Const int is handled differently at -O0 for -xc and -xc++, which can cause noise
in testsuite/libgomp.oacc-c-c++-common test-cases (which are both run for c and
c++) if const int is used for launch dimensions.
Fix this by using #defines instead.
2019-01-09 Tom de Vries <tdevries@suse.de>
PR target/88756
* testsuite/libgomp.oacc-c-c++-common/reduction-1.c (ng, nw, vl): Use
#define instead of "const int".
* testsuite/libgomp.oacc-c-c++-common/reduction-2.c (ng, nw, vl): Same.
* testsuite/libgomp.oacc-c-c++-common/reduction-3.c (ng, nw, vl): Same.
* testsuite/libgomp.oacc-c-c++-common/reduction-4.c (ng, nw, vl): Same.
* testsuite/libgomp.oacc-c-c++-common/reduction-5.c (ng, nw, vl): Same.
From-SVN: r267747
When using a compiler build with:
...
+#define PTX_DEFAULT_VECTOR_LENGTH PTX_CTA_SIZE
+#define PTX_MAX_VECTOR_LENGTH PTX_CTA_SIZE
...
and running the libgomp testsuite, we run into an execution failure in
parallel-loop-1.c, due to a cuda launch failure:
...
nvptx_exec: kernel f6_none_none$_omp_fn$0: launch gangs=480, workers=0, \
vectors=1024
libgomp: cuLaunchKernel error: invalid argument
...
because workers == 0.
The workers variable is set to 0 here in nvptx_exec:
...
workers = blocks / actual_vectors;
...
because actual_vectors is 1024, and blocks is 768:
...
cuOccupancyMaxPotentialBlockSize: grid = 10, block = 768
...
Fix this by ensuring that workers is at least one.
2019-01-09 Tom de Vries <tdevries@suse.de>
* plugin/plugin-nvptx.c (nvptx_exec): Make sure to launch with at least
one worker.
From-SVN: r267746
Slice value expression has backend type a struct of a pointer and
two ints. Make sure the len and cap are converted to int when
creating slice value expression.
Reviewed-on: https://go-review.googlesource.com/c/156897
From-SVN: r267745
Test that StdUniquePtrPrinter correctly prints std::unique_ptr objects
using the old layout, prior to the PR libstdc++/77990 changes.
The printer test for a valueless std::variant started to fail because
the PR libstdc++/87431 fix meant it no longer became valueless. Change
the test to use a type that is not trivially copyable, so that the
exception causes it to become valueless.
* testsuite/libstdc++-prettyprinters/compat.cc: Test printer support
for old std::unique_ptr layout.
* testsuite/libstdc++-prettyprinters/cxx17.cc: Fix std::variant test
to become valueless. Add filesystem::path tests.
From-SVN: r267743
When the contained value is not trivially copy (or move) constructible
the union's copy (or move) constructor will be deleted, and so the
_Optional_payload delegating constructors are invalid. G++ fails to
diagnose this because it incorrectly performs copy elision in the
delegating constructors. Clang does diagnose it (llvm.org/PR40245).
The solution is to avoid performing any copy (or move) when the
contained value's copy (or move) constructor isn't trivial. Instead the
contained value can be constructed by calling _M_construct. This is OK,
because the relevant constructor doesn't need to be constexpr when the
contained value isn't trivially copy (or move) constructible.
Additionally, this patch removes a lot of code duplication in the
_Optional_payload partial specializations and the _Optional_base partial
specialization, by hoisting it into common base classes.
The Python pretty printer for std::optional needs to be adjusted to
support the new layout. Retain support for the old layout, and add a
test to verify that the support still works.
PR libstdc++/87855
* include/std/optional (_Optional_payload_base): New class template
for common code hoisted from _Optional_payload specializations. Use
a template for the union, to allow a partial specialization for
types with non-trivial destructors. Add constructors for in-place
initialization to the union.
(_Optional_payload(bool, const _Optional_payload&)): Use _M_construct
to perform non-trivial copy construction, instead of relying on
non-standard copy elision in a delegating constructor.
(_Optional_payload(bool, _Optional_payload&&)): Likewise for
non-trivial move construction.
(_Optional_payload): Derive from _Optional_payload_base and use it
for everything except the non-trivial assignment operators, which are
defined as needed.
(_Optional_payload<false, C, M>): Derive from the specialization
_Optional_payload<true, false, false> and add a destructor.
(_Optional_base_impl::_M_destruct, _Optional_base_impl::_M_reset):
Forward to corresponding members of _Optional_payload.
(_Optional_base_impl::_M_is_engaged, _Optional_base_impl::_M_get):
Hoist common members from _Optional_base.
(_Optional_base): Make all members and base class public.
(_Optional_base::_M_get, _Optional_base::_M_is_engaged): Move to
_Optional_base_impl.
* python/libstdcxx/v6/printers.py (StdExpOptionalPrinter): Add
support for new std::optional layout.
* testsuite/libstdc++-prettyprinters/compat.cc: New test.
From-SVN: r267742
2019-01-08 Janus Weil <janus@gcc.gnu.org>
PR fortran/88047
* class.c (gfc_find_vtab): For polymorphic typespecs, the components of
the class container may not be available (in case of invalid code).
2019-01-08 Janus Weil <janus@gcc.gnu.org>
PR fortran/88047
* gfortran.dg/class_69.f90: New test case.
From-SVN: r267735
There is no need to generate vzeroupper if caller passes arguments in
AVX/AVX512 registers.
Tested on i686 and x86-64 with and without --with-arch=native.
gcc/
PR target/88717
* config/i386/i386.c (ix86_avx_u128_mode_exit): Call
ix86_avx_u128_mode_entry.
gcc/testsuite/
PR target/88717
* gcc.target/i386/pr88717.c: New test.
From-SVN: r267732
* parser.c (cp_debug_parser): Adjust printing of
local_variables_forbidden_p.
(cp_parser_new): Set local_variables_forbidden_p to 0 rather than false.
(cp_parser_primary_expression): When checking
local_variables_forbidden_p, use THIS_FORBIDDEN or
LOCAL_VARS_FORBIDDEN.
(cp_parser_lambda_body): Update the type of
local_variables_forbidden_p. Set it to 0 rather than false.
(cp_parser_condition): Adjust call to cp_parser_declarator.
(cp_parser_explicit_instantiation): Likewise.
(cp_parser_init_declarator): Likewise.
(cp_parser_declarator): New parameter. Use it.
(cp_parser_direct_declarator): New parameter. Use it to set
local_variables_forbidden_p. Adjust call to cp_parser_declarator.
(cp_parser_type_id_1): Adjust call to cp_parser_declarator.
(cp_parser_parameter_declaration): Likewise.
(cp_parser_default_argument): Update the type of
local_variables_forbidden_p. Set it to LOCAL_VARS_AND_THIS_FORBIDDEN
rather than true.
(cp_parser_member_declaration): Tell cp_parser_declarator if we saw
'static' or 'friend'.
(cp_parser_exception_declaration): Adjust call to cp_parser_declarator.
(cp_parser_late_parsing_default_args): Update the type of
local_variables_forbidden_p. Set it to LOCAL_VARS_AND_THIS_FORBIDDEN
rather than true.
(cp_parser_cache_defarg): Adjust call to cp_parser_declarator.
(cp_parser_objc_class_ivars): Likewise.
(cp_parser_objc_struct_declaration): Likewise.
(cp_parser_omp_for_loop_init): Likewise.
* parser.h (cp_parser): Change the type of local_variables_forbidden_p
to unsigned char.
(LOCAL_VARS_FORBIDDEN, LOCAL_VARS_AND_THIS_FORBIDDEN, THIS_FORBIDDEN):
Define.
* g++.dg/cpp0x/this1.C: New test.
From-SVN: r267731
2019-01-08 Martin Liska <mliska@suse.cz>
PR tree-optimization/88753
* tree-switch-conversion.c (switch_conversion::build_one_array):
Come up with local variable constructor. Convert first to
type of constructor values.
2019-01-08 Martin Liska <mliska@suse.cz>
PR tree-optimization/88753
* gcc.dg/tree-ssa/pr88753.c: New test.
From-SVN: r267728
Using #include "..." to include a header in the same directory fails if
the user compiles with -I-, so always use something like <bits/...> for
internal headers.
I haven't added tests for this, because dg-options adds options to the
end, and the position of -I- matters (if it's at the end then the tests
won't find any headers in the build tree, as they're specified by -I
options earlier in the flags). It's been manually tested though.
PR libstdc++/88066
* include/bits/locale_conv.h: Use <> for includes not "".
* include/ext/random: Likewise.
* include/ext/vstring.h: Likewise.
From-SVN: r267726
2019-01-08 Richard Biener <rguenther@suse.de>
PR tree-optimization/86554
* tree-ssa-sccvn.c (eliminate_dom_walker, rpo_elim,
rpo_avail): Move earlier.
(visit_nary_op): When value-numbering to expressions
with different overflow behavior make sure there's an
available expression on the path.
* gcc.dg/torture/pr86554-1.c: New testcase.
* gcc.dg/torture/pr86554-2.c: Likewise.
From-SVN: r267725
PR libstdc++/88749
* src/filesystem/ops.cc (last_write_time): Fix preprocessor condition
to match the one that controls whether utimbuf and utime are declared.
From-SVN: r267705
Following the discovery of regressions in GPRbuild, this reverts both
r263100 and r264608:
2019-01-08 Justin Squirek <squirek@adacore.com>
Revert:
2018-07-31 Justin Squirek <squirek@adacore.com>
gcc/ada/
* lib-writ.adb (Write_With_Lines): Modfiy the generation
of dependencies within ali files so that source unit
bodies are properly listed even if said bodies are
missing. Perform legacy behavior in GNATprove mode.
* lib-writ.ads: Modify documentation to reflect current
behavior.
and:
2018-09-26 Justin Squirek <squirek@adacore.com>
gcc/ada/
* lib-writ.adb, lib-writ.ads (Write_With_Lines): Add
documentation and an extra conditional check for RCI
units so that generated ali files will list the spec
only instead of a body when a body is not found.
From-SVN: r267680
2019-01-08 Iain Sandoe <iain@sandoe.co.uk>
gcc/testsuite/
* c-c++-common/builtin-has-attribute-3.c: Skip tests requiring symbol
alias support.
* c-c++-common/builtin-has-attribute-4.c: Likewise.
Append match for warning that ‘protected’ attribute is not supported.
From-SVN: r267674
PR c++/88554
* decl.c (finish_function): For -Wreturn-type don't add a return *this;
fixit hint if current_class_ref is NULL. Use a single if instead of
two nested ones.
* g++.dg/warn/Wreturn-type-11.C: New test.
Co-Authored-By: Jakub Jelinek <jakub@redhat.com>
From-SVN: r267672
Amongst other changes, r266077 updated value_range_base::dump so
that it additionally prints the type. This broke an assertion within
the jit testsuite, in jit.dg/test-sum-of-squares.c, which was checking
for:
": [-INF, n_"
but was now getting:
": signed int [-INF, n_"
The test is merely intended as a simple verification that we can read
dump files via gcc_jit_context_enable_dump.
This patch loosens the requirements on the dump so that it should work
with either version of value_range_base::dump.
gcc/testsuite/ChangeLog:
PR jit/88747
* jit.dg/test-sum-of-squares.c (verify_code): Update expected vrp
dump to reflect r266077.
From-SVN: r267671
PR c/88701
* c-decl.c (build_compound_literal): If not TREE_STATIC, only pushdecl
if current_function_decl is non-NULL.
* gcc.dg/pr88701.c: New test.
From-SVN: r267667
This patch started off just by adding if_then_else support in
write_attr_value to be able to write a saner expression for powerpc
tls_gdld_nomark length. Then I noticed bit-rot in functions used to
calculate insn_default_length, insn_min_length, and length_unit_log
(which are used by the shorten_branches pass). These functions
don't handle a const_int length value and return an "unknown" status
that isn't used, or in the case of or_attr_value, doesn't need to be
used. min_attr_value also attempts to return INT_MAX for the
unhandled rtl case, but this can get lost in recursive calls. I fixed
that problem by returning INT_MIN instead, and translating that to
INT_MAX in the only caller of min_attr_value.
PR target/88614
* genattrtab.c (max_attr_value, min_attr_value, or_attr_value):
Delete "unknownp" parameter. Adjust callers. Handle
CONST_INT, PLUS, MINUS, and MULT.
(attr_value_aligned): Renamed from or_attr_value.
(min_attr_value): Return INT_MIN for unhandled rtl case..
(min_fn): ..and translate to INT_MAX here.
(write_length_unit_log): Modify to cope without "unknown".
(write_attr_value): Handle IF_THEN_ELSE.
From-SVN: r267666
Bugs 88720 and 88726 report issues where a function is declared inline
in an inner scope, resulting in spurious diagnostics about it being
declared but never defined when that scope is left (possibly in some
cases also wrongly referring to the function as a nested function).
These are regressions that were introduced with the support for C99
inline semantics in 4.3 (they don't appear with 4.2; it's possible
some aspects of the bugs might have been introduced later than 4.3).
For the case of functions being wrongly referred to as nested,
DECL_EXTERNAL was not the right condition for a function being
non-nested; TREE_PUBLIC is appropriate for the case of non-nested
functions with external linkage, while !b->nested means this is the
outermost scope in which the function was declared and so avoids
catching the case of a file-scope static being redeclared inline
inside a function.
For the non-nested, external-linkage case, the code attempts to avoid
duplicate diagnostics by diagnosing only when scope != external_scope,
but actually scope == external_scope is more appropriate, as it's only
when the file and external scopes are popped that the code can
actually tell whether a function ended up being defined, and all such
functions will appear in the (GCC-internal) external scope.
Bootstrapped with no regressions on x86_64-pc-linux-gnu.
PR c/88720
PR c/88726
gcc/c:
* c-decl.c (pop_scope): Use TREE_PUBLIC and b->nested to determine
whether a function is nested, not DECL_EXTERNAL. Diagnose inline
functions declared but never defined only for external scope, not
for other scopes.
gcc/testsuite:
* gcc.dg/inline-40.c, gcc.dg/inline-41.c: New tests.
From-SVN: r267665
This is following CL 156038. doscanstackswitch uses the same
mechanism of switching goroutines as getTraceback, and so has
the same problem as described in issue golang/go#29448. This CL
applies the same fix.
Reviewed-on: https://go-review.googlesource.com/c/156697
From-SVN: r267661
This is the gccgo version of https://golang.org/cl/141822:
Only return a pointer p to the new slices backing array from makeslice.
Makeslice callers then construct sliceheader{p, len, cap} explictly
instead of makeslice returning the slice.
This change caused the GCC backend to break the runtime/pprof test by
merging together the identical functions allocateReflectTransient and
allocateTransient2M. This caused the traceback to be other than
expected. Fix that by making the functions not identical.
This is a step toward updating libgo to the Go1.12beta1 release.
Reviewed-on: https://go-review.googlesource.com/c/155937
From-SVN: r267660
Currently, when collecting a traceback for another goroutine,
getTraceback calls gogo(gp) switching to gp, which will resume in
mcall, which will call gtraceback, which will set up gp->m. There
is a gap between setting the current running g to gp and setting
gp->m. If a profiling signal arrives in between, sigtramp will
see a non-nil gp with a nil m, and will seg fault. Fix this by
setting up gp->m first.
Fixesgolang/go#29448.
Reviewed-on: https://go-review.googlesource.com/c/156038
From-SVN: r267658
The IFN_GOMP_SIMD_LANE handling in vectorizable_store tries to use MEM_REF
offsets to maintain pointer disambiguation info. This patch makes sure
that we don't try to do the same optimisation for IFN_MASK_STOREs, which
have no similar offset argument.
The patch fixes libgomp.c-c++-common/pr66199-*.c for SVE. Previously
we had an ncopies==2 store and stored both halves to the same address.
2019-01-07 Richard Sandiford <richard.sandiford@arm.com>
gcc/
* tree-vect-stmts.c (vectorizable_store): Don't use the dataref_offset
optimization for masked stores.
From-SVN: r267654
The use of "j" in:
init = permute_results[number_of_vectors - j - 1];
was out-of-sync with the new flat loop structure. Now that all that
reversing is gone, we can just use the result of duplicate_and_interleave
directly.
The other cases shouldn't be affected by postponing the insertion
of ctor_seq, since gimple_build* appends to the seq without clearing
it first (unlike some of the gimplify routines).
The ICE is already covered by gcc.dg/vect/pr63379.c.
2019-01-07 Richard Sandiford <richard.sandiford@arm.com>
gcc/
PR middle-end/88567
* tree-vect-loop.c (get_initial_defs_for_reduction): Pass the
output vector directly to duplicate_and_interleave instead of
going through a temporary. Postpone insertion of ctor_seq to
the end of the loop.
From-SVN: r267652
The C++ char_traits and ctype APIs do not disallow null pointer
arguments, so we need explicit checks to ensure we don't forward null
pointers to memcpy or memmove.
PR libstdc++/87787
* include/bits/char_traits.h (char_traits::move): Do not pass null
pointers to memmove.
* include/bits/locale_facets.h
(ctype<char>::widen(const char*, const char*, char*)): Do not
pass null pointers to memcpy.
(ctype<char>::narrow(const char*, const char*, char, char*)):
Likewise.
(ctype<char>::do_widen(const char*, const char*, char*)):
Likewise.
(ctype<char>::do_narrow(const char*, const char*, char, char*)):
Likewise.
From-SVN: r267651
Investigating PR target/86891 revealed a number of issues with the way
the AArch64 backend was handing overflow detection patterns. Firstly,
expansion for signed and unsigned types is not the same as in one form
the overflow is detected via the C flag and in the other it is done
via the V flag in the PSR. Secondly, particular care has to be taken
when describing overflow of signed types: the comparison has to be
performed conceptually on a value that cannot overflow and compared to
a value that might have overflowed.
It became apparent that some of the patterns were simply unmatchable
(they collapse to NEG in the RTL rather than subtracting from zero)
and a number of patterns were overly restrictive in terms of the
immediate constants that they supported. I've tried to address all of
these issues as well.
gcc:
PR target/86891
* config/aarch64/aarch64.c (aarch64_expand_subvti): New parameter
unsigned_p. Handle signed and unsigned overflow correction as
required.
* config/aarch64/aarch64-protos.h (aarch64_expand_subvti): Update
prototype.
* config/aarch64/aarch64.md (addv<mode>4): Use aarch64_plus_operand
for operand 2.
(add<mode>3_compareV_imm): Make this callable for expanding.
(subv<GPI:mode>4): Use register_operand for operand 1. Use
aarch64_plus_operand for operand 2.
(subv<GPI:mode>_insn): New insn pattern.
(subv<GPI:mode>_imm): Likewise.
(negv<GPI:mode>3): New expand pattern.
(negv<GPI:mode>_insn): New insn pattern.
(negv<GPI:mode>_cmp_only): Likewise.
(cmpv<GPI:mode>_insn): Likewise.
(subvti4): Use register_operand for operand 1. Update call to
aarch64_expand_subvti.
(usubvti4): Likewise.
(negvti3): New expand pattern.
(negdi_carryout): New insn pattern.
(negvdi_carryinV): New insn pattern.
(sub<mode3>_compare1_imm): Delete named insn pattern, make anonymous
version the named version.
(peepholes to convert to sub<mode3>_compare1_imm): Adjust order of
operands.
(usub<GPI:mode>3_carryinC, usub<GPI:mode>3_carryinC_z1): New insn
patterns.
(usub<GPI:mode>3_carryinC_z2, usub<GPI:mode>3_carryinC): New insn
patterns.
(sub<mode>3_carryinCV, sub<mode>3_carryinCV_z1_z2): Delete.
(sub<mode>3_carryinCV_z1, sub<mode>3_carryinCV_z2): Delete.
(sub<mode>3_carryinCV): Delete.
(sub<GPI:mode>3_carryinV): New expand pattern.
sub<mode>3_carryinV, sub<mode>3_carryinV_z2): New insn patterns.
testsuite:
* gcc.target/aarch64/subs_compare_2.c: Make '#' immediate prefix
optional in scan pattern.
From-SVN: r267650
* doc/xml/manual/spine.xml: Update copyright years.
* doc/xml/manual/status_cxx2017.xml: Adjust note about -lstdc++fs.
* doc/xml/manual/using.xml: Remove requirement to link with -lstdc++fs
for C++17 filesystem library.
* doc/html/*: Regenerate.
From-SVN: r267648
Older versions of newlib do not provide truncate so add a configure
check for it, and provide a fallback definition.
There were also some missing exports in the linker script, which went
unnoticed because there are no tests for some functions. A new link-only
test checks that every filesystem operation function is defined by the
library.
* acinclude.m4 (GLIBCXX_CHECK_FILESYSTEM_DEPS): Check for truncate.
* config.h.in: Regenerate.
* config/abi/pre/gnu.ver: Order patterns for filesystem operations
alphabetically and add missing entries for copy_symlink,
hard_link_count, rename, and resize_file.
* configure: Regenerate.
* src/c++17/fs_ops.cc (resize_file): Remove #if so posix::truncate is
used unconditionally.
* src/filesystem/ops-common.h (__gnu_posix::truncate)
[!_GLIBCXX_HAVE_TRUNCATE]: Provide fallback definition that only
supports truncating to zero length.
* testsuite/27_io/filesystem/operations/all.cc: New test.
* testsuite/27_io/filesystem/operations/resize_file.cc: New test.
From-SVN: r267647
This patch folds certain reductions of X & CST to X[I] & CST[I] if I is
the only nonzero element of CST. This includes the motivating case in
which CST[I] is -1.
We could do the same for REDUC_MAX on unsigned types, but I wasn't sure
that that special case was worth it.
2019-01-07 Richard Sandiford <richard.sandiford@arm.com>
gcc/
PR tree-optimization/88598
* tree.h (single_nonzero_element): Declare.
* tree.c (single_nonzero_element): New function.
* match.pd: Fold certain reductions of X & CST to X[I] & CST[I]
if I is the only nonzero element of CST.
gcc/testsuite/
PR tree-optimization/88598
* gcc.dg/vect/pr88598-1.c: New test.
* gcc.dg/vect/pr88598-2.c: Likewise.
* gcc.dg/vect/pr88598-3.c: Likewise.
* gcc.dg/vect/pr88598-4.c: Likewise.
* gcc.dg/vect/pr88598-5.c: Likewise.
* gcc.dg/vect/pr88598-6.c: Likewise.
From-SVN: r267646
