[ARM] Fix PR85203: cmse_nonsecure_caller returns wrong result
__builtin_cmse_nonsecure_caller implementation returns true in almost
all cases due to 2 separate bugs:
* gen_addsi is used instead of gen_andsi to retrieve the lsb
* the lsb boolean value is not negated but the specification says
the intrinsic should return true for a nonsecure caller and a
nonsecure caller is characterized with LR's lsb being 0
This was not caught due to (1) lack of runtime test and (2) the existing
RTL scan not taking into account that '.' matches newline in Tcl regular
expressions.
This commit fixes the implementation issues and improves testing of
cmse_nonsecure_caller by (1) adding a runtime test for the secure caller
case and (2) looking for an SET insn of an AND expression in the right
function. This leaves the nonsecure caller case only partly tested
since the exact value being AND and the negation are not covered by the
scan and the existing test infrastructure does not allow 2 separate
compilation and link to be performed. It is enough though to catch the
current incorrect behavior.
The commit also reorganize the scan directives in cmse-1.c to more
easily identify what function they are intended to test in the file.
2018-04-04 Thomas Preud'homme <thomas.preudhomme@arm.com>
gcc/
PR target/85203
* config/arm/arm-builtins.c (arm_expand_builtin): Change
expansion to perform a bitwise AND of the argument followed by a
boolean negation of the result.
gcc/testsuite/
PR target/85203
* gcc.target/arm/cmse/cmse-1.c: Tighten cmse_nonsecure_caller RTL scan
to match a single insn of the baz function. Move scan directives at
the end of the file below the functions they are trying to test for
better readability.
* gcc.target/arm/cmse/cmse-16.c: New testcase.
Jakub Jelinek [Wed, 4 Apr 2018 17:18:14 +0000 (19:18 +0200)]
re PR other/85161 (Test case failures in libbacktrace on powerpc64 BE starting with r253456)
PR other/85161
* elf.c (elf_zlib_fetch): Fix up predefined macro names in test for
big endian, only use 32-bit loads if endianity macros are predefined
and indicate big or little endian.
Peter Bergner [Wed, 4 Apr 2018 15:35:03 +0000 (10:35 -0500)]
re PR rtl-optimization/84878 (ICE: Segmentation fault (in add_cross_iteration_register_deps))
gcc/
PR rtl-optimization/84878
* ddg.c (add_cross_iteration_register_deps): Use DF_REF_BB to determine
the basic block. Assert the use reference is not artificial and that
it has an associated insn.
gcc/testsuite/
PR rtl-optimization/84878
* gcc.target/powerpc/pr84878.c: New test.
Michael Matz [Wed, 4 Apr 2018 14:49:06 +0000 (14:49 +0000)]
Fix -Wstringop-overflow regression
we shouldn't claim string overflows for character arrays at
end of structures; the code that tries to avoid these
accidentally passed the address of the accessed member to
array_at_struct_end_p(), but that one wants the component_ref
or array_ref itself. Needs updating of one testcase that
incorrectly expected warning to occur in this situation.
Jakub Jelinek [Wed, 4 Apr 2018 14:14:08 +0000 (16:14 +0200)]
re PR testsuite/85189 (g++.dg/inherit/override-attribs.C FAILs on 32-bit x86)
PR testsuite/85189
* g++.dg/inherit/override-attribs.C: Use dg-message instead of dg-error
for the diagnostics of overridden functions. Adjust for new wording.
Richard Biener [Wed, 4 Apr 2018 07:52:20 +0000 (07:52 +0000)]
re PR tree-optimization/85168 (ICE in tree-ssa-coalesce.c: SSA corruption: Unable to coalesce ssa_names which are marked as MUST COALESCE when -O2 is used)
[PR c++/84943] mark function as used when taking its address
fn[0]() ICEd because we would fold the INDIRECT_REF used for the
array indexing while building the address for the call, after not
finding the decl hiding there at first. But the decl would be exposed
by the folding, and then lower layers would complain we had the decl,
after all, but it wasn't one of the artificial or special functions
that could be called without being marked as used.
This patch arranges for a FUNCTION_DECL to be marked as used when
taking its address, just like we already did when taking the address
of a static function to call it as a member function (i.e. using the
obj.fn() notation). However, we shouldn't mark functions as used when
just performing overload resolution, lest we might instantiate
templates we shouldn't, as in g++.dg/overload/template1.C, so we
adjust mark_used to return early when testing conversions.
for gcc/cp/ChangeLog
PR c++/84943
* typeck.c (cp_build_addr_expr_1): Mark FUNCTION_DECL as
used.
* decl2.c (mark_used): Return without effects if tf_conv.
Bill Schmidt [Tue, 3 Apr 2018 19:06:28 +0000 (19:06 +0000)]
emmintrin.h (_mm_cvtpd_epi32): Use __vector rather than vector.
[gcc]
2018-04-03 Bill Schmidt <wschmidt@linux.ibm.com>
* config/rs6000/emmintrin.h (_mm_cvtpd_epi32): Use __vector rather
than vector.
(_mm_cvtpd_ps): Likewise.
(_mm_cvttpd_epi32): Likewise.
* config/rs6000/mmintrin.h (_mm_unpacklo_pi8): Likewise.
* config/rs6000/xmmintrin.h: For strict-ANSI C++ or C11, undefine
vector, pixel, and bool following altivec.h include.
[gcc/testsuite]
2018-04-03 Bill Schmidt <wschmidt@linux.ibm.com>
* gcc.target/powerpc/powerpc.exp: Add .C suffix for main loop.
* gcc.target/powerpc/undef-bool-1.C: New file.
* gcc.target/powerpc/undef-bool-2.c: New file.
Paolo Carlini [Tue, 3 Apr 2018 17:53:05 +0000 (17:53 +0000)]
re PR c++/84768 (ICE with failed class template argument deduction because of invalid template parameter)
/cp
2018-04-03 Paolo Carlini <paolo.carlini@oracle.com>
PR c++/84768
* pt.c (rewrite_template_parm): If the first argument is
error_mark_node return it immediately.
(build_deduction_guide): Check the return value of the
latter for error_mark_node.
(do_class_deduction): Check the return value of the latter.
/testsuite
2018-04-03 Paolo Carlini <paolo.carlini@oracle.com>
Jakub Jelinek [Tue, 3 Apr 2018 16:20:02 +0000 (18:20 +0200)]
re PR c++/85134 (ICE with invalid constexpr in 'shared' OpenMP clause)
PR c++/85134
* decl.c (cp_finish_decl): If ensure_literal_type_for_constexpr_object
fails, after clearing DECL_DECLARED_CONSTEXPR_P don't return early,
instead for static data members clear init and set DECL_EXTERNAL.
* g++.dg/gomp/pr85134.C: New test.
* g++.dg/cpp0x/constexpr-ice19.C: Expect one further error.
After middle-end changes combine now gets fed different input, from
which it makes different (but just as efficient) code. So remove the
test for particular asm output.
i386.c (emit_i387_cw_initialization): Always use logic instructions when changing rounding bits to preserve precision...
* config/i386/i386.c (emit_i387_cw_initialization): Always use logic
instructions when changing rounding bits to preserve precision bits
in the x87 control word.
gcc/
* config.gcc (nds32): Add nds32-relax-opt.o into extra_objs.
* config/nds32/constants.md (unspec_volatile_element): Add
UNSPEC_VOLATILE_RELAX_GROUP.
* config/nds32/nds32-relax-opt.c: New file.
* config/nds32/nds32-predicates.c
(nds32_symbol_load_store_p): New function.
* config/nds32/nds32-protos.h
(nds32_symbol_load_store_p): Declare function.
(make_pass_nds32_relax_opt): Declare new rtl pass function.
* config/nds32/nds32.c
(nds32_register_pass): New function to register pass.
(nds32_register_passes): New function to register passes.
* config/nds32/nds32.md (relax_group): New pattern.
* config/nds32/nds32.opt (mrelax-hint): New option.
* config/nds32/t-nds32 (nds32-relax-opt.o): New dependency.
Co-Authored-By: Kito Cheng <kito.cheng@gmail.com> Co-Authored-By: Kuan-Lin Chen <kuanlinchentw@gmail.com>
From-SVN: r258998
Jakub Jelinek [Sun, 1 Apr 2018 06:05:01 +0000 (08:05 +0200)]
re PR middle-end/85090 (wrong code with -O2 -fno-tree-dominator-opts -mavx512f -fira-algorithm=priority)
PR middle-end/85090
* config/i386/sse.md (V): Add V64QI and V32HI for TARGET_AVX512F.
(V_128_256): New mode iterator.
(*avx512dq_vextract<shuffletype>64x2_1 splitter): New define_split.
(*avx512f_vextract<shuffletype>32x4_1 splitter): Likewise.
(xop_pcmov_<mode><avxsizesuffix>): Use V_128_256 mode iterator instead
of V.
* config/i386/i386.c (ix86_expand_vector_set): Improve V32HImode and
V64QImode expansion for !TARGET_AVX512BW && TARGET_AVX512F.
* gcc.target/i386/avx512f-pr85090-1.c: New test.
* gcc.target/i386/avx512f-pr85090-2.c: New test.
* gcc.target/i386/avx512f-pr85090-3.c: New test.
* gcc.target/i386/avx512bw-pr85090-2.c: New test.
* gcc.target/i386/avx512bw-pr85090-3.c: New test.
Alexandre Oliva [Sat, 31 Mar 2018 03:44:12 +0000 (03:44 +0000)]
[PR c++/85027] deal with baselink in save_expr in instantiate_type
We use SAVE_EXPRs in conditional expressions without the middle
operand, to evaluate the first operand only once. When the conversion
of the first operand fails, we may call instantiate_type get a better
error message. We have code to peel off the SAVE_EXPR there, but then
we may end up with a BASELINK, and we're past the code that deals with
BASELINKs. Reorder the tests so that we expose the saved expr first,
and then deal with BASELINKs.
for gcc/cp/ChangeLog
PR c++/85027
* class.c (instantiate_type): Peel off SAVE_EXPR before
BASELINK.
Jakub Jelinek [Fri, 30 Mar 2018 18:38:36 +0000 (20:38 +0200)]
re PR c++/84791 (ICE with broken OpenMP reduction clause)
PR c++/84791
* semantics.c (finish_omp_reduction_clause): If
OMP_CLAUSE_REDUCTION_PLACEHOLDER is error_mark_node, return true
even if processing_template_decl.
Paul Thomas [Fri, 30 Mar 2018 12:33:49 +0000 (12:33 +0000)]
re PR fortran/84931 (Expansion of array constructor with constant implied-do-object goes sideways)
2018-03-30 Paul Thomas <pault@gcc.gnu.org>
PR fortran/84931
* simplify.c (gfc_convert_constant): Handle case of array
constructors within an array that has no iterator and improve
the conciseness of this section of code.
2018-03-30 Paul Thomas <pault@gcc.gnu.org>
PR fortran/84931
* gfortran.dg/array_constructor_53.f90: New test.
Thomas Koenig [Fri, 30 Mar 2018 09:58:35 +0000 (09:58 +0000)]
re PR fortran/85111 (ICE in min_max_choose, at fortran/simplify.c:4884 (and others))
2017-03-30 Thomas Koenig <tkoenig@gcc.gnu.org>
PR fortran/85111
* array.c (gfc_resolve_character_array_constructor): Early
exit for zero-size arrays.
* simplify.c (simplify_transformation_to_array): Exit early
if the result size is zero.
(simplify_minmaxloc_to_array): Likewise.
2017-03-30 Thomas Koenig <tkoenig@gcc.gnu.org>
PR fortran/85111
* gfortran.dg/zero_sized_10.f90: New test.
Thomas Koenig [Fri, 30 Mar 2018 09:56:46 +0000 (09:56 +0000)]
re PR fortran/85111 (ICE in min_max_choose, at fortran/simplify.c:4884 (and others))
2017-03-30 Thomas Koenig <tkoenig@gcc.gnu.org>
PR fortran/85111
* array.c (gfc_resolve_character_array_constructor): Early
exit for zero-size arrays.
* simplify.c (simplify_transformation_to_array): Exit early
if the result size is zero.
(simplify_minmaxloc_to_array): Likewise.
2017-03-30 Thomas Koenig <tkoenig@gcc.gnu.org>
PR fortran/85111
* gfortran.dg/zero_sized_10.f90: New test.
David Malcolm [Thu, 29 Mar 2018 20:43:48 +0000 (20:43 +0000)]
More #include suggestions (PR c++/84269)
PR c++/84269 reports a number of names in the C and C++ standard
libraries for which we don't yet offer #include fix-it hints.
This patch adds them (up to comment #9).
gcc/c-family/ChangeLog:
PR c++/84269
* known-headers.cc (get_stdlib_header_for_name): Add various names
from <assert.h>, <string.h>, and <memory.h>; add more names from
<stdio.h>.
gcc/cp/ChangeLog:
PR c++/84269
* name-lookup.c (get_std_name_hint): Add names from <memory>,
<tuple>, and <utility>.
gcc/testsuite/ChangeLog:
PR c++/84269
* g++.dg/lookup/missing-std-include-6.C: New test.
* g++.dg/lookup/missing-std-include.C: Add std::pair and
std::tuple tests.
* g++.dg/spellcheck-reswords.C: Expect a hint about <cstring>.
* g++.dg/spellcheck-stdlib.C: Add tests for names in <cstdio>,
<cstring>, <cassert>, and <cstdlib>.
David Malcolm [Thu, 29 Mar 2018 14:43:01 +0000 (14:43 +0000)]
More underlining of bad arguments (PR c++/85110)
As of r256448, the C++ frontend underlines many bad arguments in its
diagnostics; those where perform_overload_resolution returns a
non-NULL candidate, but there's a failure in convert_like_real.
However, for the case where perform_overload_resolution fails, but
there's a single non-viable candidate, the error is diagnosed by
cp_build_function_call_vec, and that currently doesn't underline
the bad argument:
$ g++ test.cc
test.cc: In function 'void caller(const char*)':
test.cc:6:20: error: cannot convert 'const char*' to 'const char**' for argument '2' to 'void callee(int, const char**, int)'
callee (1, fmt, 3);
^
It's going through convert_for_assignment, and
implicitly using input_location.
This patch updates convert_for_assignment for this case, using
an EXPR_LOCATION if there is one, or falling back to input_location
otherwise, underlining the argument in question:
test.cc: In function 'void caller(const char*)':
test.cc:6:14: error: cannot convert 'const char*' to 'const char**' for argument '2' to 'void callee(int, const char**, int)'
callee (1, fmt, 3);
^~~
gcc/cp/ChangeLog:
PR c++/85110
* typeck.c (convert_for_assignment): When complaining due to
conversions for an argument, attempt to use the location of the
argument.
gcc/testsuite/ChangeLog:
PR c++/85110
* g++.dg/diagnostic/param-type-mismatch-2.C: New test.
H.J. Lu [Thu, 29 Mar 2018 13:14:06 +0000 (13:14 +0000)]
i386: Enable AVX/AVX512 features only if supported by OSXSAVE
Enable AVX and AVX512 features only if their states are supported by
OSXSAVE.
PR target/85100
* config/i386/cpuinfo.c (XCR_XFEATURE_ENABLED_MASK): New.
(XSTATE_FP): Likewise.
(XSTATE_SSE): Likewise.
(XSTATE_YMM): Likewise.
(XSTATE_OPMASK): Likewise.
(XSTATE_ZMM): Likewise.
(XSTATE_HI_ZMM): Likewise.
(XCR_AVX_ENABLED_MASK): Likewise.
(XCR_AVX512F_ENABLED_MASK): Likewise.
(get_available_features): Enable AVX and AVX512 features only
if their states are supported by OSXSAVE.
Martin Liska [Thu, 29 Mar 2018 13:02:23 +0000 (15:02 +0200)]
Documentation tweaks.
2018-03-29 Martin Liska <mliska@suse.cz>
PR lto/84995.
* doc/invoke.texi: Document how LTO works with debug info.
Describe auto-load support of binutils. Mention 'x86-64'
as valid option value of -march option.
Jakub Jelinek [Thu, 29 Mar 2018 10:37:58 +0000 (12:37 +0200)]
re PR c/85094 (-g with any optimization suppresses -Wduplicated-branches)
PR c/85094
* fold-const.c (operand_equal_p): Handle DEBUG_BEGIN_STMT.
For STATEMENT_LIST, pass down OEP_LEXICOGRAPHIC and maybe
OEP_NO_HASH_CHECK for recursive call, to avoid exponential
checking.
* c-c++-common/Wduplicated-branches-14.c: New test.
projects / thirdparty / gcc.git / log
