This moves the fold rules for trunc, floor, ceil, round, nearbyint and
rint in one go, since they're tested as a group. Most of the code is
supporting the f(x)->x fold when x is known to be integer-valued.
Like with the non-negative test, this is probably more elegantly handled
by tracking range information for reals, but until that happens, I think
we should handle it analogously to tree_expr_nonnegative_p.
I've incorporated the fix for PR68031 in the new version of
integer_valued_real_p. However, it seemed confusing to test for an
SSA name at the head of the function rather than the case statement,
and not fall through to tree_simple_nonnegative_warnv_p (which
conceptually shouldn't care whether an update is in progress).
But tree_simple_nonnegative_warnv_p is a no-op for SSA names,
so I simply changed it to:
return (!name_registered_for_update_p (t)
&& depth < PARAM_VALUE (PARAM_MAX_SSA_NAME_QUERY_DEPTH)
&& gimple_stmt_nonnegative_warnv_p (SSA_NAME_DEF_STMT (t),
strict_overflow_p, depth));
and used that in the new code too.
Doing these folds later meant that IPA would start to use information
about the aborting sinf and floor in 20030125-1.c before the folds
kicked in. I changed them from noinline to weak to stop that.
Tested on x86_64-linux-gnu, aarch64-linux-gnu and arm-linux-gnueabi.
gcc/
* builtins.c (integer_valued_real_p): Move to fold-const.c.
(fold_trunc_transparent_mathfn, fold_builtin_trunc, fold_builtin_floor)
(fold_builtin_ceil, fold_builtin_round): Delete.
(fold_builtin_1): Handle constant trunc, floor, ceil and round
arguments here.
* convert.c (convert_to_real): Remove narrowing of rounding
functions.
* fold-const.h (integer_valued_real_unary_p)
(integer_valued_real_binary_p, integer_valued_real_call_p)
(integer_valued_real_single_p, integer_valued_real_p): Declare.
* fold-const.c (tree_single_nonnegative_warnv_p): Move
name_registered_for_update_p check to SSA_NAME case statement.
Don't call tree_simple_nonnegative_warnv_p for SSA names.
(integer_valued_real_unary_p, integer_valued_real_binary_p)
(integer_valued_real_call_p, integer_valued_real_single_p)
(integer_valued_real_invalid_p): New functions.
(integer_valued_real_p): Move from fold-const.c and rework
to call the functions above. Handle SSA names.
* gimple-fold.h (gimple_stmt_integer_valued_real_p): Declare.
* gimple-fold.c (gimple_assign_integer_valued_real_p)
(gimple_call_integer_valued_real_p, gimple_phi_integer_valued_real_p)
(gimple_stmt_integer_valued_real_p): New functions.
* match.pd: Fold f(f(x))->f(x) for fp->fp rounding functions f.
Fold f(x)->x for the same f if x is known to be integer-valued.
Fold f(extend(x))->extend(f'(x)) if doing so doesn't affect
the result. Canonicalize floor(x) as trunc(x) if x is
nonnegative.
gcc/testsuite/
* gcc.c-torture/execute/20030125-1.c (floor, floorf, sin, sinf):
Make weak rather than noinline.
* gcc.dg/builtins-57.c: Compile with -O.
* gcc.dg/torture/builtin-integral-1.c: Skip for -O0.
From-SVN: r229221
2015-10-23 Tom de Vries <tom@codesourcery.com>
* tree-ssa-structalias.c (create_variable_info_for_1): Add missing
setting of is_full_var in case of a single field.
From-SVN: r229218
gcc/ChangeLog:
2015-10-22 Martin Sebor <msebor@redhat.com>
PR driver/68043
* config/i386/i386.opt: Add missing periods to the ends of sentences.
* config/msp430/msp430.opt: Same.
gcc/testsuite/ChangeLog:
2015-10-22 Martin Sebor <msebor@redhat.com>
PR driver/68043
* gcc.misc-tests/help.exp: Verify that option descriptions
end in periods.
* lib/options.exp (check_for_options): Use the regexp --line option.
Print unexpected match on failure.
From-SVN: r229205
[PATCH] Fix abort in write_eligible_delay
* genattrtab.c (main): If we do not have any annul-true or annul-false
slots, then write out a dummy eligible_for_annul_true or
eligible_for_annul_false as needed.
From-SVN: r229184
* config/msp430/msp430.opt: Add -msilicon-errata and
-msilicon-errata-warn.
* config/msp430/msp430.h (ASM_SPEC): Pass new options on to
assembler.
* doc/invoke.texi: Document new options.
From-SVN: r229177
2015-10-22 Richard Biener <rguenther@suse.de>
PR tree-optimization/58497
* tree-vect-generic.c (ssa_uniform_vector_p): New helper.
(expand_vector_operations_1): Use it. Lower operations on
all uniform vectors to scalar operations if the HW supports it.
* gcc.dg/tree-ssa/vector-5.c: New testcase.
From-SVN: r229173
2015-10-22 Richard Biener <rguenther@suse.de>
PR middle-end/68046
PR middle-end/61893
* optabs.c (emit_libcall_block_1): Allow a NULL_RTX equiv.
(expand_binop): For -ftrapv optabs do not record an REG_EQUAL note.
(expand_unop): Likewise.
* gcc.dg/torture/ftrapv-2.c: New testcase.
From-SVN: r229170
libsanitizer/
PR bootstrap/68041
* configure.ac (link_sanitizer_common): Link against librt only if it
contains shm_open, required by sanitizers.
(CXX_ABI_NEEDED): Remove variable.
* configure: Regenerate.
* ubsan/Makefile.am (libubsan_la_LIBADD): Do not add -lc++abi anymore.
* ubsan/Makefile.in: Regenerate.
From-SVN: r229168
Also make build_complex_cproj available globally and use
it for the existing match.pd rules.
Tested on x86_64-linux-gnu, aarch64-linux-gnu and arm-linux-gnueabi.
gcc/
* builtins.c (fold_builtin_cproj): Delete.
(fold_builtin_1): Handle constant arguments here.
(build_complex_cproj): Move and rename to...
* tree.c: (build_complex_inf): ...this.
* tree.h (build_complex_inf): Declare.
* match.pd: Fold cproj(x)->x if x has no infinity.
Use build_complex_inf for existing cproj rules.
From-SVN: r229164
gcc/ChangeLog:
2015-10-22 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
PR target/68015
* config/s390/s390.md (mov<mode>cc): Emit compare only if we don't
already have a comparison result.
gcc/testsuite/ChangeLog:
2015-10-22 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
PR target/68015
* gcc.target/s390/pr68015.c: New test.
From-SVN: r229163
This patch addresses PR-67948 by changing the xor-and.c test,
initially written for a simplify-rtx pattern, to make it pass post
r228661 (see
https://gcc.gnu.org/ml/gcc-patches/2015-10/msg00676.html). This test
no longer triggered the simplify-rtx pattern it was written for prior
to r228661, though other optimizations did lead to the same assembly
the test checked for. The optimization added with r228661 matches the
pattern used in the test and optimizes it to a better and still valid
sequence. Being unable to easily change the test to trigger the
original simplify-rtx pattern, I chose to change it to pass with the
new produced assembly sequence.
This is correct because the transformation is valid and it yields a
more efficient pattern. However, as I pointed out before this test
doesn't test the optimization it originally was intended for.
Tested by running regression tests for armv6.
Is this OK to commit?
Thanks, Andre
2015-10-22 Andre Vieira <andre.simoesdiasvieira@arm.com>
PR testsuite/67948
* gcc.target/arm/xor-and.c: Adjust test to check for eor
instead of orr.
From-SVN: r229161
Jiong pointed out privately that there was a thinko
in the way in which the global state was being
set and reset. I don't like adding such
global state but ....
2015-10-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
PR target/63304
* config/aarch64/aarch64.c (aarch64_nopcrelative_literal_loads): New.
(aarch64_expand_mov_immediate): Use aarch64_nopcrelative_literal_loads.
(aarch64_classify_address): Likewise.
(aarch64_secondary_reload): Likewise.
(aarch64_override_options_after_change_1): Adjust.
* config/aarch64/aarch64.md (aarch64_reload_movcp<GPF_TF:mode><P:mode>):
Use aarch64_nopcrelative_literal_loads.
(aarch64_reload_movcp<VALL:mode><P:mode>): Likewise.
* config/aarch64/aarch64-protos.h (aarch64_nopcrelative_literal_loads):
Declare
2015-10-22 Jiong Wang <jiong.wang@arm.com>
Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
PR target/63304
* gcc.target/aarch64/pr63304_1.c: New test.
From-SVN: r229160
/cp
2015-10-21 Paolo Carlini <paolo.carlini@oracle.com>
PR c++/66781
* parser.c (cp_parser_enum_specifier): Upon error_at set
nested_name_specifier to error_mark_node; improve error message.
/testsuite
2015-10-21 Paolo Carlini <paolo.carlini@oracle.com>
PR c++/66781
* g++.dg/parse/enum13.C: New.
/cp
2015-10-21 Paolo Carlini <paolo.carlini@oracle.com>
PR c++/67847
* parser.c (cp_parser_enum_specifier): Reject a nested_name_specifier
which doesn't name a class or namespace.
/testsuite
2015-10-21 Paolo Carlini <paolo.carlini@oracle.com>
PR c++/67847
* g++.dg/parse/enum12.C: New.
From-SVN: r229156
2015-10-21 Steven G. Kargl <kargl@gcc.gnu.org>
PR fortran/67939
* data.c (create_character_initializer): Deal with zero length string.
2015-10-21 Steven G. Kargl <kargl@gcc.gnu.org>
PR fortran/67939
* gfortran.dg/pr67939.f90: New test.
From-SVN: r229153
When a scalar is defined in function of an array reference in the current scop,
it does variate. Graphite cannot represent the condition in scop-11.c, as a[*]
variates in the current region, and it is not an affine condition:
for (j = 0; j <= 20; j++)
a[j] = b + i;
if (a[12] == 23)
b = 3;
else
b = 1;
for (j = 0; j <= 20; j++)
a[j] = b + i;
* graphite-scop-detection.c (parameter_index_in_region): Update call to
invariant_in_sese_p_rec.
* graphite-sese-to-poly.c (extract_affine): Same.
* sese.c (invariant_in_sese_p_rec): Pass in an extra parameter has_vdefs.
(scalar_evolution_in_region): Return chrec_dont_know when the scalar variable
depends on virtual definitions in the current region.
* sese.h (invariant_in_sese_p_rec): Update declaration.
testsuite/
* gcc.dg/graphite/scop-11.c: Update pattern.
Co-Authored-By: Sebastian Pop <s.pop@samsung.com>
From-SVN: r229151
* graphite-scop-detection.c (build_scops): Do not handle scops with more
than PARAM_GRAPHITE_MAX_ARRAYS_PER_SCOP arrays.
* params.def (PARAM_GRAPHITE_MAX_ARRAYS_PER_SCOP): New.
Co-Authored-By: Sebastian Pop <s.pop@samsung.com>
From-SVN: r229150
PR ipa/67056
* ipa-polymorphic-call.c (possible_placement_new): If cur_offset
is negative we don't know the type.
(check_stmt_for_type_change): Skip constructors of non-polymorphic
types as those won't help devirutalization.
* g++.dg/ipa/pr67056.C: New testcase.
From-SVN: r229148
If all address spaces use the same modes and forms, we would
be forced to replicate these hooks in the backend. Which would
then require the creation of a new hook to replace
target_default_pointer_address_modes_p.
* targhooks.c (default_addr_space_pointer_mode): Remove check
for generic address space.
(default_addr_space_address_mode): Likewise.
(default_addr_space_valid_pointer_mode): Likewise.
(default_addr_space_legitimate_address_p): Likewise.
(default_addr_space_legitimize_address): Likewise.
From-SVN: r229143
The current default of making all undefined coversions being
set to null is not useful. It has caused all users to lie
and say that spaces are subsets when they are not, just so
that they can override the conversion.
* expr.c (expand_expr_real_2): Use convert_modes on disjoint
address spaces.
From-SVN: r229142
The fold code also expanded cabs(x+yi) to fsqrt(x*x+y*y) when optimising
for speed. tree-ssa-math-opts.c has this transformation too, but unlike
the fold code, it first checks whether the target implements the sqrt
optab. The patch simply removes the fold code and keeps the
tree-ssa-math-opts.c logic the same.
gcc.dg/lto/20110201-1_0.c was relying on us replacing cabs
with fsqrt even on targets where fsqrt is itself a library call.
The discussion leading up to that patch suggested that we only
want to test the fold on targets with a square root instruction,
so it would be OK to skip the test on other targets:
https://gcc.gnu.org/ml/gcc-patches/2011-07/msg01961.htmlhttps://gcc.gnu.org/ml/gcc-patches/2011-07/msg02036.html
The patch does that using the sqrt_insn effective target.
It's possible that removing the tree folds renders the LTO trick
unnecessary, but since the test was originally for an ICE, it seems
better to leave it as-is.
Tested on x86_64-linux-gnu, aarch64-linux-gnu and arm-linux-gnueabi.
20110201-1_0.c passes on all three.
gcc/
* builtins.c (fold_builtin_cabs): Delete.
(fold_builtin_1): Update accordingly. Handle constant arguments here.
* match.pd: Add rules previously handled by fold_builtin_cabs.
gcc/testsuite/
* gcc.dg/lto/20110201-1_0.c: Restrict to sqrt_insn targets.
Add associated options for arm*-*-*.
(sqrt): Remove dummy definition.
From-SVN: r229141
Richard Sandiford
Paulo Matos
Tom de Vries
Alan Hayward
Maxim Ostapenko
Nick Clifton
Max Ostapenko
Jason Merrill
Jason Merrill
GCC Administrator
Martin Sebor
Uros Bizjak
Martin Sebor
David Wohlferd
Mikhail Maltsev
Jeff Law
Nick Clifton
Paolo Carlini
Richard Biener
Uros Bizjak
Andreas Krebbel
Andre Vieira
Ramana Radhakrishnan
Steven G. Kargl
Jonathan Wakely
Aditya Kumar
Jan Hubicka
Eric Botcazou
Trevor Saunders
Richard Henderson