Tom de Vries [Fri, 14 Dec 2018 13:48:56 +0000 (13:48 +0000)]
[offloading] Error on missing symbols
When compiling an OpenMP or OpenACC program containing a reference in the
offloaded code to a symbol that has not been included in the offloaded code,
the offloading compiler may ICE in lto1.
Fix this by erroring out instead, mentioning the problematic symbol:
...
error: variable 'var' has been referenced in offloaded code but hasn't
been marked to be included in the offloaded code
lto1: fatal error: errors during merging of translation units
compilation terminated.
...
Build x86_64 with nvptx accelerator and reg-tested libgomp.
Build x86_64 and reg-tested libgomp.
2018-12-14 Tom de Vries <tdevries@suse.de>
* lto-cgraph.c (verify_node_partition): New function.
(input_overwrite_node, input_varpool_node): Use verify_node_partition.
* testsuite/libgomp.c-c++-common/function-not-offloaded-aux.c: New test.
* testsuite/libgomp.c-c++-common/function-not-offloaded.c: New test.
* testsuite/libgomp.c-c++-common/variable-not-offloaded.c: New test.
* testsuite/libgomp.oacc-c-c++-common/function-not-offloaded.c: New test.
* testsuite/libgomp.oacc-c-c++-common/variable-not-offloaded.c: New test.
H.J. Lu [Fri, 14 Dec 2018 12:38:04 +0000 (12:38 +0000)]
x86: Don't use get_frame_size when finalizing stack frame
get_frame_size () returns used stack slots during compilation, which
may be optimized out later. Since ix86_find_max_used_stack_alignment
is called by ix86_finalize_stack_frame_flags to check if stack frame
is required, there is no need to call get_frame_size () which may give
inaccurate final stack frame size.
Tested on AVX512 machine configured with
--with-arch=native --with-cpu=native
gcc/
PR target/88483
* config/i386/i386.c (ix86_finalize_stack_frame_flags): Don't
use get_frame_size ().
gcc/testsuite/
PR target/88483
* gcc.target/i386/stackalign/pr88483.c: New test.
Andrew Stubbs [Fri, 14 Dec 2018 11:49:28 +0000 (11:49 +0000)]
Fix LRA bug
This patch fixes an ICE building libgfortran/random.c.
The problem was an adddi3 instruction that had an eliminable frame pointer.
GCN adddi3 includes a match_scratch, which LRA substitutes with a REG, and
checks if it can be converted back to a scratch afterwards. In the meantime,
the add was converted to a move, meaning that the instruction pattern
completely changed, thus causing a segfault when the instruction is revisited
in restore_scratches.
2018-12-14 Andrew Stubbs <ams@codesourcery.com>
gcc/
* gcc/lra-int.h (lra_register_new_scratch_op): Add third parameter.
* gcc/lra-remat.c (update_scratch_ops): Pass icode to
lra_register_new_scratch_op.
* gcc/lra.c (struct sloc): Add icode field.
(lra_register_new_scratch_op): Add icode parameter.
Use icode to skip insns that have changed beyond recognition.
Eric Botcazou [Fri, 14 Dec 2018 11:08:15 +0000 (11:08 +0000)]
decl.c (rm_size): Take into account the padding in the case of a record type containing a template.
* gcc-interface/decl.c (rm_size): Take into account the padding in
the case of a record type containing a template.
* gcc-interface/trans.c (Attribute_to_gnu) <Attr_Size>: Likewise.
Do not subtract the padded size for Max_Size_In_Storage_Elements.
<Attr_Descriptor_Size>: Tweak comment.
* dwarf2out.c (analyze_discr_in_predicate): Simplify.
(analyze_variants_discr): Deal with naked boolean discriminants.
ada/
* gcc-interface/decl.c (choices_to_gnu): Directly use a naked boolean
discriminant if the value is the boolean true.
Alan Modra [Fri, 14 Dec 2018 09:03:50 +0000 (19:33 +1030)]
[RS6000] PR88311, mlongcall indirections are optimised away
Masking CALL_LONG from the cookie was done in order to simplify and
correct length attribute calculations for indirect calls at one point
in my call series tidy when the indirect patterns used alternatives
"0,n" on the cookie operand. (Leaving the CALL_LONG in place
calculated the wrong length for long calls without fp args.)
This is no longer necessary now that the indirect sysv call patterns
explicitly test for the fp arg bits in their length attribute
expressions. And without the CALL_LONG to disable insns like
call_value_local_svsv, combine merrily replaces the indirect long call
sequence with a direct call. As it should. This patch reinstates
the CALL_LONG bit.
PR rtl-optimization/88311
* config/rs6000/rs6000.c (rs6000_call_sysv): Do not mask cookie.
(rs6000_sibcall_sysv): Likewise.
Jakub Jelinek [Thu, 13 Dec 2018 23:35:10 +0000 (00:35 +0100)]
re PR tree-optimization/88444 (ICE: tree check: expected ssa_name, have integer_cst in live_on_edge, at tree-vrp.c:468; or ICE: tree check: expected ssa_name, have integer_cst in get_value_range, at vr-values.c:84)
PR tree-optimization/88444
* tree-inline.c (fold_marked_statements): Iterate up to
last_basic_block_for_fn rather than n_basic_blocks_for_fn.
Jakub Jelinek [Thu, 13 Dec 2018 23:33:57 +0000 (00:33 +0100)]
re PR rtl-optimization/88470 (ICE in maybe_record_trace_start, at dwarf2cfi.c:2354)
PR rtl-optimization/88470
* cfgcleanup.c (outgoing_edges_match): If the function is
shrink-wrapped and bb1 ends with a JUMP_INSN with a single fake
edge to EXIT, return false.
Jonathan Wakely [Thu, 13 Dec 2018 20:34:10 +0000 (20:34 +0000)]
Fix handling of POSIX paths containing a root-name
Fix path appending and concatenating to work correctly for a leading
root-name. Check a new macro, SLASHSLASH_IS_ROOT_NAME, instead of making
the behaviour depend directly on __CYGWIN__.
* src/filesystem/std-path.cc (SLASHSLASH_IS_ROOT_NAME): New macro to
control whether interpret paths with two slashes as a root-name.
(path::operator/=(const path&)) [SLASHSLASH_IS_ROOT_NAME]: Add a
root-directory when appending to a root-name.
(path::_M_append(basic_string_view<value_type>))
[SLASHSLASH_IS_ROOT_NAME]: Likewise.
(path::operator/=(const path&)) [SLASHSLASH_IS_ROOT_NAME]: Likewise.
(path::_M_concat(basic_string_view<value_type>))
[SLASHSLASH_IS_ROOT_NAME]: Likewise.
(path::lexically_normal()) [SLASHSLASH_IS_ROOT_NAME]: Use += instead
of /= to add a root-directory to the result.
* testsuite/27_io/filesystem/path/decompose/root_directory.cc: Fix
expected result for Cygwin.
Jonathan Wakely [Thu, 13 Dec 2018 20:33:55 +0000 (20:33 +0000)]
PR libstdc++/71044 optimize std::filesystem::path construction
This new implementation has a smaller footprint than the previous
implementation, due to replacing std::vector<_Cmpt> with a custom pimpl
type that only needs a single pointer. The _M_type enumeration is also
combined with the pimpl type, by using a tagged pointer, reducing
sizeof(path) further still.
Construction and modification of paths is now done more efficiently, by
splitting the input into a stack-based buffer of string_view objects
instead of a dynamically-allocated vector containing strings. Once the
final size is known only a single allocation is needed to reserve space
for it. The append and concat operations no longer require constructing
temporary path objects, nor re-parsing the entire native pathname.
This results in algorithmic improvements to path construction, and
working with large paths is much faster.
PR libstdc++/71044
* include/bits/fs_path.h (path::path(path&&)): Add noexcept when
appropriate. Move _M_cmpts instead of reparsing the native pathname.
(path::operator=(const path&)): Do not define as defaulted.
(path::operator/=, path::append): Call _M_append.
(path::concat): Call _M_concat.
(path::path(string_type, _Type): Change type of first parameter to
basic_string_view<value_type>.
(path::_M_append(basic_string_view<value_type>)): New member function.
(path::_M_concat(basic_string_view<value_type>)): New member function.
(_S_convert(value_type*, __null_terminated)): Return string view.
(_S_convert(const value_type*, __null_terminated)): Return string view.
(_S_convert(value_type*, value_type*))
(_S_convert(const value_type*, const value_type*)): Add overloads for
pairs of pointers.
(_S_convert(_InputIterator, __null_terminated)): Construct string_type
explicitly, for cases where _S_convert returns a string view.
(path::_S_is_dir_sep): Replace with non-member is_dir_sep.
(path::_M_trim, path::_M_add_root_name, path::_M_add_root_dir)
(path::_M_add_filename): Remove.
(path::_M_type()): New member function to replace _M_type data member.
(path::_List): Define new struct type instead of using std::vector.
(path::_Cmpt::_Cmpt(string_type, _Type, size_t)): Change type of
first parameter to basic_string_view<value_type>.
(path::operator+=(const path&)): Do not define inline.
(path::operator+=(const string_type&)): Call _M_concat.
(path::operator+=(const value_type*)): Likewise.
(path::operator+=(value_type)): Likewise.
(path::operator+=(basic_string_view<value_type>)): Likewise.
(path::operator/=(const path&)): Do not define inline.
(path::_M_append(path)): Remove.
* python/libstdcxx/v6/printers.py (StdPathPrinter): New printer that
understands the new path::_List type.
* src/filesystem/std-path.cc (is_dir_sep): New function to replace
path::_S_is_dir_sep.
(path::_Parser): New helper class to parse strings as paths.
(path::_List::_Impl): Define container type for path components.
(path::_List): Define members.
(path::operator=(const path&)): Define explicitly, to provide the
strong exception safety guarantee.
(path::operator/=(const path&)): Implement manually by processing
each component of the argument, rather than using _M_split_cmpts
to parse the entire string again.
(path::_M_append(string_type)): Likewise.
(path::operator+=(const path&)): Likewise.
(path::_M_concat(string_type)): Likewise.
(path::remove_filename()): Perform trim directly instead of calling
_M_trim().
(path::_M_split_cmpts()): Rewrite in terms of _Parser class.
(path::_M_trim, path::_M_add_root_name, path::_M_add_root_dir)
(path::_M_add_filename): Remove.
* testsuite/27_io/filesystem/path/append/source.cc: Test appending a
string view that aliases the path.
testsuite/27_io/filesystem/path/concat/strings.cc: Test concatenating
a string view that aliases the path.
Libgomp test-case libgomp.c/target-5.c is failing to compile when building for
x86_64 with nvptx accelerator due to missing:
- getpid
- gethostname
- isatty (pulled in by fwrite)
in the nvptx newlib.
This patch fixes the build failure by:
- adding a function gomp_print_string which limits the use of fwrite to a single
location (in affinity-fmt.c), and
- creating an nvptx version of affinity-fmt.c, which:
- overrides the configure test results HAVE_GETPID and HAVE_GETHOSTNAME, and
- implements fwrite using write.
Build and reg-tested on x86_64 with nvptx accelerator.
2018-12-13 Tom de Vries <tdevries@suse.de>
* affinity-fmt.c (gomp_print_string): New function, factored out of ...
(omp_display_affinity, gomp_display_affinity_thread): ... here, and ...
* fortran.c (omp_display_affinity_): ... here.
* libgomp.h (gomp_print_string): Declare.
* config/nvptx/affinity-fmt.c: New file. Include affinity-fmt.c,
undefining HAVE_GETPID and HAVE_GETHOSTNAME, and mapping fwrite to
write.
Jakub Jelinek [Thu, 13 Dec 2018 17:01:50 +0000 (18:01 +0100)]
re PR tree-optimization/88464 (AVX-512 vectorization of masked scatter failing with "not suitable for scatter store")
PR tree-optimization/88464
* tree-vect-stmts.c (vect_build_gather_load_calls): Handle INTEGER_TYPE
masktype if mask is non-NULL.
(vectorizable_load): Don't reject masked gather loads if masktype
in the decl is INTEGER_TYPE.
* gcc.target/i386/avx512f-pr88462-1.c: New test.
* gcc.target/i386/avx512f-pr88462-2.c: New test.
Jakub Jelinek [Thu, 13 Dec 2018 12:53:19 +0000 (13:53 +0100)]
re PR libgomp/88460 ([nvptx] FAIL: libgomp.c++/for-24.C (internal compiler error))
PR libgomp/88460
* testsuite/libgomp.c++/for-24.C (results): Include it in
omp declare target region.
(main): Use map (always, tofrom: results) instead of
map (tofrom: results).
Jakub Jelinek [Thu, 13 Dec 2018 12:52:11 +0000 (13:52 +0100)]
re PR testsuite/88454 (test case gcc.dg/tree-ssa/split-path-5.c fails after r266971)
PR testsuite/88454
* gcc.dg/tree-ssa/split-path-5.c (__ctype_ptr__): Change type from
const char * to const signed char *.
(bmhi_init): Change pattern parameter's type the same. Use
__builtin_strlen instead of undeclared strlen.
Andreas Krebbel [Thu, 13 Dec 2018 11:57:16 +0000 (11:57 +0000)]
S/390: Fix zvector vec_double builtin
The (unsigned) long int to double vector conversion instructions
expect 2 immediate parameters. One for the inexact suppression
control and another one for the rounding mode. However, the
vec_double builtin has just the vector source operand. The 2
addtional operands need to be added with an intermediate expander.
The expanders were already there but unfortunately not wired up
correctly to the builtin.
gcc/ChangeLog:
2018-12-13 Andreas Krebbel <krebbel@linux.ibm.com>
* config/s390/s390-builtins.def (s390_vec_double_s64): Map to
s390_vec_double_s64 instead of s390_vcdgb.
(s390_vec_double_u64): Map to s390_vec_double_u64 instead of
s390_vcdlgb.
gcc/testsuite/ChangeLog:
2018-12-13 Andreas Krebbel <krebbel@linux.ibm.com>
* gcc.target/s390/zvector/vec-double-1.c: New test.
* gcc.target/s390/zvector/vec-double-2.c: New test.
Jonathan Wakely [Thu, 13 Dec 2018 11:01:03 +0000 (11:01 +0000)]
Disable new tests for configurations with no libstdc++fs.a
* testsuite/27_io/filesystem/path/construct/80762.cc: Skip test if
the Filesystem TS support is not configured.
* testsuite/experimental/filesystem/path/construct/80762.cc: Likewise.
We've received reports about the -msve-vector-bits=128 bits being somewhat ambiguous.
It isn't clear whether -msve-vector-bits=128 forces vector-length-agnostic code or whether
-msve-vector-bits=scalable forces 128-bit vector-lengh-specific code.
The latter is a, perhaps unintuitive, reading that we want to exclude.
This patch makes it more explicit that -msve-vector-bits=128 is special and produces
vector-length *agnostic* code. In the end, I've rewritten the whole option documentation.
Checked make pdf that the output looks reasonable.
Jakub Jelinek [Thu, 13 Dec 2018 08:08:22 +0000 (09:08 +0100)]
re PR target/88465 (AVX512: optimize loading of constant values to kN registers)
PR target/88465
* config/i386/i386.md (*movdi_internal, *movsi_internal,
*movhi_internal, *movqi_internal): Add alternative(s) to load
0 or -1 into k registers using kxor or kxnoq instructions.
* gcc.target/i386/avx512f-pr88465.c: New test.
* gcc.target/i386/avx512dq-pr88465.c: New test.
Jakub Jelinek [Thu, 13 Dec 2018 08:00:42 +0000 (09:00 +0100)]
re PR target/88461 (AVX512: gcc should keep value in kN registers if possible)
PR target/88461
* config/i386/sse.md (VI1248_AVX512VLBW, AVX512ZEXTMASK): New
mode iterators.
(<avx512>_testm<mode>3<mask_scalar_merge_name>,
<avx512>_testnm<mode>3<mask_scalar_merge_name>): Merge patterns
with VI12_AVX512VL and VI48_AVX512VL iterators into ones with
VI1248_AVX512VLBW iterator.
(*<avx512>_testm<VI1248_AVX512VLBW:mode>3_zext,
*<avx512>_testm<VI1248_AVX512VLBW:mode>3_zext_mask,
*<avx512>_testnm<VI1248_AVX512VLBW:mode>3_zext,
*<avx512>_testnm<VI1248_AVX512VLBW:mode>3_zext_mask): New
define_insns.
Ian Lance Taylor [Wed, 12 Dec 2018 22:48:46 +0000 (22:48 +0000)]
os/signal: increase deliver time for signal testcase
This increases the time to wait for signals to be delivered in the
TestAtomicStop testcase. When running gccgo tests on ppc64 or ppc64le,
there are intermittent failures in this test because the wait time is
too small.
The testcase uses REs like {(?n)\m_*bar[.$_]constprop[.$_]0:} to find
what functions are defined. But, this also matches lines like
.L.bar.constprop.0:(which are used on powerpc64-linux).
The "(?n)" doesn't do anything here either. We should use "^" here
instead of just "\m".
PR testsuite/88318
* gcc.dg/independent-cloneids-1.c: Use ^ not \m.
Olivier Hainque [Wed, 12 Dec 2018 17:10:44 +0000 (17:10 +0000)]
Use REGNUM macros in the definitions of aarch64 PROBE_STACK regs
* config/aarch64/aarch64.c (PROBE_STACK_FIRST_REG) : Redefine as
R9_REGNUM instead of 9.
(PROBE_STACK_SECOND_REG): Redefine as R10_REGNUM instead of 10.
Jonathan Wakely [Wed, 12 Dec 2018 16:13:49 +0000 (16:13 +0000)]
Overload std::distance and std::advance for path::iterator
Although filesystem::path::iterator is only a bidirectional iterator,
the underlying sequence has random access iterators (specifically, raw
pointers). This means std::distance and std::advance can be implemented
more efficiently than the generic versions which apply ++ and --
repeatedly.
PR libstdc++/71044 (partial)
* include/bits/fs_path.h (__path_iter_distance, __path_iter_advance):
New friend functions to implement std::distance and std::advance more
efficiently.
(distance, advance): Add overloads for path::iterator.
* testsuite/27_io/filesystem/path/itr/components.cc: Test new
overload.
Joel Brobecker [Wed, 12 Dec 2018 14:22:37 +0000 (14:22 +0000)]
require profiling support for gcc.dg/lto/20100430-1_0.c test
This test currently fails unexpectedly if GCC is configured with
--disable-gcov, because it requires -fprofile-arcs. This patch
fixes the issue by requiring profiling support in order to run
this test.
Tested with two compilers, one built with --disable-gcov, resulting
in the test reporting an UNSUPPORTED result; and one built with gcov
support, resulting in 2 PASS tests.
Tom de Vries [Wed, 12 Dec 2018 13:50:51 +0000 (13:50 +0000)]
[libbacktrace] Add allocfail.sh test-case
Add test-case that forces alloc.c functions to fail, and check whether fail
handling is robust.
This is the test-case for "[libbacktrace] Fix segfault upon allocation
failure". Without that patch, this test-case fails like this:
...
allocfail.sh: line 71: 26041 Segmentation fault (core dumped) \
./allocfail $i > /dev/null 2>&1
Unallowed fail found: 13
FAIL allocfail.sh (exit status: 1)
...
This is a seperate patch because the test-case is nontrivial.
Bootstrapped and reg-tested on x86_64.
2018-12-12 Tom de Vries <tdevries@suse.de>
* Makefile.am (TESTS): Add allocfail.sh.
(check_PROGRAMS): Add allocfail.
* Makefile.in: Regenerate.
* instrumented_alloc.c: New file. Redefine malloc and realloc.
Include alloc.c.
* allocfail.c: New file.
* allocfail.sh: New file.
Rasmus Villemoes [Wed, 12 Dec 2018 10:12:36 +0000 (11:12 +0100)]
libgcc: rs6000: tramp.S: fix placement of .cfi_endproc for __trampoline_setup
Currently, .cfi_endproc and FUNC_END(__trampoline_setup) are placed
inside the #else branch of an "#if defined (__VXWORKS__) ...", so
non-pic vxworks does not get proper CFI nor a .size directive for
__trampoline_setup. I assume there's no magic reason for that (which
would warrant a comment), so move them outside.
The PR is about unnecessary saves of the pic base register, it shows on m32 Linux and m32/m64 Darwin.
The fix is to check that we are in a pic mode and that the picbase has actually been used.
As a bonus, some #ifdef’d TARGET_MACHO code is no longer required.
Iain Sandoe [Wed, 12 Dec 2018 08:42:28 +0000 (08:42 +0000)]
rs6000, darwin - fix bootstrap.
Recent changes to the lowerings for call handling in rs6000
have broken bootstrap for powerpc-darwin. This patch modifies
the calls to handle Darwin asa subset, a subsequent patch will
split Darwin out.
2018-12-12 Iain Sandoe <iain@sandoe.co.uk>
* config/rs6000/darwin.md (call_indirect_nonlocal_darwin64): Remove.
(call_nonlocal_darwin64): Remove.
(call_value_indirect_nonlocal_darwin64): Remove.
(call_value_nonlocal_darwin64): Remove.
* config/rs6000/rs6000.c (rs6000_call_template_1): Handle Darwin with
the same asm output as AIX/ELFv2. (rs6000_call_sysv): Preserve the
CALL_LONG flag when needed for Darwin.
* config/rs6000/rs6000.md (define expand “call”): Expand Darwin as
per sysv.
(define_expand “call_value”): Likewise.
(define_expand “sibcall”): Likewise.
(define_expand “sibcall_value”): Likewise.
(call_indirect_nonlocal_sysv): Mark the clobber mode P.
(call_nonlocal_sysv): Likewise.
(call_nonlocal_sysv_secure): Likewise.
(call_value_indirect_nonlocal_sysv): Likewise.
(call_value_nonlocal_sysv): Likewise.
(call_value_nonlocal_sysv_secure): Likewise.
(call_local64): Mark the clobber DI.
(call_value_local64): Likewise.
Jakub Jelinek [Wed, 12 Dec 2018 08:28:43 +0000 (09:28 +0100)]
re PR c++/88449 (__builtin_is_constant_evaluated misbehaves due to constexpr call caching)
PR c++/88449
* constexpr.c (struct constexpr_call): Add pretend_const_required
member.
(constexpr_call_hasher::equal): Return false if pretend_const_required
members differ.
(cxx_eval_call_expression): Adjust new_call initialization. Hash in
ctx->pretend_const_required.
* g++.dg/cpp2a/is-constant-evaluated1.C: Change from dg-do compile
to dg-do run.
(e): Adjust comment with correct expected value.
(main): Expect e == 1.
* g++.dg/cpp2a/is-constant-evaluated2.C: New test.
Steven G. Kargl [Tue, 11 Dec 2018 23:13:19 +0000 (23:13 +0000)]
re PR fortran/88249 (ICE in gfc_resolve_filepos, at fortran/io.c:2853)
2018-12-11 Steven G. Kargl <kargl@gcc.gnu.org>
PR fortran/88249
* gfortran.h: Update prototype for gfc_resolve_filepos().
* io.c (gfc_resolve_filepos): Check for UNIT number if ERR= is present.
Use passed in locus for error message.
* resolve.c (gfc_resolve_code): Pass locus in gfc_resolve_filepos()
call.
2018-12-11 Steven G. Kargl <kargl@gcc.gnu.org>
PR fortran/88249
* gfortran.dg/pr88249.f90: New test.
Eric Botcazou [Tue, 11 Dec 2018 23:04:39 +0000 (23:04 +0000)]
re PR ada/88429 (libada build fails with --disable-shared)
libada/
PR ada/88429
* configure.ac (default_gnatlib_target): Set to gnatlib instead of
gnatlib-plain if --disable-shared.
* configure: Regenerate.
* Makefile.in (all): Replace gnatlib prerequisite with libada.
(ADA_RTS_SUBDIR): Delete.
(libada): New target, renamed from...
(gnatlib): ...this. Merge with other library targets.
(gnatlib-plain): Delete.
(install-gnatlib): Rename to...
(install-libada): ...this.
(install): Replace install-gnatlib prerequisite with install-libada.
gcc/ada/
PR ada/88429
* gcc-interface/Makefile.in (./stamp-gnatlib1-$(RTSDIR)): Also pass
MULTISUBDIR to sub-make and add quotes around $(THREAD_KIND).
(gnatlib-shared-dual): Also pass PICFLAG_FOR_TARGET to sub-make.
(gnatlib-sjlj): Also pass MULTISUBDIR to sub-make, but do not pass
PICFLAG_FOR_TARGET.
(gnatlib-zcx): Likewise.
Jakub Jelinek [Tue, 11 Dec 2018 20:37:53 +0000 (21:37 +0100)]
re PR c++/87861 (ICE in output_constructor_regular_field, at varasm.c:5165)
PR c++/87861
* class.c (build_vtbl_initializer): For TARGET_VTABLE_USES_DESCRIPTORS
bump index for each added word.
* constexpr.c (find_array_ctor_elt): Add forward declaration.
(cxx_eval_call_expression): Handle TARGET_VTABLE_USES_DESCRIPTORS
vtable calls.
(cxx_eval_constant_expression) <case OBJ_TYPE_REF>: Divide token
by TARGET_VTABLE_USES_DESCRIPTORS if non-zero.
* constexpr.c (potential_constant_expression_1): Check want_rval
instead of checking if we have a decl.
* decl2.c (decl_maybe_constant_var_p): Don't consider volatile
constexpr variables as maybe constant.
* g++.dg/cpp0x/constexpr-volatile2.C: New test.
* g++.dg/cpp0x/pr65327.C: Add dg-error.
Paolo Carlini [Tue, 11 Dec 2018 17:47:56 +0000 (17:47 +0000)]
decl.c (grokvardecl): Add location_t parameter and use it in build_lang_decl_loc and build_decl calls.
/cp
2018-12-11 Paolo Carlini <paolo.carlini@oracle.com>
* decl.c (grokvardecl): Add location_t parameter and use it
in build_lang_decl_loc and build_decl calls.
(grokdeclarator): Move up loc declaration and use it in the
grokvardecl call too.
/testsuite
2018-12-11 Paolo Carlini <paolo.carlini@oracle.com>
* g++.dg/pr53037-4.C: Test the first two locations too.
Paolo Carlini [Tue, 11 Dec 2018 17:47:47 +0000 (17:47 +0000)]
decl.c (grokvardecl): Add location_t parameter and use it in build_lang_decl_loc and build_decl calls.
/cp
2018-12-11 Paolo Carlini <paolo.carlini@oracle.com>
* decl.c (grokvardecl): Add location_t parameter and use it
in build_lang_decl_loc and build_decl calls.
(grokdeclarator): Move up loc declaration and use it in the
grokvardecl call too.
/testsuite
2018-12-11 Paolo Carlini <paolo.carlini@oracle.com>
* g++.dg/pr53037-4.C: Test the first two locations too.
Paolo Carlini [Tue, 11 Dec 2018 17:46:42 +0000 (17:46 +0000)]
decl2.c (grokvardecl): Add location_t parameter and use it in build_lang_decl_loc and build_decl calls.
/cp
2018-12-11 Paolo Carlini <paolo.carlini@oracle.com>
* decl2.c (grokvardecl): Add location_t parameter and use it
in build_lang_decl_loc and build_decl calls.
(grokdeclarator): Move up loc declaration and use it in the
grokvardecl call too.
/testsuite
2018-12-11 Paolo Carlini <paolo.carlini@oracle.com>
* g++.dg/pr53037-4.C: Test the first two locations too.
Jakub Jelinek [Tue, 11 Dec 2018 17:16:48 +0000 (18:16 +0100)]
re PR tree-optimization/88444 (ICE: tree check: expected ssa_name, have integer_cst in live_on_edge, at tree-vrp.c:468; or ICE: tree check: expected ssa_name, have integer_cst in get_value_range, at vr-values.c:84)
PR tree-optimization/88444
* tree-vrp.c (register_edge_assert_for_2): Only register assertions
for conversions if rhs1 is a SSA_NAME.
* cfgexpand.c (asm_clobber_reg_is_valid): Also produce
error when stack pointer is clobbered.
(expand_asm_stmt): Refactor clobber check in separate function.
Richard Earnshaw [Tue, 11 Dec 2018 11:26:15 +0000 (11:26 +0000)]
[aarch64] PR target/87369 Prefer bsl/bit/bif for copysign
The copysign operations will almost always be performed on values in
floating-point registers. As such, we do not want the compiler to
simplify the operations into code sequences that can only be done
using the general-purpose register set. Unfortunately, this is what
is currently happening.
Fortunately, it seems quite unlikely that copysign() will be
subsequently followed by other logical operations on the values
involved, so I think it is acceptable to use an unspec here. This
allows us to preserve the operation in a form that allows the register
allocator to make the right choice later on, without limitation on the
final form of the operation (well, if we do end up using the gp
register bank, we get a dead constant load that we cannot easily
eliminate at a late stage).
PR target/37369
* config/aarch64/iterators.md (sizem1): Add sizes for SFmode and DFmode.
(Vbtype): Add SFmode mapping.
* config/aarch64/aarch64.md (copysigndf3, copysignsf3): Delete.
(copysign<GPF:mode>3): New expand pattern.
(copysign<GPF:mode>3_insn): New insn pattern.
Piotr Trojanek [Tue, 11 Dec 2018 11:12:41 +0000 (11:12 +0000)]
[Ada] Add "Global => null" contracts to Ada.Calendar routines
Routines in Ada.Real_Time are already annotated with Global => null
contracts to suppress spurious warnings from the flow analysis in
GNATprove. This patch adds such contracts to Ada.Calendar. No change in
runtime behavior expected.
2018-12-11 Piotr Trojanek <trojanek@adacore.com>
gcc/ada/
* libgnat/a-calend.ads: Add "Global => null" contracts to pure
routines.
Obj_26 : Task_Typ; -- ERROR
pragma Thread_Local_Storage (Obj_26);
end Pack;
----------------------------
-- Compilation and output --
----------------------------
$ gcc -c pack.adb
pack.adb:47:04: Thread_Local_Storage variable "Obj_4" is improperly
initialized
pack.adb:47:04: only allowed initialization is explicit "null", static
expression or static aggregate
pack.adb:62:04: Thread_Local_Storage variable "Obj_9" is improperly
initialized
pack.adb:62:04: only allowed initialization is explicit "null", static
expression or static aggregate
pack.adb:71:04: Thread_Local_Storage variable "Obj_12" is improperly
initialized
pack.adb:71:04: only allowed initialization is explicit "null", static
expression or static aggregate
pack.adb:74:04: Thread_Local_Storage variable "Obj_13" is improperly
initialized
pack.adb:74:04: only allowed initialization is explicit "null", static
expression or static aggregate
pack.adb:80:04: Thread_Local_Storage variable "Obj_15" is improperly
initialized
pack.adb:80:04: only allowed initialization is explicit "null", static
expression or static aggregate
pack.adb:83:04: Thread_Local_Storage variable "Obj_16" is improperly
initialized
pack.adb:83:04: only allowed initialization is explicit "null", static
expression or static aggregate
pack.adb:92:04: Thread_Local_Storage variable "Obj_19" is improperly
initialized
pack.adb:92:04: only allowed initialization is explicit "null", static
expression or static aggregate
pack.adb:101:04: Thread_Local_Storage variable "Obj_22" is improperly
initialized
pack.adb:101:04: only allowed initialization is explicit "null", static
expression or static aggregate
pack.adb:104:04: Thread_Local_Storage variable "Obj_23" is improperly
initialized
pack.adb:104:04: only allowed initialization is explicit "null", static
expression or static aggregate
pack.adb:110:04: Thread_Local_Storage variable "Obj_25" is improperly
initialized
pack.adb:110:04: only allowed initialization is explicit "null", static
expression or static aggregate
pack.adb:113:04: Thread_Local_Storage variable "Obj_26" is improperly
initialized
pack.adb:113:04: only allowed initialization is explicit "null", static
expression or static aggregate
* freeze.adb (Check_Pragma_Thread_Local_Storage): Use the
violating set to diagnose detect an illegal initialization,
rather than the complement of the OK set.
(Freeze_Object_Declaration): Factorize code in
Has_Default_Initialization.
(Has_Default_Initialization, Has_Incompatible_Initialization):
New routines.
* libgnat/g-socket.ads (Family_Type): Add new enumerated value
Family_Unspec to be able to use it in Get_Address_Info parameter
and find IPv4 together with IPv6 addresses.
(Inet_Addr_Bytes_Length): Zero length for Family_Unspec. New
IPv6 predefined constant addresses.
(IPv4_To_IPv6_Prefix): IPv4 mapped to IPv6 address prefix.
(Is_IPv4_Address): Rename from Is_IP_Address and published.
(Is_IPv6_Address): New routine.
(Image of Inet_Addr_Type): Fix description about IPv6 address
text representation.
(Level_Type): New propocol level IP_Protocol_For_IPv6_Level.
(Add_Membership_V4): New socket option equal to Add_Membership.
(Drop_Membership_V4): New socket option equal to
Drop_Membership.
(Multicast_If_V4): New socket option equal to Multicast_If.
(Multicast_Loop_V4, Add_Membership_V6, Drop_Membership_V6,
Multicast_If_V6, Multicast_Loop_V6, Multicast_Hops, IPv6_Only):
New socket option for IPv6.
(Address_Info): New record to keep address info.
(Address_Info_Array): Array to keep address info records.
(Get_Address_Info): Routine to get address info records by host
and service names.
(Host_Service): Record to keep host and service names.
(Get_Name_Info): New routine to get host and service names by
address.
(Create_Socket): Add Level parameter, IP_Protocol_For_IP_Level
default.
(Name_Array, Inet_Addr_Array): Change array index to Positive.
* libgnat/g-socket.adb (IPV6_Mreq): New record definition for
IPv6.
(Hex_To_Char): Remove.
(Short_To_Network, Network_To_Short): Move to package
GNAT.Sockets.Thin_Common.
(Is_IP_Address): Remove.
(To_In_Addr, To_Inet_Addr): Move to package
GNAT.Sockets.Thin_Common.
(Get_Socket_Option): Get value of Multicast_Loop option as
integer boolean, process IPv6 options. Don't try to get
Add_Membership_V4, Add_Membership_V6, Drop_Membership_V4, and
Drop_Membership_V6 as not supported by the socket API.
(Set_Socket_Option): Set value of Multicast_Loop option as
integer boolean, process IPv6 options.
* gsocket.h
(IPV6_ADD_MEMBERSHIP): Define from IPV6_JOIN_GROUP if necessary
for VxWorks.
(IPV6_DROP_MEMBERSHIP): Define from IPV6_LEAVE_GROUP if
necessary for VxWorks
(HAVE_INET_NTOP): New definition.
(HAVE_INET_PTON): Includes VxWorks now.
* socket.c (__gnat_getaddrinfo, __gnat_getnameinfo,
__gnat_freeaddrinfo, __gnat_gai_strerror, __gnat_inet_ntop): New
routines.
* libgnat/g-sothco.ads, libgnat/g-sothco.adb
(socklen_t, In6_Addr, To_In6_Addr): New.
(To_In_Addr, To_Inet_Addr): Move from package body GNAT.Sockets.
(To_Inet_Addr): New overload with In6_Addr type parmeter.
(In_Addr_Access_Array): Remove.
(Sockaddr): Unchecked_Union instead of Sockaddr_In and old
defined generic Sockaddr.
(Set_Address): Use it to set family, port and address into
Sockaddr.
(Get_Address): New routine to get Socket_Addr_Type from
Sockaddr.
(Addrinfo): Structure to use with getaddrinfo.
(C_Getaddrinfo, C_Freeaddrinfo, C_Getnameinfo, C_GAI_Strerror,
Inet_Ntop): New routine import.
(Short_To_Network, Network_To_Short): Move from package body
GNAT.Sockets.
* libgnat/g-stsifd__sockets.adb: Use Sockaddr instead of
Sockaddr_In.
* s-oscons-tmplt.c (AF_UNSPEC, EAI_SYSTEM, SOCK_RAW,
IPPROTO_IPV6, IP_RECVERR, SIZEOF_socklen_t, IF_NAMESIZE): New
constants.
(AI_xxxx_OFFSET): Constants to consider platform differences in
field positions and sizes for addrinfo structure.
(AI_xxxxx): Flags for getaddrinfo.
(NI_xxxxx): Flags for getnameinfo.
(IPV6_xxxxx): Socket options for IPv6.
(Inet_Ntop_Linkname): New routine.
Yannick Moy [Tue, 11 Dec 2018 11:12:26 +0000 (11:12 +0000)]
[Ada] Do not expand code inside ignored ghost bodies
While ignored ghost code is not compiled into the executable, it may
lead to compilation errors when it makes use of language features
requiring runtime support that is not available in the available runtime
library. These errors are spurious, as the executable will never call
in these runtime units.
This patch deactivates the expansion of code inside ignored ghost bodies
of subprograms and packages, so that this code is still checked for
possible semantic errors, but it does not force the presence of useless
runtime units.
There is no impact on the executable produced.
2018-12-11 Yannick Moy <moy@adacore.com>
gcc/ada/
* sem_ch6.adb (Analyze_Subprogram_Body_Helper): Deactivate
expansion in ignored ghost subprogram body.
* sem_ch7.adb (Analyze_Package_Body_Helper): Deactivate
expansion in ignored ghost package body.
Ed Schonberg [Tue, 11 Dec 2018 11:12:21 +0000 (11:12 +0000)]
[Ada] Unnesting: do not eliminate subprogram whose address is taken
2018-12-11 Ed Schonberg <schonberg@adacore.com>
gcc/ada/
* exp_unst.adb (Register_Subprogram): A subprogram whose address
is taken (through attribute Address or Access) is reachable and
cannot be eliminated, even if there is no explicit call to it:
the address may be used in an indirect call or in some
address-related operation.
Ed Schonberg [Tue, 11 Dec 2018 11:12:16 +0000 (11:12 +0000)]
[Ada] Crash on misplaced First operation for GNAT iterable type
This patch improves the handling of an improper declaaration of aspect
First for a GNAT-defined iterable type,
2018-12-11 Ed Schonberg <schonberg@adacore.com>
gcc/ada/
* sem_util.adb (Get_Actual_Subtype): Function can return type
mark.
(Get_Cursor_Type): Improve recovery and error message on a
misplaced First aspect for an iterable type.
[Ada] Volatility, validity checks, and System.Aux_DEC
This patch updates validity checks to prevent the validation of an
by-reference formal parameter because the parameter is not being read in
the process.
* checks.adb: Add with and use clauses for Sem_Mech.
(Ensure_Valid): Update the "annoying special case" to include
entry and function calls. Use Get_Called_Entity to obtain the
entry or subprogram being invoked, rather than retrieving it
manually. Parameters passed by reference do not need a validity
check.
gcc/testsuite/
* gnat.dg/valid4.adb, gnat.dg/valid4_pkg.adb,
gnat.dg/valid4_pkg.ads: New testcase.
Yannick Moy [Tue, 11 Dec 2018 11:12:03 +0000 (11:12 +0000)]
[Ada] Improve error message when named number passed as global item
When a named number is used in Global/Depends contracts as global item,
the error message could be confusing to users new to Ada. Now the
message explains that named numbers are not objects.
For instance on the following illegal code snippet:
$ gcc -c bad_global.ads
1. package Bad_Global is
2.
3. X : constant := 1;
4. Y : constant := 1.0;
5.
6. procedure P with
7. Global => (Input => X,
|
>>> global item must denote object, state or current instance
of concurrent type
>>> named number "X" is not an object
8. In_Out => Y);
|
>>> global item must denote object, state or current instance
of concurrent type
>>> named number "Y" is not an object
projects / thirdparty / gcc.git / log
