto_record_function_out_decl_state): Likewise.
10185 * lto-streamer-out.c (copy_function): Likewise.
10187 2013-06-18 Richard Biener <rguenther@suse.de>
10189 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
10190 * cgraphunit.c: Include cfgloop.h.
10191 (init_lowered_empty_function): Initialize the loop tree.
10192 (assemble_thunk): Insert new BBs into loops.
10194 2013-06-18 Richard Biener <rguenther@suse.de>
10196 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
10197 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
10198 the map from cache entry to cache index optional.
10199 (streamer_tree_cache_replace_tree): Adjust accordingly.
10200 (streamer_tree_cache_append): Likewise.
10201 (streamer_tree_cache_delete): Likewise.
10202 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
10203 streamer cache map from cache entry to cache index.
10204 * lto-streamer-out.c (create_output_block): Adjust.
10206 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
10208 * config/arm/arm.md (attribute "insn"): Move multiplication and
10209 division attributes to...
10210 (attribute "type"): ... here. Remove mult.
10211 (attribute "mul32"): New attribute.
10212 (attribute "mul64"): Add umaal.
10213 (*arm_mulsi3): Update attributes.
10214 (*arm_mulsi3_v6): Likewise.
10215 (*thumb_mulsi3): Likewise.
10216 (*thumb_mulsi3_v6): Likewise.
10217 (*mulsi3_compare0): Likewise.
10218 (*mulsi3_compare0_v6): Likewise.
10219 (*mulsi_compare0_scratch): Likewise.
10220 (*mulsi_compare0_scratch_v6): Likewise.
10221 (*mulsi3addsi): Likewise.
10222 (*mulsi3addsi_v6): Likewise.
10223 (*mulsi3addsi_compare0): Likewise.
10224 (*mulsi3addsi_compare0_v6): Likewise.
10225 (*mulsi3addsi_compare0_scratch): Likewise.
10226 (*mulsi3addsi_compare0_scratch_v6): Likewise.
10227 (*mulsi3subsi): Likewise.
10228 (*mulsidi3adddi): Likewise.
10229 (*mulsi3addsi_v6): Likewise.
10230 (*mulsidi3adddi_v6): Likewise.
10231 (*mulsidi3_nov6): Likewise.
10232 (*mulsidi3_v6): Likewise.
10233 (*umulsidi3_nov6): Likewise.
10234 (*umulsidi3_v6): Likewise.
10235 (*umulsidi3adddi): Likewise.
10236 (*umulsidi3adddi_v6): Likewise.
10237 (*smulsi3_highpart_nov6): Likewise.
10238 (*smulsi3_highpart_v6): Likewise.
10239 (*umulsi3_highpart_nov6): Likewise.
10240 (*umulsi3_highpart_v6): Likewise.
10241 (mulhisi3): Likewise.
10242 (*mulhisi3tb): Likewise.
10243 (*mulhisi3bt): Likewise.
10244 (*mulhisi3tt): Likewise.
10245 (maddhisi4): Likewise.
10246 (*maddhisi4tb): Likewise.
10247 (*maddhisi4tt): Likewise.
10248 (maddhidi4): Likewise.
10249 (*maddhidi4tb): Likewise.
10250 (*maddhidi4tt): Likewise.
10251 (divsi3): Likewise.
10252 (udivsi3): Likewise.
10253 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
10254 (thumb2_mulsi_short_compare0): Likewise.
10255 (thumb2_mulsi_short_compare0_scratch): Likewise.
10256 * config/arm/arm1020e.md (1020mult1): Update attribute change.
10257 (1020mult2): Likewise.
10258 (1020mult3): Likewise.
10259 (1020mult4): Likewise.
10260 (1020mult5): Likewise.
10261 (1020mult6): Likewise.
10262 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
10264 (cortex_a15_mult64): Likewise.
10265 (cortex_a15_sdiv): Likewise.
10266 (cortex_a15_udiv): Likewise.
10267 * config/arm/arm1026ejs.md (mult1): Update attribute change.
10273 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
10274 (pj4_ir_div): Likewise.
10275 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
10276 (11_mult2): Likewise.
10277 (11_mult3): Likewise.
10278 (11_mult4): Likewise.
10279 (11_mult5): Likewise.
10280 (11_mult6): Likewise.
10281 (11_mult7): Likewise.
10282 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
10283 (cortex_a8_mla): Likewise.
10284 (cortex_a8_mull): Likewise.
10285 (cortex_a8_smulwy): Likewise.
10286 (cortex_a8_smlald): Likewise.
10287 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
10288 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
10289 (cortex_r4_mul_3): Likewise.
10290 (cortex_r4_mla_4): Likewise.
10291 (cortex_r4_mla_3): Likewise.
10292 (cortex_r4_smlald): Likewise.
10293 (cortex_r4_mull): Likewise.
10294 (cortex_r4_sdiv): Likewise.
10295 (cortex_r4_udiv): Likewise.
10296 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
10297 (cortex_a7_idiv): Likewise.
10298 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
10299 (9_mult2): Likewise.
10300 (9_mult3): Likewise.
10301 (9_mult4): Likewise.
10302 (9_mult5): Likewise.
10303 (9_mult6): Likewise.
10304 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
10305 (cortex_a53_sdiv): Likewise.
10306 (cortex_a53_udiv): Likewise.
10307 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
10308 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
10309 (mp626_mult2): Likewise.
10310 (mp626_mult3): Likewise.
10311 (mp626_mult4): Likewise.
10312 * config/arm/fa526.md (526_mult1): Update attribute change.
10313 (526_mult2): Likewise.
10314 * config/arm/arm-generic.md (mult): Update attribute change.
10315 (mult_ldsched_strongarm): Likewise.
10316 (mult_ldsched): Likewise.
10317 (multi_cycle): Likewise.
10318 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
10319 * config/arm/fa606te.md (606te_mult1): Update attribute change.
10320 (606te_mult2): Likewise.
10321 (606te_mult3): Likewise.
10322 (606te_mult4): Likewise.
10323 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
10324 (cortex_a9_mac16): Likewise.
10325 (cortex_a9_multiply): Likewise.
10326 (cortex_a9_mac): Likewise.
10327 (cortex_a9_multiply_long): Likewise.
10328 * config/arm/fa626te.md (626te_mult1): Update attribute change.
10329 (626te_mult2): Likewise.
10330 (626te_mult3): Likewise.
10331 (626te_mult4): Likewise.
10333 2013-06-18 Richard Biener <rguenther@suse.de>
10336 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
10338 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10341 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
10342 with next_active_insn.
10344 2013-06-18 Alan Modra <amodra@gmail.com>
10346 * config/rs6000/rs6000.h (enum data_align): New.
10347 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
10348 (DATA_ABI_ALIGNMENT): Define.
10349 (CONSTANT_ALIGNMENT): Correct comment.
10350 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
10351 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
10353 2013-06-17 David Malcolm <dmalcolm@redhat.com>
10355 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
10356 ATTRIBUTE_UNUSED marking.
10358 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
10360 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
10361 alternative and update.
10362 (aarch64_dup_lanedi): Delete.
10363 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
10364 * config/aarch64/aarch64-simd-builtins.def: Update.
10366 2013-06-17 Richard Biener <rguenther@suse.de>
10368 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
10369 (lto_input_scc): Declare.
10370 (lto_input_tree_1): Likewise.
10371 (struct lto_stats_d): Add num_tree_bodies_output and
10372 num_pickle_refs_output.
10373 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
10374 (lto_read_tree_1): Split out from ...
10375 (lto_read_tree): ... this.
10376 (lto_input_scc): New function.
10377 (lto_input_tree_1): Split out from ...
10378 (lto_input_tree): ... this. Handle LTO_tree_scc.
10379 (lto_data_in_create): Create the streamer cache without hashes.
10380 * lto-streamer-out.c (create_output_block): Create the streamer
10381 cache with hashes when not doing WPA.
10382 (lto_write_tree_1): Split out from ...
10383 (lto_write_tree): ... this.
10384 (get_symbol_initial_value): New function.
10385 (lto_output_tree_1): Split out from ...
10386 (lto_output_tree): ... this. Write trees as series of SCCs
10387 using a DFS walk via DFS_write_tree.
10388 (struct sccs, struct scc_entry): New types.
10389 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
10390 (DFS_write_tree_body): New function.
10391 (DFS_write_tree): Likewise.
10392 (hash_tree): Likewise.
10393 (scc_entry_compare): Likewise.
10394 (hash_scc): Likewise.
10395 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
10396 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
10397 TREE_CHAIN as regular reference.
10398 (streamer_read_integer_cst): Remove.
10399 (streamer_get_pickled_tree): Adjust.
10400 * tree-streamer-out.c (streamer_write_chain): Disable streaming
10401 of DECL_EXTERNALs in BLOCK_VARS for now.
10402 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
10404 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
10405 Add hash value argument and record that if hashes are recorded
10407 (streamer_tree_cache_insert_1): Adjust.
10408 (streamer_tree_cache_insert): Likewise.
10409 (streamer_tree_cache_insert_at): Rename to ...
10410 (streamer_tree_cache_replace_tree): ... this and adjust.
10411 (streamer_tree_cache_append): Adjust.
10412 (record_common_node): Likewise.
10413 (streamer_tree_cache_create): Add argument whether to
10414 record hash values together with trees.
10415 (streamer_tree_cache_delete): Adjust.
10416 * tree-streamer.h (struct streamer_tree_cache_d): Add
10418 (streamer_read_integer_cst): Remove.
10419 (streamer_tree_cache_insert): Adjust.
10420 (streamer_tree_cache_append): Likewise.
10421 (streamer_tree_cache_insert_at): Rename to ...
10422 (streamer_tree_cache_replace_tree): ... this and adjust.
10423 (streamer_tree_cache_create): Add argument whether to record hashes.
10424 (streamer_tree_cache_get): Rename to ...
10425 (streamer_tree_cache_get_tree): ... this.
10426 (streamer_tree_cache_get_hash): New function.
10427 * tree.c (cache_integer_cst): New function.
10428 * tree.h (cache_integer_cst): Declare.
10429 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
10430 * lto-symtab.c (lto_varpool_replace_node): Only release
10431 DECL_INITIAL of non-prevailing decls.
10432 * varpool.c (varpool_remove_initializer): Do not release
10433 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
10435 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
10437 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
10438 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
10439 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
10440 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
10441 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
10442 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
10443 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
10444 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
10445 instead of TARGET_64BIT.
10446 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
10447 Require ISA_HAS_<D>DIV.
10449 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
10451 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
10452 (mips*-*-linux*): Move default with_llsc setting to where other
10454 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
10456 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
10457 Likewise. Remove default with_tune setting. Move default float
10458 setting to its own block. Handle with_llsc in the same block as above.
10460 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
10462 PR rtl-optimization/57425
10463 PR rtl-optimization/57569
10464 * alias.c (write_dependence_p): Add new parameters mem_mode,
10465 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
10466 Changed all callers.
10467 (canon_anti_dependence): New function.
10468 * cse.c (check_dependence): Use canon_anti_dependence.
10469 * cselib.c (cselib_invalidate_mem): Likewise.
10470 * rtl.h (canon_anti_dependence): Declare.
10472 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
10474 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
10475 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
10476 ".set mips3" for 64-bit targets.
10478 2013-06-15 Dehao Chen <dehao@google.com>
10480 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
10481 * gimple-low.c (gimple_check_call_matching_types): Likewise.
10482 (gimple_check_call_args): Likewise.
10483 * value-prof.c (check_ic_target): Likewise.
10484 * ipa-inline.c (early_inliner): Likewise.
10485 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
10486 * cgraph.c (cgraph_create_edge_1): Likewise.
10487 (cgraph_make_edge_direct): Likewise.
10489 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
10492 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
10493 rs6000_output_move_128bit to handle emitting quad memory
10494 operations. Set attribute length to 8 bytes.
10496 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
10498 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
10500 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
10501 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
10502 (aarch64_<su>mlsl<mode>): Likewise.
10504 2013-06-14 Mike Stump <mikestump@comcast.net>
10506 * Makefile.in (TARGET_H): Add insn-codes.h.
10508 2013-06-14 Alan Modra <amodra@gmail.com>
10510 PR middle-end/57134
10511 PR middle-end/57586
10512 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
10513 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
10514 bitfield expansion when EXPAND_MEMORY.
10515 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
10517 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
10519 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
10520 test for clearing quad memory on 32-bit later.
10522 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
10524 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
10525 (fold_negate_expr): Likewise.
10526 (fold_real_zero_addition_p): Handle vectors.
10527 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
10529 2013-06-14 Alan Modra <amodra@gmail.com>
10531 * varasm.c (force_const_mem): Revert 2013-06-07 change.
10533 2013-06-13 Jan Hubicka <jh@suse.cz>
10535 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
10536 Local comdats are not externally visible.
10537 * symtab.c (dump_symtab_base): Dump externally visible.
10538 (verify_symtab_base): Verify back links in the symtab hash.
10540 2013-06-13 Bin Cheng <bin.cheng@arm.com>
10542 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
10543 CONVERT_EXPR as equal nodes.
10545 2013-06-13 Bin Cheng <bin.cheng@arm.com>
10547 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
10549 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
10551 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
10552 Generalize to complex and vector.
10553 * tree.c (build_all_ones_cst): New function.
10554 * tree.h (build_all_ones_cst): Declare it.
10556 2013-06-13 Alan Modra <amodra@gmail.com>
10558 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
10559 * config/rs6000/rs6000.md (signbittf2): New insn.
10560 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
10561 (abstf2_internal, cmptf_internal2): Likewise.
10562 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
10564 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
10565 Pat Haugen <pthaugen@us.ibm.com>
10566 Peter Bergner <bergner@vnet.ibm.com>
10568 * config/rs6000/rs6000.c (emit_load_locked): Add support for
10569 power8 byte, half-word, and quad-word atomic instructions.
10570 (emit_store_conditional): Likewise.
10571 (rs6000_expand_atomic_compare_and_swap): Likewise.
10572 (rs6000_expand_atomic_op): Likewise.
10574 * config/rs6000/sync.md (larx): Add new modes for power8.
10576 (AINT): New mode iterator to include TImode as well as normal
10577 integer modes on power8.
10578 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
10579 that VSX registers are not considered. Use AINT mode iterator
10580 instead of INT1 to allow inclusion of quad word atomic operations
10582 (load_locked<mode>): Likewise.
10583 (store_conditional<mode>): Likewise.
10584 (atomic_compare_and_swap<mode>): Likewise.
10585 (atomic_exchange<mode>): Likewise.
10586 (atomic_nand<mode>): Likewise.
10587 (atomic_fetch_<fetchop_name><mode>): Likewise.
10588 (atomic_nand_fetch<mode>): Likewise.
10589 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
10591 (ATOMIC): On power8, add QImode, HImode modes.
10592 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
10593 modes that promote to SImode.
10594 (load_lockedti): Convert TImode arguments to PTImode, so that we
10595 get a guaranteed even/odd register pair.
10596 (load_lockedpti): Likewise.
10597 (store_conditionalti): Likewise.
10598 (store_conditionalpti): Likewise.
10600 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
10601 atomic load/store instructions.
10604 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
10606 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
10608 (insn_count): New attribute, with most cases extracted from...
10609 (length): ...here. Redefine most cases in terms of insn_count.
10610 (single_insn): Delete.
10611 (can_delay): Use insn_count to check for single instructions.
10612 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
10613 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
10614 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
10615 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
10616 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
10617 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
10618 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
10619 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
10620 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
10621 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
10622 rather than "length".
10623 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
10624 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
10625 Use "insn_count" rather than "length".
10626 * config/mips/mips-dsp.md
10627 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
10628 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
10631 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
10633 PR tree-optimization/57361
10634 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
10636 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
10638 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
10640 (aarch64_simd_combine<mode>): New instruction expansion.
10641 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
10642 function prototype.
10643 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
10644 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
10646 2013-06-12 Jan Hubicka <jh@suse.cz>
10648 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
10649 decl has when in streaming stage.
10650 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
10651 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
10653 2013-06-12 Roland Stigge <stigge@antcom.de>
10656 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
10658 2013-06-12 Jakub Jelinek <jakub@redhat.com>
10660 PR tree-optimization/57537
10661 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
10662 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
10664 2013-06-12 Richard Biener <rguenther@suse.de>
10666 * data-streamer.h (streamer_write_char_stream): CSE
10667 obs->current_pointer.
10668 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
10669 streamer_write_char_stream manually and optimize the resulting loop.
10670 (streamer_write_hwi_stream): Likewise.
10672 2013-06-12 Jan Hubicka <jh@suse.cz>
10674 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
10675 * cgraph.h (varpool_create_empty_node): Declare.
10676 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
10678 * symtab.c (symtab_unregister_node): Be lax about missin entries
10680 (symtab_get_node): Update comment.
10681 * varpool.c (varpool_create_empty_node): Break out from ...
10682 (varpool_node_for_decl): ... here.
10683 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
10685 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
10687 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
10688 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
10689 part. Use straight-line flow at the end.
10690 <COMPONENT_REF>: Remove superfluous else.
10691 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
10693 2013-06-12 Jakub Jelinek <jakub@redhat.com>
10696 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
10697 target hook even for !TREE_PUBLIC decls. If no resolution info
10698 is available, return false for common and external decls.
10700 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
10702 * config/rl78/constraints.md (U): New constraint.
10703 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
10706 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
10709 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
10710 to allow returning address to AT_PLATFORM name.
10712 2013-06-11 Jan Hubicka <jh@suse.cz>
10714 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
10715 * cgraph.h (symtab_node_base): Add weakref flag.
10716 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
10717 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
10718 (output_weakrefs): Use weakref flag.
10719 * fold-const.c (simple_operand_p): Handle WEAK.
10720 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
10721 * ipa.c (varpool_externally_visible_p): Drop weakref.
10722 (function_and_variable_visibility): Update comment; fix weakref
10723 sanity checks; do not clear DECL_WEAK on them.
10724 * lto-cgraph.c (lto_output_node): update.
10725 (lto_output_varpool_node): Update.
10726 (input_overwrite_node): Update.
10727 (input_node): Update.
10728 (input_varpool_node): Update.
10729 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
10730 (lto_symtab_merge_symbols): Add sanity check.
10731 (lto_symtab_prevailing_decl): Do not special case weakrefs.
10732 * passes.c (rest_of_decl_compilation): Set static flag, too.
10733 * symtab.c (dump_symtab_base): Dump weakref.
10734 (verify_symtab_base): Sanity check weakrefs.
10735 (symtab_make_decl_local): Remove duplicated code.
10736 (symtab_alias_ultimate_target): Simplify.
10737 * varpool.c (varpool_create_variable_alias): Set weakref flag.
10739 2013-06-11 Tom de Vries <tom@codesourcery.com>
10741 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
10742 sequence_vect == NULL.
10744 2013-06-11 DJ Delorie <dj@redhat.com>
10746 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
10747 (rl78_unwind_word_mode): New.
10749 2013-06-11 David Malcolm <dmalcolm@redhat.com>
10751 * final.c (debug_prefix_maps): Make static.
10753 2013-06-11 David Malcolm <dmalcolm@redhat.com>
10755 * function.c (initial_trampoline): Remove stray copy.
10757 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
10759 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
10761 2013-06-11 Martin Jambor <mjambor@suse.cz>
10763 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
10764 within bounds at the beginning of the function.
10766 2013-06-11 Alan Modra <amodra@gmail.com>
10768 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
10770 (get_named_section): Don't NULL !DECL_P decl.
10772 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
10774 * doc/invoke.texi (core-avx2): Document.
10776 (atom): Updated with MOVBE.
10778 2013-06-11 Richard Biener <rguenther@suse.de>
10780 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
10782 2013-06-11 Anton Blanchard <anton@samba.org>
10784 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
10785 correct shift value in little-endian mode.
10787 2013-06-11 Jakub Jelinek <jakub@redhat.com>
10790 * varasm.c (get_variable_align): Move #endif to the right place.
10792 2013-06-10 Cary Coutant <ccoutant@google.com>
10794 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
10795 for hash so that hash table traversal order is deterministic.
10797 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
10798 Pat Haugen <pthaugen@us.ibm.com>
10799 Peter Bergner <bergner@vnet.ibm.com>
10801 * config/rs6000/vector.md (GPR move splitter): Do not split moves
10802 of vectors in GPRS if they are direct moves or quad word load or
10805 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
10807 (direct_move_p): Likewise.
10808 (quad_load_store_p): Likewise.
10810 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
10811 classes into bins based on the physical register type.
10812 (reg_class_to_reg_type): Likewise.
10813 (IS_STD_REG_TYPE): Likewise.
10814 (IS_FP_VECT_REG_TYPE): Likewise.
10815 (reload_fpr_gpr): Arrays to determine what insn to use if we can
10816 use direct move instructions.
10817 (reload_gpr_vsx): Likewise.
10818 (reload_vsx_gpr): Likewise.
10819 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
10820 information that is a simplification of register classes. Also
10821 precalculate direct move reload helpers.
10822 (direct_move_p): New function to return true if the operation can
10823 be done as a direct move instruciton.
10824 (quad_load_store_p): New function to return true if the operation
10825 is a quad memory operation.
10826 (rs6000_legitimize_address): If quad memory, only allow register
10827 indirect for TImode addresses.
10828 (rs6000_legitimate_address_p): Likewise.
10829 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
10830 (rs6000_reload_register_type): Likewise.
10831 (register_to_reg_type): Return register type.
10832 (rs6000_secondary_reload_simple_move): New helper function for
10833 secondary reload and secondary memory needed to identify anything
10834 that is a simple move, and does not need reloading.
10835 (rs6000_secondary_reload_direct_move): New helper function for
10836 secondary reload to identify cases that can be done with several
10837 instructions via the direct move instructions.
10838 (rs6000_secondary_reload_move): New helper function for secondary
10839 reload to identify moves between register types that can be done.
10840 (rs6000_secondary_reload): Add support for quad memory operations
10841 and for direct move.
10842 (rs6000_secondary_memory_needed): Likewise.
10843 (rs6000_debug_secondary_memory_needed): Change argument names.
10844 (rs6000_output_move_128bit): New function to return the move to
10845 use for 128-bit moves, including knowing about the various
10846 limitations of quad memory operations.
10848 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
10849 memory operations. call rs6000_output_move_128bit for the actual
10850 instruciton(s) to generate.
10851 (vsx_movti_64bit): Likewise.
10853 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
10854 (UNSPEC_P8V_MTVSRWZ): Likewise.
10855 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
10856 (UNSPEC_P8V_MTVSRD): Likewise.
10857 (UNSPEC_P8V_XXPERMDI): Likewise.
10858 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
10859 (UNSPEC_FUSION_GPR): Likewise.
10860 (FMOVE128_GPR): New iterator for direct move.
10861 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
10862 (f32_sv): Likewise.
10863 (f32_dm): Likewise.
10864 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
10865 loads and direct move instructions.
10866 (zero_extendsidi2_lfiwzx): Likewise.
10867 (extendsidi2_lfiwax): Likewise.
10868 (extendsidi2_nocell): Likewise.
10869 (floatsi<mode>2_lfiwax): Likewise.
10870 (lfiwax): Likewise.
10871 (floatunssi<mode>2_lfiwzx): Likewise.
10872 (lfiwzx): Likewise.
10873 (fix_trunc<mode>_stfiwx): Likewise.
10874 (fixuns_trunc<mode>_stfiwx): Likewise.
10875 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
10876 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
10877 (parity<mode>2_cmpb): Set length/type attr.
10878 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
10879 for 'mr.' to fast_compare.
10880 (bpermd_<mode>): Change type attr to popcnt.
10881 (p8_fmrgow_<mode>): New insns for power8 direct move support.
10882 (p8_mtvsrwz_1): Likewise.
10883 (p8_mtvsrwz_2): Likewise.
10884 (reload_fpr_from_gpr<mode>): Likewise.
10885 (p8_mtvsrd_1): Likewise.
10886 (p8_mtvsrd_2): Likewise.
10887 (p8_xxpermdi_<mode>): Likewise.
10888 (reload_vsx_from_gpr<mode>): Likewise.
10889 (reload_vsx_from_gprsf): Likewise.
10890 (p8_mfvsrd_3_<mode>): LIkewise.
10891 (reload_gpr_from_vsx<mode>): Likewise.
10892 (reload_gpr_from_vsxsf): Likewise.
10893 (p8_mfvsrd_4_disf): Likewise.
10894 (multi-word GPR splits): Do not split direct moves or quad memory
10897 2013-06-10 David Malcolm <dmalcolm@redhat.com>
10899 * tree-into-ssa.c (interesting_blocks): Make static.
10901 2013-06-10 Jakub Jelinek <jakub@redhat.com>
10904 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
10905 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
10906 Use DATA_ABI_ALIGNMENT for that case instead if defined.
10907 (get_variable_align): New function.
10908 (get_variable_section, emit_bss, emit_common,
10909 assemble_variable_contents, place_block_symbol): Use
10910 get_variable_align instead of DECL_ALIGN.
10911 (assemble_noswitch_variable): Add align argument, use it
10912 instead of DECL_ALIGN.
10913 (assemble_variable): Adjust caller. Use get_variable_align
10914 instead of DECL_ALIGN.
10915 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
10917 (DATA_ABI_ALIGNMENT): Define.
10918 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
10919 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
10920 opt is false, only return the psABI mandated alignment increase.
10921 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
10922 (DATA_ABI_ALIGNMENT): ... this.
10923 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
10924 (DATA_ABI_ALIGNMENT): ... this.
10925 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
10926 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
10927 (DATA_ABI_ALIGNMENT): ... this.
10928 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
10929 * doc/tm.texi: Regenerated.
10931 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
10933 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
10934 cmp_code to construct REG_EQUAL note.
10936 2013-06-09 Jakub Jelinek <jakub@redhat.com>
10939 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
10940 that operands[2] doesn't overlap with operands[0].
10942 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
10943 Jan Hubicka <jh@suse.cz>
10945 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
10946 hack to mark symbols as used.
10948 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
10950 PR rtl-optimization/57559
10951 * lra-constraints.c (process_alt_operands): Don't discourage
10952 memory with known offset for offsetable memory constraint.
10953 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
10955 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
10957 * varasm.c (struct oc_local_state): Reorder fields.
10958 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
10959 and adjust accordingly.
10960 (output_constructor): Reorder initialization code and adjust call to
10961 output_constructor_bitfield.
10963 2013-06-07 Jan Hubicka <jh@suse.cz>
10965 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
10967 2013-06-07 David Malcolm <dmalcolm@redhat.com>
10969 * tree-object-size.c (unknown): Make const.
10971 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10973 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
10974 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
10975 for last alternative in the cpu_facility attribute.
10977 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10980 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
10981 (xordi3): Change operand 2 constraint to arm_xordi_operand.
10982 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
10983 * config/arm/constraints.md (Dg): New constraint.
10984 * config/arm/neon.md (xordi3_neon): Remove.
10985 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
10986 * config/arm/predicates.md (arm_xordi_operand): New predicate.
10988 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10990 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
10991 Clean up alternatives.
10993 2013-06-07 Alan Modra <amodra@gmail.com>
10995 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
10998 2013-06-07 Alan Modra <amodra@gmail.com>
11000 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
11002 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11004 * config/arm/constraints.md (Df): New constraint.
11005 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
11006 Correct length attribute for last two alternatives.
11008 2013-06-07 Alan Modra <amodra@gmail.com>
11010 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
11011 override user -mfp-in-toc.
11012 (offsettable_ok_by_alignment): Consider just the current access
11013 rather than the whole object, unless BLKmode. Handle
11014 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
11015 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
11016 for -mcmodel=medium.
11017 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
11018 override user -mfp-in-toc or -msum-in-toc. Default to
11019 -mno-fp-in-toc for -mcmodel=medium.
11021 2013-06-06 DJ Delorie <dj@redhat.com>
11023 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
11024 TARGET_VALID_POINTER_MODE.
11026 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
11027 Pat Haugen <pthaugen@us.ibm.com>
11028 Peter Bergner <bergner@vnet.ibm.com>
11030 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
11031 Document new power8 builtins.
11033 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
11034 condition code register, to allow 128-bit logical operations to be
11035 done in the VSX or GPR registers.
11036 (nor<mode>3): Use the canonical form for nor.
11037 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
11038 vclz*, and vpopcnt* vector instructions.
11039 (nand<mode>3): Likewise.
11040 (orc<mode>3): Likewise.
11041 (clz<mode>2): LIkewise.
11042 (popcount<mode>2): Likewise.
11044 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
11045 that only the GPRs are recognized.
11047 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
11048 support for new power8 builtins.
11050 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
11052 (xscvdpspn): Likewise.
11058 (vpopcnt): Likewise.
11059 (vpopcntb): Likewise.
11060 (vpopcnth): Likewise.
11061 (vpopcntw): Likewise.
11062 (vpopcntd): Likewise.
11064 (vmrgew): Likewise.
11065 (vmrgow): Likewise.
11067 (eqv_v16qi3): Likewise.
11068 (eqv_v8hi3): Likewise.
11069 (eqv_v4si3): Likewise.
11070 (eqv_v2di3): Likewise.
11071 (eqv_v4sf3): Likewise.
11072 (eqv_v2df3): Likewise.
11074 (nand_v16qi3): Likewise.
11075 (nand_v8hi3): Likewise.
11076 (nand_v4si3): Likewise.
11077 (nand_v2di3): Likewise.
11078 (nand_v4sf3): Likewise.
11079 (nand_v2df3): Likewise.
11081 (orc_v16qi3): Likewise.
11082 (orc_v8hi3): Likewise.
11083 (orc_v4si3): Likewise.
11084 (orc_v2di3): Likewise.
11085 (orc_v4sf3): Likewise.
11086 (orc_v2df3): Likewise.
11088 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
11089 allow power8 quad mode in 64-bit.
11090 (rs6000_builtin_vectorized_function): Add support to vectorize
11091 ISA 2.07 count leading zeros, population count builtins.
11092 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
11093 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
11094 (builtin_function_type): Add vgbbd builtin function which takes an
11096 (altivec_expand_vec_perm_const): Add support for new power8 merge
11099 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
11100 that does not include TImdoe for use with 32-bit.
11101 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
11103 (UNSPEC_VSX_CVDPSPN): Likewise.
11104 (vsx_xscvdpspn): Likewise.
11105 (vsx_xscvspdpn): Likewise.
11106 (vsx_xscvdpspn_scalar): Likewise.
11107 (vsx_xscvspdpn_directmove): Likewise.
11108 (vsx_and<mode>3): Split logical operations into 32-bit and
11109 64-bit. Add support to do logical operations on TImode as well as
11110 VSX vector types. Allow logical operations to be done in either
11111 VSX registers or in general purpose registers in 64-bit mode. Add
11112 splitters if GPRs were used. For AND, add clobber of CCmode to
11113 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
11115 (vsx_and<mode>3_32bit): Likewise.
11116 (vsx_and<mode>3_64bit): Likewise.
11117 (vsx_ior<mode>3): Likewise.
11118 (vsx_ior<mode>3_32bit): Likewise.
11119 (vsx_ior<mode>3_64bit): Likewise.
11120 (vsx_xor<mode>3): Likewise.
11121 (vsx_xor<mode>3_32bit): Likewise.
11122 (vsx_xor<mode>3_64bit): Likewise.
11123 (vsx_one_cmpl<mode>2): Likewise.
11124 (vsx_one_cmpl<mode>2_32bit): Likewise.
11125 (vsx_one_cmpl<mode>2_64bit): Likewise.
11126 (vsx_nor<mode>3): Likewise.
11127 (vsx_nor<mode>3_32bit): Likewise.
11128 (vsx_nor<mode>3_64bit): Likewise.
11129 (vsx_andc<mode>3): Likewise.
11130 (vsx_andc<mode>3_32bit): Likewise.
11131 (vsx_andc<mode>3_64bit): Likewise.
11132 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
11133 and xxlorc instructions.
11134 (vsx_eqv<mode>3_64bit): Likewise.
11135 (vsx_nand<mode>3_32bit): Likewise.
11136 (vsx_nand<mode>3_64bit): Likewise.
11137 (vsx_orc<mode>3_32bit): Likewise.
11138 (vsx_orc<mode>3_64bit): Likewise.
11140 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
11142 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
11144 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
11145 (p8_vmrgow): Likewise.
11146 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
11147 GPRs to be split under VSX.
11148 (p8v_clz<mode>2): Add power8 count leading zero support.
11149 (p8v_popcount<mode>2): Add power8 population count support.
11150 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
11153 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
11156 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
11158 (vec_nand): Likewise.
11159 (vec_vclz): Likewise.
11160 (vec_vclzb): Likewise.
11161 (vec_vclzd): Likewise.
11162 (vec_vclzh): Likewise.
11163 (vec_vclzw): Likewise.
11164 (vec_vgbbd): Likewise.
11165 (vec_vmrgew): Likewise.
11166 (vec_vmrgow): Likewise.
11167 (vec_vpopcnt): Likewise.
11168 (vec_vpopcntb): Likewise.
11169 (vec_vpopcntd): Likewise.
11170 (vec_vpopcnth): Likewise.
11171 (vec_vpopcntw): Likewise.
11173 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
11175 PR rtl-optimization/57468
11176 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
11179 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
11181 PR rtl-optimization/57459
11182 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
11183 type when setting live regs.
11185 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
11187 * config/s390/s390.opt (mlra): New option.
11188 * config/s390/s390.c (s390_decompose_address): Check displacement
11189 for all registers for LRA.
11190 (s390_secondary_reload): Don't used secondary reloads for LRA.
11191 (s390_lra_p): New function.
11192 (TARGET_LRA_P): Define.
11193 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
11194 of attribute cpu_facility to zarch for the last alternative.
11195 (*cmpmem_short): Ditto.
11197 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
11199 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
11200 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
11201 (arm_expand_prologue): Likewise.
11203 2013-06-06 Teresa Johnson <tejohnson@google.com>
11206 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
11207 as this is now done by redirect_edge_and_branch_force.
11208 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
11209 barriers, and fix interaction with splitting.
11210 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
11211 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
11212 reflect changes made in the routine.
11213 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
11214 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
11215 since this is called in cfglayout mode, and replace partition fixup
11216 with assert as that is now done by force_nonfallthru_and_redirect.
11217 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
11218 already be marked with region crossing note.
11219 (insert_section_boundary_note): Make non-static, gate on flag
11220 has_bb_partition, rewrite to also check for multiple partitions.
11221 (rest_of_handle_reorder_blocks): Remove call to
11222 insert_section_boundary_note, now done later during free_cfg.
11223 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
11224 * bb-reorder.h (insert_section_boundary_note): Declare.
11225 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
11226 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
11227 invoke insert_section_boundary_note.
11228 (try_redirect_by_replacing_jump): Remove unnecessary
11229 check for region crossing note.
11230 (fixup_partition_crossing): New function.
11231 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
11232 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
11233 in non-cfglayout mode.
11234 (force_nonfallthru_and_redirect): Fixup partition boundaries,
11235 remove old code that tried to do this. Emit barrier correctly
11236 when we are in cfglayout mode.
11237 (last_bb_in_partition): New function.
11238 (rtl_split_edge): Correctly fixup partition boundaries.
11239 (commit_one_edge_insertion): Remove old code that tried to
11240 fixup region crossing edge since this is now handled in
11241 split_block, and set up insertion point correctly since
11242 block may now end in a jump.
11243 (verify_hot_cold_block_grouping): Guard against checking when not in
11244 linearized RTL mode.
11245 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
11247 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
11248 rtl_verify_flow_info, so not called in cfglayout mode.
11249 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
11250 (fixup_reorder_chain): Remove old code that attempted to fixup region
11251 crossing note as this is now handled in force_nonfallthru_and_redirect.
11252 (duplicate_insn_chain): Don't duplicate switch section notes.
11253 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
11255 * basic-block.h (emit_barrier_after_bb): Declare.
11257 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11259 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
11260 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
11261 arm_usatsihi): Adjust alternatives for arm_restrict_it.
11263 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11265 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
11267 * config/arm/ldmstm.md: Regenerate.
11269 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11271 * config/arm/sync.md (atomic_loaddi_1):
11272 Disable predication for arm_restrict_it.
11273 (arm_load_exclusive<mode>): Likewise.
11274 (arm_load_exclusivesi): Likewise.
11275 (arm_load_exclusivedi): Likewise.
11276 (arm_load_acquire_exclusive<mode>): Likewise.
11277 (arm_load_acquire_exclusivesi): Likewise.
11278 (arm_load_acquire_exclusivedi): Likewise.
11279 (arm_store_exclusive<mode>): Likewise.
11280 (arm_store_exclusive<mode>): Likewise.
11281 (arm_store_release_exclusivedi): Likewise.
11282 (arm_store_release_exclusive<mode>): Likewise.
11284 2013-06-06 Richard Biener <rguenther@suse.de>
11286 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
11288 (lto_tag_is_tree_code_p): Adjust.
11289 (lto_tag_is_gimple_code_p): Likewise.
11290 (lto_gimple_code_to_tag): Likewise.
11291 (lto_tag_to_gimple_code): Likewise.
11292 (lto_tree_code_to_tag): Likewise.
11293 (lto_tag_to_tree_code): Likewise.
11294 * data-streamer.h (streamer_write_hwi_in_range): Use
11295 uhwi streaming to stream the normalized range.
11296 (streamer_read_hwi_in_range): Likewise.
11298 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11300 * config/arm/arm.md (enabled_for_depr_it): New attribute.
11301 (predicable_short_it): Likewise.
11302 (predicated): Likewise.
11303 (enabled): Handle above.
11304 (define_cond_exec): Set predicated attribute to yes.
11306 2013-06-05 Mike Stump <mikestump@comcast.net>
11308 * gdbinit.in (__FUNCTION__): Add.
11310 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
11312 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
11313 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
11315 2013-06-05 Jan Hubicka <jh@suse.cz>
11317 * varasm.c (mark_decl_referenced): Revert the removal until targets
11320 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
11322 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
11323 instead of mark_decl_referenced.
11325 2013-06-05 Jan Hubicka <jh@suse.cz>
11327 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
11328 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
11329 and symtab_used_from_object_file_p.
11330 (cgraph_make_node_local_1): Clear forced_by_abi.
11331 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
11332 * cgraph.h (symtab_node_base): Add forced_by_abi.
11333 (decide_is_variable_needed): Remove.
11334 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
11335 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
11336 (decide_is_symbol_needed): ... this one; handle symbols in general;
11337 always analyze virtuals; honnor forced_by_abi.
11338 (cgraph_finalize_function): Update.
11339 (varpool_finalize_decl): Update.
11340 (symbol_defined_and_needed): Remove.
11341 (analyze_functions): Update.
11342 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
11343 output_refs, input_overwrite_node): Handle forced_by_abi.
11344 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
11345 (address_taken_from_non_vtable_p): ... this one.
11346 (comdat_can_be_unshared_p_1): New function.
11347 (cgraph_comdat_can_be_unshared_p): Rename to ...
11348 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
11349 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
11350 (function_and_variable_visibility): Clear forced_by_abi as needed.
11351 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
11352 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
11353 * symtab.c (dump_symtab_base): Dump forced_by_abi.
11354 * varpool.c (decide_is_variable_needed): Remove.
11356 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11358 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
11359 (arm_option_override): Override arm_restrict_it where appropriate.
11360 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
11361 * config/arm/arm.opt (mrestrict-it): New command-line option.
11362 * doc/invoke.texi: Document -mrestrict-it.
11364 2013-06-05 David Malcolm <dmalcolm@redhat.com>
11366 * tsan.c (tsan_atomic_table): Make const.
11368 2013-06-05 Richard Biener <rguenther@suse.de>
11370 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
11371 index associated with the tree we are supposed to replace.
11372 * tree-streamer-out.c (pack_ts_base_value_fields): Output
11373 TREE_ASM_WRITTEN as zero for everything but SSA names.
11375 2013-06-05 David Malcolm <dmalcolm@redhat.com>
11377 * tree-ssa-structalias.c (call_stmt_vars): Make static.
11379 2013-06-04 Jan Hubicka <jh@suse.cz>
11381 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
11382 (input_node, input_varpool_node): Handle correctly external same
11384 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
11385 nodes at ltrans stage.
11387 2013-06-04 Jan Hubicka <jh@suse.cz>
11389 * ipa-inline.c (update_caller_keys): Fix availability test.
11390 (update_callee_keys): Likewise.
11391 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
11392 to follow ELF standard.
11394 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
11396 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
11397 (mips64r5900el-*-elf*): New configurations.
11398 * config/mips/mips-cpus.def (r5900): New processor.
11399 * config/mips/mips-tables.opt: Regenerate.
11400 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
11401 (mips_issue_rate): Handle PROCESSOR_R5900.
11402 (mips_reorg_process_insns): Force reorder mode for the R5900.
11403 * config/mips/mips.h (TARGET_MIPS5900): Define.
11404 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
11406 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
11408 * config/mips/mips.md (processor): Add r5900.
11409 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
11411 2013-06-04 Ian Bolton <ian.bolton@arm.com>
11413 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
11414 into function to generate MOVI instruction.
11415 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
11416 (aarch64_preferred_simd_mode): Turn into wrapper.
11417 (aarch64_output_scalar_simd_mov_immediate): New function.
11418 * config/aarch64/aarch64-protos.h: Add prototype for above.
11420 2013-06-04 Ian Bolton <ian.bolton@arm.com>
11422 * config/aarch64/aarch64.c (simd_immediate_info): Remove
11423 element_char member.
11424 (sizetochar): Return signed char.
11425 (aarch64_simd_valid_immediate): Remove elchar and other
11426 unnecessary variables.
11427 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
11428 Calculate element_char as required.
11429 * config/aarch64/aarch64-protos.h: Update and move prototype
11430 for aarch64_output_simd_mov_immediate.
11431 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
11434 2013-06-04 Ian Bolton <ian.bolton@arm.com>
11436 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
11437 information completed by aarch64_simd_valid_immediate.
11438 (aarch64_legitimate_constant_p): Update arguments.
11439 (aarch64_simd_valid_immediate): Work with struct rather than many
11441 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
11442 (aarch64_simd_make_constant): Update arguments.
11443 (aarch64_output_simd_mov_immediate): Work with struct rather than
11444 many pointers. Output immediate directly rather than as operand.
11445 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
11447 * config/aarch64/constraints.md (Dn): Update arguments.
11449 2013-06-04 Ian Bolton <ian.bolton@arm.com>
11451 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
11453 (aarch64_simd_immediate_valid_for_move): Remove.
11454 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
11455 (aarch64_simd_make_constant): Update call.
11456 (aarch64_output_simd_mov_immediate): Update call.
11457 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
11459 * config/aarch64/constraints.md (Dn): Update call.
11461 2013-06-04 Ian Bolton <ian.bolton@arm.com>
11463 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
11464 return type to bool for prototype.
11465 (aarch64_legitimate_constant_p): Check for true instead of not -1.
11466 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
11467 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
11469 2013-06-04 Catherine Moore <clm@codesourcery.com>
11471 * config/mips/mips.opt (meva): New.
11472 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
11473 (ASM_SPEC): Handle -meva.
11474 * doc/invoke.texi (meva): Document.
11476 2013-06-04 Alan Modra <amodra@gmail.com>
11478 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
11481 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11483 * rtl.def: Add extra fourth optional field to define_cond_exec.
11484 * gensupport.c (process_one_cond_exec): Process attributes from
11486 * doc/md.texi: Document fourth field in define_cond_exec.
11488 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
11490 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
11491 out the processing order as in store_bit_field_1.
11493 2013-06-04 Jan Hubicka <jh@suse.cz>
11495 PR middle-end/57500
11496 * cgraphunit.c (cgraph_process_same_body_aliases): Create
11497 non-VAR_DECL node if it does not exist yet.
11499 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
11501 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
11502 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
11503 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
11504 target_cpu_default setting.
11506 2013-06-03 Teresa Johnson <tejohnson@google.com>
11508 * dumpfile.c (opt_info_switch_p): Change -fopt-info
11509 default to -fopt-info=optimized instead of all.
11510 * doc/invoke.texi: Ditto.
11511 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
11512 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
11513 (execute_vect_slp): Emit BB vectorization success under
11514 MSG_OPTIMIZED_LOCATIONS.
11515 * tree-vect-slp.c (vect_slp_transform_bb): Change
11516 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
11517 * tree-vect-loop.c (vect_transform_loop): Ditto.
11519 2013-06-03 Jason Merrill <jason@redhat.com>
11522 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
11523 Use TARGET_EXPR for C++.
11525 2013-06-03 Jakub Jelinek <jakub@redhat.com>
11527 PR rtl-optimization/57268
11528 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
11529 if DEBUG_INSN_P (insn).
11532 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
11534 PR rtl-optimization/57268
11535 * sched-deps.c (sched_analyze_2): Flush dependence lists if
11536 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
11538 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
11540 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
11541 (ix86_avoid_lea_for_addr): Likewise.
11542 (exact_dependency_1): Likewise.
11543 (ix86_adjust_cost): Likewise.
11544 (swap_top_of_ready_list): Fix formatting and !reload_completed check
11546 (do_reorder_for_imul): Fix typo, formatting and
11547 !reload_completed check removed.
11548 (ix86_sched_reorder): Fix typo and formatting.
11549 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
11552 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
11554 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
11556 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
11558 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
11559 <STRING_CST>: Likewise.
11560 <VECTOR_CST>: Likewise.
11562 2013-06-01 Janus Weil <janus@gcc.gnu.org>
11563 Mikael Morin <mikael@gcc.gnu.org>
11565 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
11566 * config.in: Regenerated.
11567 * configure: Regenerated.
11569 2013-06-01 Jan Hubicka <jh@suse.cz>
11571 PR middle-end/57366
11572 * cgraphunit.c (compile): When weakref is not supported,
11573 set up transparent aliases before final output pass.
11574 * varasm.c (assemble_alias): Do not try to do it here.
11576 2013-06-01 Jan Hubicka <jh@suse.cz>
11578 PR middle-end/57467
11579 * passes.c (for_per_function): Skip unanalyzed functions.
11581 2013-06-01 Jan Hubicka <jh@suse.cz>
11583 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
11584 (lto_symtab_merge_symbols_1): ... this one.
11585 (lto_symtab_merge_cgraph_nodes): Rename to ...
11586 (lto_symtab_merge_symbols): ... this one; simplify.
11587 * cgraph.c (same_body_aliases_done): Rename to ...
11588 (cpp_implicit_aliases_done): ... this one.
11589 (cgraph_create_function_alias): Update.
11590 (cgraph_same_body_alias): Update.
11591 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
11592 (verify_edge_corresponds_to_fndecl): Simplify.
11593 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
11594 (cgraph_node): Remove same_body_alias.
11595 (varpool_node): Remove alias_of and extra_name_alias.
11596 (same_body_aliases_done): Rename to ..
11597 (cpp_implicit_aliases_done): ... this one.
11598 (symtab_alias_ultimate_target): Add default parameter.
11599 (symtab_resolve_alias): New function.
11600 (fixup_same_cpp_alias_visibility): Declare.
11601 (cgraph_function_node): Add default parameter.
11602 (cgraph_node_asm_name): Likewise.
11603 (cgraph_function_or_thunk_node): Add default parameter; do
11604 not ICE when it is NULL.
11605 (varpool_variable_node): Likewise.
11606 * tree-emutls.c (create_emultls_var): Update.
11607 (ipa_lower_emutls): Update.
11608 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
11609 (cgraph_reset_node): Reset alias info.
11610 (cgraph_finalize_function): Update.
11611 (fixup_same_cpp_alias_visibility): Move to symtab.c.
11612 (analyze_function): Simplify.
11613 (cgraph_process_same_body_aliases): Simplify.
11614 (analyze_functions): Fixup same body aliases.
11615 (handle_alias_pairs): Simplify.
11616 (assemble_thunk): Update.
11617 (assemble_thunks_and_aliases): Update.
11618 (output_weakrefs): Rewrite.
11619 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
11620 (lto_output_varpool_node): Likewise.
11621 (compute_ltrans_boundary): Remve assert.
11622 (get_alias_symbol): New functoin.
11623 (input_node): Rewrite alias handling.
11624 (input_varpool_node): Likewise.
11625 * ipa-pure-const.c (propagate_pure_const): Fix formating.
11626 * ipa.c (process_references): Handle weakrefs correctly.
11627 (symtab_remove_unreachable_nodes): Likewise.
11628 * trans-mem.c (get_cg_data): Update.
11629 (ipa_tm_create_version_alias): Update.
11630 (ipa_tm_execute): Update.
11631 * symtab.c (dump_symtab_base): Dump aliases.
11632 (verify_symtab_base): Verify aliases.
11633 (symtab_node_availability): New function.
11634 (symtab_alias_ultimate_target): Simplify.
11635 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
11636 handle all the fixup cases.
11637 (symtab_resolve_alias): New function.
11638 * passes.c (ipa_write_summaries): Handle weakrefs.
11639 * varpool.c (varpool_analyze_node): Simplify.
11640 (assemble_aliases): Update.
11641 (varpool_create_variable_alias): Simplify.
11642 (varpool_extra_name_alias): Simplify.
11643 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
11644 (lto_symtab_merge_symbols): ... this one.
11646 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
11649 PR rtl-optimization/57268
11650 * sched-deps.c (sched_analyze_2): Flush dependence lists if
11651 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
11653 2013-06-01 Tobias Burnus <burnus@net-b.de>
11655 Partially reverted:
11656 2013-05-31 Tobias Burnus <burnus@net-b.de>
11658 PR middle-end/57073
11659 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
11662 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
11664 PR rtl-optimization/57268
11665 * sched-deps.c (sched_analyze_2): Flush dependence lists if
11666 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
11668 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
11670 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
11671 unordered comparison operators when -fno-trapping-math is in effect
11673 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
11674 and implement unordered comparison operators properly on the e500.
11676 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
11678 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
11679 for constant scalar integers.
11680 (simplify_relational_operation_1): Likewise.
11682 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
11684 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
11685 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
11688 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
11689 Igor Zamyatin <igor.zamyatin@intel.com>
11691 Silvermont (SLM) architecture performance tuning.
11692 * config/i386/i386.h (enum ix86_tune_indices): Add
11693 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
11694 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
11696 * config/i386/i386.c (initial_ix86_tune_features)
11697 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
11698 (ix86_lea_outperforms): Handle Silvermont tuning.
11699 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
11701 (ix86_use_lea_for_mov): Likewise.
11702 (ix86_avoid_lea_for_addr): Likewise.
11703 (ix86_lea_for_add_ok): Likewise.
11704 (exact_dependency_1): New function.
11705 (exact_store_load_dependency): Likewise.
11706 (ix86_adjust_cost): Handle Silvermont tuning.
11707 (do_reoder_for_imul): Likewise.
11708 (swap_top_of_ready_list): New function.
11709 (ix86_sched_reorder): Changed to handle Silvermont tuning.
11711 * config/i386/i386.md (peepholes that split memory operand in fp
11714 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
11716 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
11717 Remove un-necessary braces.
11719 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
11721 * config/aarch64/aarch64.c (aarch64_classify_symbol):
11722 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
11724 2013-05-31 Tobias Burnus <burnus@net-b.de>
11726 PR middle-end/57073
11727 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
11729 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11732 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
11733 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
11734 * config/arm/neon.md (iordi3_neon): Remove.
11735 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
11736 * config/arm/predicates.md (imm_for_neon_logic_operand):
11737 Move to earlier in the file.
11738 (neon_logic_op2): Likewise.
11739 (arm_iordi_operand_neon): New predicate.
11741 2013-05-31 Richard Biener <rguenther@suse.de>
11743 PR tree-optimization/57478
11744 PR tree-optimization/57453
11745 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
11748 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
11750 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
11751 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
11753 2013-05-30 Tobias Burnus <burnus@net-b.de>
11754 Thomas Koenig <tkoenig@gcc.gnu.org>
11756 PR middle-end/57073
11757 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
11758 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
11760 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
11762 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
11764 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
11766 * target.def (register_usage_leveling_p): New hook.
11767 * targhooks.c (default_register_usage_leveling_p): New.
11768 * targhooks.h (default_register_usage_leveling_p): New prototype.
11769 * lra-assigns.c (register_usage_leveling_p): Use the hook.
11770 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
11771 * doc/tm.texi: Update.
11772 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
11774 2013-05-30 Ian Bolton <ian.bolton@arm.com>
11776 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
11777 (*insv_reg<mode>): New define_insn.
11779 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
11781 PR rtl-optimization/57439
11782 * postreload.c (move2add_valid_value_p): Check that we have
11783 a zero subreg_regno_offset when accessing the register in
11784 the requested mode.
11786 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
11787 Igor Zamyatin <igor.zamyatin@intel.com>
11789 Silvermont (SLM) architecture pipeline model, tuning and
11791 * config.gcc: Add slm config options and target.
11793 * config/i386/slm.md: New.
11795 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
11797 * config/i386/i386-c.c (ix86_target_macros_internal): New case
11799 (ix86_target_macros_internal): Likewise.
11801 * config/i386/i386.c (slm_cost): New cost.
11802 (m_SLM): New macro flag.
11803 (initial_ix86_tune_features): Set m_SLM.
11804 (x86_accumulate_outgoing_args): Likewise.
11805 (x86_arch_always_fancy_math_387): Likewise.
11806 (processor_target_table): Add slm cost.
11807 (cpu_names): Add slm cpu name.
11808 (x86_option_override_internal): Set SLM ISA.
11809 (ix86_issue_rate): New case PROCESSOR_SLM.
11810 (ia32_multipass_dfa_lookahead): Likewise.
11811 (fold_builtin_cpu): Add slm.
11813 * config/i386/i386.h (TARGET_SLM): New target macro.
11814 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
11815 (processor_type): Add PROCESSOR_SLM.
11817 * config/i386/i386.md (cpu): Add new value "slm".
11818 (slm.md): Include slm.md.
11820 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
11821 Zhenqiang Chen <zhenqiang.chen@linaro.org>
11823 * config/arm/arm-protos.h: Add and update function protos.
11824 * config/arm/arm.c (use_simple_return_p): New added.
11825 (thumb2_expand_return): Check simple_return flag.
11826 * config/arm/arm.md: Add simple_return and conditional simple_return.
11827 * config/arm/iterators.md: Add iterator for return and simple_return.
11829 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
11831 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
11832 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
11833 (arm_emit_vfp_multi_reg_pop): Likewise.
11834 (thumb2_emit_ldrd_pop): Likewise.
11835 (arm_expand_epilogue): Add misc REG_CFA notes.
11836 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
11838 2013-05-29 Lawrence Crowl <crowl@google.com>
11840 * config/arm/t-arm: Update for below.
11842 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
11843 Change type to hash_table. Update dependent calls and types.
11845 * config/i386/t-cygming: Update for below.
11847 * config/i386/t-interix: Update for below.
11849 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
11850 Change type to hash_table. Update dependent calls and types.
11851 (i386_find_on_wrapper_list::wrappers): Likewise.
11853 * config/ia64/t-ia64: Update for below.
11855 * config/ia64/ia64.c (bundle_state_table):
11856 Change type to hash_table. Update dependent calls and types.
11858 * config/mips/mips.c (mips_reorg_process_insns::htab):
11859 Change type to hash_table. Update dependent calls and types.
11861 * config/sol2.c (solaris_comdat_htab):
11862 Change type to hash_table. Update dependent calls and types.
11864 * config/t-sol2: Update for above.
11866 2013-05-29 Teresa Johnson <tejohnson@google.com>
11868 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
11869 functions are not yet marked as defined.
11871 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
11872 Pat Haugen <pthaugen@us.ibm.com>
11873 Peter Bergner <bergner@vnet.ibm.com>
11875 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
11879 (vrotl<mode>3): Likewise.
11880 (vashl<mode>3): Likewise.
11881 (vlshr<mode>3): Likewise.
11882 (vashr<mode>3): Likewise.
11884 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
11885 support for power8 V2DI builtins.
11887 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
11888 power8 V2DI builtins.
11889 (vupkhsw): Likewise.
11890 (vupklsw): Likewise.
11891 (vaddudm): Likewise.
11892 (vminsd): Likewise.
11893 (vmaxsd): Likewise.
11894 (vminud): Likewise.
11895 (vmaxud): Likewise.
11896 (vpkudum): Likewise.
11897 (vpksdss): Likewise.
11898 (vpkudus): Likewise.
11899 (vpksdus): Likewise.
11904 (vsubudm): Likewise.
11905 (vcmpequd): Likewise.
11906 (vcmpgtsd): Likewise.
11907 (vcmpgtud): Likewise.
11908 (vcmpequd_p): Likewise.
11909 (vcmpgtsd_p): Likewise.
11910 (vcmpgtud_p): Likewise.
11911 (vupkhsw): Likewise.
11912 (vupklsw): Likewise.
11913 (vaddudm): Likewise.
11914 (vmaxsd): Likewise.
11915 (vmaxud): Likewise.
11916 (vminsd): Likewise.
11917 (vminud): Likewise.
11918 (vpksdss): Likewise.
11919 (vpksdus): Likewise.
11920 (vpkudum): Likewise.
11921 (vpkudus): Likewise.
11926 (vsubudm): Likewise.
11928 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
11929 support for power8 V2DI instructions.
11931 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
11932 power8 V2DI instructions. Combine pack and unpack insns to use an
11933 iterator for each mode. Check whether a particular mode supports
11934 Altivec instructions instead of just checking TARGET_ALTIVEC.
11935 (UNSPEC_VPKUWUM): Likewise.
11936 (UNSPEC_VPKSHSS): Likewise.
11937 (UNSPEC_VPKSWSS): Likewise.
11938 (UNSPEC_VPKUHUS): Likewise.
11939 (UNSPEC_VPKSHUS): Likewise.
11940 (UNSPEC_VPKUWUS): Likewise.
11941 (UNSPEC_VPKSWUS): Likewise.
11942 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
11943 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
11944 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
11945 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
11946 (UNSPEC_VUPKHSB): Likewise.
11947 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
11948 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
11949 (UNSPEC_VUPKHSH): Likewise.
11950 (UNSPEC_VUPKLSB): Likewise.
11951 (UNSPEC_VUPKLSH): Likewise.
11953 (VI_char): Likewise.
11954 (VI_scalar): Likewise.
11955 (VI_unit): Likewise.
11957 (VP_small): Likewise.
11958 (VP_small_lc): Likewise.
11959 (VU_char): Likewise.
11960 (add<mode>3): Likewise.
11961 (altivec_vaddcuw): Likewise.
11962 (altivec_vaddu<VI_char>s): Likewise.
11963 (altivec_vadds<VI_char>s): Likewise.
11964 (sub<mode>3): Likewise.
11965 (altivec_vsubcuw): Likewise.
11966 (altivec_vsubu<VI_char>s): Likewise.
11967 (altivec_vsubs<VI_char>s): Likewise.
11968 (altivec_vavgs<VI_char>): Likewise.
11969 (altivec_vcmpbfp): Likewise.
11970 (altivec_eq<mode>): Likewise.
11971 (altivec_gt<mode>): Likewise.
11972 (altivec_gtu<mode>): Likewise.
11973 (umax<mode>3): Likewise.
11974 (smax<mode>3): Likewise.
11975 (umin<mode>3): Likewise.
11976 (smin<mode>3): Likewise.
11977 (altivec_vpkuhum): Likewise.
11978 (altivec_vpkuwum): Likewise.
11979 (altivec_vpkshss): Likewise.
11980 (altivec_vpkswss): Likewise.
11981 (altivec_vpkuhus): Likewise.
11982 (altivec_vpkshus): Likewise.
11983 (altivec_vpkuwus): Likewise.
11984 (altivec_vpkswus): Likewise.
11985 (altivec_vpks<VI_char>ss): Likewise.
11986 (altivec_vpks<VI_char>us): Likewise.
11987 (altivec_vpku<VI_char>us): Likewise.
11988 (altivec_vpku<VI_char>um): Likewise.
11989 (altivec_vrl<VI_char>): Likewise.
11990 (altivec_vsl<VI_char>): Likewise.
11991 (altivec_vsr<VI_char>): Likewise.
11992 (altivec_vsra<VI_char>): Likewise.
11993 (altivec_vsldoi_<mode>): Likewise.
11994 (altivec_vupkhsb): Likewise.
11995 (altivec_vupkhs<VU_char>): Likewise.
11996 (altivec_vupkls<VU_char>): Likewise.
11997 (altivec_vupkhsh): Likewise.
11998 (altivec_vupklsb): Likewise.
11999 (altivec_vupklsh): Likewise.
12000 (altivec_vcmpequ<VI_char>_p): Likewise.
12001 (altivec_vcmpgts<VI_char>_p): Likewise.
12002 (altivec_vcmpgtu<VI_char>_p): Likewise.
12003 (abs<mode>2): Likewise.
12004 (vec_unpacks_hi_v16qi): Likewise.
12005 (vec_unpacks_hi_v8hi): Likewise.
12006 (vec_unpacks_lo_v16qi): Likewise.
12007 (vec_unpacks_hi_<VP_small_lc>): Likewise.
12008 (vec_unpacks_lo_v8hi): Likewise.
12009 (vec_unpacks_lo_<VP_small_lc>): Likewise.
12010 (vec_pack_trunc_v8h): Likewise.
12011 (vec_pack_trunc_v4si): Likewise.
12012 (vec_pack_trunc_<mode>): Likewise.
12014 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
12016 (vec_vmaxsd): Likewise.
12017 (vec_vmaxud): Likewise.
12018 (vec_vminsd): Likewise.
12019 (vec_vminud): Likewise.
12020 (vec_vpksdss): Likewise.
12021 (vec_vpksdus): Likewise.
12022 (vec_vpkudum): Likewise.
12023 (vec_vpkudus): Likewise.
12024 (vec_vrld): Likewise.
12025 (vec_vsld): Likewise.
12026 (vec_vsrad): Likewise.
12027 (vec_vsrd): Likewise.
12028 (vec_vsubudm): Likewise.
12029 (vec_vupkhsw): Likewise.
12030 (vec_vupklsw): Likewise.
12032 2013-05-29 Jan Hubicka <jh@suse.cz>
12034 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
12035 flags; reorder rest of fields in more consistent way.
12036 (varpool_node): Remove analyzed, finalized and alias.
12037 (cgraph_ndoe): Likewise.
12038 (symtab_alias_ultimate_target): New function.
12039 (cgraph_function_node): Move offline.
12040 (cgraph_reset_node): Declare.
12041 (cgraph_comdat_can_be_unshared_p): Remove.
12042 (varpool_remove_initializer): Declare.
12043 (varpool_first_defined_variable, varpool_next_defined_variable
12044 cgraph_first_defined_function, cgraph_next_defined_function): Update.
12045 (cgraph_function_with_gimple_body_p): Update.
12046 (varpool_all_refs_explicit_p): Update.
12047 (symtab_alias_target): New function.
12048 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
12049 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
12050 (cgraph_function_or_thunk_node): Simplify using
12051 symtab_alias_ultimate_target.
12052 (varpool_variable_node): Likewise.
12053 * cgraph.c (cgraph_create_function_alias): Update.
12054 (cgraph_add_thunk): Update.
12055 (cgraph_remove_node): Update.
12056 (dump_cgraph_node): Do not dump removed flags.
12057 (cgraph_function_body_availability): Update.
12058 (cgraph_propagate_frequency): Update.
12059 (verify_cgraph_node): Check sanity of local flag.
12060 (cgraph_function_node): Move here from cgraph.h; revamp for
12061 cgraph_function_or_thunk_node.
12062 * lto-symtab.c (lto_varpool_replace_node): Update.
12063 (lto_symtab_resolve_can_prevail_p): Update.
12064 (lto_symtab_merge_cgraph_nodes): Update.
12065 * ipa-cp.c (determine_versionability, initialize_node_lattices,
12066 propagate_constants_accross_call, devirtualization_time_bonus,
12067 ipcp_propagate_stage): Update.
12068 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
12069 * ipa-inline-transform.c (clone_inlined_nodes,
12070 preserve_function_body_p): Update.
12071 * ipa-reference.c (propagate): Update.
12072 (write_node_summary_p): Update.
12073 * toplev.c (wrapup_global_declaration_2): Update.
12074 * cgraphunit.c (cgraph_analyze_function): Rename to ...
12075 (analyze_function) ... this one.
12076 (cgraph_process_new_functions): Update.
12077 (cgraph_reset_node): Export.
12078 (cgraph_finalize_function): Update.
12079 (cgraph_add_new_function): Update.
12080 (process_function_and_variable_attributes): Update.
12081 (varpool_finalize_decl): Update.
12082 (symbol_finalized): Remove.
12083 (symbol_finalized_and_needed): Rename to ...
12084 (symbol_defined_and_needed): ... update.
12085 (cgraph_analyze_functions): Update.
12086 (handle_alias_pairs): Update.
12087 (mark_functions_to_output): Update.
12088 (assemble_thunk): Update.
12089 (output_in_order): Update.
12090 (output_weakrefs): Update.
12091 (finalize_compilation_unit): Update.
12092 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
12093 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
12094 input_node, input_varpool_node): Update.
12095 * dbxout.c (dbxout_expand_expr): Update.
12096 * cgraphclones.c (cgraph_clone_node): Update.
12097 (cgraph_copy_node_for_versioning): Update.
12098 (cgraph_materialize_clone): Update.
12099 (cgraph_materialize_all_clones): Update.
12100 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
12101 propagate_pure_const, propagate_nothrow): Update.
12102 * lto-streamer-out.c (lto_output, write_symbol): Update.
12103 * ipa-utils.c (ipa_reverse_postorder): Update.
12104 * ipa-inline.c (can_inline_edge_p): Update.
12105 (update_caller_keys, ipa_inline): Update.
12106 * dwarf2out.c (reference_to_unused,
12107 premark_types_used_by_global_vars_helper): Update.
12108 * tree-eh.c (tree_could_trap_p): Update.
12109 * ipa-split.c (consider_split, execute_split_functions): Update.
12110 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
12111 has_addr_references_p): Update; move ahead in file for better
12113 (process_references): Simplify.
12114 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
12115 bodies are removed.
12116 (cgraph_comdat_can_be_unshared_p): Make static.
12117 (cgraph_externally_visible_p): Update.
12118 (varpool_externally_visible_p): Update.
12119 (function_and_variable_visibility): Update.
12120 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
12121 ipa_tm_mark_force_output_node): Update.
12122 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
12123 estimate_edge_devirt_benefit, inline_generate_summary,
12124 inline_write_summary): Update.
12125 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
12126 * ipa-prop.c (ipa_compute_jump_functions): Update.
12127 (ipa_print_node_params, ipa_prop_read_section,
12128 ipa_update_after_lto_read, read_replacements_section): Update.
12129 * varasm.c (mark_decl_referenced): Update.
12130 (assemble_alias, dump_tm_clone_pairs): Update.
12131 * tree-inline.c (copy_bb): Update.
12132 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
12134 * symtab.c (dump_symtab_base): Print new flags.
12135 (verify_symtab_base): Verify new flags.
12136 (symtab_alias_ultimate_target): New function.
12137 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
12138 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
12140 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
12142 * i386.c (ix86_get_function_versions_dispatcher,
12143 ix86_generate_version_dispatcher_body): Update.
12144 (fold_builtin_cpu): Use varpool_add_new_variable.
12145 * varpool.c (varpool_remove_initializer): Break out from ...
12146 (varpool_remove_node): ... this one.
12147 (dump_varpool_node, varpool_node_for_asm,
12148 cgraph_variable_initializer_availability, varpool_analyze_node,
12149 varpool_assemble_decl, varpool_remove_unreferenced_decls,
12150 varpool_finalize_named_section_flags, varpool_create_variable_alias):
12153 2013-05-29 Jan Hubicka <jh@suse.cz>
12155 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
12157 2013-05-29 Easwaran Raman <eraman@google.com>
12159 PR tree-optimization/57442
12160 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
12161 when control exits the main loop.
12163 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
12165 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
12167 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
12168 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
12169 * rx/t-rx: Add rx100 under multi library matches option for nofpu
12172 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12174 PR tree-optimization/57441
12175 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
12176 Don't limit size of incr_vec to number of candidates.
12178 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
12180 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
12181 and mips16 directories.
12182 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
12183 (MULTILIB_DIRNAMES): Ditto.
12184 (MULTILIB_EXCEPTIONS): Add new exceptions.
12185 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
12186 (MULTILIB_DIRNAMES): Ditto.
12187 (MULTILIB_EXCEPTIONS): Add new exceptions.
12189 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
12190 Marcus Shawcroft <marcus.shawcroft@arm.com>
12192 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
12193 SYMBOL_TINY_ABSOLUTE.
12194 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
12195 SYMBOL_TINY_ABSOLUTE.
12196 (aarch64_expand_mov_immediate): Likewise.
12197 (aarch64_classify_symbol): Likewise.
12198 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
12199 Permit SYMBOL_TINY_ABSOLUTE.
12200 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
12202 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
12203 Marcus Shawcroft <marcus.shawcroft@arm.com>
12205 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
12206 Refactor if/switch. Replace gcc_assert with if.
12208 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
12210 * config/i386/i386.c (initial_ix86_tune_features): Enable
12211 FP Reassociation for AMD bdver1 and bdver2.
12213 2013-05-29 Martin Jambor <mjambor@suse.cz>
12215 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
12216 and IMAGPART_EXPR do not occur within other handled_components.
12218 2013-05-29 Richard Biener <rguenther@suse.de>
12220 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
12221 access on whether the use is in the BB we currently try to
12223 (vect_bb_vectorization_profitable_p): Pass the BB we currently
12224 vectorize to vect_bb_slp_scalar_cost.
12226 2013-05-29 Richard Biener <rguenther@suse.de>
12228 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
12229 computing scalar cost offsetted by stmts that are kept live
12231 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
12232 for computation of scalar cost.
12234 2013-05-28 Steve Ellcey <sellcey@mips.com>
12236 * config/mips/mips-cpus.def (mips32r2): Change processor type.
12238 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
12240 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
12241 array notation built-in reduction functions.
12242 * doc/passes.texi (Passes): Added documentation about changes done
12244 * doc/invoke.texi (C Dialect Options): Added documentation about
12245 the -fcilkplus flag.
12246 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
12247 (BUILTINS_DEF): Depend on cilkplus.def.
12248 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
12249 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
12250 * cilkplus.def: New file.
12252 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
12254 PR rtl-optimization/57439
12255 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
12257 2013-05-28 Easwaran Raman <eraman@google.com>
12259 PR tree-optimization/57337
12260 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
12261 (find_insert_point): Correctly identify the insertion point
12262 when two statements with the same UID is compared.
12264 2013-05-28 Richard Biener <rguenther@suse.de>
12266 PR tree-optimization/56787
12267 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
12268 from the list of data references.
12269 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
12271 (vect_analyze_loop_operations): Likewise.
12272 (vect_transform_loop): Remove clobbers.
12274 2013-05-28 Martin Jambor <mjambor@suse.cz>
12276 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
12277 and REALPART_EXPRs have scalar type.
12279 2013-05-28 Richard Biener <rguenther@suse.de>
12281 PR tree-optimization/57411
12282 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
12284 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
12285 virtual operand propagation.
12287 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
12289 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
12290 destination register for bmasksi_vis.
12291 (vector_init_bshuffle): Likewise.
12292 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
12294 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
12296 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
12297 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
12298 mode if the instruction isn't available in the original mode.
12299 * config/sparc/sparc.opt (mfix-ut699): New option.
12300 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
12301 (divdf3): Turn into expander.
12302 (divdf3_nofix): New insn.
12303 (divdf3_fix): Likewise.
12304 (divsf3): Disable if -mfix-ut699.
12305 (sqrtdf2): Turn into expander.
12306 (sqrtdf2_nofix): New insn.
12307 (sqrtdf2_fix): Likewise.
12308 (sqrtsf2): Disable if -mfix-ut699.
12310 2013-05-27 Richard Biener <rguenther@suse.de>
12312 PR middle-end/57412
12313 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
12314 block for the new loop.
12316 2013-05-27 Richard Biener <rguenther@suse.de>
12318 PR tree-optimization/57343
12319 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
12320 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
12321 (number_of_iterations_cond): Do not build the folded tree.
12323 2013-05-27 Richard Biener <rguenther@suse.de>
12326 PR middle-end/57381
12327 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
12328 OEP_CONSTANT_ADDRESS_OF retained.
12330 PR tree-optimization/57417
12331 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
12332 for unchanged base.
12333 (set_ssa_val_to): Compare addresses using
12334 get_addr_base_and_unit_offset.
12336 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
12338 PR rtl-optimization/56833
12339 * postreload.c (move2add_record_mode): New function.
12340 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
12341 (move2add_use_add2_insn): Use move2add_record_sym_value.
12342 (move2add_use_add3_insn): Likewise.
12343 (reload_cse_move2add): Use move2add_valid_value_p and
12344 move2add_record_mode. Invalidate call-clobbered and REG_INC
12345 affected regs by setting reg_mode to VOIDmode.
12346 (move2add_note_store): Don't pretend the inside of a SUBREG is
12347 the actual destination. Invalidate single/leading registers by
12348 setting reg_mode to VOIDmode.
12349 Use move2add_record_sym_value, move2add_valid_value_p and
12350 move2add_record_mode.
12352 2013-05-27 Richard Biener <rguenther@suse.de>
12354 PR tree-optimization/57396
12355 * tree-affine.c (double_int_constant_multiple_p): Properly
12356 return false for val == 0 and div != 0.
12358 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
12360 * config/mips/mips.h: Use #elif in preprocessor conditions.
12362 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
12365 * config/mips/constraints.md (kl): New constraint.
12366 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
12367 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
12368 constraint for operand 0. Split after CSE for MIPS16. Emit a move
12369 from LO for MIPS16.
12370 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
12372 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
12375 * config/mips/mips.c (mips_can_inline_p): New function.
12376 (TARGET_CAN_INLINE_P): Define.
12378 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
12380 * sched-int.h (ds_t, dw_t): Make unsigned int.
12381 Fix documentation that describes how all the ds_t bits are used.
12382 Reserve the last bit for delayed-branch scheduling.
12383 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
12384 (BITS_PER_DEP_WEAK): Fix definition and documentation.
12385 (gen_dep_weak_1): Remove prototype.
12386 * sched-deps.c (get_dep_weak_1): Make static.
12387 * target.def (speculate_insn, needs_block_p, gen_spec_check,
12388 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
12389 * doc/tm.texi: Regenerate.
12390 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
12392 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
12395 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
12397 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
12398 Sandra Loosemore <sandra@codesourcery.com>
12400 * config.gcc (powerpc-*): Allow native for with-cpu.
12402 2013-05-24 Jeff Law <law@redhat.com>
12404 PR tree-optimization/57124
12405 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
12406 conversion feeding a condition if the range has an overflow
12407 if -fstrict-overflow. Add warnings for when we do make the
12410 2013-05-24 Dehao Chen <dehao@google.com>
12412 * tree-cfg.c (locus_discrim_map): Fix the typo.
12413 (locus_discrim_hasher): Likewise.
12414 (locus_discrim_hasher::hash): Likewise.
12415 (locus_discrim_hasher::equal): Likewise.
12417 2013-05-24 Martin Jambor <mjambor@suse.cz>
12419 PR tree-optimization/57294
12420 * cgraph.h (ipa_record_stmt_references): Declare.
12421 * cgraphbuild.c (ipa_record_stmt_references): New function.
12422 (build_cgraph_edges): Use ipa_record_stmt_references.
12423 (rebuild_cgraph_edges): Likewise.
12424 (cgraph_rebuild_references): Likewise.
12425 * ipa-prop.c (ipa_modify_call_arguments): Discard references
12426 associated with the old statement and build references from the
12427 newly built statements.
12428 * ipa-ref.c (ipa_remove_stmt_references): New function.
12429 * ipa-ref.h (ipa_remove_stmt_references): Declare.
12431 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
12433 * lra-constraints.c (emit_spill_move): Use smaller mode for
12435 (check_and_process_move): Consider mem-reg moves for secondary
12437 (curr_insn_transform): Don't lose insns emitted before for
12438 secondary memory moves.
12439 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
12440 reg set up in the current insn.
12442 2013-05-24 Dehao Chen <dehao@google.com>
12444 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
12446 (locus_descrim_hasher::equal): Likewise.
12447 (build_gimple_cfg): New discriminator assignment algorithm.
12448 (make_edges): Likewise.
12449 (next_discriminator_for_locus): Likewise.
12450 (same_line_p): Likewise.
12451 (assign_discriminators): Likewise.
12452 (make_cond_expr_edges): Likewise.
12453 (make_gimple_switch_edges): Likewise.
12454 (make_goto_expr_edges): Likewise.
12455 (make_gimple_asm_edges): Likewise.
12457 2013-05-24 Ian Bolton <ian.bolton@arm.com>
12459 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
12460 X format specifier to only display bottom 16 bits.
12461 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
12462 immediate to match for operand 2, since it will be masked.
12464 2013-05-24 Richard Biener <rguenther@suse.de>
12466 PR tree-optimization/57287
12467 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
12468 all SSA names that occur in abnormal PHIs.
12470 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
12473 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
12474 that index is not negative.
12476 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
12478 PR rtl-optimization/55177
12479 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
12480 (simplify_byte_swapping_operation): New.
12481 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
12482 (simplify_relational_operation_1): Deal with BSWAP.
12484 2013-05-23 Richard Henderson <rth@redhat.com>
12487 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
12488 (ix86_reorg): Call it.
12490 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
12493 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
12494 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
12495 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
12497 2013-05-23 Christian Bruel <christian.bruel@st.com>
12500 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
12502 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
12503 Marcus Shawcroft <marcus.shawcroft@arm.com>
12505 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
12506 * config/aarch64/constraints.md (Usa): Remove.
12507 * doc/md.texi (AArch64 Usa): Remove.
12509 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
12510 Marcus Shawcroft <marcus.shawcroft@arm.com>
12512 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
12513 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
12514 * config/aarch64/predicates.md (aarch64_const_address): Remove.
12515 (aarch64_mov_operand): Use aarch64_mov_operand_p.
12517 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
12519 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
12520 instruction (AdvSIMD).
12521 * config/aarch64/aarch64-builtins.c
12522 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
12523 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
12525 2013-05-23 Martin Jambor <mjambor@suse.cz>
12527 PR middle-end/57347
12528 * tree.h (contains_bitfld_component_ref_p): Declare.
12529 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
12530 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
12532 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
12533 not access a bit-field. Assert all final offsets are byte-aligned.
12535 2013-05-23 Richard Biener <rguenther@suse.de>
12537 PR tree-optimization/57380
12538 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
12539 least one invariant or re-used load.
12540 * passes.c (init_optimization_passes): Move pass_phiprop before
12543 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
12545 * config/aarch64/aarch64-simd.md
12546 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
12548 2013-05-23 Richard Biener <rguenther@suse.de>
12550 PR middle-end/57381
12551 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
12552 OEP_CONSTANT_ADDRESS_OF retained.
12554 2013-05-23 Jakub Jelinek <jakub@redhat.com>
12556 PR middle-end/57344
12557 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
12558 don't lower unit. Handle unit not being always BITS_PER_WORD.
12560 2013-05-23 Richard Biener <rguenther@suse.de>
12562 PR rtl-optimization/57341
12563 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
12564 instead of true_dependence.
12566 2013-05-22 David Malcolm <dmalcolm@redhat.com>
12568 * bb-reorder.c (branch_threshold): Make const.
12569 (exec_threshold): Ditto.
12571 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
12572 Pat Haugen <pthaugen@us.ibm.com>
12573 Peter Bergner <bergner@vnet.ibm.com>
12575 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
12576 documentation for the power8 crypto builtins.
12578 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
12580 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
12581 macros for defining power8 builtin functions.
12582 (BU_P8V_AV_2): Likewise.
12583 (BU_P8V_AV_P): Likewise.
12584 (BU_P8V_VSX_1): Likewise.
12585 (BU_P8V_OVERLOAD_1): Likewise.
12586 (BU_P8V_OVERLOAD_2): Likewise.
12587 (BU_CRYPTO_1): Likewise.
12588 (BU_CRYPTO_2): Likewise.
12589 (BU_CRYPTO_3): Likewise.
12590 (BU_CRYPTO_OVERLOAD_1): Likewise.
12591 (BU_CRYPTO_OVERLOAD_2): Likewise.
12592 (XSCVSPDP): Fix typo, point to the correct instruction.
12593 (VCIPHER): Add power8 crypto builtins.
12594 (VCIPHERLAST): Likewise.
12595 (VNCIPHER): Likewise.
12596 (VNCIPHERLAST): Likewise.
12597 (VPMSUMB): Likewise.
12598 (VPMSUMH): Likewise.
12599 (VPMSUMW): Likewise.
12600 (VPERMXOR_V2DI): Likewise.
12601 (VPERMXOR_V4SI: Likewise.
12602 (VPERMXOR_V8HI: Likewise.
12603 (VPERMXOR_V16QI: Likewise.
12604 (VSHASIGMAW): Likewise.
12605 (VSHASIGMAD): Likewise.
12606 (VPMSUM): Likewise.
12607 (VPERMXOR): Likewise.
12608 (VSHASIGMA): Likewise.
12610 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
12611 __CRYPTO__ if the crypto instructions are available.
12612 (altivec_overloaded_builtins): Add support for overloaded power8
12615 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
12616 support for power8 crypto builtins.
12617 (builtin_function_type): Likewise.
12618 (altivec_init_builtins): Add support for builtins that take vector
12619 long long (V2DI) arguments.
12621 * config/rs6000/crypto.md: New file, define power8 crypto
12624 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
12625 Pat Haugen <pthaugen@us.ibm.com>
12626 Peter Bergner <bergner@vnet.ibm.com>
12628 * doc/invoke.texi (Option Summary): Add power8 options.
12629 (RS/6000 and PowerPC Options): Likewise.
12631 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
12632 constraints.md instead of rs6000.h. Reorder w* constraints. Add
12633 wm, wn, wr documentation.
12635 * config/rs6000/constraints.md (wm): New constraint for VSX
12636 registers if direct move instructions are enabled.
12637 (wn): New constraint for no registers.
12638 (wq): New constraint for quad word even GPR registers.
12639 (wr): New constraint if 64-bit instructions are enabled.
12640 (wv): New constraint if power8 vector instructions are enabled.
12641 (wQ): New constraint for quad word memory locations.
12643 * config/rs6000/predicates.md (const_0_to_15_operand): New
12644 constraint for 0..15 for crypto instructions.
12645 (gpc_reg_operand): If VSX allow registers in VSX registers as well
12646 as GPR and floating point registers.
12647 (int_reg_operand): New predicate to match only GPR registers.
12648 (base_reg_operand): New predicate to match base registers.
12649 (quad_int_reg_operand): New predicate to match even GPR registers
12650 for quad memory operations.
12651 (vsx_reg_or_cint_operand): New predicate to allow vector logical
12652 operations in both GPR and VSX registers.
12653 (quad_memory_operand): New predicate for quad memory operations.
12654 (reg_or_indexed_operand): New predicate for direct move support.
12656 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
12657 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
12658 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
12659 (POWERPC_MASKS): Add power8 options.
12660 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
12663 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
12664 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
12666 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
12667 (-mpower8-fusion): New power8 options.
12668 (-mpower8-fusion-sign): Likewise.
12669 (-mpower8-vector): Likewise.
12670 (-mcrypto): Likewise.
12671 (-mdirect-move): Likewise.
12672 (-mquad-memory): Likewise.
12674 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
12675 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
12677 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
12678 (rs6000_debug_vector_unit): Add p8_vector.
12679 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
12680 definitions. Also print fusion state.
12681 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
12682 (rs6000_builtin_mask_calculate): Add power8 builtin support.
12683 (rs6000_option_override_internal): Add support for power8.
12684 (rs6000_common_init_builtins): Add debugging for skipped builtins
12685 if -mdebug=builtin.
12686 (rs6000_adjust_cost): Add power8 support.
12687 (rs6000_issue_rate): Likewise.
12688 (insn_must_be_first_in_group): Likewise.
12689 (insn_must_be_last_in_group): Likewise.
12690 (force_new_group): Likewise.
12691 (rs6000_register_move_cost): Likewise.
12692 (rs6000_opt_masks): Likewise.
12694 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
12695 power8 capable assembler, default to power7 options.
12696 (TARGET_DIRECT_MOVE): Likewise.
12697 (TARGET_CRYPTO): Likewise.
12698 (TARGET_P8_VECTOR): Likewise.
12699 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
12700 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
12701 (VECTOR_MEM_P8_VECTOR_P): Likewise.
12702 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
12703 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
12704 (TARGET_XSCVDPSPN): Likewise.
12705 (TARGET_XSCVSPDPN): Likewsie.
12706 (TARGET_SYNC_HI_QI): Likewise.
12707 (TARGET_SYNC_TI): Likewise.
12708 (MASK_CRYPTO): Likewise.
12709 (MASK_DIRECT_MOVE): Likewise.
12710 (MASK_P8_FUSION): Likewise.
12711 (MASK_P8_VECTOR): Likewise.
12712 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
12713 temporary used by some of the direct move instructions to get two FP
12714 temporary registers does not force creation of a stack frame.
12715 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
12716 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
12717 that any VSX registers are tieable, even if they are also an
12718 Altivec vector mode.
12719 (r6000_reg_class_enum): Add wm, wr, wv constraints.
12720 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
12721 (RS6000_BTM_CRYPTO): Likewise.
12722 (RS6000_BTM_COMMON): Likewise.
12724 * config/rs6000/rs6000.md (cpu attribute): Add power8.
12725 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
12726 (enum rs6000_vector): Add power8 vector support.
12728 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
12732 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
12733 (any_sibcall_could_use_r3): this and handle indirect calls.
12734 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
12736 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12738 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
12740 2013-05-22 Richard Biener <rguenther@suse.de>
12742 PR middle-end/57349
12743 * profile.c (branch_prob): Do not split blocks that are
12744 abnormally receiving from ECF_RETURNS_TWICE functions.
12746 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
12748 * recog.c (offsettable_address_addr_space_p): Fix calculation of
12749 address mode. Move pointer mode initialization to the same place.
12751 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
12753 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
12754 while it has any effect.
12756 2013-05-21 Easwaran Raman <eraman@google.com>
12758 PR tree-optimization/57322
12759 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
12760 UID of the statement added to the BB to be 1.
12762 2013-05-21 Jakub Jelinek <jakub@redhat.com>
12764 PR tree-optimization/57331
12765 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
12766 of conversion from pointer type to integral type with integer.
12768 2013-05-21 Martin Jambor <mjambor@suse.cz>
12771 * ipa-prop.c (ipa_read_node_info): Process param_used and
12772 controlled_uses in the same order as when writing.
12774 2013-05-21 Magnus Granberg <baldrick@free.fr>
12777 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
12779 2013-05-21 Richard Biener <rguenther@suse.de>
12781 PR tree-optimization/57318
12782 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
12783 estimate stmts with side-effects as likely eliminated.
12785 2013-05-21 Richard Biener <rguenther@suse.de>
12787 PR tree-optimization/57330
12788 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
12789 preserve the call stmts fntype.
12791 2013-05-21 Richard Biener <rguenther@suse.de>
12793 PR tree-optimization/57303
12794 * tree-ssa-sink.c (statement_sink_location): Improve killing
12795 stmt detection and properly handle self-assignments.
12797 2013-05-21 Christian Bruel <christian.bruel@st.com>
12799 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
12800 spanning registers. LEAF_REG_REMAP is supported only for contiguous
12801 registers. Set register size out of the PARALLEL loop.
12803 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
12806 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
12807 (*fmasf4, *fmasf4_media): New insns.
12809 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
12811 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
12812 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
12813 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
12814 (mips_idiv_insns): Update the comments to say that the returned
12815 instruction counts are in units of BASE_INSN_LENGTH.
12816 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
12817 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
12818 using 2 rather than 4 as the length of indirect MIPS16 and
12819 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
12820 length of a NOP. Don't divide MIPS16 lengths by 2.
12821 (mips16_split_long_branches): Assume a branch is long if the
12822 length is greater than 4 rather than 8.
12823 * config/mips/mips.md (length): Give MIPS16 lengths directly,
12824 rather than multiplying them by 2. Multiply instruction counts
12825 by BASE_INSN_LENGTH rather than 4.
12826 (*jump_mips16, tls_get_tp_mips16_<mode>)
12827 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
12829 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
12831 * config/mips/mips.md (extended_mips16): Remove branch case.
12832 (length): Remove duplicated extended_mips16 test.
12834 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
12836 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
12838 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
12840 * recog.h (Recog_data): Rename to...
12841 (recog_data_d): ...this.
12842 (recog_data): Update accordingly.
12843 * recog.c (recog_data): Likewise.
12844 * reload.c (save_recog_data): Likewise.
12845 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
12846 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
12848 2013-05-17 Julian Brown <julian@codesourcery.com>
12850 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
12851 found in a REG_EQUAL note, invalidate it.
12853 2013-05-17 Easwaran Raman <eraman@google.com>
12855 * tree-ssa-reassoc.c (find_insert_point): New function.
12856 (insert_stmt_after): Likewise.
12857 (get_def_stmt): Likewise.
12858 (ensure_ops_are_available): Likewise.
12859 (not_dominated_by): Likewise.
12860 (rewrite_expr_tree): Do not move statements beyond what is
12861 necessary. Remove call to swap_ops_for_binary_stmt...
12862 (reassociate_bb): ... and move it here.
12863 (build_and_add_sum): Assign UIDs for new statements.
12864 (linearize_expr): Likewise.
12865 (do_reassoc): Renumber gimple statement UIDs.
12867 2013-05-17 Jan Hubicka <jh@suse.cz>
12869 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
12871 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
12872 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
12874 (output_weakrefs): Update.
12876 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
12877 Martin Jambor <mjambor@suse.cz>
12879 PR middle-end/57276
12880 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
12881 value that corresponds to the given aggval is found in values vector.
12883 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
12885 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
12886 sse, sse2, sse3, ssse3 and sse4a flags to options.
12888 2013-05-17 David Malcolm <dmalcolm@redhat.com>
12890 * gengtype-state.c: (s_expr_writer): New class, to handle
12891 prettifying of output layout of s-expressions.
12892 (state_writer): New class, to write out gtype.state.
12893 (state_written_type_count): Move this variable into member data of
12895 (s_expr_writer::s_expr_writer): New code: constructor for new class
12896 (state_writer::state_writer(): ditto
12897 (s_expr_writer::write_new_line): New function
12898 (s_expr_writer::write_any_indent): ditto
12899 (s_expr_writer::begin_s_expr): ditto
12900 (s_expr_writer::end_s_expr): ditto
12901 (write_state_fileloc): convert to method of state_writer...
12902 (state_writer:: write_state_fileloc): ...and use methods of
12903 s_expr_writer to write indentation into the gtype.state output file
12904 to visually represent the hierarchical structure of the list
12906 (write_state_fields): ditto, renaming to...
12907 (state_writer::write_state_fields)
12908 (write_state_a_string): ditto, renaming to...
12909 (state_writer::write_state_a_string)
12910 (write_state_string_option): ditto, renaming to...
12911 (state_writer::write_state_string_option)
12912 (write_state_type_option): ditto, renaming to...
12913 (state_writer::write_state_type_option)
12914 (write_state_nested_option): ditto, renaming to...
12915 (state_writer::write_state_nested_option)
12916 (write_state_option): ditto, renaming to...
12917 (state_writer::write_state_option)
12918 (write_state_options): ditto, renaming to...
12919 (state_writer::write_state_options)
12920 (write_state_lang_bitmap): ditto, renaming to...
12921 (state_writer::write_state_lang_bitmap)
12922 (write_state_version): ditto, renaming to...
12923 (state_writer::write_state_version)
12924 (write_state_scalar_type): ditto, renaming to...
12925 (state_writer::write_state_scalar_type)
12926 (write_state_string_type): ditto, renaming to...
12927 (state_writer::write_state_string_type)
12928 (write_state_undefined_type): ditto, renaming to...
12929 (state_writer::write_state_undefined_type)
12930 (write_state_struct_union_type): ditto, renaming to...
12931 (state_writer::write_state_struct_union_type)
12932 (write_state_struct_type): ditto, renaming to...
12933 (state_writer::write_state_struct_type)
12934 (write_state_user_struct_type): ditto, renaming to...
12935 (state_writer::write_state_user_struct_type)
12936 (write_state_lang_struct_type): ditto, renaming to...
12937 (state_writer::write_state_lang_struct_type)
12938 (write_state_param_struct_type): ditto, renaming to...
12939 (state_writer::write_state_param_struct_type)
12940 (write_state_pointer_type): ditto, renaming to...
12941 (state_writer::write_state_pointer_type)
12942 (write_state_array_type): ditto, renaming to...
12943 (state_writer::write_state_array_type)
12944 (write_state_gc_used): ditto, renaming to...
12945 (state_writer::write_state_gc_used)
12946 (write_state_common_type_content): ditto, renaming to...
12947 (state_writer::write_state_common_type_content)
12948 (write_state_type): ditto, renaming to...
12949 (state_writer::write_state_type)
12950 (write_state_pair_list): ditto, renaming to...
12951 (state_writer::write_state_pair_list)
12952 (write_state_pair): ditto, renaming to...
12953 (state_writer::write_state_pair)
12954 (write_state_typedefs): ditto, renaming to...
12955 (state_writer::write_state_typedefs)
12956 (write_state_structures): ditto, renaming to...
12957 (state_writer::write_state_structures)
12958 (write_state_param_structs): ditto, renaming to...
12959 (state_writer::write_state_param_structs)
12960 (write_state_variables): ditto, renaming to...
12961 (state_writer::write_state_variables)
12962 (write_state_srcdir): ditto, renaming to...
12963 (state_writer::write_state_srcdir)
12964 (write_state_files_list): ditto, renaming to...
12965 (state_writer::write_state_files_list)
12966 (write_state_languages): ditto, renaming to...
12967 (state_writer::write_state_languages)
12968 (write_state): create a state_writer instance and use it when
12969 writing out the state file
12971 2013-05-17 Mike Stump <mikestump@comcast.net>
12973 PR rtl-optimization/57304
12974 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
12975 accessing DF_REF_REAL_LOC.
12977 2013-05-17 Jakub Jelinek <jakub@redhat.com>
12979 PR rtl-optimization/57281
12980 PR rtl-optimization/57300
12981 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
12982 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
12983 what the other splitter did if the registers are dead.
12985 2013-05-17 Richard Biener <rguenther@suse.de>
12987 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
12990 2013-05-17 Jakub Jelinek <jakub@redhat.com>
12992 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
12995 2013-05-17 Marek Polacek <polacek@redhat.com>
12997 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
12998 length when doing non-zero store of storing '\0' to '\0'.
13000 2013-05-17 Jakub Jelinek <jakub@redhat.com>
13002 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
13003 vect_external_def oprnd1 with loop_vinfo, try to emit
13004 optional cast, negation and and stmts on the loop preheader
13005 edge instead of into the pattern def seq.
13007 PR tree-optimization/57051
13008 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
13009 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
13011 2013-05-16 Nick Clifton <nickc@redhat.com>
13013 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
13014 (rl78_is_naked_func): New function.
13015 (rl78_expand_prologue): Skip prologue generation for naked functions.
13016 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
13017 * doc/extend.texi (naked): Add RL78 to the list of processors
13018 that supports this attribute.
13020 2013-05-16 Jeff Law <law@redhat.com>
13022 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
13024 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
13026 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
13027 cache parameters using detect_caches_amd also for CYRIX,
13028 NSC and TM2 signatures.
13030 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
13031 Dzianis Kahanovich <mahatma@eu.by>
13035 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
13036 VIA/Centaur processors and determine their cache parameters
13037 using detect_caches_amd.
13039 2013-05-16 Teresa Johnson <tejohnson@google.com>
13041 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
13042 (rtl_verify_edges): New function.
13043 (rtl_verify_bb_insns): Ditto.
13044 (rtl_verify_bb_pointers): Ditto.
13045 (rtl_verify_bb_insn_chain): Ditto.
13046 (rtl_verify_fallthru): Ditto.
13047 (rtl_verify_bb_layout): Ditto.
13048 (rtl_verify_flow_info_1): Outline checks into new functions.
13049 (rtl_verify_flow_info): Ditto.
13051 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
13053 * cfghooks.c (copy_bbs): Add update_dominance argument.
13054 * cfghooks.h (copy_bbs): Update prototype.
13055 * tree-cfg.c (gimple_duplicate_sese_region):
13056 Add update_dominance argument.
13057 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
13058 * tree-ssa-loop-ch.c (copy_loop_headers): Update
13059 gimple_duplicate_sese_region call.
13060 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
13061 Update copy_bbs call.
13062 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
13063 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
13065 2013-05-16 Jakub Jelinek <jakub@redhat.com>
13067 * tree-vectorizer.h (NUM_PATTERNS): Increment.
13068 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
13069 vect_recog_rotate_pattern.
13070 (vect_recog_rotate_pattern): New function.
13072 2013-05-16 Jason Merrill <jason@redhat.com>
13074 * Makefile.in (LLINKER): New variable.
13075 (mostlyclean): Remove link mutex.
13076 * configure.ac: Handle --enable-link-mutex.
13077 * lock-and-run.sh: New script.
13079 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
13082 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
13085 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
13087 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
13089 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
13091 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
13092 * config/arm/arm.c (next_consecutive_mem): New function.
13093 (gen_movmem_ldrd_strd): Likewise.
13094 * config/arm/arm.md (movmemqi): Update condition and code.
13095 (unaligned_loaddi, unaligned_storedi): New patterns.
13097 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
13099 * config.gcc: Obsolete *-*-solaris2.9*.
13100 * doc/install.texi (Specific, *-*-solaris2*): Document it.
13102 2013-05-16 Richard Biener <rguenther@suse.de>
13104 * passes.c (init_optimization_passes): Move pass_parallelize_loops
13105 earlier, after GRAPHITE transforms and IV canonicalization.
13107 2013-05-16 Jakub Jelinek <jakub@redhat.com>
13109 * omp-low.c (extract_omp_for_data): For collapsed loops,
13110 if at least one of the loops is known at compile time to
13111 iterate zero times, set count to 0.
13112 (expand_omp_regimplify_p): New function.
13113 (expand_omp_for_generic): For collapsed loops, if at least
13114 one of the loops isn't known to iterate at least once,
13115 add runtime check with setting count to 0.
13116 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
13117 For unsigned types if it isn't known at compile time that
13118 the loop will iterate at least once, add runtime check to bypass
13119 the whole loop if initial condition isn't true.
13121 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
13123 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
13125 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
13127 PR middle-end/57286
13128 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
13129 transformations to avoid an infinite loop.
13131 2013-05-16 Marek Polacek <polacek@redhat.com>
13133 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
13135 2013-05-15 Leif Ekblad <leif@rdos.net>
13137 * config/i386/i386.c (ix86_decompose_address): Use
13138 DEFAULT_TLS_SEG_REG to access TLS segment register.
13139 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
13140 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
13141 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
13143 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
13146 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
13147 sibling calls to functions that would normally be lazily bound,
13148 unless $gp is call-clobbered.
13150 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
13152 * config/i386/i386.c (ix86_option_override_internal): Update
13153 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
13154 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
13155 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
13156 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
13158 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
13160 2013-05-15 Andreas Schwab <schwab@suse.de>
13162 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
13163 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
13166 2013-05-15 Teresa Johnson <tejohnson@google.com>
13168 * loop-unroll.c (report_unroll_peel): Check decision before
13169 emitting unroll/peel message.
13171 2013-05-15 Teresa Johnson <tejohnson@google.com>
13173 * function.h (has_bb_partition): New rtl_data flag.
13174 (bb_reorder_complete): Ditto.
13175 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
13176 instead of flag_reorder_blocks_and_partition.
13177 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
13178 with some enhancements.
13179 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
13180 * bb-reorder.c (connect_traces): Check for has_bb_partition
13181 instead of flag_reorder_blocks_and_partition.
13182 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
13183 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
13184 verify_hot_cold_block_grouping.
13185 (partition_hot_cold_basic_blocks): Set has_bb_partition.
13187 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
13190 * config/arm/predicates.md (call_insn_operand): New predicate.
13191 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
13192 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
13193 if insn is not a tail call.
13194 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
13196 * config/arm/arm.h (enum reg_class): New caller save register class.
13197 (REG_CLASS_NAMES): Likewise.
13198 (REG_CLASS_CONTENTS): Likewise.
13199 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
13202 2013-05-15 Richard Biener <rguenther@suse.de>
13204 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
13205 of MSG_OPTIMIZED_LOCATIONS.
13206 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
13207 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
13209 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
13210 of MSG_OPTIMIZED_LOCATIONS.
13211 (execute_vect_slp): Likewise.
13212 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
13213 (vect_create_cond_for_alias_checks): Likewise.
13214 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
13215 (vect_recog_widen_mult_pattern): Likewise.
13216 (vect_recog_widen_sum_pattern): Likewise.
13217 (vect_recog_over_widening_pattern): Likewise.
13218 (vect_recog_widen_shift_pattern): Likewise.
13219 (vect_recog_vector_vector_shift_pattern): Likewise.
13220 (vect_recog_divmod_pattern): Likewise.
13221 (vect_recog_mixed_size_cond_pattern): Likewise.
13222 (vect_recog_bool_pattern): Likewise.
13223 (vect_pattern_recog_1): Likewise.
13225 2013-05-15 Martin Jambor <mjambor@suse.cz>
13227 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
13228 non-functions to builtin_unreachable.
13229 * ipa-inline-transform.c (inline_call): Do not assert estimates were
13230 correct when new direct edges were discovered.
13232 2013-05-15 Martin Jambor <mjambor@suse.cz>
13234 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
13235 header, print symbol order instead of node uid, print more information
13236 about indirect edge targets.
13237 (ipa_make_edge_direct_to_target): Print symbol order instead of node
13239 (ipa_make_edge_direct_to_target): Likewise.
13240 (remove_described_reference): Likewise.
13241 (propagate_controlled_uses): Likewise.
13242 (ipa_print_node_params): Also print symbol order.
13243 (ipcp_transform_function): Print symbol order instead of node uids.
13244 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
13245 (cgraph_get_create_real_symbol_node): Likewise.
13246 * ipa-cp.c (print_lattice): Likewise.
13247 (print_all_lattices): Likewise.
13248 (determine_versionability): Likewise.
13249 (initialize_node_lattices): Likewise.
13250 (estimate_local_effects): Likewise.
13251 (update_profiling_info): Likewise.
13252 (create_specialized_node): Likewise.
13253 (perhaps_add_new_callers): Likewise.
13254 (decide_about_value): Likewise.
13255 (decide_whether_version_node): Likewise.
13256 (identify_dead_nodes): Likewise.
13257 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
13258 (dump_inline_summary): Likewise.
13259 (estimate_node_size_and_time): Likewise.
13260 (inline_analyze_function): Likewise.
13261 * ipa-inline.c (report_inline_failed_reason): Likewise.
13262 (want_early_inline_function_p): Likewise.
13263 (edge_badness): Likewise.
13264 (update_edge_key): Likewise.
13265 (inline_small_functions): Likewise. Add dumping of order to two other
13267 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
13268 instead of node uids.
13269 (propagate_pure_const): Likewise.
13270 (propagate_pure_const): Likewise.
13271 * ipa-utils.c (dump_cgraph_node_set): Likewise.
13272 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
13273 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
13275 * tree-pretty-print.c (dump_function_header): Likewise.
13276 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
13277 Print symbol order instead of node uids.
13279 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13281 * config/s390/s390.c (s390_register_move_cost): Don't impose the
13282 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
13284 2013-05-15 Richard Biener <rguenther@suse.de>
13286 PR tree-optimization/57275
13287 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
13288 return value for fail to do runtime alias checks for gather loads.
13290 2013-05-15 Jan Hubicka <jh@suse.cz>
13294 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
13295 weakrefs are not external.
13296 (lto_symtab_merge_decls): Fix thinko when dealing with
13297 non-lto_symtab decls.
13298 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
13299 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
13300 * varpool.c (dump_varpool_node): Dump more flags.
13302 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
13304 * config/i386/i386.c (processor_alias_table): Add instruction
13305 FSGSBASE for AMD bdver3 architecture.
13307 2013-05-14 Jakub Jelinek <jakub@redhat.com>
13309 * tree.c (warn_deprecated_use): Print file:line using locus color.
13310 * diagnostic.c (diagnostic_report_current_module): Print file:line
13311 and file:line:column using locus color.
13313 2013-05-14 Mike Stump <mikestump@comcast.net>
13315 * gdbinit.in: Add __null.
13317 2013-05-14 Mike Stump <mikestump@comcast.net>
13319 * recog.h: Rename struct recog_data to Recog_data.
13320 * recog.c: Likewise.
13321 * reload.c (can_reload_into): Likewise.
13322 * config/picochip/picochip.c: Likewise.
13324 2013-05-14 Mike Stump <mikestump@comcast.net>
13326 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
13328 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
13330 * resource.h (struct resources): Remove unch_memory member.
13331 (CLEAR_RESOURCE): Don't clear unch_memory.
13332 * resource.c (mark_referenced_resources): Don't set it.
13333 (mark_set_resources): Likewise.
13334 (mark_target_live_regs): Don't clear it.
13335 (init_resource_info): Likewise.
13336 * reorg.c (resource_conflicts_p): Don't compare it.
13337 (redundant_insn): Don't set it.
13339 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
13341 * emit-rtl.c (next_label): Remove unused function.
13342 (skip_consecutive_labels, link_cc0_insns): Move to ...
13343 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
13344 only place where these functions are used, and make them static.
13346 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
13348 * fold-const.c (fold_negate_expr): Handle vectors.
13349 (fold_truth_not_expr): Make it static.
13350 (fold_invert_truthvalue): New static function.
13351 (invert_truthvalue_loc): Handle vectors. Do not call
13352 fold_truth_not_expr directly.
13353 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
13354 <TRUTH_NOT_EXPR>: Do not cast to boolean.
13355 (fold_comparison): Handle vector constants.
13356 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
13357 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
13358 * tree.h (fold_truth_not_expr): Remove declaration.
13360 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
13362 * config/aarch64/aarch64-simd.md
13363 (aarch64_vcond_internal<mode>): Rename to...
13364 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
13365 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
13366 float modes. Clarify all iterator modes.
13367 (vcond<mode><mode>): Use new name for vcond expanders.
13368 (vcond<v_cmp_result><mode>): Likewise.
13369 (vcondu<mode><mode>: Likewise.
13370 * config/aarch64/iterators.md (VDQF_COND): New.
13372 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
13375 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
13376 variable for the shift amount. Check that we shift by non-negative
13379 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
13382 * config/arm/arm.h (EPILOGUE_USES): Only return true
13383 for LR_REGNUM after epilogue_completed.
13385 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
13387 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
13388 is error_mark_node.
13390 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
13393 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
13394 and Solaris 11+/x86 with gld.
13395 * configure: Regenerate.
13397 2013-05-14 Jakub Jelinek <jakub@redhat.com>
13399 * expmed.c (expand_shift_1): Canonicalize rotates by
13400 constant bitsize / 2 to bitsize - 1.
13401 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
13402 case ROTATERT>: Likewise.
13405 2013-05-10 Jakub Jelinek <jakub@redhat.com>
13407 * config/i386/i386.md (rotateinv): New code attr.
13408 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
13409 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
13410 roll $31, %eax, etc.
13412 2013-05-14 Richard Biener <rguenther@suse.de>
13414 PR middle-end/57235
13415 * tree-eh.c (sink_clobbers): Give up for successors with
13416 multiple predecessors and no virtual uses.
13418 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
13420 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
13421 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
13423 2013-05-14 Jakub Jelinek <jakub@redhat.com>
13425 PR middle-end/57251
13426 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
13427 the case when both op0 and op1 have VOIDmode.
13429 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
13431 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
13432 in multiply-accumulate mode.
13434 2013-05-13 Guozhi Wei <carrot@google.com>
13436 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
13438 2013-05-13 Kai Tietz <ktietz@redhat.com>
13441 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
13442 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
13443 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
13444 * config/i386/i386.c (ix86_option_override_internal): Likewise.
13445 (ix86_expand_prologue): Likewise.
13446 (ix86_expand_split_stack_prologue): Likewise.
13447 (legitimate_pic_address_disp_p): Likewise.
13448 (legitimize_pic_address): Likewise.
13449 (legitimize_tls_address): Likewise.
13450 (legitimize_pe_coff_symbol): Likewise.
13451 (output_pic_addr_const): Likewise.
13452 (construct_plt_address): Likewise.
13453 (ix86_expand_call): Likewise.
13454 (x86_output_mi_thunk): Likewise.
13455 (x86_function_profiler): Likewise.
13457 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
13459 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
13460 similar switch cases.
13461 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
13462 (aarch64_simd_mov_to_<mode>low): Delete.
13463 (aarch64_simd_mov_to_<mode>high): Delete.
13464 (move_lo_quad_<mode>): Add w<-r alternative.
13465 (aarch64_simd_move_hi_quad_<mode>): Likewise.
13466 (aarch64_simd_mov_from_*): Update type attribute.
13467 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
13470 2013-05-13 Jan Hubicka <jh@suse.cz>
13472 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
13473 * config/i386/i386.c (ix86_compute_frame_layout,
13474 ix86_expand_epilogue, emit_i387_cw_initialization,
13475 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
13476 ix86_local_alignment): Fix use of size/speed predicates.
13478 2013-05-13 Jakub Jelinek <jakub@redhat.com>
13480 PR tree-optimization/45216
13481 PR tree-optimization/57157
13482 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
13483 the (-Y) & (B - 1) variant if OP is |.
13484 * expmed.c (expand_shift_1): For rotations by const0_rtx just
13485 return shifted. Use (-op1) & (prec - 1) as other_amount
13486 instead of prec - op1.
13488 2013-05-13 Martin Jambor <mjambor@suse.cz>
13490 PR middle-end/42371
13491 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
13492 (ipa_constant_data): New type.
13493 (ipa_jump_func): Use ipa_constant_data to hold information about
13494 constant jump functions.
13495 (ipa_get_jf_constant): Adjust to jump function type changes.
13496 (ipa_get_jf_constant_rdesc): New function.
13497 (ipa_param_descriptor): New field controlled_uses.
13498 (ipa_get_controlled_uses): New function.
13499 (ipa_set_controlled_uses): Likewise.
13500 * ipa-ref.h (ipa_find_reference): Declare.
13501 * ipa-prop.c (ipa_cst_ref_desc): New type.
13502 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
13504 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
13505 New parameter cs. Adjust all callers.
13506 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
13507 (remove_described_reference): New function.
13508 (jfunc_rdesc_usable): Likewise.
13509 (try_make_edge_direct_simple_call): Decrement controlled use count,
13510 attempt to remove reference if it hits zero.
13511 (combine_controlled_uses_counters): New function.
13512 (propagate_controlled_uses): Likewise.
13513 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
13514 (ipa_edge_duplication_hook): Duplicate reference descriptions.
13515 (ipa_print_node_params): Print described use counter.
13516 (ipa_write_jump_function): Adjust to jump function type changes.
13517 (ipa_read_jump_function): New parameter CS, pass it to
13518 ipa_set_jf_constant. Adjust caller.
13519 (ipa_write_node_info): Stream controlled use count
13520 (ipa_read_node_info): Likewise.
13521 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
13523 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
13524 count. Remove cloning-added reference if it reaches zero.
13525 * ipa-ref.c (ipa_find_reference): New function.
13527 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
13529 * config/i386/i386.c (processor_target_table): Modified default
13530 alignment values for AMD BD and BT architectures.
13532 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
13534 * tree-vect-generic.c (uniform_vector_p): Move ...
13535 * tree.c (uniform_vector_p): ... here.
13536 * tree.h (uniform_vector_p): Declare it.
13537 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
13540 2013-05-13 Jakub Jelinek <jakub@redhat.com>
13542 PR tree-optimization/57230
13543 * tree-ssa-strlen.c (handle_char_store): Record length for
13544 array store from STRING_CST.
13546 PR tree-optimization/57230
13547 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
13550 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
13552 * config/epiphany/epiphany.c (epiphany_init): Check size of
13553 NUM_MODES_FOR_MODE_SWITCHING.
13554 (epiphany_expand_prologue):
13555 Remove CONFIG_REGNUM initial value handling code.
13556 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
13557 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
13558 (emit_set_fp_mode, epiphany_mode_after): Likewise.
13559 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
13560 Don't return 1 for FP_MODE_NONE.
13561 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
13562 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
13563 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
13564 * config/epiphany/epiphany.md (save_config): New pattern.
13566 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
13568 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
13570 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
13572 * config/i386/i386.md (memory): Handle sseishft1.
13573 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
13574 (*vec_extractv2di_1): Ditto.
13576 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
13578 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
13581 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
13583 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
13584 Add mthumb/march=armv7-a multilib.
13585 Add mthumb/march=armv7-r multilib.
13586 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
13588 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
13590 * config/v850/t-rtems: Add more multilibs.
13592 2013-05-10 Richard Biener <rguenther@suse.de>
13594 PR tree-optimization/57214
13595 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
13596 not propagate from SSA names that occur in abnormal PHI nodes.
13598 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
13600 * stor-layout.c (element_precision): New function.
13601 * machmode.h (element_precision): Declare it.
13602 * tree.c (build_minus_one_cst): New function.
13603 (element_precision): Likewise.
13604 * tree.h (build_minus_one_cst): Declare new function.
13605 (element_precision): Likewise.
13606 * fold-const.c (operand_equal_p): Use element_precision.
13607 (fold_binary_loc): Handle vector types.
13608 * convert.c (convert_to_integer): Use element_precision.
13609 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
13612 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
13614 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
13615 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
13616 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
13617 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
13618 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
13619 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
13620 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
13621 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
13622 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
13623 (Uuw8): New constraints.
13624 (Usb4): Move into alphabetical order.
13625 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
13626 (sd8_operand, ub8_operand, uw8_operand): New predicates.
13627 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
13628 previously unnamed patterns.
13629 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
13630 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
13631 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
13632 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
13633 of set_attr_alternative/if_then_else. Use extended_mips16 instead
13634 of specific lengths.
13636 2013-05-10 Jakub Jelinek <jakub@redhat.com>
13638 * config/i386/i386.md (rotateinv): New code attr.
13639 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
13640 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
13641 roll $31, %eax, etc.
13643 PR tree-optimization/45216
13644 PR tree-optimization/57157
13645 * tree-ssa-forwprop.c (simplify_rotate): New function.
13646 (ssa_forward_propagate_and_combine): Call it.
13648 2013-05-10 Richard Biener <rguenther@suse.de>
13650 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
13651 disable peeling when we version for aliasing.
13652 (vector_alignment_reachable_p): Honor explicit user alignment.
13653 (vect_supportable_dr_alignment): Likewise.
13654 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
13655 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
13656 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
13657 then peeling to arrange for the cost-model check to come first.
13659 2013-05-10 Alan Modra <amodra@gmail.com>
13661 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
13662 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
13663 * configure: Regenerate.
13665 2013-05-10 Alan Modra <amodra@gmail.com>
13668 * varasm.c (default_elf_select_section): Move !DECL_P check..
13669 (get_named_section): ..to here before calling get_section_name.
13671 (default_section_type_flags): Add DECL_P check.
13672 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
13673 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
13675 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
13677 * config/epiphany/epiphany.c (epiphany_expand_prologue):
13678 When using gen_stack_adjust_str with a register offset, add a
13679 REG_FRAME_RELATED_EXPR note.
13681 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
13683 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
13684 (*vec_extractv4si_zext_mem): Ditto.
13685 (*vec_extractv2di): Add 0->x and x->x alternatives.
13686 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
13687 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
13689 2013-05-09 Jason Merrill <jason@redhat.com>
13691 N3639 C++1y VLA support
13692 * gimplify.c (gimplify_vla_decl): Don't touch an existing
13695 * tree.c (build_constructor_va): New.
13696 * tree.h: Declare it.
13698 2013-05-09 Martin Jambor <mjambor@suse.cz>
13701 * gimple-fold.c (canonicalize_constructor_val): Call
13702 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
13704 2013-05-09 Jan Hubicka <jh@suse.cz>
13705 Richard Biener <rguenther@suse.de>
13708 * symtab.c (symtab_make_decl_local): Do not add private names.
13710 2013-05-09 Jan Hubicka <jh@suse.cz>
13713 * symtab.c (insert_to_assembler_name_hash): Handle clones.
13714 (unlink_from_assembler_name_hash): Likewise.
13715 (symtab_prevail_in_asm_name_hash, symtab_register_node,
13716 symtab_unregister_node, symtab_initialize_asm_name_hash,
13717 change_decl_assembler_name): Update.
13719 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
13721 * config/aarch64/aarch64.md: New movtf split.
13722 (*movtf_aarch64): Update.
13723 (aarch64_movdi_tilow): Handle TF modes and rename to
13724 aarch64_movdi_<mode>low.
13725 (aarch64_movdi_tihigh): Handle TF modes and rename to
13726 aarch64_movdi_<mode>high
13727 (aarch64_movtihigh_di): Handle TF modes and rename to
13728 aarch64_mov<mode>high_di
13729 (aarch64_movtilow_di): Handle TF modes and rename to
13730 aarch64_mov<mode>low_di
13731 (aarch64_movtilow_tilow): Remove spurious whitespace.
13732 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
13734 (aarch64_print_operand): Update.
13736 2013-05-09 Alan Modra <amodra@gmail.com>
13738 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
13740 * configure: Regenerate.
13742 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
13744 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
13745 splitter preparation statements.
13746 * config/i386/sse.md (*vec_extract* splitters): Ditto.
13747 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
13750 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13752 * gimple-ssa-strength-reduction.c (count_candidates): Change
13753 return value to int.
13754 (analyze_candidates_and_replace): Change type of length to int.
13756 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
13758 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
13759 (*vec_extract<mode>): Use VI12_128 mode iterator.
13760 (*vec_extract<mode>_mem): Ditto.
13761 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
13764 2013-05-08 Diego Novillo <dnovillo@google.com>
13769 2012-08-17 Diego Novillo <dnovillo@google.com>
13772 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
13773 * config.in: Regenerate.
13774 * configure: Regenerate.
13775 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
13777 2013-05-08 Jan Hubicka <jh@suse.cz>
13780 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
13781 * cgraph.h (symtab_node_base): Add unique_name.
13782 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
13783 input_overwrite_node, input_varpool_node): Stream unique_name.
13784 * cgraphclones.c (cgraph_create_virtual_clone,
13785 cgraph_function_versioning): Set unique_name.
13786 * ipa.c (function_and_variable_visibility): Set unique_name.
13788 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13790 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
13791 (alloc_cand_and_find_basis): Restrict conditional candidate
13792 processing to CAND_MULTs.
13794 2013-05-08 Jan Hubicka <jh@suse.cz>
13797 lto-symtab.c (lto_symtab_symbol_p): New function.
13798 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
13799 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
13800 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
13801 Skip static symbols.
13803 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
13805 PR tree-optimization/57200
13806 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
13807 Only call inform if the preceding warning_at returns true.
13809 2013-05-07 Han Shen <shenhan@google.com>
13811 * cfgexpand.c (record_or_union_type_has_array_p): New function.
13812 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
13813 * common.opt (fstack-protector-strong): New option.
13814 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
13815 * doc/invoke.texi (Optimization Options): Document
13816 "-fstack-protector-strong".
13817 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
13819 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
13821 * config/mips/mips.c (mips_machine_reorg2): Return 0.
13823 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
13825 * ira.c (update_equiv_regs): Add insn having equiv memory even if
13826 it is not lhs of the insn.
13827 (setup_reg_equiv): Remove insn having equiv memory which it is not
13829 * lra-constraints.c (process_address): Try to improve generation
13830 code for address base + disp.
13831 (lra_constraints): Make correct the code for checking insn setting
13832 up backward equivalence. Remove insn only if it is in the init
13834 * lra-eliminations.c (update_reg_eliminate): Change return value.
13835 (lra_eliminate): Use the result.
13837 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
13839 * config/i386/sse.md (ssescalarnummask): New mode attribute.
13840 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
13841 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
13842 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
13843 register target operands.
13844 (*vec_extractv8hi_sse2): New pattern.
13845 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
13846 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
13847 (*vec_extract<mode>_mem): New insn and split pattern.
13849 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
13851 * config/arm/arm.c (arm_asan_shadow_offset): New function.
13852 (TARGET_ASAN_SHADOW_OFFSET): Define.
13853 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
13854 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
13856 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13858 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
13859 (incr_vec_index): Return -1 if increment not found.
13860 (create_add_on_incoming_edge): Assert if increment not found.
13861 (record_increment): Limit number of increments recorded.
13862 (all_phi_incrs_profitable): Return false if an increment not found.
13863 (replace_profitable_candidates): Don't process increments that were
13865 (analyze_candidates_and_replace): Limit size of incr_vec.
13867 2013-05-07 Richard Biener <rguenther@suse.de>
13869 * calls.c (special_function_p): setjmp-like functions are leaf.
13870 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
13871 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
13873 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
13875 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
13876 (aarch64_simd_mov<mode>): New expander.
13877 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
13878 (aarch64_simd_mov_to_<mode>high): Likewise.
13879 (aarch64_simd_mov_from_<mode>low): Likewise.
13880 (aarch64_simd_mov_from_<mode>high): Likewise.
13881 (aarch64_dup_lane<mode>): Update.
13882 (aarch64_dup_lanedi): New instruction pattern.
13883 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
13884 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
13886 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13888 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
13889 (replace_mult_candidate): Remove unnecessary argument; remove
13890 unnecessary parameter from call to introduce_cast_before_cand.
13891 (replace_unconditional_candidate): Remove unnecessary parameter
13892 from call to replace_mult_candidate.
13893 (replace_conditional_candidate): Likewise.
13894 (insert_initializers): Use make_temp_ssa_name.
13895 (introduce_cast_before_cand): Remove unnecessary argument; use
13896 make_temp_ssa_name.
13897 (replace_one_candidate): Remove unnecessary argument; remove
13898 unnecessary parameter from calls to introduce_cast_before_cand.
13899 (replace_profitable_candidates): Remove unnecessary parameters
13900 from calls to replace_one_candidate.
13902 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13904 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
13905 phi def as possibly hiding a basis for a CAND_ADD whose operands
13906 have been commuted in the analysis.
13907 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
13909 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
13911 * config/aarch64/aarch64.md
13912 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
13913 shift value between 0-4.
13915 2013-05-07 Richard Biener <rguenther@suse.de>
13917 * double-int.h (rshift): New overload.
13918 * double-int.c (rshift): New function.
13919 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
13920 (create_reference_ops_from_ref): Remove.
13921 (vn_reference_insert): Use shared ops for constructing the
13922 reference and copy it.
13924 2013-05-07 Richard Biener <rguenther@suse.de>
13926 PR middle-end/57190
13927 * tree-eh.c (sink_clobbers): Properly propagate
13928 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
13930 2013-05-07 Jakub Jelinek <jakub@redhat.com>
13932 PR tree-optimization/57149
13933 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
13934 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
13935 collect_phi_def_edges, execute_late_warn_uninitialized): Use
13936 uninit_undefined_value_p instead of ssa_undefined_value_p.
13939 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
13940 for modifier == EXPAND_INITIALIZER.
13942 2013-05-07 Anton Blanchard <anton@samba.org>
13944 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
13945 for powerpc64 little endian.
13946 * configure: Regenerate.
13948 2013-05-06 Graham Stott <grahams@btinternet.com>
13950 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
13951 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
13952 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
13953 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
13955 2013-05-06 Graham Stott <grahams@btinternet.com>
13957 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
13958 codes which allow non-lvalues.
13960 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
13962 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
13963 components are all 1s.
13964 (integer_minus_onep): New function.
13965 * tree.h (integer_minus_onep): Declare it.
13966 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
13967 integer_minus_onep instead of integer_all_onesp.
13969 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
13972 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
13973 variations of these patterns.
13975 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
13977 * config/i386/i386.md (isa): Add x64_sse4 member.
13978 (enabled): Handle x64_sse4.
13979 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
13980 instruction for 64bit SSE4_1 targets. Update insn attributes.
13981 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
13982 instruction for SSE4_1 targets. Update insn attributes.
13983 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
13984 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
13985 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
13987 (*vec_extractv4si): Rename from *sse4_1_pextrd.
13988 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
13989 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
13991 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
13994 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
13996 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
13998 * final.c (do_assembler_dialects): Don't handle curly braces and
13999 vertical bar escaped by % as dialect delimiters.
14000 (output_asm_insn): Print curly braces and vertical bar if escaped
14001 by % and ASSEMBLER_DIALECT defined.
14002 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
14003 * doc/tm.texi: Regenerated.
14005 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
14007 * config/mips/mips.c: Include tree-pass.h.
14008 (mips_reorg): Split in pre- and post-dbr_schedule parts.
14009 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
14010 (pass_mips_machine_reorg2): New machine specific pass.
14011 (insert_pass_mips_machine_reorg2): New pass plugin definition.
14012 (mips_option_override): Register the new pass.
14013 * rtl.h (cleanup_barriers): Remove prototype.
14014 (dbr_schedule): Likewise.
14015 * jump.c (cleanup_barriers): Make static.
14016 * reorg.c (dbr_schedule): Likewise.
14018 2013-05-06 Richard Biener <rguenther@suse.de>
14020 PR tree-optimization/57185
14021 * tree-parloops.c (add_field_for_reduction): Handle anonymous
14022 SSA names properly.
14024 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
14027 * config/i386/i386.c (add_parameter_dependencies): Add dependence
14028 between "first_arg" and "insn", not "last" and "insn".
14030 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14032 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
14033 (find_candidates_in_block): Re-enable slsr_process_phi.
14034 (create_phi_basis): Fix double counting of candidate adjustment.
14036 2013-05-06 Richard Biener <rguenther@suse.de>
14038 PR middle-end/57147
14039 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
14040 the edge is also fallthru, preserve it and just clear the
14042 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
14043 also complex, preserve that and just clear the fallthru flag.
14044 * tree-inline.c (update_ssa_across_abnormal_edges): Also
14045 update virtual operands.
14047 2013-05-06 Alan Modra <amodra@gmail.com>
14049 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
14050 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
14051 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
14052 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
14053 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
14054 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
14056 2013-05-06 Alan Modra <amodra@gmail.com>
14058 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
14059 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
14060 (DEFAULT_ASM_ENDIAN): Define.
14061 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
14062 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
14063 Update -K PIC clause from sysv4.h.
14064 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
14065 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
14067 2013-05-06 Alan Modra <amodra@gmail.com>
14069 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
14070 twice for little-endian.
14071 (ashrdi3_no_power, ashrdi3): Support little-endian.
14073 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
14076 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
14077 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
14079 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
14080 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
14081 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
14084 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
14085 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
14087 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
14088 * config/pa/pa.opt: Make mbig-switch a no-op.
14089 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
14090 (CASE_VECTOR_MODE): Always return SImode.
14091 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
14092 for the !TARGET_BIG_SWITCH case.
14093 * config/pa/pa-linux.h: Likewise.
14094 * config/pa/pa-openbsd.h: Likewise.
14095 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
14096 * config/pa/pa.md (short_jump): Remove define_insn.
14097 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
14098 (casesi0): Remove define_insn.
14099 (type): Remove btable_branch.
14100 (pa_combine_type): Likewise.
14101 (in_nullified_branch_delay): Likewise.
14102 (in_call_delay): Likewise.
14103 (define_delay): Likewise.
14104 (define_insn_reservation "Z3"): Likewise.
14105 (define_insn_reservation "Z4"): Likewise.
14106 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
14107 (pa_adjust_insn_length): Remove adjustment for btable branches.
14108 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
14111 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
14113 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
14114 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
14115 Add m->r,x alternatives.
14116 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
14117 splitters using SWI48x mode iterator.
14118 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
14119 TARGET_64BIT. Add m->x alternative.
14120 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
14121 Add o->x alternative. Enable for TARGET_SSE.
14122 (sse_storeq): Remove expander.
14123 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
14124 with memory input operand.
14125 (*vec_extractv2di_1 splitter): New.
14126 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
14127 * config/i386/i386.md (ssevecmodelower): New mode attribute.
14129 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
14131 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
14132 (INT_LOWPART): Delete.
14133 (extract_MB): Adjust.
14134 (extract_ME): Adjust.
14135 (print_operand): Adjust.
14137 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
14139 * config/rs6000/predicates.md (reg_or_add_cint_operand,
14140 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
14141 (reg_or_logical_cint_operand, easy_fp_constant,
14142 logical_const_operand): Delete "CONST_DOUBLE" case.
14143 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
14144 "HOST_BITS_PER_WIDE_INT == 64" test.
14145 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
14146 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
14147 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
14148 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
14150 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
14151 CONST_DOUBLE DImode/VOIDmode case.
14152 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
14153 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
14154 CONST_DOUBLE VOIDmode case.
14155 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
14156 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
14157 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
14158 Delete CONST_DOUBLE case.
14159 (splitters for mov FMOVE64 const_double): Delete
14160 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
14161 "HOST_BITS_PER_WIDE_INT >= 64" test.
14162 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
14164 (mov DI const_double): Delete.
14166 2013-05-04 Jakub Jelinek <jakub@redhat.com>
14168 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
14169 on op shows all bits zero in mode of a lowpart subreg, return zero.
14171 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
14174 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
14175 to save TFmode registers and DImode to save TImode registers for
14176 caller save operations.
14177 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
14178 mark being partially clobbered since they only use the first
14181 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
14182 and TDmode only use the upper 64-bits of each VSX register.
14184 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14186 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
14187 (find_candidates_in_block): Disable slsr_process_phi.
14189 2013-05-03 Guozhi Wei <carrot@google.com>
14191 * coverage.c (coverage_obj_init): Move the construction of gcov
14193 (build_init_ctor): ... here.
14195 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14197 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
14198 (slsr_cand_d): Redefine def_phi.
14199 (stride_status, phi_adjust_status, count_phis_status): New enums.
14200 (find_phi_def): New.
14201 (find_basis_for_base_expr): New.
14202 (find_basis_for_candidate): Handle hidden bases.
14203 (alloc_cand_and_find_basis): Handle phi candidates.
14204 (slsr_process_phi): New.
14205 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
14206 (create_mul_imm_cand): Likewise.
14207 (create_add_ssa_cand): Exclude phi base candidates.
14208 (create_add_imm_cand): Likewise.
14209 (slsr_process_cast): Likewise.
14210 (slsr_process_copy): Likewise.
14211 (find_candidates_in_block): Handle phi candidates.
14212 (dump_candidate): Likewise.
14213 (unconditional_cands): Delete.
14214 (unconditional_cands_with_known_stride_p): Delete.
14215 (phi_dependent_cand_p): New.
14216 (cand_increment): Handle phi-dependent candidates.
14217 (replace_dependent): Delete.
14218 (replace_mult_candidate): New.
14219 (replace_unconditional_candidate): New.
14220 (incr_vec_index): Move to avoid forward reference.
14221 (create_add_on_incoming_edge): New.
14222 (create_phi_basis): New.
14223 (replace_dependents): Delete.
14224 (replace_conditional_candidate): New.
14225 (phi_add_costs): New.
14226 (replace_uncond_cands_and_profitable_phis): New.
14227 (record_increment): Handle phi adjustments.
14228 (record_phi_increments): New.
14229 (record_increments): Handle phi adjustments.
14230 (phi_incr_cost): New.
14231 (lowest_cost_path): Handle phis.
14232 (total_savings): Likewise.
14233 (analyze_increments): Likewise.
14234 (ncd_with_phi): New.
14235 (ncd_of_cand_and_phis): New.
14236 (nearest_common_dominator_for_cands): Handle phi increments.
14237 (all_phi_incrs_profitable): New.
14238 (replace_profitable_candidates): Handle phi-dependent candidates.
14239 (analyze_candidates_and_replace): Likewise.
14241 2013-05-03 Teresa Johnson <tejohnson@google.com>
14244 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
14245 do not exceed REG_BR_PROB_BASE.
14247 2013-05-03 Jeff Law <law@redhat.com>
14249 PR tree-optimization/57144
14250 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
14251 operand of the condition will bit into the new type when eliminating
14252 a cast feeding a condition.
14254 2013-05-03 Jakub Jelinek <jakub@redhat.com>
14256 PR rtl-optimization/57130
14257 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
14258 of COMPARE as in_code to the recursive call if needed.
14260 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
14262 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
14263 (enabled): Handle new members.
14264 * config/i386/sse.md (*vec_concatv2si): Merge from
14265 *vec_concatv2si_sse2 and vec_concatv2si_sse.
14266 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
14268 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
14270 PR tree-optimization/57027
14271 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
14272 for fnms opportunity, check we got the prerequisite kind
14273 of tree / gimple before using accessor functions.
14275 2013-05-03 Richard Biener <rguenther@suse.de>
14277 * double-int.h (lshift): New overload without precision
14278 and arith argument.
14279 (operator *=, operator +=, operator -=): Move ...
14280 * double-int.c (operator *=, operator +=, operator -=): ... here
14281 and implement more efficiently.
14282 (mul_double_with_sign): Remove.
14283 (lshift_double): Adjust to take unsinged shift argument, push
14284 dispatching code to callers.
14285 (mul_double_wide_with_sign): Add early out for callers that
14286 are not interested in high parts or overflow.
14287 (lshift): New function.
14288 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
14289 dispatch code here.
14290 (lrotate, rrotate): Use logical shifts.
14291 * expr.c (get_inner_reference): Use lshift.
14292 * fixed-value.c (do_fixed_divide): Likewise.
14293 * tree-dfa.c (get_ref_base_and_extent): Likewise.
14294 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
14295 (indirect_refs_may_alias_p): Likewise.
14296 (stmt_kills_ref_p_1): Likewise.
14298 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
14300 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
14302 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
14304 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
14305 scalar form of FABD instruction.
14307 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
14309 * lra-constraints.c (process_alt_operands): Add checking alt
14310 number to choose the best alternative.
14312 2013-05-02 Richard Biener <rguenther@suse.de>
14314 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
14315 bitmap and its handling.
14316 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
14318 2013-05-02 Richard Biener <rguenther@suse.de>
14320 PR middle-end/57140
14321 * tree-inline.c (copy_loops): Properly handle removed loops.
14322 (copy_cfg_body): Mark destination loops for fixup if source
14323 loops needed fixup.
14325 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
14328 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
14329 generating simple_return for naked functions.
14331 2013-05-02 Martin Jambor <mjambor@suse.cz>
14333 PR middle-end/56988
14334 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
14335 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
14337 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
14338 ipa_agg_replacement_value structures.
14339 (known_aggs_to_agg_replacement_list): Likewise.
14340 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
14341 (read_agg_replacement_chain): Likewise.
14342 (ipcp_transform_function): Also check that by_ref flags match.
14344 2013-05-02 Richard Biener <rguenther@suse.de>
14346 * graphds.h (struct graph): Add obstack member.
14347 * graphds.c (new_graph): Initialize obstack and allocate
14349 (add_edge): Allocate edge from the obstack.
14350 (free_graph): Free the obstack instead of all edges and vertices.
14352 2013-05-02 Teresa Johnson <tejohnson@google.com>
14354 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
14356 * cfg.c (update_bb_profile_for_threading): Ditto.
14357 * tree-inline.c (copy_bb): Ditto.
14358 (copy_edges_for_bb): Ditto.
14359 (initialize_cfun): Ditto.
14360 (copy_cfg_body): Ditto.
14361 (expand_call_inline): Ditto.
14362 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
14363 (estimate_node_size_and_time): Ditto.
14364 (inline_merge_summary): Ditto.
14365 * cgraphclones.c (cgraph_clone_edge): Ditto.
14366 (cgraph_clone_node): Ditto.
14367 * sched-rgn.c (compute_dom_prob_ps): Ditto.
14368 (compute_trg_info): Ditto.
14370 2013-05-02 Ian Bolton <ian.bolton@arm.com>
14372 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
14373 S reg when fp attribute set.
14374 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
14376 2013-05-02 Ian Bolton <ian.bolton@arm.com>
14378 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
14380 (*and_one_cmplsi3_compare0_uxtw): Likewise.
14381 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
14382 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
14384 2013-05-02 Richard Biener <rguenther@suse.de>
14386 * tree-scalar-evolution.c (scev_info_hasher): Remove.
14387 (struct instantiate_cache_entry): New type.
14388 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
14389 (struct instantiate_cache_type): New type.
14390 (set_instantiated_value, get_instantiated_value): Remove.
14391 (get_instantiated_value_entry): New function.
14392 (instantiate_scev_name): Use the new cache and adjust.
14393 (instantiate_scev_poly): Adjust.
14394 (instantiate_scev_binary): Likewise.
14395 (instantiate_array_ref): Likewise.
14396 (instantiate_scev_convert): Likewise.
14397 (instantiate_scev_not): Likewise.
14398 (instantiate_scev_3): Likewise.
14399 (instantiate_scev_2): Likewise.
14400 (instantiate_scev_r): Likewise.
14401 (instantiate_scev): Likewise.
14402 (resolve_mixers): Likewise.
14404 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
14407 * lra-constraints.c (best_small_class_operands_num): Remove.
14408 (process_alt_operands): Remove small_class_operands_num. Take
14409 small classes operands into losers and only if the operand is not
14410 matched. Modify debugging output.
14411 (curr_insn_transform): Remove best_small_class_operands_num.
14414 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14416 * config/aarch64/aarch64-builtins.c
14417 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
14418 * config/aarch64/aarch64-simd-builtins.def
14419 (reduc_splus_): Add new modes.
14420 (reduc_uplus_): New.
14421 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
14422 (reduc_uplus_v4sf): Likewise.
14423 (reduc_splus_v4sf): Likewise.
14424 (aarch64_addv<mode>): Likewise.
14425 (reduc_uplus_<mode>): Likewise.
14426 (reduc_splus_<mode>): Likewise.
14427 (aarch64_addvv2di): Likewise.
14428 (reduc_uplus_v2di): Likewise.
14429 (reduc_splus_v2di): Likewise.
14430 (aarch64_addvv2si): Likewise.
14431 (reduc_uplus_v2si): Likewise.
14432 (reduc_splus_v2si): Likewise.
14433 (reduc_<sur>plus_<mode>): New.
14434 (reduc_<sur>plus_v2di): Likewise.
14435 (reduc_<sur>plus_v2si): Likewise.
14436 (reduc_<sur>plus_v4sf): Likewise.
14437 (aarch64_addpv4sf): Likewise.
14438 * config/aarch64/arm_neon.h
14439 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
14440 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
14441 add UNSPEC_SADDV, UNSPEC_UADDV.
14443 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
14445 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14447 * config/aarch64/arm_neon.h
14448 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
14450 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14452 * config/aarch64/aarch64-builtins
14453 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
14455 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14457 * config/aarch64/aarch64-simd-builtins.def
14458 (reduc_smax_): New.
14459 (reduc_smin_): Likewise.
14460 (reduc_umax_): Likewise.
14461 (reduc_umin_): Likewise.
14462 (reduc_smax_nan_): Likewise.
14463 (reduc_smin_nan_): Likewise.
14466 (smax): Update for V2SF, V4SF and V2DF modes.
14469 (smin_nan): Likewise.
14470 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
14471 (<su><maxmin><mode>3): ...This, refactor.
14472 (s<maxmin><mode>3): New.
14473 (<maxmin_uns><mode>3): Likewise.
14474 (reduc_<maxmin_uns>_<mode>): Refactor.
14475 (reduc_<maxmin_uns>_v4sf): Likewise.
14476 (reduc_<maxmin_uns>_v2si): Likewise.
14477 (aarch64_<fmaxmin><mode>: Remove.
14478 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
14480 (vmin<q>_f<32,64>): Likewise.
14481 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
14483 (su): Add mappings for smax, smin, umax, umin.
14485 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
14486 (FMAXMIN): Rename as...
14487 (FMAXMIN_UNS): ...This.
14489 (fmaxminv): Likewise.
14490 (fmaxmin): Likewise.
14492 (maxmin_uns_op): Likewise.
14494 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14496 * config/aarch64/arm_neon.h
14497 (vac<ge, gt><sd>_f<32, 64>): Rename to...
14498 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
14499 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
14501 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14503 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
14504 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
14506 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14508 * config/aarch64/aarch64-simd.md
14509 (vcond<mode>_internal): Handle special cases for constant masks.
14510 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
14511 (vcondu<mode><mode>): Likewise.
14512 (vcond<v_cmp_result><mode>): New.
14514 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14516 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
14517 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
14518 * config/aarch64/aarch64-simd-builtins.def
14519 (cmeq): Update to BUILTIN_VALLDI.
14524 * config/aarch64/arm_neon.h
14525 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
14526 to builtins or C as appropriate.
14528 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14530 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
14532 (cmhi): Rename to...
14534 * config/aarch64/aarch64-simd.md
14535 (simd_mode): Add SF.
14536 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
14537 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
14538 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
14539 (cstore<mode>_neg): ...This.
14540 * config/aarch64/iterators.md
14542 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
14543 (COMPARISONS): New.
14544 (UCOMPARISONS): Likewise.
14545 (optab): Add missing comparisons.
14551 (VCMP_S): Likewise.
14552 (VCMP_U): Likewise.
14553 (V_cmp_result): Add DF, SF modes.
14554 (v_cmp_result): Likewise.
14556 (vmtype): Likewise.
14557 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
14559 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
14561 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
14562 define_insn to define_insn_and_split.
14563 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
14564 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
14565 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
14566 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
14567 (thumb2_negscc): Likewise.
14569 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
14571 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
14573 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
14575 * config/arm/thumb2.md: Remove trailing whitespaces.
14577 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
14579 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
14580 Use gen_int_mode rather than GEN_INT.
14582 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
14584 * value-prof.c (stream_in_histogram_value): Remove the strayed
14587 2013-04-30 Richard Biener <rguenther@suse.de>
14589 PR middle-end/57122
14590 * cfghooks.c (split_edge): Properly check for the loop latch edge.
14592 2013-04-30 Richard Biener <rguenther@suse.de>
14594 PR middle-end/57107
14595 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
14597 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
14599 PR rtl-optimization/56957
14600 PR rtl-optimization/57105
14601 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
14602 variable. Use just INSN_UID for determining whether an insn
14603 should be only disconnected from the insn stream.
14604 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
14606 2013-04-30 Jakub Jelinek <jakub@redhat.com>
14608 PR tree-optimization/57104
14609 * tsan.c (instrument_expr): Don't instrument accesses to
14610 DECL_HARD_REGISTER VAR_DECLs.
14612 2013-04-30 Richard Biener <rguenther@suse.de>
14614 * function.h (loops_for_fn): New inline function.
14615 (set_loops_for_fn): Likewise.
14616 * cfgloop.h (place_new_loop): Add struct function parameter.
14617 (get_loop): Likewise.
14618 (get_loops): Likewise.
14619 (number_of_loops): Likewise.
14620 (fel_next): Adjust.
14621 (fel_init): Likewise.
14622 * cfg.c (get_loop_copy): Adjust.
14623 * cfgloop.c (flow_loops_dump): Likewise.
14624 (record_loop_exits): Likewise.
14625 (verify_loop_structure): Likewise.
14626 * cfgloopanal.c (mark_irreducible_loops): Likewise.
14627 (estimate_reg_pressure_cost): Likewise.
14628 (mark_loop_exit_edges): Likewise.
14629 * cfgloopmanip.c (place_new_loop): Likewise.
14630 (add_loop): Likewise.
14631 (duplicate_loop): Likewise.
14632 * graph.c (draw_cfg_nodes): Likewise.
14633 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
14634 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
14635 (extract_affine_chrec): Likewise.
14636 (build_scop_iteration_domain): Likewise.
14637 * graphite.c (graphite_initialize): Likewise.
14638 * ira-build.c (create_loop_tree_nodes): Likewise.
14639 (more_one_region_p): Likewise.
14640 (rebuild_regno_allocno_maps): Likewise.
14641 (mark_loops_for_removal): Likewise.
14642 (mark_all_loops_for_removal): Likewise.
14643 (remove_unnecessary_regions): Likewise.
14644 (ira_build): Likewise.
14645 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
14646 * loop-init.c (fix_loop_structure): Likewise.
14647 (gate_rtl_move_loop_invariants): Likewise.
14648 (gate_rtl_unswitch): Likewise.
14649 (gate_rtl_unroll_and_peel_loops): Likewise.
14650 (rtl_doloop): Likewise.
14651 * lto-streamer-in.c (input_cfg): Likewise.
14652 * lto-streamer-out.c (output_cfg): Likewise.
14653 * modulo-sched.c (sms_schedule): Likewise.
14654 * predict.c (tree_estimate_probability): Likewise.
14655 (tree_estimate_probability_driver): Likewise.
14656 (estimate_loops): Likewise.
14657 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
14658 (move_sese_region_to_fn): Likewise.
14659 (debug_loop_num): Likewise.
14660 * tree-chrec.c (chrec_evaluate): Likewise.
14661 (hide_evolution_in_other_loops_than_loop): Likewise.
14662 (chrec_component_in_loop_num): Likewise.
14663 (reset_evolution_in_loop): Likewise.
14664 (evolution_function_is_invariant_rec_p): Likewise.
14665 * tree-if-conv.c (main_tree_if_conversion): Likewise.
14666 * tree-inline.c (copy_loops): Likewise.
14667 (copy_cfg_body): Likewise.
14668 (tree_function_versioning): Likewise.
14669 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
14670 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
14672 (add_to_evolution_1): Likewise.
14673 (scev_const_prop): Likewise.
14674 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
14675 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
14676 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
14677 (tree_ssa_lim_initialize): Likewise.
14678 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
14679 (verify_loop_closed_ssa): Likewise.
14680 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
14681 (tree_ssa_loop_im): Likewise.
14682 (tree_ssa_loop_unswitch): Likewise.
14683 (tree_vectorize): Likewise.
14684 (check_data_deps): Likewise.
14685 (tree_ssa_loop_ivcanon): Likewise.
14686 (tree_ssa_loop_bounds): Likewise.
14687 (tree_complete_unroll): Likewise.
14688 (tree_complete_unroll_inner): Likewise.
14689 (tree_parallelize_loops): Likewise.
14690 (tree_ssa_loop_prefetch): Likewise.
14691 (tree_ssa_loop_ivopts): Likewise.
14692 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
14693 * tree-vectorizer.c (vectorize_loops): Likewise.
14695 2013-04-29 Mike Frysinger <vapier@gentoo.org>
14697 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
14698 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
14699 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
14700 with EABI_LINK_SPEC.
14702 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
14705 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
14708 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
14711 * lra-constraints.c (process_alt_operands): Discourage a bit more
14712 using memory for pseudos. Print cost dump for alternatives.
14713 Modify cost values for conflicts with early clobbers.
14714 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
14716 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
14719 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
14721 2013-04-29 Ian Bolton <ian.bolton@arm.com>
14723 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
14724 from/to S register.
14725 (movdi_aarch64): Support LDR/STR from/to D register.
14727 2013-04-29 Ian Bolton <ian.bolton@arm.com>
14729 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
14730 or higher by default.
14732 2013-04-29 Richard Biener <rguenther@suse.de>
14734 PR middle-end/57075
14735 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
14736 even if not adding abnormal edges for calls that can make
14739 2013-04-29 Richard Biener <rguenther@suse.de>
14741 PR middle-end/57103
14742 * tree-cfg.c (move_stmt_op): Fix condition under which to update
14744 (move_stmt_r): Remove redundant checking.
14746 2013-04-29 Teresa Johnson <tejohnson@google.com>
14749 * basic-block.h (apply_scale): New function.
14750 (apply_probability): Use apply_scale.
14751 * gimple-streamer-in.c (input_bb): Ditto.
14752 * lto-streamer-in.c (input_cfg): Ditto.
14753 * lto-cgraph.c (merge_profile_summaries): Ditto.
14754 * tree-optimize.c (execute_fixup_cfg): Ditto.
14755 * tree-inline.c (copy_bb): Update comment to use apply_scale.
14756 (copy_edges_for_bb): Ditto.
14757 (copy_cfg_body): Ditto.
14759 2013-04-29 Tom de Vries <tom@codesourcery.com>
14761 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
14762 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
14763 (tail_merge_optimize): Handle current_loops == NULL.
14765 2013-04-26 Jeff Law <law@redhat.com>
14767 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
14768 (simplify_cond_using_ranges): Generalize code to simplify
14769 COND_EXPRs where one argument is a constant and the other
14770 is an SSA_NAME created by an integral type conversion.
14772 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14774 * config/arm/arm.md (store_minmaxsi): Use only when
14775 optimize_insn_for_size_p.
14777 2013-04-29 Christian Bruel <christian.bruel@st.com>
14780 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
14782 2013-04-29 Richard Biener <rguenther@suse.de>
14784 PR middle-end/57089
14785 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
14786 loop tree make sure to schedule a fixup for the child as well.
14787 (expand_omp_for_generic): Properly add loops.
14788 (expand_omp_for_static_nochunk): Likewise.
14789 (expand_omp_for_static_chunk): Likewise.
14790 (expand_omp_for): For the degenerate case fixup loops.
14791 (expand_omp_sections): Fix default bb placement in loops.
14792 (expand_omp_atomic_pipeline): Properly add loops.
14794 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14796 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
14798 2013-04-29 Tom de Vries <tom@codesourcery.com>
14800 * tree-ssa-tail-merge.c: Update header comment.
14802 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14804 * config/aarch64/arm_neon.h
14805 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
14806 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
14807 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
14808 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
14809 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
14810 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
14811 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
14812 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
14814 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14816 * config/aarch64/aarch64-simd.md
14817 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
14818 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
14819 fix_trunc, fixuns_trunc.
14820 (ftrunc<VDQF:mode>2): New.
14821 * config/aarch64/iterators.md (optab): Add fix, fixuns.
14822 (fix_trunc_optab): New.
14824 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14826 * config/aarch64/aarch64-builtins.c
14827 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
14828 iceilf, lround, iroundf.
14830 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
14833 * config/i386/i386.h (enum ix86_tune_indices)
14834 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
14835 New, split from X86_TUNE_INTER_UNIT_MOVES.
14836 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
14837 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
14838 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
14839 (TARGET_INTER_UNIT_MOVES): Remove.
14840 * config/i386/i386.c (initial_ix86_tune_features): Update.
14841 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
14842 (ix86_expand_convert_uns_didf_sse): Use
14843 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
14844 (ix86_expand_vector_init_one_nonzero): Ditto.
14845 (ix86_expand_vector_init_interleave): Ditto.
14846 (inline_secondary_memory_needed): Return true for moves from SSE class
14847 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
14848 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
14849 * config/i386/constraints.md (Yi, Ym): Depend on
14850 TARGET_INTER_UNIT_MOVES_TO_VEC.
14851 (Yj, Yn): New constraints.
14852 * config/i386/i386.md (*movdi_internal): Change constraints of
14853 operand 1 from Yi to Yj and from Ym to Yn.
14854 (*movsi_internal): Ditto.
14855 (*movdf_internal): Ditto.
14856 (*movsf_internal): Ditto.
14857 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
14858 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
14859 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
14860 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
14861 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
14862 * config/i386/sse.md (movdi_to_sse): Ditto.
14863 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
14864 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
14865 TARGET_INTER_UNIT_MOVES.
14866 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
14867 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
14868 instead of TARGET_INTER_UNIT_MOVES.
14869 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
14870 operand 1 from Yi to Yj and from Ym to Yn.
14872 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14874 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
14875 (float_truncate_hi_): Likewise.
14876 (float_extend_lo_): Likewise.
14877 (float_truncate_lo_): Likewise.
14878 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
14879 (aarch64_float_extend_lo_v2df): Likewise.
14880 (vec_unpacks_hi_v4sf): Likewise.
14881 (aarch64_float_truncate_lo_v2sf): Likewise.
14882 (aarch64_float_truncate_hi_v4sf): Likewise.
14883 (vec_pack_trunc_v2df): Likewise.
14884 (vec_pack_trunc_df): Likewise.
14886 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14888 * config/aarch64/aarch64-builtins.c
14889 (aarch64_fold_builtin): Fold float conversions.
14890 * config/aarch64/aarch64-simd-builtins.def
14891 (floatv2si, floatv4si, floatv2di): New.
14892 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
14893 * config/aarch64/aarch64-simd.md
14894 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
14895 * config/aarch64/iterators.md (FLOATUORS): New.
14896 (optab): Add float, floatuns.
14897 (su_optab): Likewise.
14899 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14901 * config/aarch64/aarch64-builtins.c
14902 (aarch64_builtin_vectorized_function): Use new names for
14904 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
14905 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
14906 (fcvtzu): Split as...
14907 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
14908 (fcvtas): Split as...
14909 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
14910 (fcvtau): Split as...
14911 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
14912 (fcvtps): Split as...
14913 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
14914 (fcvtpu): Split as...
14915 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
14916 (fcvtms): Split as...
14917 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
14918 (fcvtmu): Split as...
14919 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
14920 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
14921 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
14922 (lfrintnusf, lfrintnudf): Likewise.
14923 * config/aarch64/aarch64-simd.md
14924 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
14926 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
14927 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
14928 (fcvt_pattern): Likewise.
14930 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14932 * config/aarch64/aarch64-simd.md
14933 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
14934 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
14936 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14938 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
14939 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
14940 (vrnd<a,m,n,p>_f32): Implement using builtins.
14941 (vrnd<i,x><q>_f<32, 64>): New.
14943 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14945 * config/aarch64/aarch64-builtins.c
14946 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
14947 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
14948 (frintz): Rename to...
14950 (frintp): Rename to...
14952 (frintm): Rename to...
14954 (frinti): Rename to...
14955 (nearbyint): ...this.
14956 (frintx): Rename to...
14958 (frinta): Rename to...
14960 * config/aarch64/aarch64-simd.md
14961 (aarch64_frint<frint_suffix><mode>): Delete.
14962 (<frint_pattern><mode>2): Convert to insn.
14963 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
14964 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
14965 (frint_pattern): Likewise.
14966 (frint_suffix): Likewise.
14968 2013-04-29 Richard Biener <rguenther@suse.de>
14970 PR tree-optimization/57081
14971 * loop-init.c: Include tree-flow.h.
14972 (loop_optimizer_finalize): Free number of iteration estimates.
14973 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
14975 2013-04-29 Jakub Jelinek <jakub@redhat.com>
14977 PR tree-optimization/57083
14978 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
14979 non-singleton shift count range, zero extend low_bound for uns case.
14981 * config/i386/predicates.md (general_vector_operand): New predicate.
14982 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
14983 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
14984 if they aren't nonimmediate operands. If their original values
14985 satisfy const_vector_equal_evenodd_p, don't shift them.
14986 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
14987 predicates. For the SSE4.1 case force operands[{1,2}] into registers
14988 if not nonimmediate_operand.
14989 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
14990 instead of register_operand.
14991 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
14993 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
14995 * stor-layout.c (finalize_size_functions): Allocate a structure and
14996 reset cfun before dumping the functions.
14998 2013-04-27 Jakub Jelinek <jakub@redhat.com>
15000 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
15003 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
15004 use xop_pmacsdqh if uns_p.
15005 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
15006 the immediate rotate count.
15008 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
15010 * rtl.h (struct rtx_def): Add comment for field jump.
15011 (LRA_SUBREG_P): New macro.
15012 * recog.c (register_operand): Check LRA_SUBREG_P.
15013 * lra.c (lra): Add note at the end of RTL code. Align non-empty
15015 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
15016 (lra_final_code_change): Skip subreg change for operators.
15017 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
15018 if there are no operand changes.
15019 * lra-constraints.c (curr_insn_set): New.
15020 (match_reload): Set LRA_SUBREG_P.
15021 (emit_spill_move): Ditto.
15022 (check_and_process_move): Use curr_insn_set. Process only single
15023 set insns. Don't initialize sec_mem_p and change_p.
15024 (simplify_operand_subreg): Use LRA_SUBREG_P.
15025 (reg_in_class_p): New function.
15026 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
15027 of #ifdef. Add code to remove cycling.
15028 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
15029 non-null disp. Reload inner instead of disp when base and index
15030 are null. Try to put lo_sum into register.
15031 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
15032 (check_and_process_move): Move code for move cost check to
15033 simple_move_p. Remove equiv_substitution.
15034 (simple_move_p): New function.
15035 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
15036 curr_insn_set. Call check_and_process_move only for single set
15037 insns. Use the new function. Move call of check_and_process_move
15038 after operand equiv substitution and address process.
15040 2013-04-26 Jakub Jelinek <jakub@redhat.com>
15043 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
15044 with nonlocal goto receivers or returns twice calls, ignore
15045 unininitialized values from abnormal edges to nl goto receiver
15046 or returns twice call.
15048 2013-04-26 Jakub Jelinek <jakub@redhat.com>
15050 PR tree-optimization/57051
15051 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
15052 and VEC_RSHIFT_EXPR if shift count is a multiple of element
15055 2013-04-26 Richard Biener <rguenther@suse.de>
15057 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
15058 (expand_omp_taskreg): Likewise. Mark loops for fixup.
15059 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
15060 (fixup_loop_arrays_after_move): New function.
15061 (move_sese_region_to_fn): Properly outline the loop tree parts
15062 of the SESE region.
15064 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
15066 * config/i386/i386.md (type, unit): Fix long lines.
15068 2013-04-26 Richard Biener <rguenther@suse.de>
15070 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
15071 (lto-streamer-out.o): Likewise.
15072 * cfgloop.c (init_loops_structure): Export, add struct function
15073 argument and adjust.
15074 (flow_loops_find): Adjust.
15075 * cfgloop.h (enum loop_estimation): Add EST_LAST.
15076 (init_loops_structure): Declare.
15077 * lto-streamer-in.c: Include cfgloop.h.
15078 (input_cfg): Input the loop tree.
15079 * lto-streamer-out.c: Include cfgloop.h.
15080 (output_cfg): Output the loop tree.
15081 (output_struct_function_base): Do not drop PROP_loops.
15083 2013-03-26 Richard Biener <rguenther@suse.de>
15085 * tree-cfg.c (execute_build_cfg): Build the loop tree.
15086 (pass_build_cfg): Provide PROP_loops.
15087 (move_sese_region_to_fn): Remove loops that are outlined into fn
15089 * tree-inline.c: Include cfgloop.h.
15090 (initialize_cfun): Do not drop PROP_loops.
15091 (copy_loops): New function.
15092 (copy_cfg_body): Copy loop structure.
15093 (tree_function_versioning): Initialize destination loop tree.
15094 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
15095 (pass_parallelize_loops): Do IL verification.
15096 * loop-init.c (loop_optimizer_init): Fixup loops if required.
15097 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
15098 the CFG make sure we fixup loops as well.
15099 * tree-ssa-tail-merge.c: Include cfgloop.h.
15100 (replace_block_by): When merging loop latches mark loops for fixup.
15101 * lto-streamer-out.c (output_struct_function_base): Drop
15102 PROP_loops for now.
15103 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
15104 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
15105 * ipa-split.c: Include cfgloop.h.
15106 (split_function): Add the new return block to the loop tree root.
15107 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
15108 whether we have removed the forwarder block.
15109 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
15110 * cfgloop.h (place_new_loop): Declare.
15111 * cfgloopmanip.c (place_new_loop): Export.
15112 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
15113 (tree-switch-conversion.o): Likewise.
15114 (tree-complex.o): Likewise.
15115 (tree-inline.o): Likewise.
15116 (tree-ssa-tailmerge.o): Likewise.
15117 (ipa-split.o): Likewise.
15118 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
15119 (tree-ssa-copy.o): Likewise.
15120 * tree-switch-conversion.c: Include cfgloop.h
15121 (process_switch): If we emit a bit-test cascade, schedule loops
15123 * tree-complex.c: Include cfgloop.h.
15124 (expand_complex_div_wide): Properly add new basic-blocks to loops.
15125 * asan.c: Include cfgloop.h.
15126 (create_cond_insert_point): Properly add new basic-blocks to
15127 loops, schedule loop fixup.
15128 * cfgloop.c (verify_loop_structure): Check that looks are not
15130 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
15132 (expand_omp_for_generic): Likewise.
15133 (expand_omp_sections): Likewise.
15134 (expand_omp_atomic_pipeline): Schedule loops for fixup.
15135 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
15136 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
15137 is initialized, not when loops are present.
15138 * tree-parloops.c (parallelize_loops): Remove checking here.
15139 * passes.c (init_optimization_passes): Schedule a copy-propagation
15140 pass before complete unrolling of inner loops.
15142 2013-04-26 Jakub Jelinek <jakub@redhat.com>
15144 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
15145 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
15146 (colorize_init): Add argument to _WIN32 version.
15147 * toplev.c: Include diagnostic-color.h.
15148 (process_options): Default to -fdiagnostics-color=auto if
15149 GCC_COLORS env var is in the environment.
15150 * common.opt (fdiagnostics-color=): Add Var and Init.
15151 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
15152 env var is in the environment, the default is auto rather than never.
15154 * diagnostic.h (file_name_as_prefix): Add context argument.
15155 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
15156 the string as locus.
15157 * langhooks.c (lhd_print_error_function): Adjust caller.
15159 2013-04-25 Lawrence Crowl <crowl@google.com>
15161 * var-tracking.c (shared_hash_def::htab):
15162 Change type to hash_table. Update dependent calls and types.
15164 2013-04-25 Lawrence Crowl <crowl@google.com>
15166 * Makefile.in: Update as needed below.
15168 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
15169 Move declaration to after the type's method definitons.
15171 * attribs.c (htab_t scoped_attributes::attribute_hash):
15172 Change type to hash_table. Update dependent calls and types.
15174 * bitmap.c (htab_t bitmap_desc_hash):
15175 Change type to hash_table. Update dependent calls and types.
15177 * cselib.c (htab_t cselib_hash_table):
15178 Change type to hash_table. Update dependent calls and types.
15180 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
15181 (hash_string_slot_node): Move implementation into lto-streamer.h
15182 struct string_slot_hasher.
15183 (eq_string_slot_node): Likewise.
15185 * data-streamer-out.c: Update output_block::string_hash_table
15186 dependent calls and types.
15188 * dwarf2cfi.c (htab_t trace_index):
15189 Change type to hash_table. Update dependent calls and types.
15191 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
15192 Change type to hash_table. Update dependent calls and types.
15193 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
15194 (htab_t optimize_external_refs::map): Likewise.
15195 (htab_t output_comp_unit::extern_map): Likewise.
15196 (htab_t output_comdat_type_unit::extern_map): Likewise.
15197 (htab_t output_macinfo::macinfo_htab): Likewise.
15198 (htab_t optimize_location_lists::htab): Likewise.
15199 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
15201 * except.c (htab_t ehspec_hash_type):
15202 Change type to hash_table. Update dependent calls and types.
15203 (assign_filter_values::ttypes): Likewise.
15204 (assign_filter_values::ehspec): Likewise.
15205 (sjlj_assign_call_site_values::ar_hash): Likewise.
15206 (convert_to_eh_region_ranges::ar_hash): Likewise.
15208 * gcse.c (htab_t pre_ldst_table):
15209 Change type to hash_table. Update dependent calls and types.
15211 * ggc-common.c (htab_t saving_htab):
15212 Change type to hash_table. Update dependent calls and types.
15213 (htab_t loc_hash): Likewise.
15214 (htab_t ptr_hash): Likewise.
15215 (call_count): Rename ggc_call_count.
15216 (call_alloc): Rename ggc_call_alloc.
15217 (loc_descriptor): Rename make_loc_descriptor.
15218 (add_statistics): Rename ggc_add_statistics.
15220 * ggc-common.c (saving_htab):
15221 Change type to hash_table. Update dependent calls and types.
15223 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
15224 (push_gimplify_context): Likewise.
15225 (pop_gimplify_context): Likewise.
15226 (struct gimple_temp_hash_elt): Added.
15227 (struct gimplify_hasher): Likewise.
15228 (struct gimplify_ctx.temp_htab):
15229 Change type to hash_table. Update dependent calls and types.
15231 * gimple-fold.c: Include gimplify-ctx.h.
15233 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
15234 Change type to hash_table. Update dependent calls and types.
15235 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
15236 avoid potential global name collision.
15238 * gimplify.c: Include gimplify-ctx.h.
15239 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
15240 (htab_t gimplify_ctx::temp_htab):
15241 Update dependent calls and types for new type hash_table.
15242 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
15243 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
15245 * gimplify-ctx.h: New.
15246 (struct gimple_temp_hash_elt): Move from gimplify.c.
15247 (class gimplify_hasher): New.
15248 (struct gimplify_ctx): Move from gimple.h.
15249 (htab_t gimplify_ctx::temp_htab):
15250 Change type to hash_table. Update dependent calls and types.
15252 * graphite-clast-to-gimple.c: Include graphite-htab.h.
15253 (htab_t ivs_params::newivs_index):
15254 Change type to hash_table. Update dependent calls and types.
15255 (htab_t ivs_params::params_index): Likewise.
15256 (htab_t print_generated_program::params_index): Likewise.
15257 (htab_t gloog::newivs_index): Likewise.
15258 (htab_t gloog::params_index): Likewise.
15260 * graphite.c: Include graphite-htab.h.
15261 4htab_t graphite_transform_loops::bb_pbb_mapping):
15262 Change type to hash_table. Update dependent calls and types.
15264 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
15265 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
15266 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
15268 * graphite-dependences.c: Include graphite-htab.h.
15269 (loop_is_parallel_p): Change hash table type of parameter.
15271 * graphite-htab.h: New.
15272 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
15273 (extern find_pbb_via_hash): Move from graphite-poly.h.
15274 (extern loop_is_parallel_p): Move from graphite-poly.h.
15275 (extern get_loop_body_pbbs): Move from graphite-poly.h.
15277 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
15278 (extern loop_is_parallel_p): Move to graphite-htab.h.
15279 (extern get_loop_body_pbbs): Move to graphite-htab.h.
15281 * haifa-sched.c (htab_t delay_htab):
15282 Change type to hash_table. Update dependent calls and types.
15283 (htab_t delay_htab_i2): Likewise.
15285 * ira-color.c (htab_t allocno_hard_regs_htab):
15286 Change type to hash_table. Update dependent calls and types.
15288 * ira-costs.c (htab_t cost_classes_htab):
15289 Change type to hash_table. Update dependent calls and types.
15291 * loop-invariant.c (htab_t merge_identical_invariants::eq):
15292 Change type to hash_table. Update dependent calls and types.
15294 * loop-iv.c (htab_t bivs):
15295 Change type to hash_table. Update dependent calls and types.
15297 * loop-unroll.c (htab_t opt_info::insns_to_split):
15298 Change type to hash_table. Update dependent calls and types.
15299 (htab_t opt_info::insns_with_var_to_expand): Likewise.
15301 * lto-streamer.h (struct string_slot): Move from data-streamer.h
15302 (struct string_slot_hasher): New.
15303 (htab_t output_block::string_hash_table):
15304 Change type to hash_table. Update dependent calls and types.
15306 * lto-streamer-in.c (freeing_string_slot_hasher): New.
15307 (htab_t file_name_hash_table):
15308 Change type to hash_table. Update dependent calls and types.
15310 * lto-streamer-out.c: Update output_block::string_hash_table dependent
15313 * lto-streamer.c (htab_t tree_htab):
15314 Change type to hash_table. Update dependent calls and types.
15316 * omp-low.c: Include gimplify-ctx.h.
15318 * passes.c (htab_t name_to_pass_map):
15319 Change type to hash_table. Update dependent calls and types.
15320 (pass_traverse): Rename to passes_pass_traverse.
15322 * plugin.c (htab_t event_tab):
15323 Change type to hash_table. Update dependent calls and types.
15325 * postreload-gcse.c (htab_t expr_table):
15326 Change type to hash_table. Update dependent calls and types.
15327 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
15329 * sese.c (debug_rename_map_1): Make extern.
15330 (htab_t copy_bb_and_scalar_dependences::rename_map):
15331 Change type to hash_table. Update dependent calls and types.
15333 * sese.h (extern debug_rename_map): Move to .c file.
15335 * store-motion.c (htab_t store_motion_mems_table):
15336 Change type to hash_table. Update dependent calls and types.
15338 * trans-mem.c (htab_t tm_new_mem_hash):
15339 Change type to hash_table. Update dependent calls and types.
15341 * tree-browser.c (htab_t TB_up_ht):
15342 Change type to hash_table. Update dependent calls and types.
15344 * tree-cfg.c (htab_t discriminator_per_locus):
15345 Change type to hash_table. Update dependent calls and types.
15347 * tree-complex.c: Include tree-hasher.h
15348 (htab_t complex_variable_components):
15349 Change type to hash_table. Update dependent calls and types.
15351 * tree-eh.c (htab_t finally_tree):
15352 Change type to hash_table. Update dependent calls and types.
15354 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
15355 struct int_tree_hasher.
15356 (extern int_tree_map_eq): Likewise.
15357 (uid_decl_map_hash): Removed.
15358 (extern decl_tree_map_eq): Likewise.
15360 * tree-hasher.h: New.
15361 (struct int_tree_hasher): New.
15362 (typedef int_tree_htab_type): New.
15364 * tree-inline.c: Include gimplify-ctx.h.
15366 * tree-mudflap.c: Include gimplify-ctx.h.
15368 * tree-parloops.c: Include tree-hasher.h.
15369 (htab_t eliminate_local_variables_stmt::decl_address):
15370 Change type to hash_table. Update dependent calls and types.
15371 (htab_t separate_decls_in_region::decl_copies): Likewise.
15373 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
15374 Change type to hash_table. Update dependent calls and types.
15376 * tree-sra.c (candidates):
15377 Change type to hash_table. Update dependent calls and types.
15379 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
15381 (int_tree_map_hash): Likewise.
15383 * tree-ssa-dom.c (htab_t avail_exprs):
15384 Change type to hash_table. Update dependent calls and types.
15386 * tree-ssa-live.c (var_map_base_init::tree_to_index):
15387 Change type to hash_table. Update dependent calls and types.
15389 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
15390 Change type to hash_table. Update dependent calls and types.
15392 * tree-ssa-phiopt.c (seen_ssa_names):
15393 Change type to hash_table. Update dependent calls and types.
15395 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
15396 Change type to hash_table. Update dependent calls and types.
15398 * tree-ssa-uncprop.c (equiv):
15399 Change type to hash_table. Update dependent calls and types.
15401 2013-04-25 Jakub Jelinek <jakub@redhat.com>
15403 PR rtl-optimization/57003
15404 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
15405 call note_stores with kill_clobbered_value callback again after
15406 killing regs_invalidated_by_call.
15408 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
15410 * config/aarch64/aarch64-simd.md
15411 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
15412 (aarch64_simd_bsl<mode>): Likewise.
15413 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
15415 2013-04-25 Marek Polacek <polacek@redhat.com>
15417 PR tree-optimization/57066
15418 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
15420 2013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
15422 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
15424 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
15426 * config/aarch64/aarch64-builtins.c
15427 (aarch64_fold_builtin): New.
15428 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
15429 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
15430 * config/aarch64/aarch64-simd-builtins.def (abs): New.
15431 * config/aarch64/arm_neon.h
15432 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
15434 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
15435 Tejas Belagod <tejas.belagod@arm.com>
15437 * config/aarch64/aarch64-builtins.c
15438 (aarch64_gimple_fold_builtin): New.
15439 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
15440 * config/aarch64/aarch64-simd-builtins.def (addv): New.
15441 * config/aarch64/aarch64-simd.md (addpv4sf): New.
15442 (addvv4sf): Update.
15443 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
15445 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
15447 * config/aarch64/aarch64.md
15448 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
15450 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
15452 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
15453 (*ngcsi_uxtw): New pattern.
15455 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15456 Julian Brown <julian@codesourcery.com>
15458 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
15459 (TB_DREG): Add T_V4HF.
15460 (v4hf_UP): New macro.
15461 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
15462 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
15463 Handle initialisation of V4HF. Adjust initialisation of reinterpret
15465 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
15466 (arm_vector_mode_supported_p): Handle V4HF.
15467 (arm_mangle_map): Handle V4HFmode.
15468 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
15469 * config/arm/arm_neon_builtins.def: Add entries for
15470 vcvtv4hfv4sf, vcvtv4sfv4hf.
15471 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
15472 (neon_vcvtv4hfv4sf): Likewise.
15473 * config/arm/neon-gen.ml: Handle half-precision floating point
15475 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
15476 * config/arm/arm_neon.h: Regenerate.
15477 * config/arm/neon.ml (type elts): Add F16.
15478 (type vectype): Add T_float16x4, T_floatHF.
15479 (type vecmode): Add V4HF.
15480 (type features): Add Requires_FP_bit feature.
15481 (elt_width): Handle F16.
15482 (elt_class): Likewise.
15483 (elt_of_class_width): Likewise.
15484 (mode_of_elt): Refactor.
15485 (type_for_elt): Handle F16, fix error messages.
15486 (vectype_size): Handle T_float16x4.
15487 (vcvt_sh): New function.
15488 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
15489 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
15490 (string_of_mode): Handle V4HF.
15491 * doc/arm-neon-intrinsics.texi: Regenerate.
15493 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
15495 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
15496 format specifier in 'X' case.
15498 2013-04-25 Alan Modra <amodra@gmail.com>
15501 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
15502 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
15503 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
15504 Repeat for many other rotate/shift and mask patterns using subregs.
15505 Name lshiftrt insns.
15506 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
15507 on WORDS_BIG_ENDIAN.
15509 2013-04-25 Alan Modra <amodra@gmail.com>
15511 * config.gcc: Support little-endian powerpc-linux targets.
15512 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
15513 (LINK_OS_LINUX_SPEC): Define.
15514 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
15515 Preserve MASK_LITTLE_ENDIAN.
15516 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
15517 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
15518 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
15519 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
15520 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
15521 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
15522 Correct fp word order for little-endian. Don't shift toc entries
15523 smaller than a word for little-endian.
15524 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
15525 (bswapdi2 splits): Correct low-part subreg for little-endian.
15526 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
15527 low/high where such is correct only for be.
15528 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
15529 little-endian for -mcall-aixdesc.
15531 2013-04-25 Alan Modra <amodra@gmail.com>
15533 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
15534 replace_equiv_address_nv.
15536 2013-04-25 Alan Modra <amodra@gmail.com>
15538 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
15540 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
15543 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
15544 * rtl.h (struct rtx_def): ...
15546 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
15548 PR rtl-optimizations/57046
15549 * lra-constraints (split_reg): Set up lra_risky_transformations_p
15550 for multi-reg splits.
15552 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
15554 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
15556 2013-04-24 Sterling Augustine <saugustine@google.com>
15558 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
15559 (comp_dir_string, debug_str_dwo_section): New.
15560 (DEBUG_STR_DWO_SECTION): Rename to ...
15561 (DEBUG_DWO_STR_SECTION): ... this.
15562 (DEBUG_NORM_STR_SECTION): Delete.
15563 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
15564 (DEBUG_STR_DWO_SECTION_FLAGS): New.
15565 (find_AT_string): Move most logic to ...
15566 (find_AT_string_in_table): ... here. New.
15567 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
15568 add_skeleton_AT_string. Delete logic.
15569 (output_skeleton_debug_sections): Remove call to
15570 add_top_level_skeleton_die_attrs.
15571 (add_comp_dir_attribute): Move logic to comp_dir_string.
15572 (dwarf2out_init): Initialize debug_str_dwo_section.
15573 (output_indirect_string): Call find_string_form.
15574 (output_indirect_strings): Rewrite.
15575 (prune_unused_types): Empty skeleton_debug_str_hash.
15576 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
15577 (dwarf2out_finish): Call output_indirect_strings.
15579 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
15581 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
15583 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
15585 * rtl.h (struct rtx_def): Add comment for field jump.
15586 (LRA_SUBREG_P): New macro.
15587 * recog.c (register_operand): Check LRA_SUBREG_P.
15588 * lra.c (lra): Add note at the end of RTL code. Align non-empty
15590 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
15591 (lra_final_code_change): Skip subreg change for operators.
15592 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
15593 if there are no operand changes.
15594 * lra-constraints.c (curr_insn_set): New.
15595 (match_reload): Set LRA_SUBREG_P.
15596 (emit_spill_move): Ditto.
15597 (check_and_process_move): Use curr_insn_set. Process only single
15598 set insns. Don't initialize sec_mem_p and change_p.
15599 (simplify_operand_subreg): Use LRA_SUBREG_P.
15600 (reg_in_class_p): New function.
15601 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
15602 of #ifdef. Add code to remove cycling.
15603 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
15604 non-null disp. Reload inner instead of disp when base and index
15605 are null. Try to put lo_sum into register.
15606 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
15607 (check_and_process_move): Move code for move cost check to
15608 simple_move_p. Remove equiv_substitution.
15609 (simple_move_p): New function.
15610 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
15611 curr_insn_set. Call check_and_process_move only for single set
15612 insns. Use the new function. Move call of check_and_process_move
15613 after operand equiv substitution and address process.
15615 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
15617 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
15618 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
15619 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
15621 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
15623 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
15625 2013-04-24 Marek Polacek <polacek@redhat.com>
15627 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
15628 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
15629 (select_loops_exit_conditions): Likewise.
15630 (number_of_iterations_for_all_loops): Likewise.
15631 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
15632 (scev_analysis): Likewise.
15634 2013-04-02 Catherine Moore <clm@codesourcery.com>
15635 Chao-ying Fu <fu@mips.com>
15637 * config/mips/micromips.md (jraddiusp): New pattern.
15638 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
15639 instruction if possible.
15641 2013-04-24 Alan Modra <amodra@gmail.com>
15643 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
15645 2013-04-24 Julian Brown <julian@codesourcery.com>
15646 Chung-Lin Tang <cltang@codesourcery.com>
15648 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
15649 dependency behavior in enumeration type DIE generation. Add TODO note
15650 to comments about future DW_FORM_sdata/udata re-work of related code.
15652 2013-04-23 Lawrence Crowl <crowl@google.com>
15654 * Makefile.in: Update as needed below.
15656 * hash-table.h (class hash_table):
15657 Correct many methods with parameter types compare_type to the correct
15658 value_type. (Correct code was unlikely to notice the change.)
15659 (hash_table::elements_with_deleted) New.
15660 (class hashtable::iterator): New.
15661 (hashtable::begin()): New.
15662 (hashtable::end()): New.
15663 (FOR_EACH_HASH_TABLE_ELEMENT): New.
15665 * statistics.c (statistics_hashes):
15666 Change type to hash_table. Update dependent calls and types.
15668 * tree-into-ssa.c (var_infos):
15669 Change type to hash_table. Update dependent calls and types.
15671 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
15672 Change type to hash_table. Update dependent calls and types.
15674 * tree-ssa-loop-im.c (struct mem_ref.refs):
15675 Change type to hash_table. Update dependent calls and types.
15677 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
15678 Change type to hash_table. Update dependent calls and types.
15680 * tree-ssa-sccvn.c (vn_tables_s::nary):
15681 Change type to hash_table. Update dependent calls and types.
15682 (vn_tables_s::phis): Likewise.
15683 (vn_tables_s::references): Likewise.
15685 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
15686 (vn_reference_eq): Update parameter and return types.
15688 * tree-ssa-structalias.c (pointer_equiv_class_table):
15689 Change type to hash_table. Update dependent calls and types.
15690 (location_equiv_class_table): Likewise.
15692 * tree-vect-data-refs.c: Consequential changes for making
15693 peeling a hash_table.
15695 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
15696 (destroy_loop_vec_info): Dependent hash_table update.
15698 * tree-vectorizer.h (peeling_htab):
15699 Change type to hash_table. Update dependent calls and types.
15701 2013-04-23 Shiva Chen <shiva0217@gmail.com>
15703 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
15704 to check the register content is equal or not.
15705 * lra-constraints.c (match_reload): Use lra_assign_reg_val
15706 to assign register content record.
15707 * lra-eliminations.c (update_reg_eliminate): Use
15708 lra_update_reg_val_offset to update register content offset.
15709 * lra-int.h (struct lra_reg): Add offset member.
15710 (lra_reg_val_equal_p): New static inline function.
15711 (lra_update_reg_val_offset): New static inline function.
15712 (lra_assign_reg_val): New static inline function.
15713 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
15714 to assign register content record.
15715 (initialize_lra_reg_info_element): Initial offset to zero.
15717 2013-04-23 Catherine Moore <clm@codesourcery.com>
15719 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
15720 operands. Record compression.
15722 2013-04-23 Xinliang David Li <davidxl@google.com>
15724 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
15726 2013-04-23 Richard Biener <rguenther@suse.de>
15728 PR middle-end/57036
15729 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
15730 parameter, only add abnormal goto edges from the copied body
15731 if the call could perform abnormal gotos.
15732 (copy_cfg_body): Adjust.
15734 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
15736 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
15738 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
15740 * coretypes.h (gimple_stmt_iterator): Add struct to make
15743 2013-04-23 Richard Biener <rguenther@suse.de>
15745 PR tree-optimization/57026
15746 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
15747 from SSA names occuring in abnormal PHI nodes.
15749 2013-04-22 Andi Kleen <ak@linux.intel.com>
15751 * lto/lto.c (print_lto_report_1): Fix LTO report names.
15753 2013-04-22 Andi Kleen <ak@linux.intel.com>
15755 * lto/lto.c (print_lto_report_1): Declare early.
15756 (read_cgraph_and_symbols): Call print_lto_report_1 early.
15758 2013-04-22 Andi Kleen <ak@linux.intel.com>
15760 * common.opt (-flto-report-wpa): Add.
15761 * doc/invoke.texi (-flto-report-wpa): Add.
15762 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
15765 2013-04-22 Xinliang David Li <davidxl@google.com>
15767 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
15768 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
15769 * Makefile.in: New dependency
15771 David Daney <ddaney.cavm@gmail.com>
15773 * configure.ac (gcc_cv_as_micromips_support): Use the
15774 --fatal-warnings option.
15775 * configure: Regenerate.
15777 2013-04-22 Marek Polacek <polacek@redhat.com>
15780 * tsan.c (instrument_expr): Don't instrument expression
15781 in case its size is zero.
15783 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
15787 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
15789 * config/alpha/alpha.c (TARGET_LRA_P): New define.
15791 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
15793 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
15794 (gimple_stmt_iterator): New typedef.
15795 * gimple.h (gimple_stmt_iterator): Rename to...
15796 (gimple_stmt_iterator_d): ... This.
15797 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
15798 trees be valid for GIMPLE and GENERIC.
15799 (TARGET_GIMPLE_FOLD_BUILTIN): New.
15800 * gimple-fold.c (gimple_fold_call): Call target hook
15801 gimple_fold_builtin.
15802 * hooks.c (hook_bool_gsiptr_false): New.
15803 * hooks.h (hook_bool_gsiptr_false): New.
15804 * target.def (fold_stmt): New.
15805 * doc/tm.texi: Regenerate.
15807 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
15810 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
15811 a set sp if no stack realignment.
15813 2013-04-22 Nick Clifton <nickc@redhat.com>
15815 * config.gcc (tilegx-linux): Extend extra_objs rather than
15817 (tilepro-linux): Likewise.
15819 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
15821 * config/aarch64/aarch64-builtins.c
15823 (CF0, CF1, CF2, CF3, CF4, CF10): New.
15824 (VAR<1-12>): Add MAP parameter.
15825 (BUILTIN_*): Likewise.
15826 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
15827 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
15828 (aarch64_ushl_n<mode>): Likewise.
15829 (aarch64_sshr_n<mode>): Likewise.
15830 (aarch64_ushr_n<mode>): Likewise.
15831 (aarch64_<maxmin><mode>): Likewise.
15832 (aarch64_sqrt<mode>): Likewise.
15833 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
15834 (vshr<q>_n_*): Likewise.
15836 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
15838 * config/aarch64/aarch64-builtins.c
15839 (aarch64_simd_builtin_type_mode): Handle SF types.
15841 (BUILTIN_GPF): Define.
15842 (aarch64_init_simd_builtins): Handle SF types.
15843 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
15844 (frecps): Likewise.
15845 (frecpx): Likewise.
15846 * config/aarch64/aarch64-simd.md
15847 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
15848 (aarch64_frecpe<mode>): New.
15849 (aarch64_frecps<mode>): Likewise.
15850 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
15851 (v8type): Add frecp<esx>.
15852 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
15853 (aarch64_frecps<mode>): Likewise.
15854 * config/aarch64/iterators.md (FRECP): New.
15855 (frecp_suffix): Likewise.
15856 * config/aarch64/arm_neon.h
15857 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
15859 2013-04-22 Christian Bruel <christian.bruel@st.com>
15862 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
15863 (REG_CLASS_NAMES): Idem.
15864 (REG_CLASS_CONTENTS): Idem.
15865 (REGCLASS_HAS_FP_REG): Idem.
15866 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
15867 (sh_conditional_register_usage): Idem.
15869 2013-04-21 Jeff Law <law@redhat.com>
15871 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
15872 (ssa_forward_propagate_and_combine): Use it.
15874 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
15876 * lra.c: Update the flow chart diagram.
15878 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
15880 PR rtl-optimization/56847
15881 * lra-constraints.c (process_alt_operands): Discourage alternative
15882 with non-matche doffsettable memory constraint fro memory with
15885 2013-04-19 Richard Biener <rguenther@suse.de>
15887 PR tree-optimization/56982
15888 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
15890 * gimplify.c (gimplify_call_expr): Notice special calls.
15891 (gimplify_modify_expr): Likewise.
15892 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
15893 abnormal control flow receivers.
15894 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
15895 in the same way as cfun->has_nonlocal_labels.
15896 (gimple_purge_dead_abnormal_call_edges): Likewise.
15897 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
15898 receivers start a basic-block.
15900 2013-04-19 Richard Biener <rguenther@suse.de>
15902 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
15904 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
15905 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
15906 (SLP_TREE_LOAD_PERMUTATION): Add.
15907 (vect_transform_slp_perm_load): Adjust prototype.
15908 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
15909 (vect_free_slp_instance): Likewise.
15910 (vect_create_new_slp_node): Likewise.
15911 (vect_supported_slp_permutation_p): Remove.
15912 (vect_slp_rearrange_stmts): Adjust.
15913 (vect_supported_load_permutation_p): Likewise. Inline
15914 vect_supported_slp_permutation_p here.
15915 (vect_analyze_slp_instance): Compute load permutations per
15916 slp node instead of per instance.
15917 (vect_get_slp_defs): Adjust.
15918 (vect_transform_slp_perm_load): Likewise.
15919 (vect_schedule_slp_instance): Remove redundant code.
15920 (vect_schedule_slp): Remove hack for PR56270, add it ...
15921 * tree-vect-stmts.c (vectorizable_load): ... here, do not
15922 CSE loads for SLP. Adjust.
15924 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
15926 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
15927 spelling in two comments.
15929 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
15932 * config/arm/arm.c (load_multiple_sequence): Require SP
15933 as base register for loads if SP is in the register list.
15935 2013-04-19 Martin Jambor <mjambor@suse.cz>
15937 PR tree-optimization/56718
15938 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
15939 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
15940 and made public. Adjusted all callers.
15941 (ipa_intraprocedural_devirtualization): New function.
15942 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
15943 (ipa_intraprocedural_devirtualization): Likewise.
15944 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
15946 2013-04-19 Richard Biener <rguenther@suse.de>
15948 PR tree-optimization/57000
15949 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
15951 2013-04-19 Terry Guo <terry.guo@arm.com>
15953 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
15955 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
15956 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
15957 (cortex_m4_fmacs): Use new reservations.
15958 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
15960 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
15962 PR rtl-optimization/56999
15963 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
15965 (lra_coalesce): Remove split_origin_bitmap and related code.
15966 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
15967 ranges if necessary.
15969 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
15971 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
15973 (ix86_expand_call): Remove clobbered_registers array and use
15974 x86_64_ms_sysv_extra_clobbered_registers instead.
15975 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
15977 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
15979 * config/i386/i386.md (*call_rex64_ms_sysv): Use
15980 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
15981 (*call_value_rex64_ms_sysv): Ditto.
15983 2013-04-18 Cary Coutant <ccoutant@google.com>
15985 * dwarf2out.c (output_pubnames): Check die_perennial_p of
15986 parent instead of die_mark.
15988 2013-04-18 Diego Novillo <dnovillo@google.com>
15990 * gimple.c (create_gimple_tmp): New.
15991 (get_expr_type): New.
15992 (build_assign): New.
15993 (build_type_cast): New.
15994 * gimple.h (enum ssa_mode): Define.
15995 (gimple_seq_set_location): New.
15996 * asan.c (build_check_stmt): Change some gimple_build_* calls
15997 to use build_assign and build_type_cast.
15999 2013-04-18 Richard Biener <rguenther@suse.de>
16001 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
16002 handle negative step. Remove redundant checks.
16003 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
16004 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
16005 for negative step and grouped loads fail to vectorize.
16007 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
16009 * emit-rtl.c (reset_insn_used_flags): New function.
16010 (reset_all_used_flags): Use it.
16011 (verify_insn_sharing): New function.
16012 (verify_rtl_sharing): Fix verification for SEQUENCEs.
16014 2013-04-18 Jakub Jelinek <jakub@redhat.com>
16016 PR tree-optimization/56984
16017 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
16018 and (x >> M) >= N don't register any assertion if N << M is the
16021 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
16023 * lower-subreg.c (resolve_simple_move): If called self-recursive,
16024 do not delete_insn insns that have not yet been emitted, only
16025 unlink them with remove_insn.
16026 * df-scan.c (df_insn_delete): Revert r197492.
16028 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
16030 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
16031 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
16033 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
16035 * config/arm/arm.md (movsicc_insn): Convert define_insn into
16036 define_insn_and_split.
16037 (and_scc,ior_scc,negscc): Likewise.
16038 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
16040 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
16042 * config/arm/arm.c (use_return_insn): Return 0 for targets that
16043 can benefit from using a sequence of LDRD instructions in epilogue
16044 instead of a single LDM instruction.
16046 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
16049 * doc/extend.texi: Fix typo.
16051 2013-04-17 Richard Biener <rguenther@suse.de>
16053 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
16054 (vect_build_slp_tree): ... here.
16055 (vect_build_slp_tree_1): Compute which stmts of the SLP group
16056 match. Remove special-casing of mismatched complex loads.
16057 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
16058 re-try the match with swapped commutative operands.
16059 (vect_supported_load_permutation_p): Remove special-casing of
16060 mismatched complex loads.
16061 (vect_analyze_slp_instance): Adjust.
16063 2013-04-17 Richard Biener <rguenther@suse.de>
16065 PR rtl-optimization/56921
16066 * cfgloop.h (struct loop): Add simple_loop_desc member.
16067 (struct niter_desc): Mark with GTY(()).
16068 (simple_loop_desc): Do not use aux field but simple_loop_desc.
16069 * loop-iv.c (get_simple_loop_desc): Likewise.
16070 (free_simple_loop_desc): Likewise.
16073 2013-04-16 Richard Biener <rguenther@suse.de>
16075 PR rtl-optimization/56921
16076 * loop-init.c (pass_rtl_move_loop_invariants): Add
16077 TODO_do_not_ggc_collect to todo_flags_finish.
16078 (pass_rtl_unswitch): Same.
16079 (pass_rtl_unroll_and_peel_loops): Same.
16080 (pass_rtl_doloop): Same.
16082 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
16084 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
16085 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
16086 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
16088 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
16089 * tree-streamer.c (record_common_node): Adjust reference in comment.
16091 2013-04-17 Terry Guo <terry.guo@arm.com>
16093 * config/arm/cortex-m4.md: Add a new bypass.
16095 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16097 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
16099 (*subs_<optab><mode>_multp2): New pattern.
16100 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
16101 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
16103 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16105 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
16106 (*subs_mul_imm_<mode>): New pattern.
16108 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
16111 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
16112 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
16113 (vsx_movti_32bit): Change j->wa to O->wa.
16115 2013-04-16 Richard Biener <rguenther@suse.de>
16117 PR rtl-optimization/56921
16118 * loop-init.c (pass_rtl_move_loop_invariants): Add
16119 TODO_do_not_ggc_collect to todo_flags_finish.
16120 (pass_rtl_unswitch): Same.
16121 (pass_rtl_unroll_and_peel_loops): Same.
16122 (pass_rtl_doloop): Same.
16124 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
16126 * config/arm/arm.c (emit_multi_reg_push): New declaration
16127 for an existing function.
16128 (arm_emit_strd_push): New function.
16129 (arm_expand_prologue): Used here.
16130 (arm_emit_ldrd_pop): New function.
16131 (arm_expand_epilogue): Used here.
16132 (arm_get_frame_offsets): Update condition.
16133 (arm_emit_multi_reg_pop): Add a special case for load of a single
16134 register with writeback.
16136 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
16138 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
16141 2013-04-16 Richard Biener <rguenther@suse.de>
16143 PR tree-optimization/56756
16144 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
16145 (first_mem_ref_loc): New.
16146 (execute_sm): Place the load temporarily before a previous
16147 access instead of in the latch edge to ensure its SSA dependencies
16148 are defined at points dominating the load.
16150 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
16152 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
16153 correct fix by moving header and footer insn to the footer of
16154 the merged basic block. Clear BB_END of the merged-away block.
16156 PR middle-end/43631
16157 * emit-rtl.c (make_note_raw): New function.
16158 (link_insn_into_chain): New static inline function.
16159 (add_insn): Use it.
16160 (add_insn_before, add_insn_after): Factor insn chain linking code...
16161 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
16162 using link_insn_into_chain.
16163 (note_outside_basic_block_p): New helper function for emit_note_after
16164 and emit_note_before.
16165 (emit_note_after): Use nobb variant of add_insn_after if the note
16166 should not be contained in a basic block.
16167 (emit_note_before): Use nobb variant of add_insn_before if the note
16168 should not be contained in a basic block.
16169 (emit_note_copy): Use make_note_raw.
16170 (emit_note): Likewise.
16171 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
16172 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
16173 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
16174 the moved barrier the tail of the basic block it follows.
16175 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
16177 2013-04-15 Jakub Jelinek <jakub@redhat.com>
16179 PR tree-optimization/56962
16180 * gimple-ssa-strength-reduction.c (record_increment): Only set
16181 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
16182 either rhs1 or rhs2 is equal to c->base_expr.
16184 2013-04-15 Richard Biener <rguenther@suse.de>
16186 PR tree-optimization/56933
16187 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
16189 (GROUP_READ_WRITE_DEPENDENCE): Remove.
16190 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
16191 * tree-vect-data-refs.c (vect_analyze_group_access): Move
16192 dependence check ...
16193 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
16195 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
16196 GROUP_READ_WRITE_DEPENDENCE.
16198 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16200 * emit-rtl.c (reset_all_used_flags): New function.
16201 (verify_rtl_sharing): Call reset_all_used_flags before and after
16202 performing the checks.
16204 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16206 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
16207 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
16208 * config/arm/constraints.md (De): New constraint.
16209 * config/arm/neon.md (anddi3_neon): Delete.
16210 (neon_vand<mode>): Expand to standard anddi3 pattern.
16211 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
16212 Move earlier in the file.
16213 (neon_inv_logic_op2): Likewise.
16214 (arm_anddi_operand_neon): New predicate.
16216 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16218 * configure.ac (gcc_cv_ld_as_needed): Set
16219 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
16220 Use -z ignore, -z record on *-*-solaris2*.
16221 (HAVE_LD_AS_NEEDED): Update comment.
16222 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
16223 * configure: Regenerate.
16224 * config.in: Regenerate.
16225 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
16226 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
16227 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
16228 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
16229 equivalents. Fix markup.
16230 * doc/tm.texi: Regenerate.
16232 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
16234 * config/i386/i386.opt: New option mstack-protector-guard=.
16235 * config/i386/i386-opts.h: Add enum stack_protector_guard.
16236 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
16237 TARGET_SSP_TLS_GUARD.
16238 * config/i386/i386.c (ix86_option_override_internal): Set
16239 ix86_stack_protector_guard.
16240 * config/i386/i386.md (stack_protect_set): Enable for
16241 TARGET_SSP_TLS_GUARD only.
16242 (stack_protect_set_<mode>): Ditto.
16243 (stack_protect_test): Ditto.
16244 (stack_protect_test_<mode>): Ditto.
16245 * doc/invoke.texi (i386 Option): Document.
16247 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
16250 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
16251 (S_MODES): Set H_MODE bit.
16252 (SF_MODES): Set only S_MODE and SF_MODE bits.
16253 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
16254 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
16255 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
16256 <MODE_FLOAT>: Likewise.
16258 2013-04-15 Joey Ye <joey.ye@arm.com>
16260 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
16262 2013-04-15 Joey Ye <joey.ye@arm.com>
16264 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
16266 (thumb_far_jump_used_p): Count instruction size and set
16269 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
16271 * reorg.c (fill_simple_delay_slots): Reindent block of code.
16272 * resource.c (mark_target_live_regs): Reformat conditional block.
16274 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
16276 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
16277 notes, they are emitted only just before final.
16278 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
16280 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
16282 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
16283 * cfgrtl.c (delete_insn): Call it here instead.
16284 * lra-spills.c (lra_final_code_change): Use delete_insn.
16285 * haifa-sched.c (sched_remove_insn): Likewise.
16286 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
16287 returning to the nop pool.
16288 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
16289 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
16291 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
16293 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
16294 * doc/tm.texi: Regenerated.
16296 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
16298 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
16301 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
16303 * df-core.c (df_find_def): Compare register numbers.
16304 (df_find_use): Likewise.
16306 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
16309 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
16310 lra_in_progress for return.
16312 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
16314 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
16315 define_insn into define_insn_and_split and emit movsicc patterns.
16317 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
16319 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
16321 2013-04-12 Richard Biener <rguenther@suse.de>
16323 * tree-pass.h (TODO_do_not_ggc_collect): New.
16324 * passes.c (execute_one_ipa_transform_pass): Honor
16325 TODO_do_not_ggc_collect.
16326 (execute_one_pass): Likewise.
16329 2013-04-10 Richard Biener <rguenther@suse.de>
16331 * passes.c (init_optimization_passes): Remove reload pass.
16332 * ira.c (do_reload): Merge into ...
16334 (rest_of_handle_reload): Remove.
16335 (pass_reload): Likewise.
16336 * config/i386/i386.c (ix86_option_override): Refer to ira instead
16337 of reload for vzeroupper pass placement.
16339 2013-04-12 Jakub Jelinek <jakub@redhat.com>
16341 PR tree-optimization/56918
16342 PR tree-optimization/56920
16343 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
16344 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
16345 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
16346 use wide_mul_with_sign method.
16348 2013-04-12 Richard Biener <rguenther@suse.de>
16350 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
16351 not be considered a gimple constant.
16353 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
16355 * fold-const.c (const_binop): Handle vector shifts by a scalar.
16356 (fold_binary_loc): Call const_binop also for mixed vector-scalar
16359 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
16360 Jakub Jelinek <jakub@redhat.com>
16362 * opts.c: Include diagnostic-color.h.
16363 (common_handle_option): Handle OPT_fdiagnostics_color_.
16364 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
16365 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
16366 (diagnostic-color.o): New.
16367 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
16368 (diagnostic_color_rule): New enum.
16369 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
16370 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
16371 the location string.
16372 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
16373 either NULL, or color kind.
16374 * diagnostic-color.c: New file.
16375 * diagnostic-color.h: New file.
16376 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
16378 * doc/invoke.texi (-fdiagnostics-color): Document.
16379 * pretty-print.h (pp_show_color): Define.
16380 (struct pretty_print_info): Add show_color field.
16381 * diagnostic.c: Include diagnostic-color.h.
16382 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
16383 macros. Colorize error:, warning: etc. strings and also the location
16385 (diagnostic_show_locus): Colorize the caret line.
16386 * pretty-print.c: Include diagnostic-color.h.
16387 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
16388 inside of %< %> quotes or quoted through q format modifier.
16390 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16392 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
16394 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16396 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
16397 code in CC_NZ mode.
16398 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
16401 2013-04-11 Marek Polacek <polacek@redhat.com>
16403 PR tree-optimization/48184
16404 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
16406 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
16408 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
16409 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
16410 (skip_simple_arithmetic): Tidy up.
16411 * tree.h (skip_simple_constant_arithmetic): Declare.
16413 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16415 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
16417 2013-04-11 Richard Biener <rguenther@suse.de>
16419 * tree-vect-loop.c (get_initial_def_for_induction): Properly
16420 generate vector constants.
16422 2013-04-11 Richard Biener <rguenther@suse.de>
16424 PR tree-optimization/56878
16425 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
16426 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
16428 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
16429 Prefer to align the DR with the most invariant base address.
16431 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
16433 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
16436 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
16438 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
16439 floating-point vector comparisons against 0.
16441 2013-04-11 Jakub Jelinek <jakub@redhat.com>
16443 PR tree-optimization/56899
16444 * fold-const.c (extract_muldiv_1): Apply distributive law
16445 only if TYPE_OVERFLOW_WRAPS (ctype).
16447 2013-04-11 Bin Cheng <bin.cheng@arm.com>
16450 * ira-costs.c (scan_one_insn): Check whether the source rtx of
16451 loading has side effect.
16453 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
16455 * config/sparc/sparc.c: Include tree-pass.h.
16456 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
16457 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
16458 head of file. Change return type. Split off gate function.
16459 (sparc_gate_work_around_errata): New function.
16460 (pass_work_around_errata): New pass definition.
16461 (insert_pass_work_around_errata) New pass insert definition to
16462 insert pass_work_around_errata just after delayed-branch scheduling.
16463 (sparc_option_override): Insert the pass.
16464 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
16466 2013-04-10 David S. Miller <davem@davemloft.net>
16468 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
16469 or -mcpu=hypersparc.
16471 * target.def (cstore_mode): New hook.
16472 * target.h: Include insn-codes.h
16473 * targhooks.c: Likewise.
16474 (default_cstore_mode): New function.
16475 * targhooks.h: Declare it.
16476 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
16477 * doc/tm.texi: Rebuild.
16478 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
16479 target hook, rather than inspecting the insn_data.
16480 * config/sparc/sparc.c (sparc_cstore_mode): New function.
16481 (TARGET_CSTORE_MODE): Redefine.
16482 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
16484 * config/sparc/predicates.md (cstore_result_operand): New special
16486 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
16487 Use it for operand 0.
16488 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
16489 (*snesi_special): Likewise.
16490 (*snesi_zero): Likewise.
16491 (*seqsi_zero): Likewise.
16492 (*sltu_insn): Likewise.
16493 (*sgeu_insn): Likewise.
16494 (*seqdi_special): Make operand 0 and comparison operation be of
16496 (*snedi_special): Likewise.
16497 (*snedi_special_vis3): Likewise.
16498 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
16499 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
16500 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
16501 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
16502 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
16503 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
16504 (*sltu_extend_sp64): Likewise.
16505 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
16506 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
16507 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
16508 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
16509 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
16511 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
16513 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
16514 (aarch64_start_file): Use the new function.
16516 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
16517 Jason Merrill <jason@redhat.com>
16519 * common.opt: Add -gdwarf.
16520 * opts.c (common_handle_option): Handle it.
16521 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
16523 2013-04-10 Richard Biener <rguenther@suse.de>
16525 * passes.c (execute_todo): Do not call ggc_collect conditional here.
16526 (execute_one_ipa_transform_pass): But unconditionally here.
16527 (execute_one_pass): And here.
16528 (init_optimization_passes): Remove reload pass.
16529 * tree-pass.h (TODO_ggc_collect): Remove.
16530 (pass_reload): Likewise.
16531 * ira.c (do_reload): Merge into ...
16533 (rest_of_handle_reload): Remove.
16534 (pass_reload): Likewise.
16535 * config/i386/i386.c (ix86_option_override): Refer to ira instead
16536 of reload for vzeroupper pass placement.
16537 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
16538 and todo_flags_finish of all passes.
16540 2013-04-10 Richard Biener <rguenther@suse.de>
16542 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
16543 first_const_oprnd field, rename first_def_type to first_op_type.
16544 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
16545 (vect_get_and_check_slp_defs): Always use the type of the
16546 operand. Allow mixed vect_external_def, vect_constant_def types.
16547 (vect_get_constant_vectors): Handle mixed vect_external_def,
16548 vect_constant_def types.
16550 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
16552 PR tree-optimization/55524
16553 * tree-ssa-math-opts.c
16554 (convert_mult_to_fma): Don't use an fms construct
16555 when we don't have an fms operation, but fnma, and it looks
16556 likely that we'll be able to use the latter.
16558 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
16560 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
16562 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
16563 inline fail caused by overwritable functions.
16565 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
16567 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
16568 unnecessary bits in the constant power of two case.
16570 2013-04-10 Richard Biener <rguenther@suse.de>
16572 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
16573 broken code swapping operands.
16574 (vect_build_slp_tree): Do not compute load permutations here.
16575 (vect_analyze_slp_instance): Compute load permutations here,
16576 after building the SLP tree.
16578 2013-04-09 Christian Bruel <christian.bruel@st.com>
16580 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
16581 of next/prev_real_insn.
16583 2013-04-09 Jan Hubicka <jh@suse.cz>
16585 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
16586 Drop aliased parameter.
16587 (function_and_variable_visibility): Do not handle alias pairs.
16588 * cgraph.c (varpool_externally_visible_p): Update prototype.
16589 * varpool.c (varpool_add_new_variable): Update.
16591 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16593 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
16595 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
16597 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
16599 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
16601 2013-04-09 Marek Polacek <polacek@redhat.com>
16603 PR tree-optimization/48762
16604 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
16606 2013-04-09 Richard Biener <rguenther@suse.de>
16608 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
16610 (vect_build_slp_tree): Likewise.
16611 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
16612 calculating the cost of a SLP instance.
16613 (vect_analyze_slp_instance): Use it from here, after building
16616 2013-04-09 Jakub Jelinek <jakub@redhat.com>
16618 PR middle-end/56883
16619 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
16620 expand_omp_for_static_chunk): Use simple_p = true in
16621 force_gimple_operand_gsi calls when assigning to addressable decls.
16623 2013-04-09 Jeff Law <law@redhat.com>
16625 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
16626 when the boolean was created by converting a wider object which
16627 had a boolean range.
16629 2013-04-09 Richard Biener <rguenther@suse.de>
16631 * tree-vectorizer.h (slp_void_p): Remove.
16632 (slp_tree): Typedef before _slp_tree declaration.
16633 (struct _slp_tree): Use a vector of slp_tree as children.
16634 (vect_get_place_in_interleaving_chain): Remove.
16635 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
16637 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
16639 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
16640 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
16641 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
16642 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
16643 Use slp_node instead of slp_void_p and adjust.
16645 2013-04-09 Richard Biener <rguenther@suse.de>
16647 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
16648 work that is not necessary.
16650 2013-04-09 Jakub Jelinek <jakub@redhat.com>
16652 PR tree-optimization/56854
16653 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
16654 forward into clobber stmts if it would change MEM_REF lhs into
16657 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
16659 * tree.c (type_hash_lookup, type_hash_add): Make static.
16660 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
16662 2013-04-09 Richard Biener <rguenther@suse.de>
16664 * tree.h (unsave_expr_now): Remove.
16665 * tree-inline.c (mark_local_for_remap_r): Remove.
16666 (unsave_expr_1): Likewise.
16667 (unsave_r): Likewise.
16668 (unsave_expr_now): Likewise.
16669 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
16670 (propagate_tree_value): Likewise.
16672 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
16674 * doc/rtl.texi (sequence): Rewrite documentation to match the
16675 current use of SEQUENCE rtl objects.
16676 * rtl.def (SEQUENCE): Likewise.
16678 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
16679 Update documentation.
16680 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
16681 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
16683 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
16685 2013-04-08 Teresa Johnson <tejohnson@google.com>
16687 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
16688 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
16690 (estimate_edge_size_and_time): Add comment to suggest using rounding
16692 (estimate_node_size_and_time): Ditto.
16693 (remap_edge_change_prob): Use helper rounding divide methods.
16694 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
16695 (gimple_mod_pow2_value_transform): Ditto.
16696 (gimple_mod_subtract_transform): Ditto.
16697 (gimple_ic_transform): Ditto.
16698 (gimple_stringops_transform): Ditto.
16699 * stmt.c (conditional_probability): Ditto.
16700 (emit_case_dispatch_table): Ditto.
16701 * lto-cgraph.c (merge_profile_summaries): Ditto.
16702 * tree-optimize.c (execute_fixup_cfg): Ditto.
16703 * cfgcleanup.c (try_forward_edges): Ditto.
16704 * cfgloopmanip.c (scale_loop_profile): Ditto.
16706 (duplicate_loop_to_header_edge): Ditto.
16707 (lv_adjust_loop_entry_edge): Ditto.
16708 * tree-vect-loop.c (vect_transform_loop): Ditto.
16709 * profile.c (compute_branch_probabilities): Ditto.
16710 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
16711 * lto-streamer-in.c (input_cfg): Ditto.
16712 * gimple-streamer-in.c (input_bb): Ditto.
16713 * ipa-cp.c (update_profiling_info): Ditto.
16714 (update_specialized_profile): Ditto.
16715 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
16716 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
16718 * sched-rgn.c (compute_dom_prob_ps): Ditto.
16719 (compute_trg_info): Ditto.
16720 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
16721 (purge_dead_edges): Ditto.
16722 * loop-unswitch.c (unswitch_loop): Ditto.
16723 * cgraphclones.c (cgraph_clone_edge): Ditto.
16724 (cgraph_clone_node): Ditto.
16725 * tree-inline.c (copy_bb): Ditto.
16726 (copy_edges_for_bb): Ditto.
16727 (initialize_cfun): Ditto.
16728 (copy_cfg_body): Ditto.
16729 (expand_call_inline): Ditto.
16731 2013-04-08 Kai Tietz <ktietz@redhat.com>
16733 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
16734 TARGET_CYGWIN64 by TARGET_64BIT.
16736 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
16738 * config/epiphany/epiphany.md (GPR_1): New constant.
16739 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
16740 * config/epiphany/epiphany.c (gen_compare_reg):
16741 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
16742 is already in place.
16743 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
16744 Don't require being called during rtl expansion; If y operlaps r0,
16746 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
16747 (epiphany_expand_epilogue): Likewise.
16749 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
16750 Don't use CC_FPmode for ORDERED / UNORDERED.
16751 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
16753 * config/epiphany/constraints.md (CnL): New constraint.
16754 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
16755 * config/epiphany/predicates.md (add_operand): Allow 1024.
16757 * config/epiphany/epiphany.md (logical_op): New code iterator.
16758 (op_mnc): New code attribute.
16759 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
16760 (mov_f+1, mov_f+2): New peephole2 patterns.
16762 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
16763 (cstoresi4): Also allow re-use of zero result when doing a NE
16764 comparison to a non-zero operand.
16765 Use (clobber (scratch)) for first insn if the gpr output is not needed.
16767 * config/epiphany/epiphany.md (<insn_opname>v2si3):
16768 Use gen_addsi3_i / gen_subsi3_i.
16770 2013-04-08 Jakub Jelinek <jakub@redhat.com>
16774 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
16775 contain anything but clobbers, at most one __builtin_stack_restore,
16776 optionally debug stmts and final resx, and if it has at least one
16777 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
16778 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
16779 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
16780 which isn't defaut definition, remove them.
16781 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
16782 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
16783 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
16784 with MEM_REF LHS with SSA_NAME address.
16786 2013-04-08 Jeff Law <law@redhat.com>
16788 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
16790 2013-04-08 Richard Biener <rguenther@suse.de>
16792 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
16794 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
16795 determined vector type.
16796 (vect_analyze_data_refs): Likewise.
16797 (vect_get_new_vect_var): Adjust.
16798 (vect_create_destination_var): Preserve SSA name versions.
16799 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
16800 not dump anything here.
16802 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
16804 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
16805 Add member lr_slot_known.
16806 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
16808 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
16809 Remove code that sets lr_slot_offset according to what a previous
16810 version of epiphany_emit_save_restore used to do.
16811 (epiphany_emit_save_restore): When doing an lr save or restore,
16812 set/verify lr_slot_known and lr_slot_offset.
16814 2013-04-08 Xinyu Qi <xyqi@marvell.com>
16817 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
16820 2013-04-08 Richard Biener <rguenther@suse.de>
16822 * alias.c (find_base_term): Fix thinko in previous change.
16824 2013-04-08 Jakub Jelinek <jakub@redhat.com>
16826 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
16827 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
16828 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
16829 if possible to compute val.
16830 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
16831 For QImode integers don't require anything about precision. Use
16832 const_with_all_bytes_same to find out if the constant doesn't have
16833 repeated bytes in it.
16835 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16837 * config/s390/s390.c (s390_expand_insv): Only accept insertions
16840 2013-04-08 Marek Polacek <polacek@redhat.com>
16842 PR rtl-optimization/48182
16843 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
16846 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
16849 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
16850 nuses, make sure we have a label.
16852 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16855 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
16856 (rs6000_emit_swdiv_low_precision): Remove.
16857 (rs6000_emit_swdiv): Rewrite to handle between one and four
16858 iterations of Newton-Raphson generally; modify required number of
16859 iterations for some cases.
16860 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
16862 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
16864 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
16865 set-but-unused variable.
16867 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
16868 basic blocks of released function bodies garbage-collectable.
16870 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
16871 (struct rtl_opt_pass): Add TODO_df_finish.
16873 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
16875 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16877 * config/arm/constraints.md (q): New constraint.
16878 * config/arm/ldrdstrd.md: New file.
16879 * config/arm/arm.md (ldrdstrd.md) New include.
16880 (arm_movdi): Use "q" instead of "r" constraint
16881 for double-word memory access.
16882 (movdf_soft_insn): Likewise.
16883 * config/arm/vfp.md (movdi_vfp): Likewise.
16884 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
16885 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
16886 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
16887 (mem_ok_for_ldrd_strd): Likewise.
16888 (output_move_double): Update assertion.
16890 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16892 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
16894 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16896 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
16897 define_insn_and_split.
16898 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
16900 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16902 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
16903 define_insn_and_split.
16904 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
16905 (shiftsi3_compare): New pattern.
16906 (rrx): New pattern.
16907 * config/arm/unspecs.md (UNSPEC_RRX): New.
16909 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16911 * config/arm/arm.md (negdi_extendsidi): New pattern.
16912 (negdi_zero_extendsidi): Likewise.
16914 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16916 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
16917 define_insn_and_split.
16918 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
16919 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
16921 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16923 * config/arm/arm.md (arm_subdi3): Convert define_insn into
16924 define_insn_and_split.
16925 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
16926 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
16928 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16930 * config/arm/arm.md (subsi3_carryin): New pattern.
16931 (subsi3_carryin_const): Likewise.
16932 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
16933 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
16935 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16937 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
16939 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16941 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
16942 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
16944 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16946 * config/arm/arm.c (arm_expand_builtin): Change fcode
16947 type to unsigned int.
16949 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16951 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
16953 2013-04-04 Ian Lance Taylor <iant@google.com>
16955 * doc/standards.texi (Standards): The Go frontend supports the Go 1
16958 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
16960 PR middle-end/56729
16961 * df-scan.c (df_insn_delete): Disable failing assert.
16963 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16965 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
16966 New function prototype.
16967 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
16968 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
16969 (arm_builtin_vectorized_function): New function.
16971 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16973 * config/arm/arm_neon_builtins.def: New file.
16974 * config/arm/arm.c (neon_builtin_data): Move contents to
16975 arm_neon_builtins.def.
16976 (enum arm_builtins): Include neon builtin definitions.
16977 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
16978 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
16980 2013-04-04 Marek Polacek <polacek@redhat.com>
16982 PR tree-optimization/48186
16983 * predict.c (maybe_hot_frequency_p): Return false if
16984 HOT_BB_FREQUENCY_FRACTION is 0.
16985 (cgraph_maybe_hot_edge_p): Likewise.
16987 2013-04-04 Richard Biener <rguenther@suse.de>
16989 PR tree-optimization/56826
16990 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
16993 2013-04-04 Richard Biener <rguenther@suse.de>
16995 PR tree-optimization/56213
16996 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
16997 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
16999 2013-04-04 Richard Biener <rguenther@suse.de>
17001 PR tree-optimization/56837
17002 * tree-loop-distribution.c (classify_partition): For non-zero
17003 values require that the value has the same precision as its
17004 mode to be useful as memset value.
17006 2013-04-03 Nick Clifton <nickc@redhat.com>
17008 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
17009 (fmssf4): Use fmsf.s on E3V5 architectures.
17010 (fnmasf4): Use fnmaf.s on E3V5 architectures.
17011 (fnmssf4): Use fnmsf.s on E3V5 architectures.
17013 2013-04-03 Jeff Law <law@redhat.com>
17015 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
17016 (lra-eliminations.o): Likewise.
17018 2013-04-03 Teresa Johnson <tejohnson@google.com>
17020 * gcov-io.c (compute_working_sets): Moved most of body of old
17021 compute_working_sets here from profile.c.
17022 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
17023 (gcov_working_set_t): Moved typedef here from basic-block.h
17024 (compute_working_set): Declare.
17025 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
17026 (get_working_sets): Renamed from compute_working_set,
17027 replace most of body with call to new compute_working_sets.
17028 (get_exec_counts): Replace call to compute_working_sets
17029 to get_working_sets.
17030 * profile.h (get_working_sets): Renamed from compute_working_set.
17031 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
17032 to get_working_sets.
17033 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
17034 * gcov-dump.c (dump_working_sets): New function.
17036 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
17038 * hwint.c (sext_hwi, zext_hwi): New functions.
17039 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
17040 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
17041 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
17042 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
17043 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
17044 (sext_hwi, zext_hwi): New functions.
17046 2013-04-03 Jeff Law <law@redhat.com>
17048 PR tree-optimization/56799
17049 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
17050 back test for widening conversion erroneously dropped in prior change.
17052 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17055 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
17056 instead of next_real_insn.
17058 2013-04-03 Marek Polacek <polacek@redhat.com>
17061 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
17063 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17066 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
17068 (thumb1_output_casesi): Likewise.
17069 (thumb2_output_casesi): Likewise.
17071 2013-04-03 Richard Biener <rguenther@suse.de>
17073 PR tree-optimization/56817
17074 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
17076 (tree_unroll_loops_completely_1): ... new function to manually
17077 walk the loop tree, properly defering outer loops of unrolled
17078 loops to later iterations.
17080 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
17082 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
17083 (vectorizable_load): Likewise.
17084 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
17085 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
17087 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
17089 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
17092 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
17094 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
17096 2013-04-03 Bin Cheng <bin.cheng@arm.com>
17098 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
17100 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
17102 PR tree-optimization/56790
17103 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
17106 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
17108 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
17110 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
17111 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
17114 2013-04-03 Jakub Jelinek <jakub@redhat.com>
17117 * tree.h (force_folding_builtin_constant_p): New decl.
17118 * builtins.c (force_folding_builtin_constant_p): New variable.
17119 (fold_builtin_constant_p): Fold immediately also if
17120 force_folding_builtin_constant_p.
17122 2013-04-03 Richard Biener <rguenther@suse.de>
17124 PR tree-optimization/56812
17125 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
17126 DRs of the same interleaving chain are independent.
17128 2013-04-02 Jason Merrill <jason@redhat.com>
17130 * gdbinit.in (pbb): Use debug fn.
17132 2013-04-02 Lawrence Crowl <crowl@google.com>
17134 * sese.h (struct ivtype_map_elt_s): Remove unused.
17135 (extern debug_ivtype_map): Remove unused.
17136 (extern eq_ivtype_map_elts): Remove unused.
17137 * sese.c (debug_ivtype_map): Removed unused.
17138 (debug_ivtype_map_1): Removed unused.
17139 (debug_ivtype_elt): Remove unused.
17140 (eq_ivtype_map_elts): Remove unused.
17143 2013-04-02 Kai Tietz <ktietz@redhat.com>
17146 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
17147 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
17148 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
17150 (legitimize_pe_coff_symbol): Likewise.
17151 (is_imported_p): New helper-function.
17152 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
17153 for Windows x64 targets.
17154 (ix86_expand_prologue): Optimize for pe-coff targets.
17155 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
17156 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
17157 medium/large code-model.
17158 (legitimize_pic_address): Likewise.
17159 (legitimize_tls_address): Likewise.
17160 (ix86_expand_call): Likewise.
17161 (x86_output_mi_thunk): Likewise.
17162 (get_dllimport_decl): Add new beimport argument.
17163 (construct_plt_address): Don't assert for x64 pe-coff targets.
17164 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
17166 (SYMBOL_FLAG_STUBVAR): New macro.
17167 (SYMBOL_REF_STUBVAR_P): Likewise.
17168 * config/i386/winnt.c (stub_list): New structure.
17169 (stub_head): New local variable.
17170 (i386_pe_record_stub): New function.
17171 (i386_pe_file_end): Emit refptr-stubs.
17173 2013-04-02 Jakub Jelinek <jakub@redhat.com>
17175 PR rtl-optimization/56745
17176 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
17177 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
17180 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
17181 and both of them are MEM_REFs, just compare first argument for
17182 equality and attempt to deal even with differing offsets.
17185 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
17186 of gimple_clobber_p to be MEM_REF.
17187 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
17188 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
17189 after gimplification.
17190 * asan.c (get_mem_ref_of_assignment): Don't instrument
17191 gimple_clobber_p stmts.
17192 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
17193 gimple_clobber_p stmt if they have MEM_REF lhs and
17194 are dead because of another gimple_clobber_p stmt.
17195 * tree-ssa-live.c (clear_unused_block_pointer): Treat
17196 gimple_clobber_p stmts like debug stmts.
17197 (remove_unused_locals): Remove clobbers with MEM_REF lhs
17198 that refer to unused VAR_DECLs or uninitialized values.
17199 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
17200 gimple_clobber_p stmts if they refer to removed parameters.
17201 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
17204 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
17206 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
17207 using SWI48 mode attribute.
17209 2013-04-02 Wei Mi <wmi@google.com>
17211 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
17212 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
17213 *<rotate_insn><mode>3_mask in i386.md.
17215 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
17217 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
17219 2013-04-02 Richard Biener <rguenther@suse.de>
17221 PR tree-optimization/56778
17222 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
17223 Runtime alias tests are not supported for gather loads.
17224 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
17225 stmts referenced from SSA operands before updating SSA form.
17227 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
17228 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17230 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
17231 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
17232 * config/arm/cortex-a53.md: New file.
17233 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
17234 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
17235 * config/arm/arm.c (arm_issue_rate): Likewise.
17236 * config/arm/arm-tune.md: Regenerate
17237 * config/arm/arm-tables.opt: Regenerate.
17238 * config/arm/arm-cores.def: Add cortex-a53.
17240 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
17242 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
17245 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
17247 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
17248 scalar load/store operations using B/H registers.
17249 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
17251 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
17253 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
17255 * config/aarch64/aarch64.c
17256 (aarch64_simd_scalar_immediate_valid_for_move): New.
17257 * config/aarch64/aarch64-protos.h
17258 (aarch64_simd_scalar_immediate_valid_for_move): New.
17259 * config/aarch64/constraints.md (Dh, Dq): New.
17260 * config/aarch64/iterators.md (hq): New.
17262 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
17264 * reorg.c (get_branch_condition): Deal with conditional returns.
17265 (fill_simple_delay_slots): Remove dead code dealing with jumps.
17267 2013-04-01 Wei Mi <wmi@google.com>
17269 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
17270 Truncate operand 2 using %b asm operand modifier.
17271 (*<shift_insn><mode>3_mask): Ditto.
17272 (*<rotate_insn><mode>3_mask): Ditto.
17274 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
17276 PR middle-end/56798
17277 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
17279 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
17281 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
17283 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
17285 2013-03-30 Lawrence Crowl <crowl@google.com>
17287 * dse.c (clear_alias_sets): Remove never set.
17288 (disqualified_clear_alias_sets): Remove never set.
17289 (clear_alias_mode_pool): Remove never set.
17290 (dse_step0): Remove condition that is never true.
17291 (canon_address): Remove condition that is never true.
17292 (dse_step7): Remove condition that is never true.
17293 (rest_of_handle_dse): Remove condition that is never true.
17294 (rest_of_handle_dse::did_global): Remove never read from above.
17295 (dse_step2_spill): Remove never called from above.
17296 (dse_step5_spill): Remove never called from above.
17298 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
17300 * doc/md.texi (Standard Names) <casesi>: Update documentation for
17301 JUMP_TABLE_DATA changes.
17302 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
17303 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
17304 (Insns) <jump_table_data>: New entry.
17305 * doc/tm.texi: Regenerate.
17307 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
17309 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
17310 for table jump at the end of a basic block using tablejump_p.
17311 * targhooks.c (default_invalid_within_doloop): Likewise.
17312 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
17313 target hook implementation that is identical to the default hook.
17314 (rs6000_invalid_within_doloop): Remove.
17316 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
17317 unused variable from tablejump_p call.
17319 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
17320 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
17321 (INSN_DELETED_P): Likewise.
17322 (emit_jump_table_data): New prototype.
17323 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
17324 after 4th as unused.
17325 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
17326 * sched-vis.c (print_insn): Likewise.
17327 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
17328 insn for compatibility with back ends that use next_active_insn to
17329 identify jump table data.
17330 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
17331 (remove_insn): Likewise.
17332 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
17334 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
17335 (emit_jump_table_data): New function.
17337 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
17338 basic block, a JUMP_TABLE_DATA never is.
17339 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
17340 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
17341 off from code handling real insns.
17342 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
17343 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
17344 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
17345 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
17346 is not a NONDEBUG_INSN_P.
17347 * ira-costs.c (scan_one_insn): Likewise.
17348 * jump.c (mark_all_labels): Likewise.
17349 (mark_jump_label_1): Likewise.
17350 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
17351 * lra.c (get_insn_freq): Expect all insns reaching here to be in
17353 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
17354 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
17355 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
17356 JUMP_TABLE_DATA_P insns.
17357 (calculate_elim_costs_all_insns): Likewise.
17358 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
17359 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
17360 (delete_output_reload): Code style fixups.
17361 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
17362 insn flags on this non-insn.
17363 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
17364 as scheduling barriers, for pre-change compatibility.
17365 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
17366 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
17368 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
17369 redundant JUMP_TABLE_DATA_P test.
17370 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
17371 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
17372 (frv_for_each_packet): Likewise.
17373 * config/i386/i386.c (min_insn_size): Likewise.
17374 (ix86_avoid_jump_mispredicts): Likewise.
17375 * config/m32r/m32r.c (m32r_is_insn): Likewise.
17376 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
17377 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
17378 (mips16_insn_length): Robustify.
17379 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
17380 (mips16_split_long_branches): Likewise.
17381 * config/pa/pa.c (pa_combine_instructions): Likewise.
17382 * config/rs6000/rs6000.c (get_next_active_insn): Treat
17383 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
17384 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
17385 as contributing to pool range lengths.
17386 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
17387 Remove redundant JUMP_TABLE_DATA_P test.
17388 (sh_loop_align): Likewise.
17389 (split_branches): Likewise.
17390 (sh_insn_length_adjustment): Likewise.
17391 * config/spu/spu.c (get_branch_target): Likewise.
17393 2013-03-29 Jan Hubicka <jh@suse.cz>
17395 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
17396 gcov streaming; stream hot bb threshold to ltrans.
17397 * predict.c (get_hot_bb_threshold): Break out from ....
17398 (maybe_hot_count_p): ... here.
17399 (set_hot_bb_threshold): New function.
17400 * lto-section-in.c (lto_section_name): Add profile.
17401 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
17402 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
17403 and data-streamer.h
17404 (histogram_entry): New structure.
17405 (histogram, histogram_pool): New global vars.
17406 (histogram_hash): New structure.
17407 (histogram_hash::hash): New method.
17408 (histogram_hash::equal): Likewise.
17409 (account_time_size): New function.
17410 (cmp_counts): New function.
17411 (dump_histogram): New function.
17412 (ipa_profile_generate_summary): New function.
17413 (ipa_profile_write_summary): New function.
17414 (ipa_profile_read_summary): New function.
17415 (ipa_profile): Decide on threshold.
17416 (pass_ipa_profile): Add ipa_profile_write_summary and
17417 ipa_profile_read_summary.
17418 * Makefile.in (ipa.o): Update dependencies.
17419 * lto-streamer.h (LTO_section_ipa_profile): New section.
17421 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
17423 * tree.h (VAR_P): New.
17425 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
17428 * doc/invoke.texi ([-fwhole-program]): Fix typo.
17430 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
17432 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
17433 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
17434 (control_flow_insn_p): Likewise.
17435 * cfgrtl.c (duplicate_insn_chain): Likewise.
17436 * final.c (get_attr_length_1): Likewise.
17437 (shorten_branches): Likewise.
17438 (final_scan_insn): Likewise.
17439 * function.c (instantiate_virtual_regs): Likewise.
17440 * gcse.c (insert_insn_end_basic_block): Likewise.
17441 * ira-costs.c (scan_one_insn): Likewise.
17442 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
17443 * lra.c (check_rtl): Likewise.
17444 * reload1.c (elimination_costs_in_insn): Likewise.
17445 * reorg.c (follow_jumps): Likewise.
17447 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
17448 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
17449 (thumb_far_jump_used_p): Likewise.
17450 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
17451 (workaround_speculation): Likewise.
17452 (add_sched_insns_for_speculation): Likewise.
17453 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
17454 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
17455 (frv_for_each_packet): Likewise.
17456 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
17457 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
17458 (final_emit_insn_group_barriers): Likewise.
17459 * config/m32r/m32r.c (m32r_is_insn): Likewise.
17460 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
17461 (mips16_insn_length): Likewise.
17462 * config/pa/pa.c (pa_reorg): Likewise.
17463 (pa_combine_instructions): Likewise.
17464 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
17465 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
17466 (sh_reorg): Likewise.
17467 (split_branches): Likewise.
17468 * config/spu/spu.c (get_branch_target): Likewise.
17470 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
17473 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
17475 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
17476 Fix declaration name.
17478 2013-03-28 Lawrence Crowl <crowl@google.com>
17480 * graphds.h (struct graph.indicies): Remove unused.
17481 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
17482 (SCOP_ORIGINAL_PDDRS): Remove unused.
17483 * sese.h (extern insert_loop_close_phis): Removed unused.
17484 (extern insert_guard_phis): Removed unused.
17485 (extern ivtype_map_elt_info): Removed unused.
17486 (new_ivtype_map_elt): Removed unused.
17487 * sese.c (ivtype_map_elt_info): Removed unused.
17489 2013-03-28 Lawrence Crowl <crowl@google.com>
17491 * Makefile.in: Add several missing include dependences.
17493 (test-dump.o): New. This object is not added to any executable,
17494 but is present for ad-hoc testing.
17496 (debug (const bitmap_head_def &)): New.
17497 (debug (const bitmap_head_def *)): New.
17499 (extern debug (const bitmap_head_def &)): New.
17500 (extern debug (const bitmap_head_def *)): New.
17502 (debug (edge_def &)): New.
17503 (debug (edge_def *)): New.
17505 (debug (basic_block_def &)): New.
17506 (debug (basic_block_def *)): New.
17508 (dump_node (const_tree, int, FILE *)): Correct source file.
17510 (debug (die_struct &)): New.
17511 (debug (die_struct *)): New.
17513 (extern debug (die_struct &)): New.
17514 (extern debug (die_struct *)): New.
17515 * gimple-pretty-print.c
17516 (debug (gimple_statement_d &)): New.
17517 (debug (gimple_statement_d *)): New.
17518 * gimple-pretty-print.h
17519 (extern debug (gimple_statement_d &)): New.
17520 (extern debug (gimple_statement_d *)): New.
17522 (debug (ira_allocno_copy &)): New.
17523 (debug (ira_allocno_copy *)): New.
17524 (debug (ira_allocno &)): New.
17525 (debug (ira_allocno *)): New.
17527 (extern debug (ira_allocno_copy &)): New.
17528 (extern debug (ira_allocno_copy *)): New.
17529 (extern debug (ira_allocno &)): New.
17530 (extern debug (ira_allocno *)): New.
17532 (debug (live_range &)): New.
17533 (debug (live_range *)): New.
17535 (debug (lra_live_range &)): New.
17536 (debug (lra_live_range *)): New.
17538 (debug (lra_live_range &)): New.
17539 (debug (lra_live_range *)): New.
17541 (debug (omega_pb_d &)): New.
17542 (debug (omega_pb_d *)): New.
17544 (extern debug (omega_pb_d &)): New.
17545 (extern debug (omega_pb_d *)): New.
17547 (debug (const rtx_def &)): New.
17548 (debug (const rtx_def *)): New.
17550 (debug_tree (tree): Move within file.
17551 (debug_raw (const tree_node &)): New.
17552 (debug_raw (const tree_node *)): New.
17553 (dump_tree_via_hooks (const tree_node *, int)): New.
17554 (debug (const tree_node &)): New.
17555 (debug (const tree_node *)): New.
17556 (debug_verbose (const tree_node &)): New.
17557 (debug_verbose (const tree_node *)): New.
17558 (debug_head (const tree_node &)): New.
17559 (debug_head (const tree_node *)): New.
17560 (debug_body (const tree_node &)): New.
17561 (debug_body (const tree_node *)): New.
17562 (debug_vec_tree (tree): Move and reimplement in terms of dump.
17563 (debug (vec<tree, va_gc> &)): New.
17564 (debug (vec<tree, va_gc> *)): New.
17566 (extern debug (const rtx_def &)): New.
17567 (extern debug (const rtx_def *)): New.
17569 (debug_raw (simple_bitmap_def &)): New.
17570 (debug_raw (simple_bitmap_def *)): New.
17571 (debug (simple_bitmap_def &)): New.
17572 (debug (simple_bitmap_def *)): New.
17574 (extern debug (simple_bitmap_def &)): New.
17575 (extern debug (simple_bitmap_def *)): New.
17576 (extern debug_raw (simple_bitmap_def &)): New.
17577 (extern debug_raw (simple_bitmap_def *)): New.
17579 (debug (vinsn_def &)): New.
17580 (debug (vinsn_def *)): New.
17581 (debug_verbose (vinsn_def &)): New.
17582 (debug_verbose (vinsn_def *)): New.
17583 (debug (expr_def &)): New.
17584 (debug (expr_def *)): New.
17585 (debug_verbose (expr_def &)): New.
17586 (debug_verbose (expr_def *)): New.
17587 (debug (vec<rtx> &)): New.
17588 (debug (vec<rtx> *)): New.
17590 (extern debug (vinsn_def &)): New.
17591 (extern debug (vinsn_def *)): New.
17592 (extern debug_verbose (vinsn_def &)): New.
17593 (extern debug_verbose (vinsn_def *)): New.
17594 (extern debug (expr_def &)): New.
17595 (extern debug (expr_def *)): New.
17596 (extern debug_verbose (expr_def &)): New.
17597 (extern debug_verbose (expr_def *)): New.
17598 (extern debug (vec<rtx> &)): New.
17599 (extern debug (vec<rtx> *)): New.
17601 (_list_iter_cond_expr): Make inline instead of static.
17603 (debug (sreal &)): New.
17604 (debug (sreal *)): New.
17606 (extern debug (sreal &)): New.
17607 (extern debug (sreal *)): New.
17609 (extern debug_raw (const tree_node &)): New.
17610 (extern debug_raw (const tree_node *)): New.
17611 (extern debug (const tree_node &)): New.
17612 (extern debug (const tree_node *)): New.
17613 (extern debug_verbose (const tree_node &)): New.
17614 (extern debug_verbose (const tree_node *)): New.
17615 (extern debug_head (const tree_node &)): New.
17616 (extern debug_head (const tree_node *)): New.
17617 (extern debug_body (const tree_node &)): New.
17618 (extern debug_body (const tree_node *)): New.
17619 (extern debug (vec<tree, va_gc> &)): New.
17620 (extern debug (vec<tree, va_gc> *)): New.
17622 (debug (struct loop &)): New.
17623 (debug (struct loop *)): New.
17624 (debug_verbose (struct loop &)): New.
17625 (debug_verbose (struct loop *)): New.
17626 * tree-dump.c: Add header dependence.
17628 (extern debug (struct loop &)): New.
17629 (extern debug (struct loop *)): New.
17630 (extern debug_verbose (struct loop &)): New.
17631 (extern debug_verbose (struct loop *)): New.
17633 (debug (data_reference &)): New.
17634 (debug (data_reference *)): New.
17635 (debug (vec<data_reference_p> &)): New.
17636 (debug (vec<data_reference_p> *)): New.
17637 (debug (vec<ddr_p> &)): New.
17638 (debug (vec<ddr_p> *)): New.
17640 (extern debug (data_reference &)): New.
17641 (extern debug (data_reference *)): New.
17642 (extern debug (vec<data_reference_p> &)): New.
17643 (extern debug (vec<data_reference_p> *)): New.
17644 (extern debug (vec<ddr_p> &)): New.
17645 (extern debug (vec<ddr_p> *)): New.
17647 (debug (pt_solution &)): New.
17648 (debug (pt_solution *)): New.
17650 (extern debug (pt_solution &)): New.
17651 (extern debug (pt_solution *)): New.
17653 (debug (_var_map &)): New.
17654 (debug (_var_map *)): New.
17655 (debug (tree_live_info_d &)): New.
17656 (debug (tree_live_info_d *)): New.
17658 (extern debug (_var_map &)): New.
17659 (extern debug (_var_map *)): New.
17660 (extern debug (tree_live_info_d &)): New.
17661 (extern debug (tree_live_info_d *)): New.
17663 2013-03-28 Jan Hubicka <jh@suse.cz>
17665 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
17667 2013-03-28 Ian Bolton <ian.bolton@arm.com>
17669 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
17670 record only when desired or required.
17672 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
17674 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
17675 *vec_extractv2di_1_rex64. Use x64 isa attribute.
17677 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17679 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
17680 (*andsi3_compare0_uxtw): New pattern.
17681 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
17682 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
17684 2013-03-28 Jan Hubicka <jh@suse.cz>
17686 * data-streamer-in.c (streamer_read_gcov_count): New function.
17687 * gimple-streamer-out.c: Include value-prof.h.
17688 (output_gimple_stmt): Output histogram.
17689 (output_bb): Use streamer_write_gcov_count.
17690 * value-prof.c: Include data-streamer.h
17691 (dump_histogram_value): Add HIST_TYPE_MAX.
17692 (stream_out_histogram_value): New function.
17693 (stream_in_histogram_value): New function.
17694 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
17695 (stream_out_histogram_value, stream_in_histogram_value): Declare.
17696 * data-streamer-out.c (streamer_write_gcov_count): New function.
17697 (streamer_write_gcov_count_stream): New function.
17698 * lto-cgraph.c (lto_output_edge): Update counter streaming.
17699 (lto_output_node): Likewise.
17700 (input_node, input_edge): Likewise.
17701 * lto-streamer-out.c (output_cfg): Update streaming.
17702 * lto-streamer-in.c (input_cfg): Likewise.
17703 * data-streamer.h (streamer_write_gcov_count,
17704 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
17705 * gimple-streamer-in.c: Include value-prof.h
17706 (input_gimple_stmt): Input histograms.
17707 (input_bb): Update profile streaming.
17709 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
17711 * genmodes.c (emit_max_int): New function.
17712 (emit_insn_modes_h): Added call to emit_max_function.
17713 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
17715 * machmode.def: Fixed comment.
17717 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
17719 * combine.c (try_combine): Removed useless assert.
17720 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
17722 2013-03-28 Marek Polacek <polacek@redhat.com>
17723 Richard Biener <rguenther@suse.de>
17725 PR tree-optimization/56695
17726 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
17727 build signed result of a vector comparison.
17728 * tree-cfg.c (verify_gimple_comparison): Check that a result
17729 of a vector comparison has signed type.
17731 2013-03-28 Richard Biener <rguenther@suse.de>
17733 PR tree-optimization/37021
17734 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
17735 do not restrict gaps between groups.
17736 * tree-vect-stmts.c (vectorizable_load): Properly account for
17737 a gap between groups.
17739 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
17741 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
17742 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
17745 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
17747 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
17748 * doc/extend.texi (Named Address Spaces): Ditto.
17749 (Variable Attributes): Ditto.
17751 2013-03-27 Kai Tietz <ktietz@redhat.com>
17753 * config.build: Add support for cygwin x64 target.
17754 * config.gcc: Likewise.
17755 * config.host: Likewise.
17756 * configure.ac: Likewise
17757 * configure: Regenerated.
17759 2013-03-27 Kai Tietz <ktietz@redhat.com>
17761 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
17762 * config/i386/t-cygwin-w64: New file.
17763 * config/i386/cygwin-w64.h: New file.
17764 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
17765 and add support for x64-cygwin target.
17766 (CPP_SPEC): Likewise.
17767 (CXX_WRAP_SPEC_LIST): Undefine before define.
17768 (LIBGCJ_SONAME): Use 15 as version.
17770 2013-03-27 Richard Biener <rguenther@suse.de>
17772 PR tree-optimization/56716
17773 * tree-ssa-structalias.c (perform_var_substitution): Adjust
17774 dumping for ref nodes.
17776 2013-03-27 Martin Jambor <mjambor@suse.cz>
17778 PR tree-optimization/55334
17779 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
17780 restricted pointers to arrays.
17782 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
17784 * Makefile.in (.SUFFIXES): Add .cc.
17785 (.c.o): Apply same recipe for implicit rule .cc.o.
17787 2013-03-27 Richard Biener <rguenther@suse.de>
17789 PR tree-optimization/37021
17790 * tree-vect-data-refs.c (vect_check_strided_load): Allow
17791 REALPART/IMAGPART_EXPRs around the supported refs.
17792 * tree-ssa-structalias.c (find_func_aliases): Assume that
17793 floating-point values are not used to transfer pointers.
17795 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
17797 * target.def (TARGET_HAS_IFUNC_P): New target hook.
17798 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
17799 * doc/tm.texi: Regenerate.
17800 * targhooks.h (default_has_ifunc_p): New.
17801 * targhooks.c (default_has_ifunc_p): Ditto.
17802 * config/linux-protos.h: New file.
17803 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
17804 hook for linux which disables support of indirect functions in android.
17805 * config/linux-android.c: New file.
17806 * config/t-linux-android.c: Ditto.
17807 * config.gcc: Added new object file linux-android.o.
17808 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
17809 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
17810 * varasm.c (do_assemble_alias): Likewise.
17811 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
17812 doesn't support indirect functions.
17813 * configure: Regenerate.
17815 2013-03-27 Bin Cheng <bin.cheng@arm.com>
17818 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
17819 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
17822 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17824 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
17826 2013-03-27 Terry Guo <terry.guo@arm.com>
17828 * config/arm/arm-cores.def: Added core cortex-r7.
17829 * config/arm/arm-tune.md: Regenerated.
17830 * config/arm/arm-tables.opt: Regenerated.
17831 * doc/invoke.texi: Added entry for core cortex-r7.
17833 2013-03-27 Walter Lee <walt@tilera.com>
17835 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
17836 double-decrement of next_scratch_regno.
17838 2013-03-27 Walter Lee <walt@tilera.com>
17840 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
17842 (insn_v1mulus): Ditto.
17843 (insn_v2muls): Ditto.
17845 2013-03-27 Walter Lee <walt@tilera.com>
17847 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
17848 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
17850 2013-03-27 Walter Lee <walt@tilera.com>
17852 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
17853 (*sibcall_value): Ditto.
17855 2013-03-27 Walter Lee <walt@tilera.com>
17857 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
17858 (insn_mnz_v8qi): ... this ...
17859 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
17861 (insn_v<n>mnz): Replaced by ...
17862 (insn_v1mnz): ... this ...
17863 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
17865 (insn_mz_<mode>): Replaced by ...
17866 (insn_mz_v8qi): ... this ...
17867 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
17869 (insn_v<n>mz): Replaced by ...
17870 (insn_v1mz): ... this ...
17871 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
17874 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
17876 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
17878 2013-03-26 Roland McGrath <mcgrathr@google.com>
17880 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
17881 than fprintf with a non-constant, non-format string.
17883 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
17885 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
17886 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
17887 operand 0 predicate.
17888 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
17889 attribute. Use general_x64nomem_operand as operand 1 predicate.
17890 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
17891 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
17892 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
17893 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
17894 (mov<mode>_insv_1): Remove expander. Merge insn with
17895 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
17896 Use general_x64nomem_operand as operand 1 predicate.
17897 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
17898 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
17899 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
17900 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
17901 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
17902 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
17903 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
17904 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
17905 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
17906 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
17907 (general_x64nomem_operand): Ditto.
17909 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
17911 * config/rtems.opt: Add -pthread option.
17913 2013-03-26 Richard Biener <rguenther@suse.de>
17915 * alias.c (find_base_term): Avoid redundant and not used recursion.
17916 (base_alias_check): Get the initial base term from the caller.
17917 (true_dependence_1): Compute and pass base terms to base_alias_check.
17918 (write_dependence_p): Likewise.
17919 (may_alias_p): Likewise.
17921 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
17923 * config/aarch64/aarch64.c (aarch64_classify_address): Support
17924 PC-relative load in SI modes and above only.
17926 2013-03-26 Xinyu Qi <xyqi@marvell.com>
17928 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
17929 * config/arm/iwmmxt.md (WCGR0): Update.
17930 (WCGR1, WCGR2, WCGR3): Likewise.
17932 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
17934 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
17935 Use x64 and nox64 isa attributes.
17937 2013-03-26 Richard Biener <rguenther@suse.de>
17939 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
17940 alignment computations and rely on get_object_alignment_1
17941 for the !TYPE_P case.
17942 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
17944 2013-03-26 Walter Lee <walt@tilera.com>
17946 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
17947 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
17949 2013-03-25 Jeff Law <law@redhat.com>
17951 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
17952 check for INTEGRAL_TYPE_P that was missing due to checking in
17953 wrong version of prior patch.
17955 2013-03-25 Walter Lee <walt@tilera.com>
17957 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
17958 TILEGX_INSN_SHUFFLEBYTES1.
17959 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
17961 (tilegx_builtins): Ditto.
17962 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
17964 2013-03-25 Walter Lee <walt@tilera.com>
17966 * config/tilegx/tilegx.md (floatsisf2): New pattern.
17967 (floatunssisf2): New pattern.
17968 (floatsidf2): New pattern.
17969 (floatunssidf2): New pattern.
17971 2013-03-25 Walter Lee <walt@tilera.com>
17973 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
17974 tests for constraint J, K, N, P.
17976 2013-03-25 Walter Lee <walt@tilera.com>
17978 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
17979 Use indirect/pcrel encoding.
17980 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
17983 2013-03-25 Steve Ellcey <sellcey@mips.com>
17985 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
17986 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
17987 * config/mips/mips.c (mips_option_override): Set IMADD default.
17988 * config/mips/mips.h (PTF_AVOID_IMADD): New.
17989 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
17990 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
17991 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
17992 * doc/invoke.texi (-mimadd/-mno-imadd): New.
17994 2013-03-25 Jeff Law <law@redhat.com>
17996 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
17997 slightly to avoid creating and folding useless trees. Simplify
17998 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
18000 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
18002 * config/i386/i386.md (*zero_extendsidi2): Merge with
18003 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
18004 * config/i386/predicates.md (x86_64_zext_operand): Rename from
18005 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
18006 targets. Clarify comment.
18008 2013-03-25 Martin Jambor <mjambor@suse.cz>
18010 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
18011 pass-through jump functions differently.
18012 (ipa_read_jump_function): Likewise. Also use setter functions to set
18015 2013-03-25 Martin Jambor <mjambor@suse.cz>
18017 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
18018 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
18020 (ipa_get_indirect_edge_target): New function.
18021 (devirtualization_time_bonus): New parameter known_aggs, pass it to
18022 ipa_get_indirect_edge_target. Update all callers.
18023 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
18024 ipa_get_indirect_edge_target_1 instead of calling
18025 ipa_get_indirect_edge_target.
18026 (create_specialized_node): Pass aggvlas to
18027 ipcp_discover_new_direct_edges.
18029 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18031 * config/arm/arm.md (f_sels, f_seld): New types.
18032 (*cmov<mode>): New pattern.
18033 * config/arm/predicates.md (arm_vsel_comparison_operator): New
18036 2013-03-25 Kai Tietz <ktietz@redhat.com>
18038 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
18039 POSIX-printf for mingw-hosted builds.
18041 2013-03-25 Richard Biener <rguenther@suse.de>
18043 PR middle-end/56694
18044 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
18045 must-not-throw stmt location.
18047 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18049 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
18050 Emit load-acquire versions when acq is true.
18051 (arm_emit_store_exclusive): Add rel parameter.
18052 Emit store-release versions when rel is true.
18053 (arm_split_compare_and_swap): Use acquire-release instructions
18055 of barriers when appropriate.
18056 (arm_split_atomic_op): Likewise.
18057 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
18058 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
18059 (VUNSPEC_SLX): Likewise.
18060 (VUNSPEC_LDA): Likewise.
18061 (VUNSPEC_STL): Likewise.
18062 * config/arm/sync.md (atomic_load<mode>): New pattern.
18063 (atomic_store<mode>): Likewise.
18064 (arm_load_acquire_exclusive<mode>): Likewise.
18065 (arm_load_acquire_exclusivesi): Likewise.
18066 (arm_load_acquire_exclusivedi): Likewise.
18067 (arm_store_release_exclusive<mode>): Likewise.
18069 2013-03-25 Catherine Moore <clm@codesourcery.com>
18071 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
18072 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
18073 * config/mip/predicates.md (lwsp_swsp_operand,
18074 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
18075 sb16_operand, db4_operand, db7_operand, ib3_operand,
18076 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
18077 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
18078 andi16_operand): New predicates.
18079 * config/mips/mips.md (compression): New attribute.
18080 (enabled): New attribute.
18081 (length): Consider compression in computing length.
18082 (shift_compression): New code attribute.
18083 (*add<mode>3): New operands. Record compression.
18084 (sub<mode>3): Likewise.
18085 (one_cmpl<mode>2): Likewise.
18086 (*and<mode>3): Likewise.
18087 (*ior<mode>3): Likewise.
18088 (unnamed pattern for xor): Likewise.
18089 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
18090 (*<optab><mode>3): Likewise.
18091 (*mov<mode>_internal: Likewise.
18092 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
18093 (mips_unsigned_immediate_p): New.
18094 (umips_lwsp_swsp_address_p): New.
18095 (m16_based_address_p): New.
18096 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
18097 (mips_unsigned_immediate_p): New prototype.
18098 (lwsp_swsp_address_p): New prototype.
18099 (m16_based_address_p): New prototype.
18100 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
18101 (mips_signed_immediate_p): New function.
18102 (m16_based_address_p): New function.
18103 (lwsp_swsp_address_p): New function.
18104 (mips_print_operand_punctuation): Recognize short delay slot insns
18105 for microMIPS.add<mode>3"
18107 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18110 * config/arm/iterators.md (v_cmp_result): New mode attribute.
18111 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
18113 2013-03-25 Richard Biener <rguenther@suse.de>
18115 PR tree-optimization/56689
18116 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
18119 2013-03-25 Richard Biener <rguenther@suse.de>
18121 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
18123 (memory_references): Likewise.
18124 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
18125 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
18126 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
18127 (gather_mem_refs_in_loops): Fold into ...
18128 (analyze_memory_references): ... this. Move initialization
18129 to tree_ssa_lim_initialize.
18130 (fill_always_executed_in): Rename to ...
18131 (fill_always_executed_in_1): ... this.
18132 (fill_always_executed_in): Move contains_call computation to
18133 this new function from ...
18134 (tree_ssa_lim_initialize): ... here.
18135 (tree_ssa_lim): Call fill_always_executed_in.
18137 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
18139 * postreload.c (reload_combine): Fix code detecting returns.
18141 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
18143 * function.c (emit_use_return_register_into_block): On cc0 targets,
18144 do not emit the sequence between cc0 setter and user.
18146 2013-03-25 Kai Tietz <ktietz@redhat.com>
18148 * config/i386/predicates.md (local_symbolic_operand): Interpret
18149 dll-imported symbols as none-local.
18151 2013-03-25 Richard Biener <rguenther@suse.de>
18153 * tree-ssa-loop-im.c (struct depend): Remove.
18154 (struct lim_aux_data): Make depends a vec of gimples.
18155 (free_lim_aux_data): Adjust.
18156 (add_dependency): Likewise.
18157 (set_level): Likewise.
18159 2013-03-25 Richard Biener <rguenther@suse.de>
18161 PR middle-end/56434
18162 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
18163 the pointer returned by calls with ECF_MALLOC set.
18165 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
18167 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
18169 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
18171 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
18172 using MMXMODE mode iterator.
18173 (*move<mode>_internal): Merge with *movv2sf_internal and
18174 *movv2sf_internal_rex64 using MMXMODE mode iterator.
18176 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
18178 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
18179 (record_last_mem_set_info): Likewise.
18181 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
18182 of XNEWVEC followed by memset.
18183 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
18185 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
18187 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
18188 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
18189 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
18190 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
18191 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
18192 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
18193 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
18194 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
18195 BARRIER_P instead of GET_CODE.
18197 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
18199 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
18200 inaccuracy in the probing code.
18202 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
18203 (ctrapdi4): Likewise.
18205 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
18207 * calls.c (expand_call): Add missing guard to code handling return
18208 of non-BLKmode structures in MSB.
18209 * function.c (expand_function_end): Likewise.
18211 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
18213 * combine.c (try_combine): Adjust comment. Do not add the set of
18214 insn #0 if the destination indirectly is set or dies in insn #2.
18215 Tidy up code to distribute a new note.
18217 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
18219 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
18220 also for alternatives 16 and 17.
18222 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
18224 * config/i386/sse.md (*mov<mode>_internal): Merge with
18225 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
18226 Emit insn template depending on type attribute. Use
18227 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
18228 movd instead of movq mnemonic for interunit moves. Rewrite mode
18229 attribute calculation. Remove unit attribute calculation.
18230 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
18231 Set prefix_data16 attribute for DImode ssemov types.
18232 Use Ym instead of y for SSE-MMX conversion alternatives.
18233 Reorder operand constraints.
18235 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
18237 * df.h (df_insn_delete): Adjust prototype.
18238 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
18239 and let it decide whether mark the basic block dirty.
18240 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
18241 * df-scan.c (df_insn_info_delete): New helper function, split
18242 off from df_insn_delete.
18243 (df_scan_free_bb_info): Use it.
18244 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
18246 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
18247 that the insn is actually an insn and it has a non-NULL basic block.
18248 Do not mark basic block dirty if only deleting a DEBUG_INSN.
18250 2013-03-22 Richard Biener <rguenther@suse.de>
18252 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
18254 (mem_ref_alloc): Do not allocate them.
18255 (refs_independent_p): Do not query or maintain a cache.
18257 2013-03-22 Richard Biener <rguenther@suse.de>
18259 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
18260 (gather_mem_refs_in_loops): Do not compute it.
18261 (analyze_memory_references): Do not allocate it.
18262 (tree_ssa_lim_finalize): Do not free it.
18263 (for_all_locs_in_loop): Do not query all_refs_in_loop.
18265 2013-03-22 Richard Biener <rguenther@suse.de>
18267 * is-a.h (as_a): Use gcc_checking_assert.
18269 2013-03-22 Ian Bolton <ian.bolton@arm.com>
18271 * config/aarch64/aarch64.c (aarch64_print_operand): New
18272 format specifier for printing a constant in hex.
18273 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
18274 format specifier for printing second operand.
18276 2013-03-22 Richard Biener <rguenther@suse.de>
18278 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
18280 (gather_mem_refs_in_loops): Perform store accumulation here.
18281 (create_vop_ref_mapping_loop): Remove.
18282 (create_vop_ref_mapping): Likewise.
18283 (analyze_memory_references): Initialize refs_stored_in_loop.
18284 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
18285 (record_indep_loop): Remove.
18286 (record_dep_loop): New function.
18287 (ref_indep_loop_p_1): Adjust to only walk over references
18288 in the loop, not its subloops.
18289 (ref_indep_loop_p): Rename to ...
18290 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
18291 maintaining a more fine-grained cache.
18292 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
18293 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
18295 2013-03-22 Richard Biener <rguenther@suse.de>
18297 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
18298 (struct mem_ref): Make accesses_in_loop a vec of a vec of
18299 aggregate mem_ref_loc.
18300 (free_mem_ref_locs): Inline into ...
18301 (memref_free): ... this and adjust.
18302 (mem_ref_alloc): Adjust.
18303 (mem_ref_locs_alloc): Remove.
18304 (record_mem_ref_loc): Adjust.
18305 (get_all_locs_in_loop): Rewrite into ...
18306 (for_all_locs_in_loop): ... this iterator.
18307 (rewrite_mem_ref_loc): New functor.
18308 (rewrite_mem_refs): Use for_all_locs_in_loop.
18309 (sm_set_flag_if_changed): New functor.
18310 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
18311 (ref_always_accessed): New functor.
18312 (ref_always_accessed_p): Use for_all_locs_in_loop.
18314 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
18316 * tree-pass.h (PROP_gimple_lvec): New.
18317 * passes.c (dump_properties): Handle PROP_gimple_lvec.
18318 (init_optimization_passes): Move pass_lower_vector.
18319 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
18321 (pass_lower_vector): Provide PROP_gimple_lvec.
18322 (pass_lower_vector_ssa): Likewise.
18323 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
18325 2013-03-21 Mark Wielaard <mjw@redhat.com>
18327 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
18329 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
18331 * config/i386/i386.md (*movdi_internal): Disparage slightly
18332 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
18333 conversion alternatives.
18335 2013-03-21 Jakub Jelinek <jakub@redhat.com>
18337 PR middle-end/48087
18338 * diagnostic.def (DK_WERROR): New kind.
18339 * diagnostic.h (werrorcount): Define.
18340 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
18341 promoted to DK_ERROR, increment DK_WERROR counter instead of
18343 * toplev.c (toplev_main): Call print_ignored_options even if
18344 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
18345 even if just werrorcount is non-zero.
18348 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
18350 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
18351 (string_cst_pool_decl): New function.
18352 (optimize_one_addr_into_implicit_ptr): New function.
18353 (resolve_addr_in_expr): Optimize DWARF location expression
18354 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
18355 which doesn't live in memory, but has DW_AT_location or
18356 DW_AT_const_value, or refers to a string literal, into
18357 DW_OP_GNU_implicit_pointer.
18358 (optimize_location_into_implicit_ptr): New function.
18359 (resolve_addr): If removing DW_AT_location of a variable because
18360 it was DW_OP_addr of address of the variable, but the variable doesn't
18361 live in memory, try to emit const value attribute for the initializer.
18363 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
18365 * tree.h (VECTOR_TYPE_P): New macro.
18366 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
18367 TYPE_MODE): Use it.
18368 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
18369 VEC_COND_EXPR cannot be lvalues.
18370 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
18372 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
18374 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
18375 Restrict the transformation to equal modes.
18377 2013-03-21 Richard Biener <rguenther@suse.de>
18379 PR tree-optimization/39326
18380 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
18381 (MEM_ANALYZABLE): Adjust.
18382 (record_mem_ref_loc): Move bitmap ops ...
18383 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
18384 unanalyzable refs, do not record locations for it.
18385 (analyze_memory_references): Allocate ref zero as shared
18387 (refs_independent_p): Do not test for unanalyzed mems here.
18388 (ref_indep_loop_p_1): Special-case disambiguation against
18389 the unanalyzed ref.
18390 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
18392 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
18394 * config/arm/arm-protos.h (tune_params): Add
18395 prefer_neon_for_64bits field.
18396 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
18397 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
18398 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
18399 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
18400 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
18401 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
18402 (arm_option_override): Handle -mneon-for-64bits new option.
18403 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
18404 (prefer_neon_for_64bits): Declare new variable.
18405 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
18406 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
18407 (arch_enabled): Handle new arch types. Remove support for onlya8
18409 (one_cmpldi2): Use new arch names.
18410 (zero_extend<mode>di2, extend<mode>di2): Ditto.
18411 * config/arm/arm.opt (mneon-for-64bits): Add option.
18412 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
18413 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
18414 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
18416 * doc/invoke.texi (-mneon-for-64bits): Document.
18418 2013-03-21 Richard Biener <rguenther@suse.de>
18420 PR tree-optimization/39326
18421 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
18422 (sort_bbs_in_loop_postorder_cmp): New function.
18423 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
18425 2013-03-21 Richard Biener <rguenther@suse.de>
18427 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
18428 (vect_insert_into_interleaving_chain): Likewise.
18429 (vect_drs_dependent_in_basic_block): Inline ...
18430 (vect_slp_analyze_data_ref_dependence): ... here. New function,
18432 (vect_analyze_data_ref_dependence): ... here. Simplify.
18433 (vect_check_interleaving): Simplify.
18434 (vect_analyze_data_ref_dependences): Likewise. Split out ...
18435 (vect_slp_analyze_data_ref_dependences): ... this new function.
18436 (dr_group_sort_cmp): New function.
18437 (vect_analyze_data_ref_accesses): Compute data-reference groups
18438 here instead of in vect_analyze_data_ref_dependence. Use
18439 a more efficient algorithm.
18440 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
18441 vect_slp_analyze_data_ref_dependences. Call
18442 vect_analyze_data_ref_accesses earlier.
18443 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
18444 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
18445 (vect_slp_analyze_data_ref_dependences): New prototype.
18447 2013-03-21 Richard Biener <rguenther@suse.de>
18449 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
18450 ref is stored in the loop.
18451 (find_refs_for_sm): Walk only over all stores.
18452 (store_motion_loop): Allocate from lim_bitmap_obstack.
18453 (store_motion): Likewise.
18455 2013-03-21 Richard Biener <rguenther@suse.de>
18457 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
18458 Update virtual SSA form.
18460 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
18462 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
18463 * configure: Regenerate.
18464 * config.in: Regenerate.
18465 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
18466 if !HAVE_LD_EH_FRAME_CIEV3.
18468 2013-03-21 Richard Biener <rguenther@suse.de>
18470 * tree-cfg.c (verify_expr_no_block): New function.
18471 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
18472 nor DECL_VALUE_EXPR have locations with associated blocks.
18473 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
18474 (clear_unused_block_pointer): Remove code dealing with
18475 blocks in DECL_DEBUG_EXPR locations.
18477 2013-03-21 Richard Biener <rguenther@suse.de>
18479 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
18480 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
18481 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
18482 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
18483 instead of DECL_DEBUG_EXPR_IS_FROM.
18484 * gimplify.c (gimplify_modify_expr): Likewise.
18485 * tree-cfg.c (verify_expr_location_1): Likewise.
18486 * tree-complex.c (create_one_component_var): Likewise.
18487 * tree-sra.c (create_access_replacement): Likewise.
18488 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
18489 (clear_unused_block_pointer): Likewise.
18490 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
18491 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
18492 * var-tracking.c (var_debug_decl): Likewise.
18493 (track_expr_p): Likewise.
18494 * tree-inline.c (add_local_variables): Likewise. Set
18495 DECL_HAS_DEBUG_EXPR_P after copying it.
18496 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
18497 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
18499 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
18502 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
18503 * configure: Regenerate.
18504 * config.in: Regenerate.
18505 * config/i386/i386.md (*movdf_internal): Use
18506 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
18507 movd instead of movq mnemonic for interunit moves.
18508 (*movdi_internal): Ditto.
18510 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
18512 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
18513 (abd<mode>_3): New pattern.
18514 (aba<mode>_3): New pattern.
18515 (fabd<mode>_3): New pattern.
18517 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
18519 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
18520 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
18521 occurrence of REGISTER_PREFIX as its empty string.
18523 2013-03-20 Jeff Law <law@redhat.com>
18525 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
18526 addititional equivalences for equality comparisons between an SSA_NAME
18527 and a constant where the SSA_NAME was set from a widening conversion.
18529 2013-03-20 Walter Lee <walt@tilera.com>
18531 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
18533 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
18535 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
18536 depending on type attribute.
18537 (*movti_internal): Ditto.
18538 (*movtf_internal): Ditto.
18539 (*movxf_internal): Ditto.
18540 (*movdf_internal): Ditto.
18541 (*movsf_internal): Ditto.
18543 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
18545 * config/i386/i386.md (*movti_internal): Set prefix attribute to
18546 maybe_vex for sselog1 and ssemov types.
18547 (*movdi_internal): Reorder operand constraints.
18548 (*movsi_internal): Ditto. Set prefix attribute to
18549 maybe_vex for sselog1 and ssemov types.
18550 (*movtf_internal): Set prefix attribute to maybe_vex
18551 for sselog1 and ssemov types.
18552 (*movdf_internal): Ditto. Set prefix_data16 attribute for
18553 DImode ssemov types. Reorder operand constraints.
18554 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
18555 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
18556 attribute for SImode ssemov types. Reorder operand constraints.
18558 2013-03-20 Martin Jambor <mjambor@suse.cz>
18560 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
18561 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
18563 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
18565 * config/rs6000/predicates.md (indexed_address, update_address_mem
18566 update_indexed_address_mem): New predicates.
18567 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
18568 attribute for load/store instructions.
18569 * config/rs6000/dfp.md (movsd_store): Likewise.
18570 (movsd_load): Likewise.
18571 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
18572 (unnamed HI->DI extend define_insn): Likewise.
18573 (unnamed SI->DI extend define_insn): Likewise.
18574 (unnamed QI->SI extend define_insn): Likewise.
18575 (unnamed QI->HI extend define_insn): Likewise.
18576 (unnamed HI->SI extend define_insn): Likewise.
18577 (unnamed HI->SI extend define_insn): Likewise.
18578 (extendsfdf2_fpr): Likewise.
18579 (movsi_internal1): Likewise.
18580 (movsi_internal1_single): Likewise.
18581 (movhi_internal): Likewise.
18582 (movqi_internal): Likewise.
18583 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
18584 attribute for load/store instructions.
18585 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
18587 (mov<mode>_softfloat): Likewise.
18588 (mov<mode>_hardfloat32): Likewise.
18589 (mov<mode>_hardfloat64): Likewise.
18590 (mov<mode>_softfloat64): Likewise.
18591 (movdi_internal32): Likewise.
18592 (movdi_internal64): Likewise.
18593 (probe_stack_<mode>): Likewise.
18595 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
18597 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
18598 floating point, and decimal floating point to reload iterator.
18600 * config/rs6000/constraints.md (wl constraint): New constraints to
18601 return FLOAT_REGS if certain options are used to reduce the number
18602 of separate patterns that exist in the file.
18603 (wx constraint): Likewise.
18604 (wz constraint): Likewise.
18606 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
18607 -mdebug=reg, print wg, wl, wx, and wz constraints.
18608 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
18609 Initialize the reload functions for 64-bit binary/decimal floating
18611 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
18612 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
18613 create the buffer on the stack to overcome not having a 32-bit
18615 (rs6000_emit_move): Likewise.
18616 (rs6000_secondary_memory_needed_rtx): Likewise.
18617 (rs6000_alloc_sdmode_stack_slot): Likewise.
18618 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
18619 via xxlxor, just like DFmode 0.0.
18621 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
18622 define as 1 if we are running on a power7 or newer.
18623 (enum r6000_reg_class_enum): Add new constraints.
18625 * config/rs6000/dfp.md (movsd): Delete, combine with binary
18626 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
18627 with other moves by using conditional constraits (wg). Use LFIWZX
18628 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
18629 (movsd splitter): Likewise.
18630 (movsd_hardfloat): Likewise.
18631 (movsd_softfloat): Likewise.
18633 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
18634 binary and decimal floating point moves.
18635 (fmove_ok): New attributes to combine binary and decimal floating
18636 point moves, and to combine power6x (mfpgpr) moves along normal
18638 (real_value_to_target): Likewise.
18639 (f32_lr): Likewise.
18640 (f32_lm): Likewise.
18641 (f32_li): Likewise.
18642 (f32_sr): Likewise.
18643 (f32_sm): Likewise.
18644 (f32_si): Likewise.
18645 (movsf): Combine binary and decimal floating point moves. Combine
18646 power6x (mfpgpr) moves with other moves by using conditional
18647 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
18648 (mov<mode> for SFmode/SDmode); Likewise.
18649 (SFmode/SDmode splitters): Likewise.
18650 (movsf_hardfloat): Likewise.
18651 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
18652 (movsf_softfloat): Likewise.
18653 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
18655 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
18656 wx and wz constraints.
18658 * config/rs6000/constraints.md (wg constraint): New constraint to
18659 return FLOAT_REGS if -mmfpgpr (power6x) was used.
18661 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
18664 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
18665 -mdebug=reg, print wg, wl, wx, and wz constraints.
18666 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
18667 Initialize the reload functions for 64-bit binary/decimal floating
18669 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
18670 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
18671 create the buffer on the stack to overcome not having a 32-bit
18673 (rs6000_emit_move): Likewise.
18674 (rs6000_secondary_memory_needed_rtx): Likewise.
18675 (rs6000_alloc_sdmode_stack_slot): Likewise.
18676 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
18677 via xxlxor, just like DFmode 0.0.
18679 * config/rs6000/dfp.md (movdd): Delete, combine with binary
18680 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
18681 with other moves by using conditional constraits (wg). Use LFIWZX
18682 and STFIWX for loading SDmode on power7.
18683 (movdd splitters): Likewise.
18684 (movdd_hardfloat32): Likewise.
18685 (movdd_softfloat32): Likewise.
18686 (movdd_hardfloat64_mfpgpr): Likewise.
18687 (movdd_hardfloat64): Likewise.
18688 (movdd_softfloat64): Likewise.
18690 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
18691 64-bit binary and decimal floating point moves.
18692 (FMOVE64X): Likewise.
18693 (movdf): Combine 64-bit binary and decimal floating point moves.
18694 Combine power6x (mfpgpr) moves with other moves by using
18695 conditional constraits (wg).
18696 (mov<mode> for DFmode/DDmode): Likewise.
18697 (DFmode/DDmode splitters): Likewise.
18698 (movdf_hardfloat32): Likewise.
18699 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
18700 (movdf_softfloat32): Likewise.
18701 (movdf_hardfloat64_mfpgpr): Likewise.
18702 (movdf_hardfloat64): Likewise.
18703 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
18704 (movdf_softfloat64): Likewise.
18705 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
18706 (reload_<mode>_load): Move to later in the file so they aren't in
18707 the middle of the floating point move insns.
18708 (reload_<mode>_store): Likewise.
18710 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
18713 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
18714 constraint if -mdebug=reg.
18715 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
18716 Enable using dd reload support if needed.
18718 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
18719 binary and decimal floating point moves in rs6000.md.
18720 (movtd_internal): Likewise.
18722 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
18723 decimal floating point moves.
18725 (movtf_internal): Likewise.
18726 (mov<mode>_internal, TDmode/TFmode): Likewise.
18727 (movtf_softfloat): Likewise.
18728 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
18730 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
18731 movdi_internal64, using wg constraint for move direct operations.
18732 (movdi_internal64): Likewise.
18734 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
18735 MODES_TIEABLE_P for selected modes. Print the numerical value of
18736 the various virtual registers. Use GPR/FPR first/last values,
18737 instead of hard coding the register numbers. Print which modes
18738 have reload functions registered.
18739 (rs6000_option_override_internal): If -mdebug=reg, trace the options
18740 settings before/after setting cpu, target and subtarget settings.
18741 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
18742 and for secondary reload failures in rs6000_secondary_reload_inner.
18743 (rs6000_secondary_reload_fail): Likewise.
18744 (rs6000_secondary_reload_inner): Likewise.
18746 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
18747 macros for first/last GPR and FPR registers.
18748 (LAST_GPR_REGNO): Likewise.
18749 (FIRST_FPR_REGNO): Likewise.
18750 (LAST_FPR_REGNO): Likewise.
18752 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
18753 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
18754 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
18755 (vcond<mode><mode>): Likewise.
18756 (vcondu<mode><mode>): Likewise.
18757 (vector_gtu<mode>): Likewise.
18758 (vector_gte<mode>): Likewise.
18759 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
18760 to prevent the compiler from converting DImode operations to TImode.
18761 (ior<mode>3): Likewise.
18762 (and<mode>3): Likewise.
18763 (one_cmpl<mode>2): Likewise.
18764 (nor<mode>3): Likewise.
18765 (andc<mode>3): Likewise.
18767 * config/rs6000/constraints.md (wt constraint): New constraint
18768 that returns VSX_REGS if TImode is allowed in VSX registers.
18770 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
18771 constant under VSX.
18773 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
18774 similar to TImode, but it is restricted to being in the GPRs.
18776 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
18777 TImode to occupy a single VSX register.
18779 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
18780 -mvsx-timode for power7/power8.
18781 (power7 cpu): Likewise.
18782 (power8 cpu): Likewise.
18784 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
18785 sure that TFmode/TDmode take up two registers if they are ever
18786 allowed in the upper VSX registers.
18787 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
18789 (rs6000_init_hard_regno_mode_ok): Likewise.
18790 (rs6000_debug_reg_global): Add debugging for PTImode and wt
18791 constraint. Print if LRA is turned on.
18792 (rs6000_option_override_internal): Give an error if -mvsx-timode
18793 and VSX is not enabled.
18794 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
18795 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
18796 to reg+offset addressing. Use PTImode when checking offset
18797 addresses for validity.
18798 (reg_offset_addressing_ok_p): Likewise.
18799 (rs6000_legitimate_offset_address_p): Likewise.
18800 (rs6000_legitimize_address): Likewise.
18801 (rs6000_legitimize_reload_address): Likewise.
18802 (rs6000_legitimate_address_p): Likewise.
18803 (rs6000_eliminate_indexed_memrefs): Likewise.
18804 (rs6000_emit_move): Likewise.
18805 (rs6000_secondary_reload): Likewise.
18806 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
18807 reloads to fpr registers to continue to use reg+offset addressing,
18808 but 64-bit reloads to altivec registers need reg+reg addressing.
18809 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
18810 it. Treat LO_SUM like a PLUS operation.
18811 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
18812 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
18813 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
18814 registers to share a register with a smaller sized type, since VSX
18815 puts scalars in the upper 64-bits.
18816 (print_operand): Add support for PTImode.
18817 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
18818 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
18819 registers, but don't have arithmetic support.
18820 (rs6000_memory_move_cost): Add test for VSX.
18821 (rs6000_opt_masks): Add -mvsx-timode.
18823 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
18826 (VSr): Use wt constraint for TImode.
18827 (VSv): Drop TImode support.
18828 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
18829 (vsx_movti_64bit): Likewise.
18830 (vsx_movti_32bit): Likewise.
18831 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
18832 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
18833 one '?' on the appropriate output constraint. Do not allow TImode
18834 logical operations on 32-bit systems.
18835 (vsx_ior<mode>3): Likewise.
18836 (vsx_xor<mode>3): Likewise.
18837 (vsx_one_cmpl<mode>2): Likewise.
18838 (vsx_nor<mode>3): Likewise.
18839 (vsx_andc<mode>3): Likewise.
18840 (vsx_concat_<mode>): Likewise.
18841 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
18843 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
18844 OPTION_MASK_VSX_TIMODE.
18845 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
18846 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
18848 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
18849 (TI2 iterator): New iterator for TImode, PTImode.
18850 (wd mode attribute): Add values for vector types.
18851 (movti_string): Replace TI move operations with operations for TImode
18852 and PTImode. Add support for TImode being allowed in VSX registers.
18853 (mov<mode>_string, TImode/PTImode): Likewise.
18854 (movti_ppc64): Likewise.
18855 (mov<mode>_ppc64, TImode/PTImode): Likewise.
18856 (TI mode splitters): Likewise.
18858 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
18861 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
18863 PR tree-optimization/56355
18864 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
18865 Also handle integers with undefined overflow.
18867 2013-03-20 Catherine Moore <clm@codesourcery.com>
18868 Maciej W. Rozycki <macro@codesourcery.com>
18869 Tom de Vries <tom@codesourcery.com>
18870 Nathan Sidwell <nathan@codesourcery.com>
18871 Iain Sandoe <iain@codesourcery.com>
18872 Nathan Froyd <froydnj@codesourcery.com>
18873 Chao-ying Fu <fu@mips.com>
18875 * doc/extend.texi: (micromips, nomicromips, nocompression):
18876 Document new function attributes.
18877 * doc/invoke.texi (minterlink-compressed, mmicromips,
18878 m14k, m14ke, m14kec): Document new options.
18879 (minterlink-mips16): Update documentation.
18880 * doc/md.texi (ZC, ZD): Document new constraints.
18881 * configure.ac (gcc_cv_as_micromips): Check if linker
18882 supports the .set micromips directive.
18883 * configure: Regenerate.
18884 * config.in: Regenerate.
18885 * config/mips/mips-tables.opt: Regenerate.
18886 * config/mips/micromips.md: New file.
18887 * constraints.md (ZC, ZD): New constraints.
18888 * config/mips/predicates.md (movep_src_register): New predicate.
18889 (movep_src_operand): New predicate.
18890 (non_volatile_mem_operand): New predicate.
18891 * config/mips/mips.md (multimem): New type.
18892 (length): Differentiate between 17-bit and 18-bit branch offsets.
18893 (MOVEP1, MOVEP2): New mode iterator.
18894 (mov_<load>l): Use ZC constraint.
18895 (mov_<load>r): Likewise.
18896 (mov_<store>l): Likewise.
18897 (mov_<store>r): Likewise.
18898 (*branch_equality<mode>_inverted): Add microMIPS support.
18899 (*branch_equality<mode>): Likewise.
18900 (*jump_absolute): Likewise.
18901 (indirect_jump_<mode>): Likewise.
18902 (tablejump_<mode>): Likewise.
18903 (<optab>_internal): Likewise.
18904 (sibcall_internal): Likewise.
18905 (sibcall_value_internal): Likewise.
18906 (prefetch): Use constraint ZD.
18907 * config/mips/mips.opt (minterlink-compressed): New option.
18908 (minterlink-mips16): Now an alias for minterlink-compressed.
18909 (mmicromips): New option.
18910 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
18911 (compare_and_swap_12): Likewise.
18912 (sync_add<mode>): Likewise.
18913 (sync_<optab>_12): Likewise.
18914 (sync_old_<optab>_12): Likewise.
18915 (sync_new_<optab>_12): Likewise.
18916 (sync_nand_12): Likewise.
18917 (sync_old_nand_12): Likewise.
18918 (sync_new_nand_12): Likewise.
18919 (sync_sub<mode>): Likewise.
18920 (sync_old_add<mode>): Likewise.
18921 (sync_old_sub<mode>): Likewise.
18922 (sync_new_add<mode>): Likewise.
18923 (sync_new_sub<mode>): Likewise.
18924 (sync_<optab><mode>): Likewise.
18925 (sync_old_<optab><mode>): Likewise.
18926 (sync_new_<optab><mode>): Likewise.
18927 (sync_nand<mode>): Likewise.
18928 (sync_old_nand<mode>): Likewise.
18929 (sync_new_nand<mode>): Likewise.
18930 (sync_lock_test_and_set<mode>): Likewise.
18931 (test_and_set_12): Likewise.
18932 (atomic_compare_and_swap<mode>): Likewise.
18933 (atomic_exchange<mode>_llsc): Likewise.
18934 (atomic_fetch_add<mode>_llsc): Likewise.
18935 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
18936 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
18937 (umips_save_restore_pattern_p): Likewise.
18938 (umips_load_store_pair_p): Likewise.
18939 (umips_output_load_store_pair): Likewise.
18940 (umips_movep_target_p): Likewise.
18941 (umips_12bit_offset_address_p): Likewise.
18942 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
18943 (mips_base_mips16): Rename this...
18944 (mips_base_compression_flags): ...to this. Update all uses.
18945 (mips_attribute_table): Add micromips, nomicromips and nocompression.
18946 (mips_mips16_decl_p): Delete.
18947 (mips_nomips16_decl_p): Delete.
18948 (mips_get_compress_on_flags): New function.
18949 (mips_get_compress_off_flags): New function.
18950 (mips_get_compress_mode): New function.
18951 (mips_get_compress_on_name): New function.
18952 (mips_get_compress_off_name): New function.
18953 (mips_insert_attributes): Support multiple compression types.
18954 (mips_merge_decl_attributes): Likewise.
18955 (umips_12bit_offset_address_p): New function.
18956 (mips_start_function_definition): Emit .set micromips directive.
18957 (mips_call_may_need_jalx_p): New function.
18958 (mips_function_ok_for_sibcall): Add microMIPS support.
18959 (mips_print_operand_punctuation): Support short delay slots and
18961 (umips_swm_mask, umips_swm_encoding): New.
18962 (umips_build_save_restore): New function.
18963 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
18964 (was_mips16_p): Remove.
18965 (old_compression_mode): New.
18966 (mips_set_compression_mode): New function.
18967 (mips_set_current_function): Add microMIPS support.
18968 (mips_option_override): Likewise.
18969 (umips_save_restore_pattern_p): New function.
18970 (umips_output_save_restore): New function.
18971 (umips_load_store_pair_p_1): New function.
18972 (umips_load_store_pair_p): New function.
18973 (umips_output_load_store_pair_1): New function.
18974 (umips_output_load_store_pair): New function.
18975 (umips_movep_target_p) New function.
18976 (mips_prepare_pch_save): Add microMIPS support.
18977 * config/mips/mips.h (TARGET_COMPRESSION): New.
18978 (TARGET_CPU_CPP_BUILTINS): Update macro
18979 to use new compression flags and to support microMIPS.
18980 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
18981 (MIPS_ARCH_FLOAT_SPEC): Likewise.
18982 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
18983 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
18984 (ASM_SPEC): Support mmicromips and mno-micromips.
18985 (M16STORE_REG_P): New macro.
18986 (MIPS_CALL): Support TARGET_MICROMIPS.
18987 (MICROMIPS_J): New macro.
18988 (mips_base_mips16): Rename this...
18989 (mips_base_compression_flags): ...to this.
18990 (UMIPS_12BIT_OFFSET_P): New macro.
18991 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
18992 (MULTILIB_DIRNAMES): Likewise.
18993 2013-03-20 Richard Biener <rguenther@suse.de>
18995 PR tree-optimization/56661
18996 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
18997 the result does not have to be distinct.
18999 2013-03-20 Richard Biener <rguenther@suse.de>
19001 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
19004 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
19005 Steven Bosscher <steven@gcc.gnu.org>
19007 PR rtl-optimization/56605
19008 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
19010 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
19013 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
19014 that require movd instead of movq.
19016 2013-03-20 Richard Biener <rguenther@suse.de>
19018 * tree-ssa-structalias.c (struct variable_info): Add pointer
19019 to the first field of an aggregate with sub-vars. Make
19020 this and the pointer to the next subfield its ID.
19021 (vi_next): New function.
19022 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
19023 storedanything_id, integer_id): Increment by one.
19024 (new_var_info, get_call_vi, lookup_call_clobber_vi,
19025 get_call_clobber_vi): Adjust.
19026 (solution_set_expand): Simplify and speedup.
19027 (solution_set_add): Inline into ...
19028 (set_union_with_increment): ... this. Adjust accordingly.
19029 (do_sd_constraint): Likewise.
19030 (do_ds_constraint): Likewise.
19031 (do_complex_constraint): Simplify.
19032 (build_pred_graph): Adjust.
19033 (solve_graph): Likewise. Simplify and speedup.
19034 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
19035 get_constraint_for_component_ref, get_constraint_for_1,
19036 first_vi_for_offset, first_or_preceding_vi_for_offset,
19037 create_function_info_for, create_variable_info_for_1,
19038 create_variable_info_for, intra_create_variable_infos): Adjust.
19039 (init_base_vars): Push NULL for ID zero.
19040 (compute_points_to_sets): Adjust.
19042 2013-03-20 Richard Biener <rguenther@suse.de>
19044 * cfgloop.c (verify_loop_structure): Streamline and avoid
19045 ICEing on corrupt loop tree.
19046 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
19049 2013-03-20 Richard Biener <rguenther@suse.de>
19051 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
19052 check whether an SSA update is needed.
19054 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
19056 * config/mips/constraints.md (T): Rename to...
19060 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
19061 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
19063 2013-03-19 Ian Bolton <ian.bolton@arm.com>
19065 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
19066 (*subsi3_carryin_uxtw): Likewise.
19068 2013-03-19 Ian Bolton <ian.bolton@arm.com>
19070 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
19071 (*rorsi3_insn_uxtw): Likewise.
19073 2013-03-19 Ian Bolton <ian.bolton@arm.com>
19075 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
19076 (*extrsi5_insn_uxtw): Likewise.
19078 2013-03-19 Richard Biener <rguenther@suse.de>
19080 PR tree-optimization/56273
19081 * passes.c (init_optimization_passes): Move second VRP after DOM.
19083 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
19085 * config/i386/i386.md (*movti_internal): Merge from
19086 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
19087 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
19088 nox64 isa attributes.
19090 2013-03-18 Richard Biener <rguenther@suse.de>
19092 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
19094 (merge_node_constraints): Likewise.
19095 (build_succ_graph): Likewise.
19096 (valid_graph_edge): Inline into single caller.
19097 (unify_nodes): Likewise. Use bitmap_set_bit return value
19099 (scc_visit): Fix formatting and variable use.
19100 (do_sd_constraint): Use gcc_checking_assert.
19101 (do_ds_constraint): Likewise.
19102 (do_complex_constraint): Likewise.
19103 (condense_visit): Likewise. Cleanup.
19104 (dump_pred_graph): New function.
19105 (perform_var_substitution): Dump the pred-graph before
19106 variable substitution.
19107 (find_equivalent_node): Use gcc_checking_assert.
19108 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
19110 2013-03-18 Richard Biener <rguenther@suse.de>
19112 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
19113 Remove cond_expr_stmt_list argument and do not gimplify the
19115 (vect_loop_versioning): Adjust.
19116 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
19117 Cleanup to use less temporaries.
19118 (vect_create_data_ref_ptr): Cleanup.
19120 2013-03-18 Jakub Jelinek <jakub@redhat.com>
19122 PR tree-optimization/56635
19123 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
19124 require types_compatible_p types.
19126 2013-03-18 Nick Clifton <nickc@redhat.com>
19128 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
19129 spurious backslash.
19131 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
19132 Add missing line to comment describing function.
19134 2013-03-18 Richard Biener <rguenther@suse.de>
19136 PR tree-optimization/56210
19137 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
19138 Handle string / character search functions.
19139 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
19141 2013-03-18 Richard Biener <rguenther@suse.de>
19143 PR middle-end/56483
19144 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
19145 and implement properly.
19146 * gimple.h (gimple_cond_single_var_p): Remove.
19148 2013-03-18 Richard Biener <rguenther@suse.de>
19150 * tree-data-ref.h (find_data_references_in_loop): Declare.
19151 * tree-data-ref.c (get_references_in_stmt): Use a stack
19152 vector pre-allocated in the callers.
19153 (find_data_references_in_stmt): Adjust.
19154 (graphite_find_data_references_in_stmt): Likewise.
19155 (create_rdg_vertices): Likewise.
19156 (find_data_references_in_loop): Export.
19157 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
19158 Compute dependences here...
19159 (vect_analyze_data_refs): ...not here. When we encounter
19160 a non-vectorizable data reference in basic-block vectorization
19161 truncate the data reference vector. Do not bother to
19162 fixup data-dependence information for gather loads.
19163 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
19164 of data references, as reported.
19166 2013-03-18 Richard Biener <rguenther@suse.de>
19168 PR tree-optimization/3713
19169 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
19170 has_constants and expr.
19171 (stmt_has_constants): Properly valueize SSA names when deciding
19172 whether the stmt has constants.
19174 2013-03-18 Richard Biener <rguenther@suse.de>
19176 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
19177 whole function when there is nothing to do.
19178 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
19179 * tree-vectorizer.c (vectorize_loops): Update virtual and
19180 loop-closed SSA once.
19181 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
19183 2013-03-18 Richard Biener <rguenther@suse.de>
19185 PR middle-end/56113
19186 * domwalk.c (bb_postorder): New global static.
19187 (cmp_bb_postorder): New function.
19188 (walk_dominator_tree): Replace scheme imposing an order for
19189 visiting dominator sons by one sorting them at the time they
19190 are pushed on the stack.
19192 2013-03-18 Richard Biener <rguenther@suse.de>
19194 PR tree-optimization/39326
19195 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
19196 (struct mem_ref): Replace mem member with ao_ref typed member.
19197 (MEM_ANALYZABLE): Adjust.
19198 (memref_eq): Likewise.
19199 (mem_ref_alloc): Likewise.
19200 (gather_mem_refs_stmt): Likewise.
19201 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
19202 (execute_sm_if_changed_flag_set): Adjust.
19203 (execute_sm): Likewise.
19204 (ref_always_accessed_p): Likewise.
19205 (refs_independent_p): Likewise.
19206 (can_sm_ref_p): Likewise.
19208 2013-03-18 Jakub Jelinek <jakub@redhat.com>
19211 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
19212 return 1 even for !unsignedp.
19214 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
19216 * config/i386/i386.md (isa): Add x64 and nox64.
19217 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
19218 (*pushtf): Enable *roF alternative for x64 isa only.
19219 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
19220 mode attribute of integer alternatives to DImode for TARGET_64BIT.
19221 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
19222 (*movtf_internal): Merge from *movtf_internal_rex64 and
19223 *movtf_internal_sse. Use x64 and nox64 isa attributes.
19224 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
19225 nox64 isa attributes.
19226 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
19227 nox64 isa attributes.
19228 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
19230 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
19232 * config/alpha/alpha.c (TARGET_LRA_P): New define.
19234 2013-03-17 Jakub Jelinek <jakub@redhat.com>
19237 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
19238 class names. Remove trailing comma after "ALL_REGS".
19240 2013-03-16 Jan Hubicka <jh@suse.cz>
19242 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
19243 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
19244 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
19245 of cgraph_get_create_node.
19246 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
19248 2013-03-16 Jason Merrill <jason@redhat.com>
19251 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
19252 with DW_AT_default_value.
19254 2013-03-16 Jakub Jelinek <jakub@redhat.com>
19256 * BASE-VER: Set to 4.9.0.
19258 2013-03-14 Andi Kleen <ak@linux.intel.com>
19261 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
19262 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
19263 Document _x* TSX intrinsics.
19265 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
19266 David Holsgrove <david.holsgrove@xilinx.com>
19268 * configure.ac: Add MicroBlaze TLS support detection.
19269 * configure: Regenerate.
19270 * config/microblaze/microblaze-protos.h
19271 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
19272 symbol_mentioned_p, label_mentioned_p): Add prototypes.
19273 * config/microblaze/microblaze.c (microblaze_address_type): Add
19274 ADDRESS_TLS and tls_reloc address types.
19275 (microblaze_address_info): Add tls_reloc.
19276 (TARGET_HAVE_TLS): Define.
19277 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
19278 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
19279 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
19280 load_tls_operand, microblaze_call_tls_get_addr,
19281 microblaze_legitimize_tls_address): New functions.
19282 (microblaze_classify_unspec): Handle UNSPEC_TLS.
19283 (get_base_reg): Use microblaze_tls_symbol_p.
19284 (microblaze_classify_address): Handle TLS.
19285 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
19286 label_mentioned_p and microblaze_tls_referenced_p.
19287 (microblaze_legitimize_address): Handle TLS.
19288 (microblaze_address_insns): Handle ADDRESS_TLS.
19289 (pic_address_needs_scratch): Handle TLS.
19290 (print_operand_address): Handle TLS.
19291 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
19292 (microblaze_expand_move): Handle TLS.
19293 (microblaze_legitimate_constant_p): Check
19294 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
19295 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
19296 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
19297 (PIC_OFFSET_TABLE_REGNUM): Set.
19298 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
19299 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
19300 (addsi3, movsi_internal2, movdf_internal): Update constraints
19301 * config/microblaze/predicates.md (arith_plus_operand): Define
19302 (move_operand): Redefine as move_src_operand,
19303 check microblaze_tls_referenced_p.
19305 2013-03-14 Ian Bolton <ian.bolton@arm.com>
19307 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
19308 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
19310 2013-03-14 Ian Bolton <ian.bolton@arm.com>
19312 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
19315 2013-03-14 Jakub Jelinek <jakub@redhat.com>
19317 PR tree-optimization/53265
19318 * common.opt (Waggressive-loop-optimizations): New option.
19319 * tree-ssa-loop-niter.c: Include tree-pass.h.
19320 (do_warn_aggressive_loop_optimizations): New function.
19321 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
19322 if number_of_latch_executions returned constant.
19323 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
19324 early. If number_of_latch_executions returned constant, set
19325 nb_iterations_upper_bound back to it.
19326 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
19328 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
19329 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
19331 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
19332 (MULTILIB_OSDIRNAMES): Set.
19333 * genmultilib: If defaultosdirname doesn't start with :: , set
19334 defaultosdirname2 instead, clear it and emit two . multilib_raw
19335 entries instead of just one.
19337 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
19339 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
19340 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
19341 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
19342 (SUBTARGET_OVERRIDE_OPTIONS): New.
19344 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
19347 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
19348 (musermode): Convert to Var(TARGET_USERMODE).
19349 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
19350 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
19351 * config/sh/sh.c (sh_option_override): Use
19352 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
19353 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
19355 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
19357 (udivsi3_i4_single, divsi3_i4_single): Use
19358 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
19360 2013-03-13 Dave Korn <dave.korn.cygwin@....>
19362 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
19365 2013-03-13 Richard Biener <rguenther@suse.de>
19367 PR tree-optimization/56608
19368 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
19369 calls when vectorizing basic-blocks.
19371 2013-03-13 Jakub Jelinek <jakub@redhat.com>
19374 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
19377 2013-03-12 Jakub Jelinek <jakub@redhat.com>
19379 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
19381 2013-03-11 Jan Hubicka <jh@suse.cz>
19384 * lto-streamer-out.c (output_symbol_p): Skip references from
19385 constructors of external variables.
19387 2013-03-11 Jan Hubicka <jh@suse.cz>
19389 PR middle-end/56571
19390 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
19392 * emit-rtl.c (verify_rtx_sharing): Likewise.
19393 (copy_insn_1): Likewise.
19394 * rtl.c (copy_rtx): Likewise.
19396 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
19399 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
19400 output_operand_lossage message.
19402 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
19405 * arm.c (shift_op): Validate RTL pattern on the fly.
19406 (arm_print_operand, case 'S'): Don't use shift_operator to validate
19409 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
19412 * config/pa/pa.md (call_value): Check for calls to powf and direct to
19413 new call patterns that clobber %fr12.
19414 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
19415 split and postreload patterns.
19416 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
19417 registers %fr12 and %fr12R as call used.
19419 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
19421 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
19422 (canon_address, record_store, replace_read, check_mem_read_rtx,
19423 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
19424 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
19425 rest_of_handle_dse): Likewise.
19427 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
19429 PR middle-end/56524
19430 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
19432 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
19433 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
19434 (save_optabs_if_changed): Replace with...
19435 (init_tree_optimization_optabs): ...this.
19436 * optabs.c (save_optabs_if_changed): Rename to...
19437 (init_tree_optimization_optabs): ...this. Take the optimization node
19438 as argument. Do nothing if the base optabs are already correct.
19439 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
19440 to recompute optabs.
19441 * function.h (function): Remove optabs field.
19442 * function.c (invoke_set_current_function_hook): Call
19443 init_tree_optimization_optabs. Use the result to initialize
19446 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
19448 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
19449 if GTMA_HAS_NO_INSTRUMENTATION.
19450 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
19451 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
19452 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
19453 * gimple-pretty-print.c (dump_gimple_transaction): Handle
19454 GTMA_HAS_NO_INSTRUMENTATION.
19456 2013-03-08 Jakub Jelinek <jakub@redhat.com>
19458 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
19461 2013-03-08 Marek Polacek <polacek@redhat.com>
19462 Jakub Jelinek <jakub@redhat.com>
19464 PR tree-optimization/56478
19465 * predict.c (is_comparison_with_loop_invariant_p): Change the
19466 type of loop_step to tree.
19467 (predict_loops): Adjust.
19468 (predict_iv_comparison): Perform the computations on double_ints.
19470 2013-03-08 Richard Biener <rguenther@suse.de>
19472 PR tree-optimization/56570
19473 * tree-cfg.c (verify_expr_location_1): Verify locations for
19475 * tree-sra.c (create_access_replacement): Strip locations
19476 from DECL_DEBUG_EXPRs.
19478 2013-03-08 Richard Biener <rguenther@suse.de>
19480 * tree-inline.c (expand_call_inline): Do not associate
19481 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
19482 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
19484 2013-03-08 Richard Biener <rguenther@suse.de>
19486 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
19487 or block changes with -Og. Fix for location / block encoding
19488 changes and PHI arguments with locations.
19490 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
19492 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
19494 (struct output_info): Likewise.
19495 (register_overhead): Remove bad gcc_assert.
19496 (bitmap_find_bit): If there is only a single bitmap element, do not
19497 count a miss as a search.
19498 (print_statistics): Update for counter type changes.
19499 (dump_bitmap_statistics): Likewise. Print headers such that they
19500 are properly lined up with the printed counters.
19502 2013-03-07 Jakub Jelinek <jakub@redhat.com>
19504 PR tree-optimization/56559
19505 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
19506 check that it has only a single use.
19508 2013-03-07 Richard Biener <rguenther@suse.de>
19510 * doc/invoke.texi (fwhole-program): Discourage use in combination
19513 2013-03-06 Jakub Jelinek <jakub@redhat.com>
19515 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
19517 PR tree-optimization/56539
19518 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
19519 instead of GSI_CONTINUE_LINKING as last argument to
19520 force_gimple_operand_gsi. Adjust function comment.
19522 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
19525 PR middle-end/56548
19526 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
19527 promoted mode, convert the result back to the original mode.
19529 2013-03-06 Richard Biener <rguenther@suse.de>
19531 PR middle-end/56294
19532 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
19533 (insert_updated_phi_nodes_compare_uids): New function.
19534 (update_ssa): Sort symbols_to_rename after UID before
19535 traversing it to insert PHI nodes.
19537 2013-03-06 Richard Biener <rguenther@suse.de>
19539 PR middle-end/50494
19540 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
19541 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
19544 2013-02-13 Richard Biener <rguenther@suse.de>
19547 * varasm.c (output_constant_def_1): Get the decl representing
19548 the constant as argument.
19549 (output_constant_def): Wrap output_constant_def_1.
19550 (make_decl_rtl): Use output_constant_def_1 with the decl
19551 representing the constant.
19552 (build_constant_desc): Optionally re-use a decl already
19553 representing the constant.
19554 (tree_output_constant_def): Adjust.
19556 2013-03-06 Joey Ye <joey.ye@arm.com>
19559 * gcc.c (convert_white_space): New function.
19560 (main): Handles white space in function name.
19562 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
19565 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
19566 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
19567 to SH_DIV_CALL_TABLE for TARGET_SH2.
19568 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
19570 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
19571 call-table options.
19573 2013-03-05 Sterling Augustine <saugustine@google.com>
19574 Cary Coutant <ccoutant@google.com>
19577 * dwarf2out.c (resolve_addr): Don't call
19578 remove_loc_list_addr_table_entries a second time for the same
19581 2013-03-05 Jakub Jelinek <jakub@redhat.com>
19584 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
19585 (avoid_complex_debug_insns): New function.
19586 (expand_debug_locations): Call it.
19588 PR rtl-optimization/56484
19589 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
19590 lifetimes of hard registers on small register class machines.
19592 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
19594 * config/microblaze/microblaze-protos.h: Rename
19595 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
19596 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
19598 (microblaze_fast_interrupt_function_p): New function.
19599 (microblaze_is_interrupt_handler): Rename to
19600 microblaze_is_interrupt_variant and add fast_interrupt check.
19601 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
19602 (save_restore_insns): Likewise.
19603 (compute_frame_size): Likewise.
19604 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
19605 (microblaze_globalize_label): Likewise.
19606 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
19607 * config/microblaze/microblaze.md: Use wrapper
19608 microblaze_is_interrupt_variant.
19610 2013-03-05 Kai Tietz <ktietz@redhat.com>
19612 * sdbout.c (sdbout_one_type): Switch to current function's section
19613 supporting cold/hot.
19615 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
19617 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
19620 2013-03-05 Jakub Jelinek <jakub@redhat.com>
19622 PR middle-end/56461
19623 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
19624 if VALGRIND_GET_VBITS is defined, temporarily make object
19625 memory all defined, and restore previous valgrind addressability
19626 and definability afterwards. Free this_object at the end.
19628 PR middle-end/56461
19629 * lra.c (lra): Call lra_clear_live_ranges if live_p,
19630 right before calling lra_create_live_ranges, also call it
19631 when clearing live_p. Only call lra_clear_live_ranges
19632 at the end if live_p.
19634 PR middle-end/56461
19635 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
19637 2013-03-05 Richard Biener <rguenther@suse.de>
19639 PR tree-optimization/56521
19640 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
19643 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
19646 * except.h (remove_unreachable_eh_regions): New prototype.
19647 * except.c (remove_eh_handler_splicer): New function, split out
19648 of remove_eh_handler.
19649 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
19650 warning about running it on many EH regions one at a time.
19651 (remove_unreachable_eh_regions_worker): New function, walk the
19652 EH tree in depth-first order and remove non-marked regions.
19653 (remove_unreachable_eh_regions): New function.
19654 * tree-eh.c (mark_reachable_handlers): New function, split out
19655 from remove_unreachable_handlers.
19656 (remove_unreachable_handlers): Use mark_reachable_handlers and
19657 remove_unreachable_eh_regions.
19658 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
19659 and remove_unreachable_eh_regions.
19661 2013-03-05 Richard Biener <rguenther@suse.de>
19663 PR middle-end/56525
19664 * loop-init.c (fix_loop_structure): Remove loops in two stages,
19665 not freeing them until the end.
19667 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19669 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
19671 2013-03-05 Richard Biener <rguenther@suse.de>
19673 PR tree-optimization/56270
19674 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
19675 of loads after scheduling an SLP instance.
19677 2013-03-05 Jakub Jelinek <jakub@redhat.com>
19679 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
19681 (check_gcc_parallelize): Run guality.exp as a separate job from
19682 vect.exp with unsorted.exp and $(dg_target_exps) separately from
19683 struct-layout-1.exp with stackalign.exp.
19685 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
19687 PR middle-end/56461
19688 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
19689 load_index sbitmap even if some bit in it isn't set.
19691 PR middle-end/56461
19692 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
19693 (discover_iteration_bound_by_body_walk): Change queues to
19694 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
19695 spelling in comment. Call safe_push on queues[bound_index] directly.
19696 Release queues[queue_index] in every iteration unconditionally.
19697 Release bounds vector.
19699 PR middle-end/56461
19700 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
19701 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
19702 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
19705 2013-03-05 Richard Biener <rguenther@suse.de>
19708 * tree-inline.c (remap_blocks_to_null): New function.
19709 (expand_call_inline): When expanding a call stmt without
19710 an associated block inline remap all callee blocks to NULL.
19712 2013-03-05 Jakub Jelinek <jakub@redhat.com>
19714 PR rtl-optimization/56494
19715 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
19716 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
19717 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
19719 PR middle-end/56461
19720 * sel-sched-ir.c (free_sched_pools): Release
19721 succs_info_pool.stack[succs_info_pool.max_top] vectors too
19722 if succs_info_pool.max_top isn't -1.
19725 * opts.c (opts_obstack, opts_concat): Moved to...
19726 * opts-common.c (opts_obstack, opts_concat): ... here.
19728 2013-03-04 Jakub Jelinek <jakub@redhat.com>
19730 PR middle-end/56461
19731 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
19733 2013-03-04 Martin Jambor <mjambor@suse.cz>
19735 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
19736 all appropriate places.
19738 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
19740 PR tree-optimization/56424
19741 * ipa-split.c (split_function): Do not set the RSO flag if result is
19742 not by reference and its type is a register type.
19744 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
19746 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
19747 (microblaze_legitimate_pic_operand): Likewise
19748 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
19749 new function microblaze_legitimate_pic_operand
19750 * config/microblaze/microblaze-protos.h
19751 (microblaze_legitimate_pic_operand): Declare.
19753 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
19755 * config/microblaze/predicates.md (call_insn_simple_operand):
19756 New predicate for supported rtx code types.
19757 * config/microblaze/microblaze.md (call_internal1): Use
19758 call_insn_simple_operand predicate.
19760 2013-03-04 Jakub Jelinek <jakub@redhat.com>
19762 PR middle-end/56461
19763 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
19764 partitions.ordered_remove.
19766 PR middle-end/56461
19767 * tree-vect-stmts.c (vectorizable_conversion): Don't call
19768 vec_oprnds0.create (1) for modifier == NONE.
19770 PR middle-end/56461
19771 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
19772 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
19773 vec_oprnds1 right before pushing anything to it for
19776 PR middle-end/56461
19777 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
19778 set nbbs to 0 instead of having separate code path.
19779 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
19780 instead of false as last argument if returning NULL.
19782 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
19784 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
19785 the attribute is now called "target" instead of "option".
19786 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
19787 * doc/tm.texi.in (Target Attributes): Likewise document the correct
19788 attribute/pragma name for TARGET_OPTION_VALID_P and
19789 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
19790 * doc/tm.texi: Regenerated.
19792 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
19794 * config/microblaze/microblaze.c:
19795 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
19796 * config/microblaze/microblaze.h: Add -mxl-reorder to
19798 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
19799 instructions emitted if TARGET_REORDER.
19800 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
19801 or 0 for -m/-mno case, but initialises as 2 to detect default use case
19804 2013-03-01 Xinliang David Li <davidxl@google.com>
19806 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
19809 2013-03-01 Jakub Jelinek <jakub@redhat.com>
19811 PR middle-end/56461
19812 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
19813 vector even when returning true. Fix up function comment formatting.
19815 PR middle-end/56461
19816 * ira-build.c (ira_loop_nodes_count): New variable.
19817 (create_loop_tree_nodes): Initialize it.
19818 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
19820 PR middle-end/56461
19821 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
19822 method on dr_chain and result_chain.
19823 * tree-vect-stmts.c (vectorizable_store): Only call
19824 result_chain.create if j == 0.
19826 PR middle-end/56461
19827 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
19828 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
19829 before overwriting it.
19831 2013-03-01 Tobias Burnus <burnus@net-b.de>
19833 * doc/extended.texi (C Extensions): Change order in @menu
19835 (Other MIPS Built-in Functions): Move last MIPS entry before
19836 "picoChip Built-in Functions".
19837 (SH Built-in Functions): Move after RX Built-in Functions.
19838 * doc/gcc.texi (Introduction): Change order in @menu
19840 * doc/md.texi (Constraints): Ditto.
19841 * gty.texi (Type Information): Ditto.
19842 (User-provided marking routines for template types): Make
19844 * doc/invoke.texi (AArch64 Options): Move before
19845 "Adapteva Epiphany Options".
19847 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
19848 Jakub Jelinek <jakub@redhat.com>
19851 * asan.c (gate_asan): Lookup no_sanitize_address instead of
19852 no_address_safety_analysis attribute.
19853 * doc/extend.texi (no_address_safety_attribute): Rename to
19854 no_sanitize_address attribute, mention no_address_safety_analysis
19855 attribute as deprecated alias.
19857 2013-02-28 Jakub Jelinek <jakub@redhat.com>
19859 PR middle-end/56461
19860 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
19861 type to vec<vec<tree> > *.
19862 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
19863 to be vec<tree> instead of vec<tree> *, set vec_defs
19864 to vNULL and call vec_defs.create (number_of_vects), adjust other
19866 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
19867 vectorizable_condition): Adjust vect_get_slp_defs callers.
19869 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
19871 * config/aarch64/aarch64.c
19872 (aarch64_float_const_representable): Remove unused variable.
19874 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
19876 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
19878 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
19880 * config/aarch64/aarch64-builtins.c
19881 (aarch64_init_simd_builtins): Make static.
19883 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
19885 * config/aarch64/aarch64.c
19886 (aarch64_simd_make_constant): Make static.
19888 2013-02-28 Martin Jambor <mjambor@suse.cz>
19890 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
19891 with no initialization to the RHS of debug statements.
19893 2013-02-28 Martin Jambor <mjambor@suse.cz>
19895 PR tree-optimization/56294
19896 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
19898 (get_access_replacement): Do not call create_access_replacement.
19899 Assert a replacement exists.
19900 (get_repl_default_def_ssa_name): Create the replacement declaration
19903 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
19905 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
19906 final_end_function.
19908 2013-02-28 Marek Polacek <polacek@redhat.com>
19910 PR rtl-optimization/56466
19911 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
19912 if we're changing a loop.
19913 (peel_loops_completely): Likewise.
19915 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
19918 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
19920 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
19923 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
19924 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
19925 INTX_FTYPE_FX, FX_FTYPE_INTX.
19926 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
19928 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
19930 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
19931 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
19932 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
19933 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
19934 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
19935 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
19936 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
19937 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
19938 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
19939 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
19940 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
19941 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
19942 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
19943 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
19944 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
19945 (avrxmega6): Increase max flash segments from 5 to 6.
19946 * config/avr/t-multilib: Regenerate.
19947 * config/avr/avr-tables.opt: Regenerate.
19948 * doc/avr-mmcu.texi: Regenerate.
19950 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
19952 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
19953 (avr_device_to_arch): Rename to avr_device_to_ld.
19954 (avr_device_to_as): New prototype.
19955 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
19956 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
19957 * config/avr/driver-avr.c (avr_device_to_as): New.
19958 (avr_device_to_arch): Rename to avr_device_to_ld.
19960 2013-02-27 Jakub Jelinek <jakub@redhat.com>
19962 PR middle-end/56461
19963 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
19964 method on dr_chain and result_chain.
19966 PR middle-end/56461
19967 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
19968 pointer_set_destroy on not_executed_last_iteration.
19970 PR middle-end/56461
19971 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
19973 PR middle-end/56461
19974 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
19975 FOR_EACH_DEFINED_FUNCTION when freeing state.
19977 PR middle-end/56461
19978 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
19980 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
19983 PR middle-end/56461
19984 * ipa-cp.c (decide_whether_version_node): Call vec_free on
19985 known_aggs[i].items and release known_aggs vector.
19987 PR middle-end/56461
19988 * ipa-reference.c (propagate): Free node_info even for alias nodes.
19990 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
19992 * config/microblaze/microblaze.c (microblaze_emit_compare):
19993 Use xor for EQ/NE comparisions.
19994 * config/microblaze/microblaze.md (cstoresf4): Add constraints
19995 (cbranchsf4): Adjust operator to comparison_operator.
19997 2013-02-27 Jakub Jelinek <jakub@redhat.com>
19999 PR middle-end/56461
20000 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
20002 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
20003 vec_safe_push, always update *slot.
20004 (redirect_edge_var_map_clear): Use vec_free.
20005 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
20006 (free_var_map_entry): Use vec_free.
20007 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
20008 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
20010 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
20012 PR middle-end/45472
20013 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
20014 when the may_trap_p bit of the exprs being merged differs.
20015 Reorder tests for speculativeness in the logical and operator.
20017 2013-02-27 Jakub Jelinek <jakub@redhat.com>
20019 * incpath.c (add_standard_paths): Use reconcat instead of concat
20020 where appropriate and avoid leaking memory.
20022 * opts.h: Include obstack.h.
20023 (opts_concat): New prototype.
20024 (opts_obstack): New declaration.
20025 * opts.c (opts_concat): New function.
20026 (opts_obstack): New variable.
20027 (init_options_struct): Call gcc_init_obstack on opts_obstack.
20028 (finish_options): Use opts_concat instead of concat
20029 and XOBNEWVEC instead of XNEWVEC.
20030 * opts-common.c (generate_canonical_option, decode_cmdline_option,
20031 generate_option): Likewise.
20032 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
20033 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
20036 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
20037 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
20039 2013-02-26 Jakub Jelinek <jakub@redhat.com>
20041 PR middle-end/56461
20042 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
20044 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
20046 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
20047 (arm_block_move_unaligned_straight): Likewise.
20048 (arm_adjust_block_mem): Likewise.
20050 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
20053 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
20054 temp, cond and label.
20055 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
20058 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
20059 * config/c6x/c6x.h (dbx_register_map): Update declaration.
20062 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
20063 of prologue/epilogue functions.
20066 * config/tilegx/tilegx.c (tilegx_expand_prologue):
20067 Remove unused variable cfa_offset.
20068 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
20071 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
20072 type promotion to unsigned.
20075 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
20076 for HOST_WIDE_INT of 32 bit / same size as int.
20077 (arm_block_move_unaligned_straight): Likewise.
20078 (arm_adjust_block_mem): Likewise.
20081 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
20084 2013-02-26 Marek Polacek <polacek@redhat.com>
20086 PR tree-optimization/56426
20087 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
20089 2013-02-26 Richard Biener <rguenther@suse.de>
20092 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
20093 unused variable loops.
20095 2013-02-26 Jakub Jelinek <jakub@redhat.com>
20097 PR tree-optimization/56448
20098 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
20099 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
20100 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
20101 later operands of the references, or even first operand for
20102 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
20104 PR tree-optimization/56443
20105 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
20106 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
20107 to type_for_mode langhook.
20109 2013-02-25 Matt Turner <mattst88@gmail.com>
20111 * doc/invoke.texi: Document r4700.
20113 2013-02-25 Richard Biener <rguenther@suse.de>
20115 PR tree-optimization/56175
20116 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
20118 (simplify_bitwise_binary): ... here. Also guard the conversion
20119 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
20121 2013-02-25 Catherine Moore <clm@codesourcery.com>
20124 2013-02-24 Catherine Moore <clm@codesourcery.com>
20125 Maciej W. Rozycki <macro@codesourcery.com>
20126 Tom de Vries <tom@codesourcery.com>
20127 Nathan Sidwell <nathan@codesourcery.com>
20128 Iain Sandoe <iain@codesourcery.com>
20129 Nathan Froyd <froydnj@codesourcery.com>
20130 Chao-ying Fu <fu@mips.com>
20132 * doc/extend.texi: (micromips, nomicromips, nocompression):
20133 Document new function attributes.
20134 * doc/invoke.texi (minterlink-compressed, mmicromips,
20135 m14k, m14ke, m14kec): Document new options.
20136 (minterlink-mips16): Update documentation.
20137 * doc/md.texi (ZC, ZD): Document new constraints.
20138 * configure.ac (gcc_cv_as_micromips): Check if linker
20139 supports the .set micromips directive.
20140 * configure: Regenerate.
20141 * config.in: Regenerate.
20142 * config/mips/mips-tables.opt: Regenerate.
20143 * config/mips/micromips.md: New file.
20144 * constraints.md (ZC, AD): New constraints.
20145 * config/mips/predicates.md (movep_src_register): New predicate.
20146 (movep_src_operand): New predicate.
20147 (non_volatile_mem_operand): New predicate.
20148 * config/mips/mips.md (multimem): New type.
20149 (length): Differentiate between 17-bit and 18-bit branch offsets.
20150 (MOVEP1, MOVEP2): New mode iterator.
20151 (mov_<load>l): Use ZC constraint.
20152 (mov_<load>r): Likewise.
20153 (mov_<store>l): Likewise.
20154 (mov_<store>r): Likewise.
20155 (*branch_equality<mode>_inverted): Add microMIPS support.
20156 (*branch_equality<mode>): Likewise.
20157 (*jump_absolute): Likewise.
20158 (indirect_jump_<mode>): Likewise.
20159 (tablejump_<mode>): Likewise.
20160 (<optab>_internal): Likewise.
20161 (sibcall_internal): Likewise.
20162 (sibcall_value_internal): Likewise.
20163 (prefetch): Use constraint ZD.
20164 * config/mips/mips.opt (minterlink-compressed): New option.
20165 (minterlink-mips16): Now an alias for minterlink-compressed.
20166 (mmicromips): New option.
20167 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
20168 (compare_and_swap_12): Likewise.
20169 (sync_add<mode>): Likewise.
20170 (sync_<optab>_12): Likewise.
20171 (sync_old_<optab>_12): Likewise.
20172 (sync_new_<optab>_12): Likewise.
20173 (sync_nand_12): Likewise.
20174 (sync_old_nand_12): Likewise.
20175 (sync_new_nand_12): Likewise.
20176 (sync_sub<mode>): Likewise.
20177 (sync_old_add<mode>): Likewise.
20178 (sync_old_sub<mode>): Likewise.
20179 (sync_new_add<mode>): Likewise.
20180 (sync_new_sub<mode>): Likewise.
20181 (sync_<optab><mode>): Likewise.
20182 (sync_old_<optab><mode>): Likewise.
20183 (sync_new_<optab><mode>): Likewise.
20184 (sync_nand<mode>): Likewise.
20185 (sync_old_nand<mode>): Likewise.
20186 (sync_new_nand<mode>): Likewise.
20187 (sync_lock_test_and_set<mode>): Likewise.
20188 (test_and_set_12): Likewise.
20189 (atomic_compare_and_swap<mode>): Likewise.
20190 (atomic_exchange<mode>_llsc): Likewise.
20191 (atomic_fetch_add<mode>_llsc): Likewise.
20192 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
20193 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
20194 (umips_save_restore_pattern_p): Likewise.
20195 (umips_load_store_pair_p): Likewise.
20196 (umips_output_load_store_pair): Likewise.
20197 (umips_movep_target_p): Likewise.
20198 (umips_12bit_offset_address_p): Likewise.
20199 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
20200 (mips_base_mips16): Rename this...
20201 (mips_base_compression_flags): ...to this. Update all uses.
20202 (mips_attribute_table): Add micromips, nomicromips and nocompression.
20203 (mips_mips16_decl_p): Delete.
20204 (mips_nomips16_decl_p): Delete.
20205 (mips_get_compress_on_flags): New function.
20206 (mips_get_compress_off_flags): New function.
20207 (mips_get_compress_mode): New function.
20208 (mips_get_compress_on_name): New function.
20209 (mips_get_compress_off_name): New function.
20210 (mips_insert_attributes): Support multiple compression types.
20211 (mips_merge_decl_attributes): Likewise.
20212 (umips_12bit_offset_address_p): New function.
20213 (mips_start_function_definition): Emit .set micromips directive.
20214 (mips_call_may_need_jalx_p): New function.
20215 (mips_function_ok_for_sibcall): Add microMIPS support.
20216 (mips_print_operand_punctuation): Support short delay slots and
20218 (umips_swm_mask, umips_swm_encoding): New.
20219 (umips_build_save_restore): New function.
20220 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
20221 (was_mips16_p): Remove.
20222 (old_compression_mode): New.
20223 (mips_set_compression_mode): New function.
20224 (mips_set_current_function): Add microMIPS support.
20225 (mips_option_override): Likewise.
20226 (umips_save_restore_pattern_p): New function.
20227 (umips_output_save_restore): New function.
20228 (umips_load_store_pair_p_1): New function.
20229 (umips_load_store_pair_p): New function.
20230 (umips_output_load_store_pair_1): New function.
20231 (umips_output_load_store_pair): New function.
20232 (umips_movep_target_p) New function.
20233 (mips_prepare_pch_save): Add microMIPS support.
20234 * config/mips/mips.h (TARGET_COMPRESSION): New.
20235 (TARGET_CPU_CPP_BUILTINS): Update macro
20236 to use new compression flags and to support microMIPS.
20237 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
20238 (MIPS_ARCH_FLOAT_SPEC): Likewise.
20239 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
20240 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
20241 (ASM_SPEC): Support mmicromips and mno-micromips.
20242 (M16STORE_REG_P): New macro.
20243 (MIPS_CALL): Support TARGET_MICROMIPS.
20244 (MICROMIPS_J): New macro.
20245 (mips_base_mips16): Rename this...
20246 (mips_base_compression_flags): ...to this.
20247 (UMIPS_12BIT_OFFSET_P): New macro.
20248 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
20249 (MULTILIB_DIRNAMES): Likewise.
20251 2013-02-25 Tom de Vries <tom@codesourcery.com>
20253 PR rtl-optimization/56131
20254 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
20255 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
20256 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
20258 2013-02-25 Tobias Burnus <burnus@net-b.de>
20260 * doc/invoke.texi (-fsanitize=): Move from optimization
20261 to debugging options.
20263 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
20265 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
20267 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
20268 Alexander Monakov <amonakov@ispras.ru>
20270 PR middle-end/56077
20271 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
20272 flush pending lists also on non-jumps. Adjust comment.
20274 2013-02-24 Catherine Moore <clm@codesourcery.com>
20275 Maciej W. Rozycki <macro@codesourcery.com>
20276 Tom de Vries <tom@codesourcery.com>
20277 Nathan Sidwell <nathan@codesourcery.com>
20278 Iain Sandoe <iain@codesourcery.com>
20279 Nathan Froyd <froydnj@codesourcery.com>
20280 Chao-ying Fu <fu@mips.com>
20282 * doc/extend.texi: (micromips, nomicromips, nocompression):
20283 Document new function attributes.
20284 * doc/invoke.texi (minterlink-compressed, mmicromips,
20285 m14k, m14ke, m14kec): Document new options.
20286 (minterlink-mips16): Update documentation.
20287 * doc/md.texi (ZC, ZD): Document new constraints.
20288 * configure.ac (gcc_cv_as_micromips): Check if linker
20289 supports the .set micromips directive.
20290 * configure: Regenerate.
20291 * config.in: Regenerate.
20292 * config/mips/mips-tables.opt: Regenerate.
20293 * config/mips/micromips.md: New file.
20294 * constraints.md (ZC, AD): New constraints.
20295 * config/mips/predicates.md (movep_src_register): New predicate.
20296 (movep_src_operand): New predicate.
20297 (non_volatile_mem_operand): New predicate.
20298 * config/mips/mips.md (multimem): New type.
20299 (length): Differentiate between 17-bit and 18-bit branch offsets.
20300 (MOVEP1, MOVEP2): New mode iterator.
20301 (mov_<load>l): Use ZC constraint.
20302 (mov_<load>r): Likewise.
20303 (mov_<store>l): Likewise.
20304 (mov_<store>r): Likewise.
20305 (*branch_equality<mode>_inverted): Add microMIPS support.
20306 (*branch_equality<mode>): Likewise.
20307 (*jump_absolute): Likewise.
20308 (indirect_jump_<mode>): Likewise.
20309 (tablejump_<mode>): Likewise.
20310 (<optab>_internal): Likewise.
20311 (sibcall_internal): Likewise.
20312 (sibcall_value_internal): Likewise.
20313 (prefetch): Use constraint ZD.
20314 * config/mips/mips.opt (minterlink-compressed): New option.
20315 (minterlink-mips16): Now an alias for minterlink-compressed.
20316 (mmicromips): New option.
20317 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
20318 (compare_and_swap_12): Likewise.
20319 (sync_add<mode>): Likewise.
20320 (sync_<optab>_12): Likewise.
20321 (sync_old_<optab>_12): Likewise.
20322 (sync_new_<optab>_12): Likewise.
20323 (sync_nand_12): Likewise.
20324 (sync_old_nand_12): Likewise.
20325 (sync_new_nand_12): Likewise.
20326 (sync_sub<mode>): Likewise.
20327 (sync_old_add<mode>): Likewise.
20328 (sync_old_sub<mode>): Likewise.
20329 (sync_new_add<mode>): Likewise.
20330 (sync_new_sub<mode>): Likewise.
20331 (sync_<optab><mode>): Likewise.
20332 (sync_old_<optab><mode>): Likewise.
20333 (sync_new_<optab><mode>): Likewise.
20334 (sync_nand<mode>): Likewise.
20335 (sync_old_nand<mode>): Likewise.
20336 (sync_new_nand<mode>): Likewise.
20337 (sync_lock_test_and_set<mode>): Likewise.
20338 (test_and_set_12): Likewise.
20339 (atomic_compare_and_swap<mode>): Likewise.
20340 (atomic_exchange<mode>_llsc): Likewise.
20341 (atomic_fetch_add<mode>_llsc): Likewise.
20342 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
20343 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
20344 (umips_save_restore_pattern_p): Likewise.
20345 (umips_load_store_pair_p): Likewise.
20346 (umips_output_load_store_pair): Likewise.
20347 (umips_movep_target_p): Likewise.
20348 (umips_12bit_offset_address_p): Likewise.
20349 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
20350 (mips_base_mips16): Rename this...
20351 (mips_base_compression_flags): ...to this. Update all uses.
20352 (mips_attribute_table): Add micromips, nomicromips and nocompression.
20353 (mips_mips16_decl_p): Delete.
20354 (mips_nomips16_decl_p): Delete.
20355 (mips_get_compress_on_flags): New function.
20356 (mips_get_compress_off_flags): New function.
20357 (mips_get_compress_mode): New function.
20358 (mips_get_compress_on_name): New function.
20359 (mips_get_compress_off_name): New function.
20360 (mips_insert_attributes): Support multiple compression types.
20361 (mips_merge_decl_attributes): Likewise.
20362 (umips_12bit_offset_address_p): New function.
20363 (mips_start_function_definition): Emit .set micromips directive.
20364 (mips_call_may_need_jalx_p): New function.
20365 (mips_function_ok_for_sibcall): Add microMIPS support.
20366 (mips_print_operand_punctuation): Support short delay slots and
20368 (umips_swm_mask, umips_swm_encoding): New.
20369 (umips_build_save_restore): New function.
20370 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
20371 (was_mips16_p): Remove.
20372 (old_compression_mode): New.
20373 (mips_set_compression_mode): New function.
20374 (mips_set_current_function): Add microMIPS support.
20375 (mips_option_override): Likewise.
20376 (umips_save_restore_pattern_p): New function.
20377 (umips_output_save_restore): New function.
20378 (umips_load_store_pair_p_1): New function.
20379 (umips_load_store_pair_p): New function.
20380 (umips_output_load_store_pair_1): New function.
20381 (umips_output_load_store_pair): New function.
20382 (umips_movep_target_p) New function.
20383 (mips_prepare_pch_save): Add microMIPS support.
20384 * config/mips/mips.h (TARGET_COMPRESSION): New.
20385 (TARGET_CPU_CPP_BUILTINS): Update macro
20386 to use new compression flags and to support microMIPS.
20387 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
20388 (MIPS_ARCH_FLOAT_SPEC): Likewise.
20389 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
20390 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
20391 (ASM_SPEC): Support mmicromips and mno-micromips.
20392 (M16STORE_REG_P): New macro.
20393 (MIPS_CALL): Support TARGET_MICROMIPS.
20394 (MICROMIPS_J): New macro.
20395 (mips_base_mips16): Rename this...
20396 (mips_base_compression_flags): ...to this.
20397 (UMIPS_12BIT_OFFSET_P): New macro.
20398 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
20399 (MULTILIB_DIRNAMES): Likewise.
20401 2013-02-24 Jakub Jelinek <jakub@redhat.com>
20404 * target-globals.c (save_target_globals): For init_reg_sets and
20405 target_reinit remporarily set this_fn_optabs to this_target_optabs.
20407 2013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
20409 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
20410 * config/aarch64/t-aarch64
20411 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
20413 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
20415 PR inline-asm/56148
20416 * lra-constraints.c (process_alt_operands): Reload operand
20417 conflicting with earlier clobber only if no more other conflicting
20420 2013-02-22 Jakub Jelinek <jakub@redhat.com>
20423 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
20424 if not linking a shared library.
20426 2013-02-22 Seth LaForge <sethml@google.com>
20428 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
20430 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
20432 * config/arm/arm.md (split for extendsidi): Update condition.
20433 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
20434 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
20435 (qhs_zextenddi_cstr): Likewise.
20437 2013-02-21 Jakub Jelinek <jakub@redhat.com>
20439 PR middle-end/56420
20440 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
20441 avoid signed wrapping.
20442 (expand_mult): Handle properly multiplication by
20443 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
20444 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
20445 in the compiler if coeff is HOST_WIDE_INT_MIN.
20446 (expand_divmod): Don't make ext_op1 static, change it's type to
20447 uhwi. Avoid undefined behavior in -INTVAL (op1).
20449 PR rtl-optimization/50339
20450 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
20452 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
20453 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
20454 into splitting_ashiftrt field.
20455 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
20457 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
20460 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
20462 PR middle-end/56108
20463 * trans-mem.c (execute_tm_mark): Do not expand transactions that
20464 are sure to go irrevocable.
20466 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
20468 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
20469 scalars are valid operands.
20471 2013-02-21 Martin Jambor <mjambor@suse.cz>
20473 PR tree-optimization/56310
20474 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
20475 only matching indices and non-negative final offsets.
20476 (intersect_aggregates_with_edge): Pass src_idx to
20477 agg_replacements_to_vector. Pass src_idx insstead of index to
20478 intersect_with_agg_replacements.
20480 2013-02-21 Martin Jambor <mjambor@suse.cz>
20482 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
20483 instead of hard-wired defaults.
20485 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
20487 * doc/invoke.texi (MIPS Options): Update documentation of the
20488 floating-point multiply-accumulate instruction restrictions.
20490 2013-02-21 Kostya Serebryany <kcc@google.com>
20492 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
20493 asan_shadow_offset on x86_64 linux.
20495 2013-02-21 Richard Biener <rguenther@suse.de>
20497 PR tree-optimization/56415
20499 2013-02-11 Richard Biener <rguenther@suse.de>
20501 PR tree-optimization/56273
20502 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
20505 2013-02-21 Jakub Jelinek <jakub@redhat.com>
20508 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
20511 PR inline-asm/56405
20512 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
20513 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
20515 2013-02-20 Jan Hubicka <jh@suse.cz>
20517 PR tree-optimization/56265
20518 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
20519 when target is referenced for first time.
20521 2013-02-20 Richard Biener <rguenther@suse.de>
20523 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
20524 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
20525 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
20526 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
20527 not return anything.
20528 (rename_ssa_copies): Do not remove unused locals.
20529 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
20530 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
20531 * passes.c (execute_function_todo): Do not schedule unused locals
20532 removal if cleanup_tree_cfg did something.
20533 * tree-ssa-live.c (remove_unused_locals): Dump statistics
20534 about the number of removed locals.
20536 2013-02-20 Richard Biener <rguenther@suse.de>
20538 PR tree-optimization/56398
20539 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
20541 2013-02-20 Martin Jambor <mjambor@suse.cz>
20543 PR tree-optimization/55334
20544 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
20545 restricted pointers to arrays.
20547 2013-02-20 Richard Biener <rguenther@suse.de>
20548 Jakub Jelinek <jakub@redhat.com>
20550 PR tree-optimization/56396
20551 * tree-ssa-ccp.c (n_const_val): New static variable.
20552 (get_value): Return NULL for SSA names we don't have a lattice
20554 (ccp_initialize): Initialize n_const_val.
20555 * tree-ssa-copy.c (n_copy_of): New static variable.
20556 (init_copy_prop): Initialize n_copy_of.
20557 (get_value): Return NULL_TREE for SSA names we don't have a
20560 2013-02-20 Martin Jambor <mjambor@suse.cz>
20562 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
20564 2013-02-20 Richard Biener <rguenther@suse.de>
20566 * genpreds.c (write_lookup_constraint): Do not compare first
20567 letter of the constraint again.
20569 2013-02-20 Richard Biener <rguenther@suse.de>
20571 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
20573 (get_use_iv_cost): Terminate hashtable walk when coming across
20576 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
20578 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
20579 reassociation for avx2 targets.
20581 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20583 * config/microblaze/microblaze.c: microblaze_has_clz = 0
20584 Add version check for v8.10.a to enable microblaze_has_clz
20585 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
20586 version and TARGET_PATTERN_COMPARE check
20587 * config/microblaze/microblaze.md: New clzsi2 instruction
20589 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20591 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
20592 function before branching.
20594 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
20596 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
20598 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
20600 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
20602 PR middle-end/55889
20603 * sel-sched.c: Include ira.h.
20604 (implicit_clobber_conflict_p): New function.
20605 (moveup_expr): Use it.
20606 * Makefile.in (sel-sched.o): Depend on ira.h.
20608 2013-02-19 Richard Biener <rguenther@suse.de>
20610 PR tree-optimization/56384
20611 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
20612 (vn_hash_type): Split out from ...
20613 (vn_hash_constant_with_type): ... here.
20614 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
20615 (vn_phi_eq): Compare types from vn_phi_s structure.
20616 (vn_phi_lookup): Populate vn_phi_s type.
20617 (vn_phi_insert): Likewise.
20619 2013-02-19 Jakub Jelinek <jakub@redhat.com>
20621 PR tree-optimization/56350
20622 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
20623 if haven't found reduction or nested cycle operand, rather than
20624 asserting we must find it.
20626 PR tree-optimization/56381
20627 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
20630 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
20631 Jakub Jelinek <jakub@redhat.com>
20634 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
20635 (swap_optab_enable): Same.
20636 (init_all_optabs): Use argument instead of global.
20637 * tree.h (struct tree_optimization_option): New field target_optabs.
20638 * expr.h (init_all_optabs): Add argument to prototype.
20639 (TREE_OPTIMIZATION_OPTABS): New.
20640 (save_optabs_if_changed): Protoize.
20641 * optabs.h: Declare this_fn_optabs.
20642 * optabs.c (save_optabs_if_changed): New.
20643 Declare this_fn_optabs.
20644 (init_optabs): Add argument to init_all_optabs() call.
20645 * function.c (invoke_set_current_function_hook): Handle per
20647 * function.h (struct function): New field optabs.
20648 * config/mips/mips.c (mips_set_mips16_mode): Handle when
20649 optimization_current_node has changed.
20650 * target-globals.h (save_target_globals_default_opts): Protoize.
20651 * target-globals.c (save_target_globals_default_opts): New.
20653 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20656 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
20657 registers %fr12 and %fr12R as call used.
20660 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
20661 and HImode, require all displacements to be an integer multiple of
20663 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
20664 only allow QImode and HImode when reload is in progress and strict is
20665 true. Likewise for symbolic addresses. Use base14_operand to check
20666 displacements in REG+BASE addresses.
20668 2013-02-18 Richard Biener <rguenther@suse.de>
20670 PR tree-optimization/56366
20671 * tree-vect-loop.c (get_initial_def_for_induction): Properly
20672 handle sign-conversion of outer-loop initial induction value.
20674 2013-02-18 Richard Biener <rguenther@suse.de>
20676 PR middle-end/56349
20677 * cfghooks.c (merge_blocks): If we merge a latch into another
20678 block adjust references to it.
20679 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
20680 (verify_loop_structure): Verify that a recorded latch is in fact
20683 2013-02-18 Richard Biener <rguenther@suse.de>
20685 PR tree-optimization/56321
20686 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
20687 order SSA name release and virtual operand unlinking.
20689 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20691 * config/microblaze/microblaze.md (save_stack_block): Define.
20692 (restore_stack_block): Likewise.
20694 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20696 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
20697 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
20698 * config/microblaze/microblaze.c (microblaze_option_override):
20699 Bail out early for PIC modes when target does not support PIC.
20701 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20703 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
20704 Replace with a microblaze version.
20705 (microblaze_trampoline_init): Adapt for microblaze.
20706 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
20709 2013-02-16 Jakub Jelinek <jakub@redhat.com>
20710 Dodji Seketeli <dodji@redhat.com>
20713 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
20714 (instrument_mem_region_access): Do not forget to always put
20715 instrumentation of the of 'base' and 'base + len' in a "if (len !=
20716 0) statement, even for cases where either 'base' or 'base + len'
20717 are not instrumented -- because they have been previously
20718 instrumented. Simplify the logic by putting all the statements
20719 instrument 'base + len' inside a sequence, and then insert that
20720 sequence right before the current insertion point. Then, to
20721 instrument 'base + len', just get an iterator on that statement.
20722 And do not forget to update the pointer to iterator the function
20723 received as argument.
20725 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
20727 PR rtl-optimization/56348
20728 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
20730 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
20732 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
20733 (clean_graph_dump_file): Pass base to start_graph_dump.
20735 2013-02-14 Richard Henderson <rth@redhat.com>
20738 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
20740 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
20742 * collect2-aix.h: Define F_LOADONLY.
20744 2013-02-14 Richard Biener <rguenther@suse.de>
20747 * varasm.c (output_constant_def_1): Get the decl representing
20748 the constant as argument.
20749 (output_constant_def): Wrap output_constant_def_1.
20750 (make_decl_rtl): Use output_constant_def_1 with the decl
20751 representing the constant.
20752 (build_constant_desc): Optionally re-use a decl already
20753 representing the constant.
20754 (tree_output_constant_def): Adjust.
20756 2013-02-14 Dodji Seketeli <dodji@redhat.com>
20759 * asan.c (instrument_builtin_call): Really put the length of the
20760 second source argument into src1_len.
20762 2013-02-13 Jakub Jelinek <jakub@redhat.com>
20764 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
20765 argument. If it is false, don't create edge from then_bb to
20767 (insert_if_then_before_iter): Pass true to it.
20768 (build_check_stmt): Pass false to it.
20769 (transform_statements): Flush hash table only on extended basic
20770 block boundaries, rather than at the beginning of every bb.
20771 Don't flush hash table on nonfreeing_call_p calls.
20772 * tree-flow.h (nonfreeing_call_p): New prototype.
20773 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
20775 2013-02-13 David S. Miller <davem@davemloft.net>
20777 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
20779 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
20782 * ira.c (max_regno_before_ira): Move from ...
20784 (fix_reg_equiv_init): Use max_regno_before_ira instead of
20787 2013-02-13 Jakub Jelinek <jakub@redhat.com>
20789 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
20791 2013-02-13 Richard Biener <rguenther@suse.de>
20794 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
20795 globals in MEM_REFs.
20797 2013-02-13 Richard Biener <rguenther@suse.de>
20799 * loop-init.c (loop_optimizer_init): Clear loop state when
20800 re-initializing preserved loops.
20801 * loop-unswitch.c (unswitch_single_loop): Return whether
20802 we unswitched the loop. Do not verify loop state here.
20803 (unswitch_loops): When we unswitched a loop discover new loops.
20805 2013-02-13 Kostya Serebryany <kcc@google.com>
20807 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
20809 * sanitizer.def: Rename __asan_init to __asan_init_v1.
20811 2013-02-12 Dodji Seketeli <dodji@redhat.com>
20813 Avoid instrumenting duplicated memory access in the same basic block
20814 * Makefile.in (asan.o): Add new dependency on hash-table.h
20815 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
20816 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
20817 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
20818 (free_mem_ref_resources, has_mem_ref_been_instrumented)
20819 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
20820 (get_mem_ref_of_assignment): New functions.
20821 (get_mem_refs_of_builtin_call): Extract from
20822 instrument_builtin_call and tweak a little bit to make it fit with
20824 (instrument_builtin_call): Use the new
20825 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
20826 of is_gimple_builtin_call.
20827 (instrument_derefs, instrument_mem_region_access): Insert the
20828 instrumented memory reference into the hash table.
20829 (maybe_instrument_assignment): Renamed instrument_assignment into
20830 this, and change it to advance the iterator when instrumentation
20831 actually happened and return true in that case. This makes it
20832 homogeneous with maybe_instrument_assignment, and thus give a
20833 chance to callers to be more 'regular'.
20834 (transform_statements): Clear the memory reference hash table
20835 whenever we enter a new BB, when we cross a function call, or when
20836 we are done transforming statements. Use
20837 maybe_instrument_assignment instead of instrumentation. No more
20838 need to special case maybe_instrument_assignment and advance the
20839 iterator after calling it; it's now handled just like
20840 maybe_instrument_call. Update comment.
20842 2013-02-13 Richard Biener <rguenther@suse.de>
20844 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
20845 Fix loop discovery code.
20847 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
20849 PR inline-asm/56148
20850 * lra-constraints.c (process_alt_operands): Match early clobber
20851 operand with itself. Check conflicts with earlyclobber only if
20852 the operand is not reloaded. Prefer to reload conflicting operand
20853 if earlyclobber and matching operands are the same.
20855 2013-02-12 Richard Biener <rguenther@suse.de>
20858 * lto-streamer-out.c (write_symbol): Do not output symbols
20859 for hard register variables.
20861 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
20864 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
20865 (umulsidi3_insn, mulsidi3_insn): New insns.
20867 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
20869 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
20870 (struct tune_params): Add vec_costs field.
20871 * config/arm/arm.c (arm_builtin_vectorization_cost)
20872 (arm_add_stmt_cost): New functions.
20873 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
20874 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
20875 (arm_default_vec_cost): New struct of type cpu_vec_costs.
20876 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
20877 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
20878 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
20879 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
20881 2013-02-12 Richard Biener <rguenther@suse.de>
20884 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
20885 decls again if possible.
20887 2013-02-12 Richard Biener <rguenther@suse.de>
20889 PR middle-end/56288
20890 * tree-ssa.c (verify_ssa_name): Fix check, move
20891 SSA_NAME_IN_FREE_LIST check up.
20893 2013-02-12 Jakub Jelinek <jakub@redhat.com>
20894 Steven Bosscher <steven@gcc.gnu.org>
20896 PR rtl-optimization/56151
20897 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
20898 equal to op0 or op1, and last_insn pattern is CODE operation
20899 with MEM dest and one of the operands matches that MEM.
20901 2013-02-11 Sriraman Tallam <tmsriramgoogle.com>
20903 * doc/extend.texi: Document Function Multiversioning and "default"
20904 parameter string to target attribute.
20905 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
20906 target attribute parameter is "default".
20907 (ix86_compare_version_priority): Remove checks for target attribute.
20908 (ix86_mangle_function_version_assembler_name): Change error to sorry.
20909 Remove check for target attribute equal to NULL. Add assert.
20910 (ix86_generate_version_dispatcher_body): Change error to sorry.
20912 2013-02-11 Iain Sandoe <iain@codesourcery.com>
20913 Jack Howarth <howarth@bromo.med.uc.edu>
20914 Patrick Marlier <patrick.marlier@gmail.com>
20917 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
20918 define ENDFILE_SPEC as TM_DESTRUCTOR.
20919 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
20921 2013-02-11 Alexander Potapenko <glider@google.com>
20922 Jack Howarth <howarth@bromo.med.uc.edu>
20923 Jakub Jelinek <jakub@redhat.com>
20926 * config/darwin.c (cdtor_record): Rename ctor_record.
20927 (sort_cdtor_records): Rename sort_ctor_records.
20928 (finalize_dtors): New routine to sort destructors by
20929 priority before use in assemble_integer.
20930 (machopic_asm_out_destructor): Use finalize_dtors if needed.
20932 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
20934 PR rtl-optimization/56275
20935 * simplify-rtx.c (avoid_constant_pool_reference): Check that
20936 offset is non-negative and less than cmode size before
20937 calling simplify_subreg.
20939 2013-02-11 Richard Biener <rguenther@suse.de>
20941 PR tree-optimization/56264
20942 * cfgloop.h (fix_loop_structure): Adjust prototype.
20943 * loop-init.c (fix_loop_structure): Return the number of
20944 newly discovered loops.
20945 * tree-cfgcleanup.c (repair_loop_structures): When new loops
20946 are discovered, do a full loop-closed SSA rewrite.
20948 2013-02-11 Richard Biener <rguenther@suse.de>
20950 PR tree-optimization/56273
20951 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
20953 (check_array_ref): Fix missing newline in dumps.
20954 (search_for_addr_array): Likewise.
20956 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
20958 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
20960 2013-02-09 Jakub Jelinek <jakub@redhat.com>
20963 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
20965 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
20967 PR rtl-optimization/56246
20968 * lra-constraints.c (simplify_operand_subreg): Try to reuse
20970 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
20971 constraints are satisfied.
20973 2013-02-08 Jeff Law <law@redhat.com>
20976 * emit-rtl.c (reg_is_parm_p): New function.
20977 * regs.h (reg_is_parm_p): New prototype.
20978 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
20979 callee-clobbered registers.
20981 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
20984 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
20985 If there is no implicit builtin declaration, just return NULL.
20987 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
20989 * config/i386/sse.md (FMAMODEM): New mode iterator.
20990 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
20991 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
20993 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
20995 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
20996 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
20997 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
20999 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
21001 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
21002 (microblaze*-*-elf): Likewise.
21003 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
21005 * config/microblaze/microblaze-c.c: Add builtin defines for
21006 _LITTLE_ENDIAN and _BIG_ENDIAN.
21007 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
21008 add to TARGET_DEFAULT flags.
21009 Expand ASM_SPEC and LINK_SPEC.
21010 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
21011 * config/microblaze/microblaze.md: Update extendsidi2 and
21012 movdi_internal instructions to use low-order / high-order reg
21014 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
21015 options and inversemask / mask of LITTLE_ENDIAN.
21016 * config/microblaze/t-microblaze: Expand multilib options to
21017 include mlittle-endian (le) and update exceptions patterns.
21019 2013-02-08 Jakub Jelinek <jakub@redhat.com>
21021 PR rtl-optimization/56195
21022 * lra-constraints.c (get_reload_reg): Don't reuse regs
21023 if they have smaller mode than requested, if they have
21024 wider mode than requested, try to return a SUBREG.
21026 PR tree-optimization/56250
21027 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
21028 if type is unsigned and code isn't MULT_EXPR.
21030 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
21032 PR tree-optimization/56064
21033 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
21034 bits according to mode.
21035 * fixed-value.h (fixed_from_double_int)
21036 (const_fixed_from_double_int): Adjust comments.
21038 2013-02-08 Richard Biener <rguenther@suse.de>
21041 * lto-streamer.h (struct data_in): Remove current_file, current_line
21042 and current_col members.
21043 * lto-streamer-out.c (lto_output_location): Stream changed bits
21044 en-block for efficiency.
21045 * lto-streamer-in.c (clear_line_info): Remove.
21046 (lto_input_location): Cache current file, line and column
21047 globally via local statics. Read changed bits en-block.
21048 (input_function): Do not call clear_line_info.
21049 (lto_read_body): Likewise.
21050 (lto_input_toplevel_asms): Likewise.
21052 2013-02-08 Michael Matz <matz@suse.de>
21054 PR tree-optimization/52448
21055 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
21056 (nt_call_phase): New static.
21057 (add_or_mark_expr): Only mark accesses with newer phase than any
21059 (nonfreeing_call_p): New.
21060 (nt_init_block): Update nt_call_phase, mark blocks as visited.
21061 (nt_fini_block): Keep blocks marked as visited.
21062 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
21064 2013-02-08 Richard Biener <rguenther@suse.de>
21066 * ira.c (ira): Free broken dominator information.
21068 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
21070 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
21072 2013-02-08 Marek Polacek <polacek@redhat.com>
21074 * cfgloop.c (verify_loop_structure): Add more checking of headers.
21076 2013-02-08 Richard Biener <rguenther@suse.de>
21078 PR middle-end/56181
21079 * cfgloop.h (flow_loops_find): Adjust.
21080 (bb_loop_header_p): Declare.
21081 * cfgloop.c (bb_loop_header_p): New function split out from ...
21082 (flow_loops_find): ... here. Adjust function signature,
21083 support incremental loop structure update.
21084 (verify_loop_structure): Cleanup. Verify a loop is a loop.
21085 * cfgloopmanip.c (fix_loop_structure): Move ...
21086 * loop-init.c (fix_loop_structure): ... here.
21087 (apply_loop_flags): Split out from ...
21088 (loop_optimizer_init): ... here.
21089 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
21090 in incremental mode, only remove dead loops here.
21092 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
21095 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
21096 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
21097 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
21098 (*round<mode>3.libgcc): New insns for fixed-modes.
21099 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
21100 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
21101 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
21102 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
21103 implementations. Define to __builtin_avr_absFX,
21104 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
21105 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
21106 __builtin_avr_countlsFX, respectively.
21107 * config/avr/avr-c.c (target.h): Include it.
21108 (enum avr_builtin_id): New enum.
21109 (avr_resolve_overloaded_builtin): New static function.
21110 (avr_register_target_pragmas): Use it to set
21111 targetm.resolve_overloaded_builtin.
21112 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
21113 tree nodes used by DEF_BUILTIN.
21114 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
21115 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
21116 <AVR_BUILTIN_xxBITS>: Same.
21118 2013-02-08 Richard Biener <rguenther@suse.de>
21120 * cfgloop.c (verify_loop_structure): Properly handle
21121 a loop exiting to another loop header.
21122 * ira-int.h (ira_loops): Remove.
21123 * ira.c (ira_loops): Remove.
21124 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
21125 (do_reload): Use loop_optimizer_finalize.
21126 * ira-build.c (create_loop_tree_nodes): Use get_loops and
21127 number_of_loops to access the loop tree.
21128 (more_one_region_p): Likewise.
21129 (finish_loop_tree_nodes): Likewise.
21130 (rebuild_regno_allocno_maps): Likewise.
21131 (mark_loops_for_removal): Likewise.
21132 (mark_all_loops_for_removal): Likewise.
21133 (remove_unnecessary_regions): Likewise.
21134 (ira_build): Likewise.
21135 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
21137 2013-02-08 Richard Biener <rguenther@suse.de>
21139 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
21140 * ipa-pure-const.c (analyze_function): Avoid calling
21141 mark_irreducible_loops twice.
21142 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
21144 2013-02-07 David S. Miller <davem@davemloft.net>
21146 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
21148 * var-tracking.c (vt_add_function_parameter): Test the presence of
21149 HAVE_window_save properly and do not remap argument registers when
21150 we have a leaf function.
21152 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
21155 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
21157 * config/i386/i386.c (ix86_print_operand): Ditto.
21159 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
21161 * lra-constraints.c (process_alt_operands): Fix recently added comment.
21163 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
21165 PR rtl-optimization/56225
21166 * lra-constraints.c (process_alt_operands): Check that reload hard
21167 reg can hold value for strict_low_part.
21169 2013-02-07 Jakub Jelinek <jakub@redhat.com>
21172 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
21173 dwarf2out_end_function.
21174 (in_first_function_p, maybe_at_text_label_p,
21175 first_loclabel_num_not_at_text_label): New variables.
21176 (dwarf2out_var_location): In the first function find out
21177 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
21178 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
21181 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
21183 PR rtl-optimization/56178
21184 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
21185 SUBREG of a register. Tidy up related block of code.
21186 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
21187 note if the source is a register or a SUBREG of a register.
21189 2013-02-07 Jakub Jelinek <jakub@redhat.com>
21192 * config/rs6000/rs6000.md (ptrm): New mode attr.
21193 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
21194 call_value_indirect_aix<pttrsize>,
21195 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
21198 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
21200 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
21201 if -bnortl. Convert to strcmp and strncmp.
21203 2013-02-07 Alan Modra <amodra@gmail.com>
21206 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
21207 addresses won't wrap when offsetting.
21208 (rs6000_secondary_reload): Provide secondary reloads needed for
21209 wrapping LO_SUM addresses.
21211 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
21213 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
21214 MACH, just __MACH__.
21216 2013-02-06 Richard Biener <rguenther@suse.de>
21218 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
21219 instead of calling fix_loop_structure.
21221 2013-02-06 Jakub Jelinek <jakub@redhat.com>
21223 PR middle-end/56217
21224 * omp-low.c (use_pointer_for_field): Return false if
21225 lower_send_shared_vars doesn't generate any copy-out code.
21227 2013-02-06 Tom de Vries <tom@codesourcery.com>
21229 PR rtl-optimization/56131
21230 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
21231 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
21232 of the label is NULL. Add comment.
21234 2013-02-05 Jakub Jelinek <jakub@redhat.com>
21236 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
21239 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
21240 (STATIC_LIBTSAN_LIBS): Likewise.
21241 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
21242 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
21243 is defined, don't add anything else beyond that.
21244 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
21245 (LINK_COMMAND_SPEC): Use them.
21247 PR tree-optimization/56205
21248 * tree-stdarg.c (check_all_va_list_escapes): Return true if
21249 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
21250 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
21252 2013-02-05 Richard Biener <rguenther@suse.de>
21254 PR tree-optimization/53342
21255 PR tree-optimization/53185
21256 * tree-vectorizer.h (vect_check_strided_load): Remove.
21257 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
21258 not disallow peeling for vectorized strided loads.
21259 (vect_check_strided_load): Make static and simplify.
21260 (vect_analyze_data_refs): Adjust.
21261 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
21262 correctly when vectorizing strided loads.
21264 2013-02-05 Richard Biener <rguenther@suse.de>
21266 * doc/install.texi: Refer to ISL, not PPL.
21268 2013-02-05 Jan Hubicka <jh@suse.cz>
21270 PR tree-optimization/55789
21271 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
21273 2013-02-05 Jan Hubicka <jh@suse.cz>
21275 PR tree-optimization/55789
21276 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
21277 the dead call anyway.
21279 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
21282 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
21284 2013-02-04 Alexander Potapenko <glider@google.com>
21285 Jack Howarth <howarth@bromo.med.uc.edu>
21286 Jakub Jelinek <jakub@redhat.com>
21289 * config/darwin.c (sort_ctor_records): Stabilized qsort
21290 on constructor priority by using original position.
21291 (finalize_ctors): New routine to sort constructors by
21292 priority before use in assemble_integer.
21293 (machopic_asm_out_constructor): Use finalize_ctors if needed.
21295 2013-02-04 Jakub Jelinek <jakub@redhat.com>
21298 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
21299 about visibility on artificial decls.
21300 * config/sol2.c (solaris_assemble_visibility): Likewise.
21302 2013-02-04 Kai Tietz <ktietz@redhat.com>
21305 * config/i386/i386.c (function_value_ms_64): Add additional valtype
21306 argument and improve checking of return-argument types for 16-byte
21308 (ix86_function_value_1): Add additional valtype argument on call
21309 of function_value_64.
21310 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
21311 handling infunction_value_64 function.
21313 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
21315 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
21317 2013-02-04 Richard Biener <rguenther@suse.de>
21319 PR tree-optimization/56188
21320 * tree-ssa-structalias.c (label_visit): Consider case with
21321 initially non-empty points-to set.
21322 (perform_var_substitution): Dump node mapping and clean up.
21324 2013-02-04 Richard Guenther <rguenther@suse.de>
21327 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
21328 node prevail as last resort.
21329 (lto_symtab_merge_decls): Remove guard on LTRANS here.
21330 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
21332 2013-02-04 Richard Biener <rguenther@suse.de>
21334 PR tree-optimization/56113
21335 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
21337 (equiv_class_lookup_or_add): ... this.
21338 (label_visit): Adjust and fix error in previous patch.
21339 (perform_var_substitution): Adjust.
21341 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
21343 * config/sh/divtab.c: Fix formatting and comments throughout the file.
21344 * config/sh/sh4-300.md: Likewise.
21345 * config/sh/sh4a.md: Likewise.
21346 * config/sh/constraints.md: Likewise.
21347 * config/sh/sh.md: Likewise.
21348 * config/sh/netbsd-elf.h: Likewise.
21349 * config/sh/predicates.md: Likewise.
21350 * config/sh/sh-protos.h: Likewise.
21351 * config/sh/ushmedia.h: Likewise.
21352 * config/sh/linux.h: Likewise.
21353 * config/sh/sh.c: Likewise.
21354 * config/sh/superh.h: Likewise.
21355 * config/sh/elf.h: Likewise.
21356 * config/sh/sh4.md: Likewise.
21357 * config/sh/sh.h: Likewise.
21359 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21361 * config/pa/constraints.md: Adjust unused letters. Change "T"
21362 constraint to match_test floating_point_store_memory_operand().
21363 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
21364 (base14_operand): New.
21365 (floating_point_store_memory_operand): New.
21366 (integer_store_memory_operand): Revise to use base14_operand and
21367 reg_plus_base_memory_operand.
21368 (move_dest_operand): Allow symbolic_memory_operands.
21369 (symbolic_memory_operand): Check for LO_SOM.
21370 (symbolic_operand): Change default case to break.
21371 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
21372 CONST_DOUBLE values to be reloaded by putting them into memory when
21373 the destination is a floating point register.
21374 (movdf): Remove code to handle CONST_DOUBLE.
21376 (reload_indf_r1): New.
21377 (reload_insf_r1): New.
21378 Consistently use "Q" and "T" constraints with integer and floating
21379 point move instructions, respectively.
21380 (movdi): Remove FAIL.
21381 Change predicate for source operand unamed DImode move from
21382 general_operand to move_src_operand.
21383 (umulsidi3): Change predicate for destination operand to
21385 Likewise for similar unamed patterns.
21386 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
21387 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
21388 (hppa_legitimize_address): Simplify mask calculation.
21389 (pa_emit_move_sequence): Revised handling of secondary reloads from
21390 REG+D addresses for floating point loads and stores. Directly handle
21391 loading CONST0_RTX (mode) to a floating point register.
21392 (pa_secondary_reload): Handle reloading DF and SFmode constant values
21393 to floating point registers. Don't restrict secondary reloads to
21394 floating point registers to integer modes. Revise some comments and
21396 (TARGET_LEGITIMATE_ADDRESS_P): Define.
21397 (pa_legitimate_address_p): New.
21398 (pa_legitimize_reload_address): New.
21399 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
21400 (STRICT_REG_OK_FOR_BASE_P): New.
21401 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
21402 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
21404 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
21405 Andrew Dixie <andrewd@gentrack.com>
21407 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
21410 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
21412 * expmed.c (extract_bit_field_1): Pass the full width of the
21413 structure to get_best_reg_extraction_insn.
21415 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
21418 * configure.ac (use_cxa_atexit): Add AIX.
21419 * configure: Regenerate.
21421 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
21423 2013-02-01 Jakub Jelinek <jakub@redhat.com>
21426 * final.c (need_profile_function): New variable.
21427 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
21428 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
21429 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
21430 notes, targetm.asm_out.function_prologue doesn't emit anything,
21431 HAVE_prologue and profiler should be emitted before prologue,
21432 set need_profile_function instead of emitting it.
21433 (final_scan_insn): If need_profile_function, emit
21434 profile_function on the first NOTE_INSN_BASIC_BLOCK or
21435 NOTE_INSN_FUNCTION_BEG note.
21437 2013-02-01 Richard Henderson <rth@redhat.com>
21439 * config/rs6000/rs6000.md (smulditi3): New.
21442 * config/alpha/alpha.md (umulditi3): New.
21444 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
21446 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
21447 (ASM_OUTPUT_ALIGNED_LOCAL): New.
21449 2013-02-01 Richard Biener <rguenther@suse.de>
21451 PR tree-optimization/56113
21452 * tree-ssa-structalias.c (label_visit): Reduce work for
21453 single-predecessor nodes.
21455 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
21457 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
21458 range isn't testing for zero.
21460 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
21462 PR middle-end/56113
21463 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
21465 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
21466 Nick Clifton <nickc@redhat.com>
21468 * config/v850/constraints.md (Q): Define as a memory constraint.
21469 * config/v850/predicates.md (label_ref_operand): New predicate.
21470 (e3v5_shift_operand): New predicate.
21471 (ior_operator): New predicate.
21472 * config/v850/t-v850: Add e3v5 multilib.
21473 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
21474 (v850_gen_movdi): Prototype.
21475 * config/v850/v850.c: Add support for e3v5 architecture.
21476 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
21478 (construct_save_jarl): Add e3v5 long JARL support.
21479 (v850_adjust_insn_length): New function. Adjust length of call
21480 insns when using e3v5 instructions.
21481 (v850_gen_movdi): New function: Generate instructions to move a
21483 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
21484 (CPP_SPEC): Define __v850e3v5__ as appropriate.
21485 (TARGET_USE_FPU): Enable for e3v5.
21486 (CONST_OK_FOR_W): New macro.
21487 (ADJUST_INSN_LENGTH): Define.
21488 * config/v850/v850.md (UNSPEC_LOOP): Define.
21489 (attr cpu): Add v850e3v5.
21490 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
21491 (movdi): New pattern.
21492 (movdi_internal): New pattern.
21493 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
21494 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
21495 (cstoresf4): Likewise.
21496 (cstoredf4): Likewise.
21497 (insv): New pattern.
21498 (rotlso3_a): New pattern.
21499 (rotlsi3_b): New pattern
21500 (rotlsi3_v850e3v5): New pattern.
21501 (doloop_begin): New pattern.
21502 (fix_loop_counter): New pattern.
21503 (doloop_end): New pattern.
21504 (branch_normal): Add e3v5 long branch support.
21505 (branch_invert): Likewise.
21506 (branch_z_normal): Likewise.
21507 (branch_z_invert): Likewise.
21508 (branch_nz_normal): Likewise.
21509 (branch_nz_invert): Likewise.
21510 (call_internal_short): Add e3v5 register-indirect JARL support.
21511 (call_internal_long): Likewise.
21512 (call_value_internal_short): Likewise.
21513 (call_value_internal_long): Likewise.
21514 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
21515 (mloop): New option.
21516 * config.gcc: Add support for configuring v840e3v5 target.
21517 * doc/invoke.texi: Document new v850 specific command line options.
21519 2013-01-31 Paul Koning <ni1d@arrl.net>
21523 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
21524 children if parent is a class.
21525 (prune_unused_types_prune): Don't add DW_AT_declaration.
21527 2013-01-31 Richard Biener <rguenther@suse.de>
21529 PR tree-optimization/56157
21530 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
21531 match up operand with SLP child.
21533 2013-01-31 Jason Merrill <jason@redhat.com>
21536 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
21537 parameters the first time.
21538 (gen_scheduled_generic_parms_dies): Check completeness here.
21540 2013-01-31 Richard Biener <rguenther@suse.de>
21542 PR middle-end/53073
21543 * common.opt (faggressive-loop-optimizations): New flag,
21544 enabled by default.
21545 * doc/invoke.texi (faggressive-loop-optimizations): Document.
21546 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
21547 infer_loop_bounds_from_undefined by it.
21549 2013-01-31 Richard Biener <rguenther@suse.de>
21551 PR tree-optimization/56150
21552 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
21553 visit virtual operands.
21554 (find_uses_to_rename_bb): Likewise.
21556 2013-01-31 Richard Biener <rguenther@suse.de>
21558 PR tree-optimization/56150
21559 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
21560 mixed store non-store stmts.
21562 2013-01-30 Jakub Jelinek <jakub@redhat.com>
21565 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
21566 LIBASAN_EARLY_SPEC is defined.
21567 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
21568 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
21570 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
21573 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
21574 invalid args instead of ICEing on it.
21575 (ix86_valid_target_attribute_tree): Return error_mark_node if
21576 ix86_valid_target_attribute_inner_p failed.
21577 (ix86_valid_target_attribute_p): Return false only if
21578 ix86_valid_target_attribute_tree returned error_mark_node. Allow
21579 target("default") attribute.
21580 (sorted_attr_string): Change argument from const char * to tree,
21581 merge in all target attribute arguments rather than just one.
21582 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
21583 instead of free. Avoid using strcat.
21584 (ix86_mangle_function_version_assembler_name): Mangle
21585 target("default") as if no target attribute is present. Adjust
21586 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
21587 instead of xmalloc and XDELETEVEC instead of free.
21588 (ix86_function_versions): Don't return true if one of the decls
21589 doesn't have target attribute. If they don't and one of the decls
21590 is DECL_FUNCTION_VERSIONED, report an error. Adjust
21591 sorted_attr_string caller. Use XDELETEVEC instead of free.
21592 (ix86_supports_function_versions): Remove.
21593 (make_name): Fix up formatting.
21594 (make_dispatcher_decl): Remove resolver_name and its initialization.
21595 Avoid leaking memory.
21596 (is_function_default_version): Return true if there is
21597 target("default") attribute rather than no target attribute at all.
21598 (make_resolver_func): Avoid leaking memory.
21599 (ix86_generate_version_dispatcher_body): Likewise.
21600 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
21601 * target.def (supports_function_versions): Remove.
21602 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
21603 * doc/tm.texi: Regenerated.
21605 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
21607 PR rtl-optimization/56144
21608 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
21609 for values with side effects.
21611 2013-01-30 Richard Biener <rguenther@suse.de>
21613 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
21614 (sparseset_pop): Likewise.
21615 * cfganal.c (compute_idf): Likewise. Increase work-stack size
21616 to be able to use quick_push in the worker loop.
21618 2013-01-30 Marek Polacek <polacek@redhat.com>
21620 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
21622 2013-01-30 Richard Biener <rguenther@suse.de>
21625 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
21627 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
21629 PR tree-optimization/56064
21630 * fixed-value.c (fixed_from_double_int): New function.
21631 * fixed-value.h (fixed_from_double_int): New prototype.
21632 (const_fixed_from_double_int): New static inline function.
21633 * fold-const.c (native_interpret_fixed): New static function.
21634 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
21635 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
21636 (native_encode_fixed): New static function.
21637 (native_encode_expr) <FIXED_CST>: Use it.
21638 (native_interpret_int): Move double_int worker code to...
21639 * double-int.c (double_int::from_buffer): ...this new static method.
21640 * double-int.h (double_int::from_buffer): Prototype it.
21642 2013-01-30 Richard Biener <rguenther@suse.de>
21644 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
21645 New pointer-map and obstack.
21646 (init_alias_vars): Allocate pointer-map and obstack.
21647 (delete_points_to_sets): Free them.
21648 (find_what_var_points_to): Cache result.
21649 (find_what_p_points_to): Adjust for changed interface of
21650 find_what_var_points_to.
21651 (compute_points_to_sets): Likewise.
21652 (ipa_pta_execute): Likewise.
21654 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
21656 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
21657 * configure: Regenerate.
21658 * config.in: Regenerate.
21659 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
21660 #nobits/#progbits if supported.
21662 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
21665 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
21666 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
21667 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
21669 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
21671 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
21672 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
21674 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
21676 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
21677 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
21679 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
21681 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
21683 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
21684 * config/arm/cortex-a7.md: New bypasses using
21685 arm_mac_accumulator_is_result.
21687 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
21689 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
21690 (cortex_a7_neon_mla): Likewise.
21691 (cortex_a7_fpfmad): New reservation.
21692 (cortex_a7_fpmacs): Use ffmas and update required units.
21693 (cortex_a7_fpmuld): Update required units and latency.
21694 (cortex_a7_fpmacd): Likewise.
21695 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
21696 (cortex_a7_neon). Likewise.
21697 (bypass) Update participating units.
21699 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
21701 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
21702 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
21704 * config/arm/vfp11.md (vfp_farith): Use ffmas.
21705 (vfp_fmul): Use ffmad.
21706 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
21707 (cortex_r4_fmacd): Use ffmad.
21708 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
21709 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
21710 (cortex_a9_fmacd): Use ffmad.
21711 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
21712 (cortex_a8_vfp_macd): Use ffmad.
21713 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
21714 (cortex_a5_fpmacd): Use ffmad.
21715 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
21716 (cortex_a15_vfp_macd): Use ffmad.
21717 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
21719 2013-01-29 Jason Merrill <jason@redhat.com>
21722 * varasm.c (default_assemble_visibility): Don't warn about
21723 visibility on artificial decls.
21725 2013-01-29 Richard Biener <rguenther@suse.de>
21727 PR tree-optimization/56113
21728 * tree-ssa-structalias.c (equiv_class_lookup): Also return
21730 (label_visit): Free duplicate bitmaps and record the leader instead.
21731 (perform_var_substitution): Adjust.
21733 2013-01-29 Richard Biener <rguenther@suse.de>
21735 PR tree-optimization/55270
21736 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
21737 the CFG, schedule loops for fixup.
21739 2013-01-29 Nick Clifton <nickc@redhat.com>
21741 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
21744 2013-01-28 Leif Ekblad <leif@rdos.net>
21746 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
21747 * config/i386/i386.h (TARGET_RDOS): New macro.
21748 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
21749 * config/i386/i386.c (ix86_option_override_internal): For 64bit
21750 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
21751 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
21752 DEFAULT_LARGE_SECTION_THRESHOLD.
21753 * config/i386/i386.md (R14_REG, R15_REG): New constants.
21754 * config/i386/rdos.h: New file.
21755 * config/i386/rdos64.h: New file.
21757 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
21760 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
21763 2013-01-28 Jakub Jelinek <jakub@redhat.com>
21765 PR rtl-optimization/56117
21766 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
21767 call cselib_lookup_from_insn on the MEM before calling
21768 add_insn_mem_dependence.
21770 2013-01-28 Richard Biener <rguenther@suse.de>
21772 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
21773 to a stmt that didn't have one.
21774 (copy_phis_for_bb): Likewise for PHI arguments.
21775 (copy_debug_stmt): Likewise for debug stmts.
21777 2013-01-28 Richard Biener <rguenther@suse.de>
21779 PR tree-optimization/56034
21780 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
21781 (partition_builtin_p): Adjust.
21782 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
21783 it is the last partition.
21784 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
21785 up the vertex for the definition.
21786 (classify_partition): Classify whether a partition is a
21787 PKIND_REDUCTION, thus has uses outside of the loop.
21788 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
21789 Merge all PKIND_REDUCTION partitions into the last partition.
21790 (tree_loop_distribution): Seed partitions from reductions as well.
21792 2013-01-28 Jakub Jelinek <jakub@redhat.com>
21794 PR tree-optimization/56125
21795 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
21796 pow(x,c) into sqrt(x) * powi(x, n/2) or
21797 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
21798 optimizing for size.
21799 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
21800 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
21803 PR tree-optimization/56094
21804 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
21805 to UNKNOWN_LOCATION while gimplifying expr.
21807 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
21810 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
21811 operand 0 in movabs insn template for -masm=intel asm alternative.
21812 (*movabs<mode>_2): Ditto for operand 1.
21814 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
21817 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
21820 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
21822 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
21825 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
21827 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
21828 Undef to avoid warning.
21830 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
21832 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
21833 * configure: Regenerate.
21835 2013-01-25 Jakub Jelinek <jakub@redhat.com>
21837 PR tree-optimization/56098
21838 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
21839 for stmts with volatile ops.
21840 (cond_store_replacement): Don't optimize if assign has volatile ops.
21841 (cond_if_else_store_replacement_1): Don't optimize if either
21842 then_assign or else_assign have volatile ops.
21843 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
21846 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
21848 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
21850 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
21852 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
21853 missing ':' in asm example.
21855 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
21857 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
21858 entries into lane and laneq entries.
21859 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
21860 Remove AdvSIMD scalar modes.
21861 (aarch64_sq<r>dmulh_laneq<mode>): New.
21862 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
21864 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
21865 builtin implementations to relfect changes in RTL in aarch64-simd.md.
21866 * config/aarch64/iterators.md (VCOND): New.
21869 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
21872 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
21873 Add NULL LIBNAME argument to existing definitions.
21874 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
21875 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
21876 * config/avr/avr.c (DEF_BUILTIN): Same.
21877 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
21878 (avr_expand_builtin): Expand to a vanilla call if a libgcc
21879 implementation is available (DECL_ASSEMBLER_NAME is set).
21880 (avr_fold_absfx): New static function.
21881 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
21882 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
21883 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
21884 AVR_BUILTIN_ABSLLK.
21885 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
21886 (abshk, absk, abslk, absllk): Provide as static inline functions.
21888 2013-01-25 Marek Polacek <polacek@redhat.com>
21890 PR tree-optimization/56035
21891 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
21893 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
21895 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
21896 (*movtf_internal_rex64): Add (!o,C) alternative
21897 (*movxf_internal_rex64): Ditto.
21898 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
21900 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
21902 * doc/invoke.texi: fix typo.
21903 * doc/objc.texi: fix typo.
21905 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
21907 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
21908 for the first two alternatives.
21910 2013-01-24 Diego Novillo <dnovillo@google.com>
21912 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
21913 (ggc-zone.o): Remove.
21914 * configure.ac: Remove option --with-gc.
21915 * configure: Re-generate.
21916 * doc/install.texi: Remove documentation for --with-gc.
21917 * gengtype.c (write_enum_defn): Remove. Update all users.
21918 (write_Types_process_field): Remove generation of gt_e_* argument.
21919 (output_type_enum): Remove. Update all users.
21920 (write_enum_defn): Remove. Update all users.
21921 (enum alloc_zone): Remove. Update all users.
21922 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
21923 * ggc-common.c (ggc_splay_alloc): Remove first argument.
21924 Update all callers.
21925 (struct ptr_data): Remove field TYPE. Update all users.
21926 (gt_pch_note_object): Remove argument TYPE. Update all users.
21927 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
21929 * ggc-none.c (ggc_alloc_typed_stat): Remove.
21930 (struct alloc_zone): Remove.
21931 (ggc_internal_alloc_zone_stat): Remove.
21932 (ggc_internal_cleared_alloc_zone_stat): Remove.
21933 * ggc-page.c (ggc_alloc_typed_stat): Remove.
21934 (ggc_pch_count_object): Remove last argument. Update all users.
21935 (ggc_pch_alloc_object): Remove last argument. Update all users.
21936 (struct alloc_zone): Remove.
21937 * ggc-zone.c: Remove.
21938 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
21939 (struct alloc_zone): Remove.
21940 (ggc_alloc_typed_stat): Remove.
21941 (ggc_alloc_typed): Remove.
21942 (ggc_splay_alloc): Remove first argument.
21943 (rtl_zone): Remove. Update all users.
21944 (tree_zone): Remove. Update all users.
21945 (tree_id_zone): Remove. Update all users.
21946 (ggc_internal_zone_alloc_stat): Remove. Update all users.
21947 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
21948 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
21949 * tree-ssanames.c: Remove references to zone allocator in comments.
21951 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
21953 * config/avr/avr.c (avr_out_fract): Make register numbers that
21954 might be outside of source operand signed.
21956 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
21958 * config/i386/constraints.md (Yf): New constraint.
21959 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
21960 of f constraint to conditionaly disable x87 register preferences.
21961 (*movdf_internal): Ditto.
21962 (*movsf_internal): Ditto.
21964 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
21966 PR inline-asm/55934
21967 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
21968 that have operands with impossible constraints.
21969 Add a FIXME for a speed-up opportunity.
21970 * lra-constraints.c (process_alt_operands): Verify that a class
21971 selected from constraints on asms is valid for the operand mode.
21972 (curr_insn_transform): Remove incorrect comment.
21974 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
21976 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
21977 TOC operand is a valid symbol ref in the constant pool.
21979 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
21981 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
21983 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
21986 * config/avr/stdfix.h: New file.
21987 * t-avr (stdfix-gcc.h): New rule to build it.
21988 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
21990 2013-01-23 Kostya Serebryany <kcc@google.com>
21992 * config/darwin.h: remove dependency on
21993 CoreFoundation (asan on Mac OS).
21995 2013-01-23 Jakub Jelinek <jakub@redhat.com>
21998 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
21999 instead of cmpdi_operand for first comparison operand.
22000 Don't assert that comparison operands aren't both constants.
22002 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
22004 * doc/install.texi (Downloading the Source): Update references to
22005 downloading separate components.
22007 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
22009 * doc/extend.texi (__int128): Improve grammar.
22011 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
22014 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
22015 alternative to (o,r).
22016 (*movdi_internal_rex64): Remove (!o,n) alternative.
22017 (DImode immediate->memory splitter): Remove.
22018 (DImode immediate->memory peephole2): Remove.
22019 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
22020 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
22021 alternative to (!o,*r).
22022 (*movtf_internal_sse): New pattern.
22023 (*movxf_internal_rex64): New pattern.
22024 (*movxf_internal): Disable for TARGET_64BIT.
22025 (*movdf_internal_rex64): Remove (!o,F) alternative.
22027 2013-01-22 Jakub Jelinek <jakub@redhat.com>
22029 PR middle-end/56074
22030 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
22031 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
22032 * tree-vect-loop-manip.c (find_loop_location): Also ignore
22033 stmt locations where LOCATION_LOCUS of the stmt location is
22034 UNKNOWN_LOCATION or BUILTINS_LOCATION.
22037 * config/i386/i386.md (UNSPEC_STOS): New.
22038 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
22039 *strsetqi_1): Add UNSPEC_STOS.
22041 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
22044 * doc/invoke.texi: Remove left over -Wsynth example.
22046 2013-01-21 Jakub Jelinek <jakub@redhat.com>
22048 PR tree-optimization/56051
22049 * fold-const.c (fold_binary_loc): Don't fold
22050 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
22051 a narrowing conversion, or widening conversion from signed
22054 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
22056 PR rtl-optimization/56023
22057 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
22058 dependent on debug instruction.
22060 2013-01-21 Martin Jambor <mjambor@suse.cz>
22062 PR middle-end/56022
22063 * function.c (allocate_struct_function): Call
22064 invoke_set_current_function_hook earlier.
22066 2013-01-21 Jakub Jelinek <jakub@redhat.com>
22068 * reload1.c (init_reload): Only initialize reload_obstack
22069 during the first call.
22071 2013-01-21 Marek Polacek <polacek@redhat.com>
22073 * cfgloop.c (verify_loop_structure): Fix up grammar.
22075 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
22077 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
22078 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
22080 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
22083 * config/arm/marvell-pj4.md: Update copyright year.
22084 Fix up use of alu to alu_reg and simple_alu_imm.
22086 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
22088 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
22090 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
22093 * lra-constraints.c (curr_insn_transform): Don't reuse original
22094 insn for secondary memory move when memory mode should be different.
22096 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
22098 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
22099 atomic_storedi_1): New patterns.
22101 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
22103 btver2 pipeline descriptions.
22104 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
22106 * config/i386/i386.md (btver2_decode): New type attributes.
22107 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
22109 * config/i386/btver2.md: New file describing btver2 pipelines.
22111 2013-01-19 Andrew Pinski <apinski@cavium.com>
22113 PR tree-optimization/52631
22114 * tree-ssa-sccvn (visit_use): Before looking up the original
22115 statement, try looking up the simplified expression.
22117 2013-01-19 Anthony Green <green@moxielogic.com>
22119 * config/moxie/moxie.c (moxie_expand_prologue): Set
22120 current_function_static_stack_size.
22122 2013-01-18 Jakub Jelinek <jakub@redhat.com>
22124 PR tree-optimization/56029
22125 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
22126 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
22128 2013-01-18 Sharad Singhai <singhai@google.com>
22130 PR tree-optimization/55995
22131 * dumpfile.c (dump_loc): Print location only if available.
22132 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
22134 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
22137 * lra-constraints.c (curr_insn_transform): Reuse original insn for
22138 secondary memory move.
22139 (inherit_reload_reg): Use rclass instead of cl for
22140 check_secondary_memory_needed_p.
22142 2013-01-18 Jakub Jelinek <jakub@redhat.com>
22144 PR middle-end/56015
22145 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
22146 the case where writing real complex part of target modifies op1.
22148 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
22150 * config/aarch64/aarch64-simd.md
22151 (aarch64_vcond_internal<mode>): Handle unordered cases.
22152 * config/aarch64/iterators.md (v_cmp_result): New.
22154 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
22155 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
22157 * config/arm/marvell-pj4.md: New file.
22158 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
22159 * config/arm/arm.md (generic_sched): Add marvell_pj4.
22160 (generic_vfp): Likewise.
22161 * config/arm/arm-cores.def: Add marvell-pj4.
22162 * config/arm/arm-tune.md: Regenerate.
22163 * config/arm/arm-tables.opt: Regenerate.
22164 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
22165 * doc/invoke.texi: Document marvell-pj4.
22167 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
22169 * config/aarch64/arm_neon.h: Map scalar types to standard types.
22171 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
22176 * var-tracking.c (negative_power_of_two_p): New.
22177 (global_get_addr_cache, local_get_addr_cache): New.
22178 (get_addr_from_global_cache, get_addr_from_local_cache): New.
22179 (vt_canonicalize_addr): Rewrite using the above. Adjust the
22181 (vt_stack_offset_p): Remove.
22182 (vt_canon_true_dep): Always canonicalize loc's address.
22183 (clobber_overlapping_mems): Make sure we have a MEM.
22184 (local_get_addr_clear_given_value): New.
22185 (val_reset): Clear local cached entries.
22186 (compute_bb_dataflow): Create and release the local cache.
22187 Disable duplicate MEMs clobbering.
22188 (emit_notes_in_bb): Clobber MEMs likewise.
22189 (vt_emit_notes): Create and release the local cache.
22190 (vt_initialize, vt_finalize): Create and release the global
22191 cache, respectively.
22192 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
22194 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
22196 PR libmudflap/53359
22197 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
22198 not found in the symtab.
22200 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
22203 PR rtl-optimization/55547
22204 PR rtl-optimization/53827
22207 * alias.c (offset_overlap_p): New, factored out of...
22208 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
22209 the conservative special case for symbolic constants. Don't
22210 adjust zero sizes on alignment.
22212 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
22214 PR rtl-optimization/52573
22215 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
22216 REG_UNUSED for the same register.
22218 2013-01-17 Richard Biener <rguenther@suse.de>
22219 Marek Polacek <polacek@redhat.com>
22221 PR rtl-optimization/55833
22222 * loop-unswitch.c (unswitch_loops): Move loop verification...
22223 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
22224 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
22225 Set it to true when we're removing a loop from hierarchy tree in
22226 an irreducible region.
22227 (fix_bb_placements): Adjust caller.
22228 (fix_loop_placements): Likewise.
22230 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
22232 * config/avr/builtins.def (DEF_BUILTIN): Factor out
22233 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
22234 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
22235 Remove ID. Adjust comments.
22236 * config/avr/avr-c.c (avr_builtin_name): Remove.
22237 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
22238 * config/avr/avr.c (avr_tolower): New static function.
22239 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
22240 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
22241 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
22244 2013-01-17 Jan Hubicka <jh@suse.cz>
22246 PR tree-optimization/55273
22247 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
22249 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
22252 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
22253 store through atomic_store<mode>_1.
22254 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
22256 2013-01-17 Martin Jambor <mjambor@suse.cz>
22258 PR tree-optimizations/55264
22259 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
22260 for virtual methods.
22261 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
22262 virtual methods before inlining is over.
22263 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
22265 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
22268 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
22270 PR rtl-optimization/56005
22271 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
22272 pending reads for prefetch.
22274 2013-01-16 Ian Bolton <ian.bolton@arm.com>
22276 * config/aarch64/aarch64.md
22277 (*cstoresi_neg_uxtw): New pattern.
22278 (*cmovsi_insn_uxtw): New pattern.
22279 (*<optab>si3_uxtw): New pattern.
22280 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
22281 (*<optab>si3_insn_uxtw): New pattern.
22282 (*bswapsi2_uxtw): New pattern.
22284 2013-01-16 Richard Biener <rguenther@suse.de>
22286 * tree-inline.c (tree_function_versioning): Remove set but
22287 never used variable.
22289 2013-01-16 Richard Biener <rguenther@suse.de>
22291 PR tree-optimization/55964
22292 * tree-flow.h (rename_variables_in_loop): Remove.
22293 (rename_variables_in_bb): Likewise.
22294 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
22295 (copy_loop_before): Adjust and delete update-ssa status.
22296 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
22297 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
22298 (rename_variables_in_loop): Remove.
22299 (slpeel_update_phis_for_duplicate_loop): Likewise.
22300 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
22301 use available cfg machinery instead of duplicating it.
22302 Update PHI nodes and perform poor-mans SSA update here.
22303 (slpeel_tree_peel_loop_to_edge): Adjust.
22305 2013-01-16 Richard Biener <rguenther@suse.de>
22307 PR tree-optimization/54767
22308 PR tree-optimization/53465
22309 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
22310 (vrp_visit_phi_node): For PHI arguments coming via backedges
22311 drop all symbolical range information.
22312 (execute_vrp): Compute backedges.
22314 2013-01-16 Richard Biener <rguenther@suse.de>
22316 * doc/install.texi: Update CLooG and ISL requirements to
22319 2013-01-16 Christian Bruel <christian.bruel@st.com>
22322 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
22323 (broken_move): Handle UNSPECV_SP_SWITCH_B.
22324 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
22326 2013-01-16 DJ Delorie <dj@redhat.com>
22328 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
22329 (UNSPECV_SP_SWITCH_E): New.
22330 (sp_switch_1): Change to an unspec.
22331 (sp_switch_2): Change to an unspec. Don't use post-inc when we
22334 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
22336 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
22337 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
22338 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
22339 (expand_mem_thread_fence): Ditto.
22340 (expand_mem_signal_fence): Ditto.
22341 (expand_atomic_load): Ditto.
22342 (expand_atomic_store): Ditto.
22344 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
22346 PR rtl-optimization/55547
22347 PR rtl-optimization/53827
22350 * alias.c (memrefs_conflict_p): Set sizes to negative after
22353 2013-01-15 Jakub Jelinek <jakub@redhat.com>
22356 * function.c (thread_prologue_and_epilogue_insns): Always
22357 add crtl->drap_reg to set_up_by_prologue.set, even if
22358 stack_realign_drap is false.
22360 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
22362 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
22363 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
22364 *call): Fix indention.
22366 2013-01-15 Tom de Vries <tom@codesourcery.com>
22369 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
22372 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
22374 PR rtl-optimization/55153
22375 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
22377 2013-01-15 Martin Jambor <mjambor@suse.cz>
22379 PR tree-optimization/55920
22380 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
22381 accesses as grp_to_be_debug_replaced.
22383 2013-01-15 Jakub Jelinek <jakub@redhat.com>
22385 PR tree-optimization/55920
22386 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
22387 there is non-useless type conversion needed from debug rhs to lhs,
22388 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
22390 2013-01-15 Joseph Myers <joseph@codesourcery.com>
22391 Mikael Pettersson <mikpe@it.uu.se>
22394 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
22396 (ASM_OUTPUT_CASE_LABEL): Remove.
22397 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
22398 * final.c (shorten_branches): Update alignment of labels before
22399 jump tables if CASE_VECTOR_SHORTEN_MODE.
22401 2013-01-15 Richard Biener <rguenther@suse.de>
22404 * system.h: Do not include gmp.h for building host tools.
22406 2013-01-15 Richard Biener <rguenther@suse.de>
22408 PR middle-end/55882
22409 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
22410 account for bitpos when computing alignment.
22412 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
22414 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
22415 (ix86_target_macros_internal): Likewise.
22417 * config/i386/i386.c (m_CORE2I7): Removed.
22418 (m_CORE_HASWELL): New macro.
22419 (m_CORE_ALL): Likewise.
22420 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
22421 (initial_ix86_arch_features): Likewise.
22422 (processor_target_table): Initializations for Core avx2.
22423 (cpu_names): New names "core-avx2".
22424 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
22425 PROCESSOR_CORE_HASWELL.
22426 (ix86_issue_rate): New case.
22427 (ia32_multipass_dfa_lookahead): Likewise.
22428 (ix86_sched_init_global): Likewise.
22430 * config/i386/i386.h (TARGET_HASWELL): New macro.
22431 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
22432 (processor_type): New PROCESSOR_HASWELL.
22434 2013-01-15 Jakub Jelinek <jakub@redhat.com>
22436 PR tree-optimization/55955
22437 * tree-vect-loop.c (vectorizable_reduction): Give up early on
22438 *SHIFT_EXPR and *ROTATE_EXPR codes.
22440 PR tree-optimization/48766
22441 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
22442 -ftrapv disable -fwrapv.
22444 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
22447 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
22448 etc. to 1 and not to __flash.
22449 Use LL suffix for __INT24_MAX__ with -mint8.
22450 Use ULL suffix for __UINT24_MAX__ with -mint8.
22452 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
22454 * config/avr/avr-arch.h
22455 (struct base_arch_s): Use typedef avr_arch_t instead.
22456 (struct arch_info_s): Use typedef avr_arch_info_t instead.
22457 (struct mcu_type_s): Use typedef avr_mcu_t instead.
22458 * config/avr/avr.c: Same.
22459 * config/avr/avr-devices.c: Same.
22460 * config/avr/driver-avr.c: Same.
22461 * config/avr/gen-avr-mmcu-texi.c: Same.
22462 * config/avr/avr-mcus.def: Adjust comment.
22464 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
22466 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
22467 * config/aarch64/iterators.md (VALLDI): New.
22469 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
22470 Andi Kleen <ak@linux.intel.com>
22473 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
22474 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
22477 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
22479 * config/avr/avr-stdint.h: Remove trailing blanks.
22480 * config/avr/avr-log.h: Same.
22481 * config/avr/avr-arch.h: Same.
22482 * config/avr/avr-devices.c: Same.
22483 * config/avr/avr-dimode.md: Same.
22484 * config/avr/predicates.md: Same.
22485 * config/avr/avr-c.c: Same. And fix typo.
22487 * config/avr/avr-protos.h: Same. And:
22488 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
22489 (init_cumulative_args): Rename to avr_init_cumulative_args.
22490 (expand_prologue): Rename to avr_expand_prologue.
22491 (expand_epilogue): Rename to avr_expand_epilogue.
22492 (adjust_insn_length): Rename to avr_adjust_insn_length.
22493 (notice_update_cc): Rename to avr_notice_update_cc.
22494 (final_prescan_insn): Rename to avr_final_prescan_insn.
22495 * config/avr/avr.c: Same.
22496 * config/avr/avr.h: Same.
22497 * config/avr/avr.md: Remove trailing blanks.
22498 (prologue): Use avr_expand_prologue.
22499 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
22501 2013-01-14 Richard Biener <rguenther@suse.de>
22503 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
22504 verify_location, collect_subblocks): New functions.
22505 (verify_gimple_in_cfg): Verify that locations only reference
22506 BLOCKs in the functions BLOCK tree.
22508 2013-01-14 Richard Biener <rguenther@suse.de>
22510 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
22512 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
22514 (insert_out_of_ssa_copy_on_edge): Likewise.
22515 (rewrite_close_phi_out_of_ssa): Likewise.
22516 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
22518 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
22519 propagated constants.
22520 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
22523 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
22525 * config/avr/avr-modes.def: Add GPL copyright notice.
22527 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
22529 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
22530 MEMMODEL_MASK to determine memory model.
22531 (atomic_store<mode>): Ditto from operands[2].
22532 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
22534 2013-01-13 Jakub Jelinek <jakub@redhat.com>
22537 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
22538 (fold_gimple_assign): Don't call unshare_expr here.
22539 (fold_ctor_reference): Call unshare_expr.
22541 2013-01-13 Terry Guo <terry.guo@arm.com>
22543 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
22544 * doc/fragments.texi: Document MULTILIB_REUSE.
22545 * gcc.c (multilib_reuse): New internal spec.
22546 (set_multilib_dir): Also search multilib from multilib_reuse.
22547 * genmultilib (tmpmultilib3): Refactor code.
22548 (tmpmultilib4): Ditto.
22549 (multilib_reuse): New multilib argument.
22551 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
22553 * Makefile.in: Update copyright.
22555 2013-01-12 Tom de Vries <tom@codesourcery.com>
22557 PR middle-end/55890
22558 * calls.c (expand_call): Check if arg_nr is valid.
22560 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
22562 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
22563 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
22564 documentation. Add missing '__' in front of
22565 __builtin_ia32_packssdw256.
22567 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
22570 * config/s390/s390.c (s390_preferred_reload_class): Do not return
22571 NO_REGS for larl operands.
22572 (s390_reload_larl_operand): Use s390_load_address instead of
22575 2013-01-11 Richard Biener <rguenther@suse.de>
22577 * tree-cfg.c (verify_node_sharing_1): Split out from ...
22578 (verify_node_sharing): ... here.
22579 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
22581 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
22583 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
22584 Substitute TREECHECKING.
22585 * configure: Regenerate.
22586 * Makefile.in (TREECHECKING): New.
22588 2013-01-11 Richard Guenther <rguenther@suse.de>
22590 PR tree-optimization/44061
22591 * tree-vrp.c (extract_range_basic): Compute zero as
22592 value-range for __builtin_constant_p of function parameters.
22594 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
22596 Update copyright years.
22598 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
22600 PR rtl-optimization/55672
22601 * lra-eliminations.c (mark_not_eliminable): Permit addition with
22602 const to be eliminable.
22604 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
22606 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
22607 * configure: Regenerate.
22609 2013-01-10 Richard Biener <rguenther@suse.de>
22611 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
22613 2013-01-10 Richard Biener <rguenther@suse.de>
22616 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
22617 locations for virtual PHI arguments.
22618 (rewrite_update_phi_arguments): Likewise.
22620 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
22622 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
22625 2013-01-10 Jakub Jelinek <jakub@redhat.com>
22627 PR tree-optimization/55921
22628 * tree-complex.c (expand_complex_asm): New function.
22629 (expand_complex_operations_1): Call it for GIMPLE_ASM.
22631 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
22634 * config/s390/s390.c (s390_symref_operand_p)
22635 (s390_loadrelative_operand_p): Merge the two functions.
22636 (s390_check_qrst_address, print_operand_address): Add parameters
22637 to s390_loadrelative_operand_p invokation.
22638 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
22639 (s390_reload_larl_operand, s390_secondary_reload): Use
22640 s390_loadrelative_operand_p instead of s390_symref_operand_p.
22641 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
22643 2013-01-09 Mike Stump <mikestump@comcast.net>
22645 * dse.c (record_store): Remove unnecessary assert.
22647 2013-01-09 Jan Hubicka <jh@suse.cz>
22649 PR tree-optimization/55569
22650 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
22651 * cfgloop.h (scale_loop_profile): Likewise.
22653 2013-01-09 Jan Hubicka <jh@suse.cz>
22656 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
22658 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
22660 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
22662 PR middle-end/55114
22663 * expr.h (maybe_emit_group_store): Declare.
22664 * expr.c (maybe_emit_group_store): New function.
22665 * builtins.c (expand_builtin_int_roundingfn): Call it.
22666 (expand_builtin_int_roundingfn_2): Likewise.
22668 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
22670 PR rtl-optimization/55829
22671 * lra-constraints.c (match_reload): Add code for absent output.
22672 (curr_insn_transform): Add code for reloads of matched inputs
22675 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
22677 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
22678 attribute of movddup insn to DF.
22679 (*vec_interleave_lowv2df): Ditto.
22680 (vec_dupv2df): Ditto.
22682 2013-01-09 Jan Hubicka <jh@suse.cz>
22684 PR tree-optimiation/55875
22685 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
22686 EVERY_ITERATION parameter.
22687 (number_of_iterations_exit): Check if exit is executed every iteration.
22688 (idx_infer_loop_bounds): Similarly here.
22689 (n_of_executions_at_most): Simplify
22690 to only test for cases where statement is dominated by the
22691 particular bound; handle correctly the "postdominance" test.
22692 (scev_probably_wraps_p): Use max loop iterations info
22693 as a global bound first.
22695 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
22696 Nick Clifton <nickc@redhat.com>
22698 * config/v850/v850.md (cbranchsf4): New pattern.
22699 (cstoresf4): New pattern.
22700 (cbranchdf4): New pattern.
22701 (cstoredf4): New pattern.
22702 (movsicc): Disallow floating point comparisons.
22703 (cmpsf_le_insn): Fix order of operators.
22704 (cmpsf_lt_insn): Likewise.
22705 (cmpsf_eq_insn): Likewise.
22706 (cmpdf_le_insn): Likewise.
22707 (cmpdf_lt_insn): Likewise.
22708 (cmpdf_eq_insn): Likewise.
22709 (cmpsf_ge_insn): Use LE comparison.
22710 (cmpdf_ge_insn): Likewise.
22711 (cmpsf_gt_insn): Use LT comparison.
22712 (cmpdf_gt_insn): Likewise.
22713 (cmpsf_ne_insn): Delete pattern.
22714 (cmpdf_ne_insn): Delete pattern.
22715 * config/v850/v850.c (v850_gen_float_compare): Use
22716 gen_cmpdf_eq_insn for NE comparison.
22717 (v850_float_z_comparison_operator)
22718 (v850_float_nz_comparison_operator): Move from here ...
22719 * config/v850/predicates.md: ... to here. Move GT and GE
22720 comparisons into v850_float_z_comparison_operator.
22721 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
22723 (v850_float_nz_comparison_operator): Likewise.
22725 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
22727 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
22728 with calls to gen_insvsi/gen_insvdi.
22730 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
22732 * config/i386/i386.c (initial_ix86_tune_features): Set up
22733 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
22735 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
22736 Jakub Jelinek <jakub@redhat.com>
22738 PR tree-optimization/48189
22739 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
22740 If nitercst is 0, don't predict the exit edge.
22742 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22744 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
22745 in asm_fprintf with reg_names.
22746 (aarch64_print_operand_address): Likewise.
22747 (aarch64_return_addr): Likewise.
22748 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
22750 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
22752 * config/pa/pa.h (VAL_U6_BITS_P): Define.
22753 (INT_U6_BITS): Likewise.
22754 * config/pa/predicates.md (uint6_operand): New predicate.
22755 (shift5_operand, shift6_operand): Likewise.
22756 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
22758 (lshrdi3): Use shift6_operand.
22759 (shrpsi4, shrpdi4): New insn patterns.
22760 (extzv): Delete expander.
22761 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
22762 predicates in unamed zero extract patterns. Tighten common constraint.
22763 (extv): Delete expander.
22764 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
22765 predicates in unamed sign extract patterns. Tighten common constraint.
22766 (insv): Delete expander.
22767 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
22768 predicates in unamed insert patterns. Tighten common constraint.
22769 Change uint32_operand predicate to uint6_operand predicate in unamed
22770 DImode pattern to insert constant values of type 1...1xxxx.
22772 2013-01-04 Jan Hubicka <jh@suse.cz>
22774 PR tree-optimization/55823
22775 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
22778 2013-01-08 Jakub Jelinek <jakub@redhat.com>
22779 Uros Bizjak <ubizjak@gmail.com>
22781 PR rtl-optimization/55845
22782 * df-problems.c (can_move_insns_across): Stop scanning at
22783 volatile_insn_p source instruction or give up if
22784 across_from .. across_to range contains any volatile_insn_p
22787 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
22789 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
22790 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
22792 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
22793 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
22795 2013-01-08 Jakub Jelinek <jakub@redhat.com>
22798 * asan.c (asan_clear_shadow): New function.
22799 (asan_emit_stack_protection): Use it.
22801 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
22803 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
22804 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
22805 with tab instead of space.
22807 2013-01-08 Nick Clifton <nickc@redhat.com>
22809 * config/rl78/rl78.c (rl78_expand_prologue): Always select
22810 register bank 0 at the start of an interrupt handler.
22811 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
22814 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
22816 * config/aarch64/aarch64-simd.md
22817 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
22818 (aarch64_simd_bsl): Likewise.
22819 (aarch64_vcond_internal<mode>): Likewise.
22820 (vcond<mode><mode>): Likewise.
22821 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
22822 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
22824 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
22826 * config/aarch64/aarch64-builtins.c
22827 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
22829 2013-01-08 Martin Jambor <mjambor@suse.cz>
22832 * tree-sra.c (analyze_access_subtree): Return true also after
22833 potentially creating a debug-only replacement.
22835 2013-01-08 Jakub Jelinek <jakub@redhat.com>
22837 PR middle-end/55890
22838 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
22840 PR tree-optimization/54120
22841 * tree-vrp.c (range_fits_type_p): Don't allow
22842 src_precision < precision from signed vr to unsigned_p
22843 if vr->min or vr->max is negative.
22844 (simplify_float_conversion_using_ranges): Test can_float_p
22845 against CODE_FOR_nothing.
22847 2013-01-08 Jakub Jelinek <jakub@redhat.com>
22848 Richard Biener <rguenther@suse.de>
22850 PR middle-end/55851
22851 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
22852 types instead of just INTEGER_TYPE types.
22854 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
22856 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
22859 2013-01-07 Steve Ellcey <sellcey@mips.com>
22862 * config/mips/mips.opt: Change mad to mmad to match documentation.
22864 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
22867 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
22868 .progmemx.data now.
22870 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
22873 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
22874 (avr_addrspace_t): Add .section_name field.
22875 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
22877 (avr_addrspace): Same. Initialize .section_name. Remove last
22878 NULL entry. Put __memx into .progmemx.data.
22879 (progmem_section_prefix): Remove.
22880 (avr_asm_init_sections): No need to initialize progmem_section.
22881 (avr_asm_named_section): Use avr_addrspace[].section_name to get
22882 section name prefix.
22883 (avr_asm_select_section): Ditto. And use get_unnamed_section to
22884 retrieve the progmem section.
22885 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
22886 boundary to run over avr_addrspace[].
22887 (avr_register_target_pragmas): Ditto.
22889 2013-01-06 Jakub Jelinek <jakub@redhat.com>
22891 * varasm.c (output_constant_def_contents): For asan_protect_global
22892 protected strings, adjust DECL_ALIGN if needed, before testing for
22894 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
22895 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
22897 (output_object_block): For asan protected decls, emit asan padding
22898 after their contents.
22899 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
22900 (asan_finish_file): Test it here instead.
22902 2013-01-07 Nick Clifton <nickc@redhat.com>
22903 Matthias Klose <doko@debian.org>
22904 Doug Kwan <dougkwan@google.com>
22905 H.J. Lu <hongjiu.lu@intel.com>
22908 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
22910 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
22912 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
22914 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
22916 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
22918 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
22921 * doc/install.texi (Cross-Compiler-Specific Options): Document
22924 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
22926 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
22927 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
22928 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
22929 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
22930 vqmovun_high_s64): Fix source operand number and update copyright.
22932 2013-01-07 Richard Biener <rguenther@suse.de>
22934 PR middle-end/55890
22935 * gimple.h (gimple_call_builtin_p): New overload.
22936 * gimple.c (validate_call): New function.
22937 (gimple_call_builtin_p): Likewise.
22938 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
22939 Use gimple_call_builtin_p.
22940 (find_func_clobbers): Likewise.
22941 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
22942 (strlen_optimize_stmt): Likewise.
22944 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
22946 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
22947 (vld1q_dup_*): Likewise.
22948 (vld1_*): Likewise.
22949 (vld1q_*): Likewise.
22950 (vld1_lane_*): Likewise.
22951 (vld1q_lane_*): Likewise.
22953 2013-01-07 Richard Biener <rguenther@suse.de>
22955 * lto-streamer.h (LTO_minor_version): Bump to 2.
22957 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
22959 * config/aarch64/aarch64-protos.h
22960 (aarch64_const_double_zero_rtx_p): Rename to...
22961 (aarch64_float_const_zero_rtx_p): ...this.
22962 (aarch64_float_const_representable_p): New.
22963 (aarch64_output_simd_mov_immediate): Likewise.
22964 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
22965 move immediate case.
22966 * config/aarch64/aarch64.c
22967 (aarch64_const_double_zero_rtx_p): Rename to...
22968 (aarch64_float_const_zero_rtx_p): ...this.
22969 (aarch64_print_operand): Allow printing of new constants.
22970 (aarch64_valid_floating_const): New.
22971 (aarch64_legitimate_constant_p): Check for valid floating-point
22973 (aarch64_simd_valid_immediate): Likewise.
22974 (aarch64_vect_float_const_representable_p): New.
22975 (aarch64_float_const_representable_p): Likewise.
22976 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
22977 (aarch64_output_simd_mov_immediate): New.
22978 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
22979 (*movdf_aarch64): Likewise.
22980 * config/aarch64/constraints.md (Ufc): New.
22981 (Y): call aarch64_float_const_zero_rtx.
22982 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
22984 2013-01-07 Richard Biener <rguenther@suse.de>
22986 PR tree-optimization/55888
22987 PR tree-optimization/55862
22988 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
22989 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
22990 not if it is contained therein.
22992 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
22994 * config/avr/t-avr: Typo.
22996 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
22999 * config/avr/t-avr: Don't automatically rebuild
23000 $(srcdir)/config/avr/t-multilib
23001 $(srcdir)/config/avr/avr-tables.opt
23002 $(srcdir)/doc/avr-mmcu.texi
23003 (avr-mcus): New phony target to build them on request.
23004 (s-avr-mlib, s-avr-mmcu-texi): Remove.
23005 * avr/avr-mcus.def: Adjust comments.
23007 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
23009 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
23011 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
23013 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
23015 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
23017 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
23019 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
23021 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
23022 to generate profiling.
23023 * config/rs6000/aix64.h (LIB_SPEC): Same.
23025 2013-01-04 Andrew Pinski <apinski@cavium.com>
23027 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
23029 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
23031 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
23033 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
23035 (ix86_expand_move): Ditto.
23036 (ix86_zero_extend_to_Pmode): Ditto.
23037 (ix86_expand_call): Ditto.
23038 (ix86_expand_special_args_builtin): Ditto.
23039 (ix86_expand_builtin): Ditto.
23041 2013-01-04 Richard Biener <rguenther@suse.de>
23043 PR tree-optimization/55862
23044 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
23045 translating them through PHI nodes.
23047 2013-01-04 Martin Jambor <mjambor@suse.cz>
23049 PR tree-optimization/55755
23050 * tree-sra.c (sra_modify_assign): Do not check that an access has no
23051 children when trying to avoid producing a VIEW_CONVERT_EXPR.
23053 2013-01-04 Marek Polacek <polacek@redhat.com>
23055 PR middle-end/55859
23056 * opts.c (default_options_optimization): Clarify error message.
23058 2013-01-04 Richard Biener <rguenther@suse.de>
23060 PR middle-end/55863
23061 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
23064 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23067 * config/pa/pa.md (movsi): Revert previous change.
23068 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
23071 2013-01-03 Richard Henderson <rth@redhat.com>
23073 * config/i386/i386.c (ix86_expand_move): Always assign to op1
23074 after eliminating TLS symbols.
23076 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
23079 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
23080 * graphite-poly.c (debug_gmp_value): Likewise.
23082 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
23085 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
23086 selected code model, define __code_mode_small__, __code_model_medium__,
23087 __code_model_large__, __code_model_32__ or __code_model_kernel__.
23088 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
23089 xchg temporary register with %k. Declare temporary register as
23091 [__x86_64__]: For medium and large code models, preserve %rbx register.
23093 2013-01-03 Richard Biener <rguenther@suse.de>
23095 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
23096 (dump_subscript): Adjust.
23097 (finalize_ddr_dependent): Do not dump redundant info.
23098 (analyze_siv_subscript): Adjust.
23099 (subscript_dependence_tester): Likewise.
23100 (compute_affine_dependence): Likewise.
23102 2013-01-03 Richard Biener <rguenther@suse.de>
23105 2013-01-03 Richard Biener <rguenther@suse.de>
23107 PR tree-optimization/55857
23108 * tree-vect-stmts.c (vectorizable_load): Do not setup
23109 re-alignment for invariant loads.
23111 2013-01-02 Richard Biener <rguenther@suse.de>
23113 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
23114 invariant load do not generate a vector load from the scalar location.
23116 2013-01-03 Richard Biener <rguenther@suse.de>
23118 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
23119 for not vectorizing.
23120 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
23121 not build INDIRECT_REFs, call get_name once only.
23122 (vect_create_data_ref_ptr): Likewise. Dump base object kind
23123 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
23125 2013-01-03 Richard Biener <rguenther@suse.de>
23127 PR tree-optimization/55857
23128 * tree-vect-stmts.c (vectorizable_load): Do not setup
23129 re-alignment for invariant loads.
23131 2013-01-03 Richard Biener <rguenther@suse.de>
23134 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
23135 prefer a built-in decl.
23137 2013-01-03 Jakub Jelinek <jakub@redhat.com>
23139 * gcc.c (process_command): Update copyright notice dates.
23140 * gcov.c (print_version): Likewise.
23141 * gcov-dump.c (print_version): Likewise.
23143 PR rtl-optimization/55838
23144 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
23145 iv0.step, iv1.step and step.
23147 2013-01-03 Jakub Jelinek <jakub@redhat.com>
23148 Marc Glisse <marc.glisse@inria.fr>
23150 PR tree-optimization/55832
23151 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
23152 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
23153 integer_{one,zero}_node.
23155 2013-01-03 Jakub Jelinek <jakub@redhat.com>
23158 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
23159 * var-tracking.c (reverse_op): Don't add reverse ops to
23160 VALUEs that have already
23161 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
23163 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
23165 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
23167 2013-01-02 Teresa Johnson <tejohnson@google.com>
23169 * dumpfile.c (dump_loc): Print filename with location.
23170 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
23171 new location_t parameter to emit complete unroll message with
23172 new dump framework.
23173 (canonicalize_loop_induction_variables): Compute loops location
23174 and pass to try_unroll_loop_completely.
23175 * loop-unroll.c (report_unroll_peel): New function.
23176 (peel_loops_completely): Use new dump format with location
23177 for main dumpfile message, and invoke report_unroll_peel on success.
23178 (decide_unrolling_and_peeling): Ditto.
23179 (decide_peel_once_rolling): Remove old dumpfile message subsumed
23180 by report_unroll_peel.
23181 (decide_peel_completely): Ditto.
23182 (decide_unroll_constant_iterations): Ditto.
23183 (decide_unroll_runtime_iterations): Ditto.
23184 (decide_peel_simple): Ditto.
23185 (decide_unroll_stupid): Ditto.
23186 * cfgloop.c (get_loop_location): New function.
23187 * cfgloop.h (get_loop_location): Declare.
23189 2013-01-02 Sriraman Tallam <tmsriram@google.com>
23191 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
23194 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23196 PR middle-end/55198
23197 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
23198 BLKmode objects when EXPAND_MEMORY is specified.
23200 2013-01-02 Sriraman Tallam <tmsriram@google.com>
23202 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
23204 (fold_builtin_cpu): Do not share cpu model decls across statements.
23206 2013-01-02 Jason Merrill <jason@redhat.com>
23209 * tree.c (build_array_type_1): Revert earlier change.
23211 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
23213 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
23215 * config/aarch64/aarch64-tune.md: Re-generate.
23217 2013-01-02 Richard Biener <rguenther@suse.de>
23219 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
23220 invariant load do not generate a vector load from the scalar location.
23222 2013-01-02 Richard Biener <rguenther@suse.de>
23225 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
23226 * configure: Regenerate.
23228 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
23230 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
23231 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
23232 (expand_builtin_int_roundingfn_2): Keep the original target around
23233 for the fallback case.
23235 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
23237 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
23238 to be clear for sign changes.
23240 2013-01-01 Jan Hubicka <jh@suse.cz>
23242 * ipa-inline-analysis.c: Fix formatting.
23244 2013-01-01 Jakub Jelinek <jakub@redhat.com>
23246 PR tree-optimization/55831
23247 * tree-vect-loop.c (get_initial_def_for_induction): Use
23248 gsi_after_labels instead of gsi_start_bb.
23250 Copyright (C) 2013 Free Software Foundation, Inc.
23252 Copying and distribution of this file, with or without modification,
23253 are permitted in any medium without royalty provided the copyright
23254 notice and this notice are preserved.