[gcc]
2016-08-12 Michael Meissner <meissner@linux.vnet.ibm.com>
* config/rs6000/vsx.md (vsx_concat_<mode>): Add support for the
ISA 3.0 MTVSRDD instruction.
(vsx_splat_<mode>): Change cpu type of MTVSRDD instruction to
vecperm.
[gcc/testsuite]
2016-08-12 Michael Meissner <meissner@linux.vnet.ibm.com>
* gcc.target/powerpc/vec-init-1.c: New tests to test various
vector initialization options.
* gcc.target/powerpc/vec-init-2.c: Likewise.
* gcc.target/powerpc/vec-init-3.c: New test to make sure MTVSRDD
is generated on ISA 3.0.
From-SVN: r239428
2016-08-12 Bernd Edlinger <bernd.edlinger@hotmail.de>
PR tree-optimization/71083
* tree-predcom.c (ref_at_iteration): Use a COMPONENT_REF for the
bitfield access when possible.
From-SVN: r239426
gcc/ChangeLog:
PR middle-end/71654
* match.pd ((T)A CMP (T)B -> A CMP B): Allow (T)A to be a
sign-changing cast from a shorter unsigned type to a wider
signed type.
gcc/testsuite/ChangeLog:
PR middle-end/71654
* gcc.dg/c-c++-common/pr71654.c: New test.
* gcc.dg/tree-ssa/vrp23: Add -fno-tree-forwprop to
dg-options.
* gcc.dg/tree-ssa/vrp24: Likewise.
From-SVN: r239421
PR c/67410
* c-typeck.c (set_nonincremental_init_from_string): Use / instead of
% to determine val element to change. Assert that
wchar_bytes * charwidth fits into val array.
* gcc.dg/pr67410.c: New test.
Co-Authored-By: Martin Liska <mliska@suse.cz>
From-SVN: r239419
* config/i386/sse.md (avx512dq_vextract<shuffletype>64x2_1_maskm,
vec_extract_hi_<mode>_mask, vec_extract_lo_<mode>_mask,
vec_extract_hi_<mode>_mask): Use Yk constraint instead of k.
From-SVN: r239418
PR tree-optimization/69848
* tree-vectorizer.h (enum vect_def_type): New condition reduction
type CONST_COND_REDUCTION.
* tree-vect-loop.c (vectorizable_reduction): Support new condition
reudction type CONST_COND_REDUCTION.
gcc/testsuite
PR tree-optimization/69848
* gcc.dg/vect/vect-pr69848.c: New test.
From-SVN: r239416
PR testsuite/71008
* gcc.target/i386/avx512f-vprord-1.c: Scan also for vprold.
* gcc.target/i386/avx512f-vprorq-1.c: Scan also for vprolq.
* gcc.target/i386/avx512vl-vprord-1.c: Scan also for vprold.
* gcc.target/i386/avx512vl-vprorq-1.c: Scan also for vprolq.
From-SVN: r239415
2016-08-12 Richard Biener <rguenther@suse.de>
PR tree-optimization/57326
* tree-ssa-pre.c (fully_constant_expression): Handle simplification
returning an SSA name.
(phi_translate_1): When fully_constant_expression returns a NAME
make sure we have a leader for it.
* gcc.dg/tree-ssa/ssa-pre-32.c: New testcase.
* gcc.dg/tree-ssa/loadpre14.c: Adjust.
* gcc.dg/tree-ssa/pr35287.c: Likewise.
* gcc.target/i386/pr45685.c: Likewise.
* gcc.dg/tree-ssa/predcom-1.c: Disable PRE.
* gcc.dg/tree-ssa/predcom-2.c: Likewise.
* gcc.dg/tree-ssa/predcom-3.c: Likewise.
* gcc.dg/tree-ssa/ssa-sink-10.c: Likewise.
* gfortran.dg/pr34163.f90: Likewise.
From-SVN: r239414
* g++.dg/gcov/gcov-16.C: New test.
* lib/gcov.exp: Support new argument for run-gcov function.
Adam Fineman <afineman@afineman.com>
* gcov.c (process_file): Create .gcov file when .gcda
file is missing.
From-SVN: r239413
2016-08-12 Richard Biener <rguenther@suse.de>
* tree-vrp.c (vrp_visit_phi_node): Allow a last iteration if
the currently executable edges have fixed ranges. Always
go through update_value_range.
From-SVN: r239404
Although C++ reference types, denoted by DW_TAG_reference_type in
DWARFv2+ debug info, are unchangeable, we output names of reference type
with DW_TAG_const_type, because internally we mark such variables as
TREE_READONLY. That's an internal implementation detail that shouldn't
leak to debug information. This patch fixes this.
for gcc/ChangeLog
PR debug/55641
* dwarf2out.c (decl_quals): Don't map TREE_READONLY to
TYPE_QUAL_CONST in reference-typed decls.
for gcc/testsuite/ChangeLog
PR debug/55641
* g++.dg/debug/dwarf2/ref-1.C: New.
From-SVN: r239402
We used to emit, in debug information, the values bound to pointer to
member function template parameters only when they were NULL or
virtual member functions, because those can be represented with
DW_AT_const_value.
In order to represent the symbolic pointer to member function
constants for non-virtual member functions, we'd need to be able to
emit relocations for part of DW_AT_const_value, which we don't. The
more viable alternative is to use DW_AT_location to represent such
values, as slated for inclusion in DWARFv5, according to
<URL:http://www.dwarfstd.org/ShowIssue.php?issue=130412.1>.
With this patch, when we can't emit a DW_AT_const_value, we emit each
"member" of the pointer to member function "record" as a
DW_OP_stack_value DW_OP_piece, as long as the referenced member
function is output in the same translation unit, otherwise we'd get
relocations to external symbols, something to avoid in debug sections.
for gcc/ChangeLog
PR debug/49366
* dwarf2out.c (loc_list_from_tree_1): Expand some CONSTRUCTORs
in DW_OP_pieces, just enough to handle pointers to member
functions.
(gen_remaining_tmpl_value_param_die_attribute): Use a location
expression on DWARFv5 if a constant value doesn't work.
for gcc/testsuite/ChangeLog
PR debug/49366
* g++.dg/debug/dwarf2/template-params-12.H: New.
* g++.dg/debug/dwarf2/template-params-12f.C: New.
* g++.dg/debug/dwarf2/template-params-12g.C: New.
* g++.dg/debug/dwarf2/template-params-12n.C: New.
* g++.dg/debug/dwarf2/template-params-12s.C: New.
* g++.dg/debug/dwarf2/template-params-12u.C: New.
* g++.dg/debug/dwarf2/template-params-12v.C: New.
* g++.dg/debug/dwarf2/template-params-12w.C: New.
From-SVN: r239401
2016-08-11 Richard Biener <rguenther@suse.de>
Bill Schmidt <wschmidt@linux.vnet.ibm.com>
PR rtl-optimization/72855
* df-core.c (df_verify): Turn off DF_VERIFY_SCHEDULED at end.
Co-Authored-By: Bill Schmidt <wschmidt@linux.vnet.ibm.com>
From-SVN: r239395
We want to build dir_regfile.go if not GNU/linux, and not solaris/386,
and not solaris/sparc. The latter two conditions were incorrect. To
write ! solaris/386 we have to write !solaris !386. I forgot De
Morgan's Law.
Reviewed-on: https://go-review.googlesource.com/26870
From-SVN: r239393
Support TImode CONST_WIDE_INT store generated from piecewise store.
Need to verify performance impact before enabling TImode CONST_INT
store for __int128.
gcc/
* config/i386/i386.c (timode_scalar_to_vector_candidate_p): Allow
TImode CONST_WIDE_INT store.
(timode_scalar_chain::convert_insn): Handle CONST_WIDE_INT store.
gcc/testsuite/
* gcc.target/i386/pieces-strcpy-1.c: New test.
* gcc.target/i386/pieces-strcpy-2.c: Likewise.
From-SVN: r239383
Use TImode for piecewise move in 64-bit mode. We should use TImode in
32-bit mode and use OImode or XImode if they are available. But since
by_pieces_ninsns determines the widest mode with MAX_FIXED_MODE_SIZE,
we can only use TImode in 64-bit mode.
gcc/
* config/i386/i386.h (MOVE_MAX_PIECES): Use TImode in 64-bit
mode if unaligned SSE load and store are optimal.
gcc/testsuite/
* gcc.target/i386/pieces-memcpy-1.c: New test.
* gcc.target/i386/pieces-memcpy-2.c: Likewise.
* gcc.target/i386/pieces-memcpy-3.c: Likewise.
* gcc.target/i386/pieces-memcpy-4.c: Likewise.
* gcc.target/i386/pieces-memcpy-5.c: Likewise.
* gcc.target/i386/pieces-memcpy-6.c: Likewise.
From-SVN: r239378
PR c/72816
* c-decl.c (grokdeclarator): When adding TYPE_DOMAIN for flexible
array member through typedef, for orig_qual_indirect == 0 clear
orig_qual_type.
* gcc.dg/pr72816.c: New test.
From-SVN: r239371
gcc/ChangeLog:
2016-08-11 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
* config/s390/s390-builtins.def: Mark last operand of s390_vlvg*
and s390_vlgv* builtins as element selector.
From-SVN: r239360
2016-08-11 Richard Biener <rguenther@suse.de>
PR tree-optimization/72772
* cfgloopmanip.c (create_preheader): Use split_edge if there
is a single loop entry, avoiding degenerate PHIs.
* gcc.dg/graphite/pr35356-1.c: Adjust.
* gcc.dg/tree-ssa/pr59597.c: Likewise.
From-SVN: r239357
frontend:
2016-08-11 Janne Blomqvist <jb@gcc.gnu.org>
* check.c (gfc_check_random_seed): Use new seed size in check.
* intrinsic.texi (RANDOM_NUMBER): Updated documentation.
(RANDOM_SEED): Likewise.
testsuite:
2016-08-11 Janne Blomqvist <jb@gcc.gnu.org>
* gfortran.dg/random_7.f90: Take into account that the last seed
value is the special p value.
* gfortran.dg/random_seed_1.f90: Seed size is now constant.
libgfortran:
2016-08-11 Janne Blomqvist <jb@gcc.gnu.org>
* intrinsics/random.c: Replace KISS with xorshift1024* using
per-thread state.
* runtime/main.c (init): Don't call random_seed_i4.
From-SVN: r239356
2016-08-11 Richard Biener <rguenther@suse.de>
* tree-ssa-threadbackward.c (pass_data_thread_jumps): Remove
unconditional TODO_cleanup_cfg.
(pass_thread_jumps::execute): Initialize loops, perform a CFG
cleanup only if we threaded a jump.
* gcc.dg/tree-ssa/ssa-dom-thread-7.c: Adjust.
From-SVN: r239355
pr71680.c -m64 -O1 -mlra, ira output showing two problem insns.
(insn 7 5 26 3 (set (reg:SI 159 [ a ])
(mem/c:SI (reg/f:DI 158) [1 a+0 S4 A8])) pr71680.c:13 464 {*movsi_internal1}
(expr_list:REG_EQUIV (mem/c:SI (reg/f:DI 158) [1 a+0 S4 A8])
(nil)))
(insn 26 7 27 3 (set (reg:DI 162)
(unspec:DI [
(fix:SI (subreg:SF (reg:SI 159 [ a ]) 0))
] UNSPEC_FCTIWZ)) pr71680.c:13 372 {fctiwz_sf}
(expr_list:REG_DEAD (reg:SI 159 [ a ])
(nil)))
Insn 26 requires that reg 159 be of class FLOAT_REGS.
first lra action:
deleting insn with uid = 7.
Changing pseudo 159 in operand 1 of insn 26 on equiv [r158:DI]
Creating newreg=164, assigning class ALL_REGS to subreg reg r164
26: r162:DI=unspec[fix(r164:SI#0)] 7
REG_DEAD r159:SI
Inserting subreg reload before:
30: r164:SI=[r158:DI]
[snip]
Change to class FLOAT_REGS for r164
Well, that didn't do much. lra tried the equiv mem, found that didn't
work, and had to reload. Effectively getting back to the two original
insns but r159 replaced with r164. simplify_operand_subreg did not do
anything in this case because SLOW_UNALIGNED_ACCESS was true (wrongly
for power8, but that's beside the point). So now we have, using
abbreviated rtl notation:
r164:SI=[r158:DI]
r162:DI=unspec[fix(r164:SI)]
The problem here is that the first insn isn't valid, due to the rs6000
backend not supporting SImode in fprs, and r164 must be an fpr to make
the second insn valid.
next lra action:
Creating newreg=165 from oldreg=164, assigning class GENERAL_REGS to r165
30: r165:SI=[r158:DI]
Inserting insn reload after:
31: r164:SI=r165:SI
so now we have
r165:SI=[r158:DI]
r164:SI=r165:SI
r162:DI=unspec[fix(r164:SI)]
This ought to be good on power8, except for one little thing.
r165 is GENERAL_REGS so the first insn is good, a gpr load from mem.
r164 is FLOAT_REGS, making the last insn good, a fctiwz.
The second insn ought to be a sldi, mtvsrd, xscvspdpn combination, but
that is only supported for SFmode. So lra continue on reloading the
second insn, but in vain because it never tries anything other than
SImode and as noted above, SImode is not valid in fprs.
What this patch does is arrange to emit the two reloads needed for the
SLOW_UNALIGNED_ACCESS case at once, moving the subreg to the second
insn in order to switch modes, producing:
r164:SI=[r158:DI]
r165:SF=r164:SI#0
r162:DI=unspec[fix(r165:SF)]
I've also tidied a couple of other things:
1) "old" is unnecessary as it duplicated "operand".
2) Rejecting mem subregs due to SLOW_UNALIGNED_ACCESS only makes sense
if the original mode was not slow.
PR target/71680
* lra-constraints.c (simplify_operand_subreg): Allow subreg
mode for mem when SLOW_UNALIGNED_ACCESS if inner mode is also
slow. Emit two reloads for slow mem case, first loading in
fast innermode, then converting to required mode.
testsuite/
* gcc.target/powerpc/pr71680.c: New.
From-SVN: r239342
* cp-tree.h (IF_STMT_CONSTEXPR_P): New.
* name-lookup.c (push_to_top_level, pop_from_top_level_1): Handle it.
* parser.h (struct cp_parser): Add in_discarded_stmt field.
* parser.c (cp_parser_selection_statement): Handle 'if constexpr'.
(cp_parser_jump_statement): Avoid deducing from a discarded return.
* pt.c (tsubst_expr): Only instantiate taken branch of constexpr if.
* semantics.c (begin_if_stmt): Set the binding level this_entity.
(finish_if_stmt_cond): Require the condition of a
constexpr if to be constant.
* decl.c (level_for_constexpr_if): New.
(named_label_entry): Add in_constexpr_if field.
(poplevel_named_label_1): Set it.
(check_goto): Check it.
(check_previous_goto_1): Check level_for_constexpr_if.
From-SVN: r239338
