Uros Bizjak [Tue, 26 Feb 2019 17:21:43 +0000 (18:21 +0100)]
extend.texi (__builtin_object_size): Use @pxref instead of @xref inside parenthesis.
* doc/extend.texi (__builtin_object_size):
Use @pxref instead of @xref inside parenthesis.
(__builtin_has_attribute): Add missing comma after @xref.
(__builtin_object_size): Ditto.
* doc/md.texi (cond_*{mode}): Use @samp instead of @var around op1[i].
* fortran/invoke.texi (-ffpe-trap): Use @var for every item
in the list.
Tom de Vries [Tue, 26 Feb 2019 16:09:40 +0000 (16:09 +0000)]
[libbacktrace] Fix btest-lto for older gcc
With gcc 4.8.5, btest_lto ends up with a privatized name for the variable with
name 'global':
...
$ nm btest_lto | grep ' d ' | grep global 0000000000617150 d global.2530
...
which makes test5 fail:
...
test5: unexpected syminfo name got global.2530 expected global
...
Fix this failure by accepting this type of name as a valid name in btest_lto.
2019-02-26 Tom de Vries <tdevries@suse.de>
* btest.c (test5): Allow global.* as minimal symbol name for global.
Richard Biener [Tue, 26 Feb 2019 14:09:19 +0000 (14:09 +0000)]
re PR tree-optimization/89505 (LibreOffice miscompilation starting with r260383)
2019-02-26 Richard Biener <rguenther@suse.de>
PR tree-optimization/89505
* tree-ssa-structalias.c (compute_dependence_clique): Make sure
to handle restrict pointed-to vars with multiple subvars
correctly.
Jakub Jelinek [Tue, 26 Feb 2019 10:08:29 +0000 (11:08 +0100)]
re PR target/89474 (ice in df_reg_chain_verify_unmarked, at df-scan.c:3995)
PR target/89474
* config/i386/i386.c (remove_partial_avx_dependency): Call
df_analyze etc. before creation of the v4sf_const0 pseudo, rather than
after changing possibly many instructions to use that pseudo. Fix up
insertion of v4sf_const0 setter at the start of bb.
Tom de Vries [Tue, 26 Feb 2019 07:54:57 +0000 (07:54 +0000)]
[libbacktrace] Require dwz for b3test_dwz_buildid
If dwz is not available, then the libbacktrace test b3test_dwz_buildid fails
like this:
...
gmake[4]: *** No rule to make target 'b3test_dwz_buildid'
...
Fix this by guarding the test with HAVE_DWZ.
Tested on x86_64 with and without dwz installed.
2019-02-26 Tom de Vries <tdevries@suse.de>
* Makefile.am (TESTS): Only add b3test_dwz_buildid if HAVE_DWZ.
* Makefile.in: Regenerate.
Jakub Jelinek [Mon, 25 Feb 2019 22:58:45 +0000 (23:58 +0100)]
re PR c/77754 (internal compiler error : tree code 'call_expr' is not supported in LTO streams)
PR c/77754
* gcc.c-torture/compile/pr77754-1.c: New test.
* gcc.c-torture/compile/pr77754-2.c: New test.
* gcc.c-torture/compile/pr77754-3.c: New test.
* gcc.c-torture/compile/pr77754-4.c: New test.
* gcc.c-torture/compile/pr77754-5.c: New test.
* gcc.c-torture/compile/pr77754-6.c: New test.
Sandra Loosemore [Mon, 25 Feb 2019 20:38:42 +0000 (15:38 -0500)]
c.opt (Wmissing-attributes): Clean up doc string.
2019-02-25 Sandra Loosemore <sandra@codesourcery.com>
Martin Sebor <msebor@gmail.com>
gcc/
* c-family/c.opt (Wmissing-attributes): Clean up doc string.
* common.opt (Wattribute-alias): Likewise.
* doc/invoke.texi (Option Summary): List general form of
-Wattribute-alias=. List positive form of -Wmissing-attributes.
(-Wmissing-attributes): Invert entry, rewrite and correct default.
Add cross-references.
(-Wattribute-alias): Rewrite and correct default. Mention
considered attributes (same as for -Wmissing-attributes).
Co-Authored-By: Martin Sebor <msebor@gmail.com>
From-SVN: r269196
Paul A. Clarke [Mon, 25 Feb 2019 19:36:05 +0000 (19:36 +0000)]
[rs6000] PR89338, PR89339: Fix compat vector intrinsics for BE and 32-bit
Test FAILS: sse2-cvtpd2dq-1, sse2-cvtpd2ps, sse2-cvttpd2dq on powerpc64
(big-endian).
_mm_cvtpd_epi32, _mm_cvtpd_ps, _mm_cvttpd_epi32: Type conversion from
vector doubleword type to vector word type leaves the results in even
lanes in big endian mode.
Test FAILS: sse-cvtss2si-1, sse-cvtss2si-2, sse-movmskb-1 on powerpc
(32-bit big-endian).
Incorrect type for interpreting the result from mfvsrd instruction leads
to incorrect results. Also, mfvsrd instruction only works as expected in
64-bit mode or for 32-bit quantities in 32-bit mode. A more general,
if slower, solution is needed for 32-bit mode.
2019-02-25 Paul A. Clarke <pc@us.ibm.com>
[gcc]
* config/rs6000/emmintrin.h (_mm_cvtpd_epi32): Fix big endian.
(_mm_cvtpd_ps): Likewise.
(_mm_cvttpd_epi32): Likewise.
PR target/89338
* config/rs6000/xmmintrin.h (_mm_cvtss_f32): Fix type mismatch.
(_mm_cvt_ss2si): Fix type mismatch and 32-bit.
Tamar Christina [Mon, 25 Feb 2019 17:57:01 +0000 (17:57 +0000)]
AArch64: Fix command line options canonicalization version #2. (PR target/88530)
Commandline options on AArch64 don't get canonicalized into the smallest
possible set before output to the assembler. This means that overlapping feature
sets are emitted with superfluous parts.
Normally this isn't an issue, but in the case of crypto we have retro-actively
split it into aes and sha2. We need to emit only +crypto to the assembler
so old assemblers continue to work.
Because of how -mcpu=native and -march=native work they end up enabling all
feature bits. Instead we need to get the smallest possible set, which would also
fix the problem with older the assemblers and the retro-active split.
The function that handles this is called quite often. It is called for any
push/pop options or attribute that changes arch, cpu etc. In order to not make
this search for the smallest set too expensive we sort the options based on the
number of features (bits) they enable. This allows us to process the list
linearly instead of quadratically (Once we have enabled a feature, we know that
anything else that enables it can be ignored. By sorting we'll get the biggest
groups first and thus the smallest combination of commandline flags).
The Option handling structures have been extended to have a boolean to indicate
whether the option is synthetic, with that I mean if the option flag itself
enables a new feature.
e.g. +crypto isn't an actual feature, it just enables other features, but others
like +rdma enable multiple dependent features but is itself also a feature.
There are two ways to solve this.
1) Either have the options that are feature bits also turn themselves on, e.g.
change rdma to turn on FP, SIMD and RDMA as dependency bits.
2) Make a distinction between these two different type of features and have the
framework handle it correctly.
Even though it's more code I went for the second approach, as it's the one
that'll be less fragile (people can't forget it) and gives the least surprises.
Effectively this patch changes the following:
The values before the => are the old compiler and after the => the new code.
PR target/88530
* common/config/aarch64/aarch64-common.c
(struct aarch64_option_extension): Add is_synthetic.
(all_extensions): Use it.
(TARGET_OPTION_INIT_STRUCT): Define hook.
(struct gcc_targetm_common): Moved to end.
(all_extensions_by_on): New.
(opt_ext_cmp, typedef opt_ext): New.
(aarch64_option_init_struct): New.
(aarch64_contains_opt): New.
(aarch64_get_extension_string_for_isa_flags): Output smallest set.
* config/aarch64/aarch64-option-extensions.def
(AARCH64_OPT_EXTENSION): Explicitly include AES and SHA2 in crypto.
(fp, simd, crc, lse, fp16, rcpc, rdma, dotprod, aes, sha2, sha3,
sm4, fp16fml, sve, profile, rng, memtag, sb, ssbs, predres):
Set is_synthetic to false.
(crypto): Set is_synthetic to true.
* config/aarch64/driver-aarch64.c (AARCH64_OPT_EXTENSION): Add
SYNTHETIC.
gcc/testsuite/ChangeLog:
PR target/88530
* gcc.target/aarch64/options_set_1.c: New test.
* gcc.target/aarch64/options_set_2.c: New test.
* gcc.target/aarch64/options_set_3.c: New test.
* gcc.target/aarch64/options_set_4.c: New test.
* gcc.target/aarch64/options_set_5.c: New test.
* gcc.target/aarch64/options_set_6.c: New test.
* gcc.target/aarch64/options_set_7.c: New test.
* gcc.target/aarch64/options_set_8.c: New test.
* gcc.target/aarch64/options_set_9.c: New test.
Jakub Jelinek [Mon, 25 Feb 2019 10:32:38 +0000 (11:32 +0100)]
re PR target/89434 (wrong code with -Og and __builtin_mul_overflow())
PR target/89434
* config/arm/arm.md (*subsi3_carryin_const): Use
arm_neg_immediate_operand predicate instead of
arm_not_immediate_operand, "L" constraint instead of "K" and
print it using %n2 instead of %B2.
(*subsi3_carryin_const0): New define_insn.
(*subsi3_carryin_compare_const): Use const_int_I_operand predicate
instead of arm_not_operand and "I" constraint instead of "K" and
print it using %n3 instead of %B2. Instead of using match_dup 2 add
another match_operand and in the condition check that it is negation
of operands[2].
(*subsi3_carryin_compare_const0): New define_ins.
(*subdi_di_zesidi): Adjust to use *subsi3_carryin_const0 instead of
*subsi3_carryin_const.
(*arm_cmpdi_insn): Fix splitting into *subsi3_carryin_compare_const,
split into *subsi3_carryin_compare_const0 if the highpart is zero.
Thomas Koenig [Sun, 24 Feb 2019 22:49:47 +0000 (22:49 +0000)]
re PR fortran/89174 (Allocation segfault with CLASS(*) MOLD)
2019-02-24 Thomas Koenig <tkoenig@gcc.gnu.org>
PR fortran/89174
* trans-expr.c (gfc_find_and_cut_at_last_class_ref): Add is_mold
to garguments. If we are dealing with a MOLD, call
gfc_expr_to_initialize().
* trans-stmt.c (gfc_trans_allocate): For MOLD, pass is_mold=true
to gfc_find_and_cut_at_last_class_ref.
* trans.h (gfc_find_and_cut_at_last_class_ref): Add optional
argument is_mold with default false.
2019-02-24 Thomas Koenig <tkoenig@gcc.gnu.org>
PR fortran/89174
* gfortran.dg/allocate_with_mold_3.f90: New test.
Jakub Jelinek [Sun, 24 Feb 2019 19:23:51 +0000 (20:23 +0100)]
re PR rtl-optimization/89445 (_mm512_maskz_loadu_pd "forgets" to use the mask)
PR rtl-optimization/89445
* simplify-rtx.c (simplify_ternary_operation): Don't use
simplify_merge_mask on operands that may trap.
* rtlanal.c (may_trap_p_1): Use FLOAT_MODE_P instead of
SCALAR_FLOAT_MODE_P checks. For integral division by zero, if
second operand is CONST_VECTOR, check if any element could be zero.
Don't expect traps for VEC_{MERGE,SELECT,CONCAT,DUPLICATE} unless
their operands can trap.
Jonathan Wakely [Sat, 23 Feb 2019 21:19:00 +0000 (21:19 +0000)]
P0340R2 Making std::underlying_type SFINAE-friendly
* include/std/type_traits (__underlying_type_impl): New helper to
make underlying_type SFINAE-friendly.
(underlying_type): Derive from __underlying_type_impl.
* testsuite/20_util/underlying_type/requirements/typedefs-3.cc: New
test.
Martin Sebor [Sat, 23 Feb 2019 20:46:04 +0000 (20:46 +0000)]
builtin-sprintf-10.c: Cast remaining wchar_t to wint_t to avoid a bogus -Wformat warning in ILP32...
gcc/testsuite/ChangeLog:
* gcc.dg/tree-ssa/builtin-sprintf-10.c: Cast remaining
wchar_t to wint_t to avoid a bogus -Wformat warning in
ILP32 (bug 77970).
Martin Sebor [Sat, 23 Feb 2019 20:25:57 +0000 (20:25 +0000)]
gimple-ssa-sprintf.c (target_strtol): Rename...
gcc/ChangeLog:
* gimple-ssa-sprintf.c (target_strtol): Rename...
(target_strtohwi): ...to this. Handle values up to HOST_WIDE_INT_MAX.
(parse_directive): Adjust to name change. Use HOST_WIDE_INT_MAX to
check for range error.
H.J. Lu [Sat, 23 Feb 2019 19:39:35 +0000 (19:39 +0000)]
driver: Also prune joined switches with negation
When -march=native is passed to host_detect_local_cpu to the backend,
it overrides all command lines after it. That means
$ gcc -march=native -march=skylake-avx512
is the treated as
$ gcc -march=skylake-avx512 -march=native
Prune joined switches with Negative and RejectNegative to allow
-march=skylake-avx512 to override previous -march=native on command-line.
gcc/
PR driver/69471
* opts-common.c (prune_options): Also prune joined switches
with Negative and RejectNegative.
* config/i386/i386.opt (march=): Add Negative(march=).
(mtune=): Add Negative(mtune=).
* doc/options.texi: Document Negative used together with Joined
and RejectNegative.
Paul Thomas [Sat, 23 Feb 2019 13:18:47 +0000 (13:18 +0000)]
re PR fortran/88117 (ICE in gimplify_var_or_parm_decl, at gimplify.c:2697)
2019-02-23 Paul Thomas <pault@gcc.gnu.org>
PR fortran/88117
* resolve.c (deferred_op_assign): Return if the lhs expression
has the pointer attribute.
* trans-expr.c (gfc_trans_assignment_1): Do not fix the string
length if the lhs expression has the pointer attribute.
2019-02-23 Paul Thomas <pault@gcc.gnu.org>
PR fortran/88117
* gfortran.dg/deferred_character_32.f90 : New test
Paul Thomas [Sat, 23 Feb 2019 12:18:44 +0000 (12:18 +0000)]
re PR fortran/89385 (Incorrect members of C descriptor for an allocatable object)
2019-02-23 Paul Thomas <pault@gcc.gnu.org>
PR fortran/89385
PR fortran/89366
* decl.c (gfc_verify_c_interop_param): Restriction on string
length being one is lifted for F2018.
* trans-expr.c (gfc_conv_gfc_desc_to_cfi_desc): For scalar
characters with intent in, make a temporary and copy the result
of the expression evaluation into it.
(gfc_conv_procedure_call): Set a flag for character formal args
having a character length that is not unity. If the procedure
is bind C, call gfc_conv_gfc_desc_to_cfi_desc in this case.
Also, extend bind C calls to unconditionally convert both
pointers and allocatable expressions.
2019-02-23 Paul Thomas <pault@gcc.gnu.org>
PR fortran/89385
* gfortran.dg/ISO_Fortran_binding_1.f90 : Correct test for
previously incorrect lbound for allocatable expressions. Also
correct stop values to avoid repetition.
* gfortran.dg/ISO_Fortran_binding_5.f90 : New test
* gfortran.dg/ISO_Fortran_binding_5.c : Support previous test.
PR fortran/89366
* gfortran.dg/ISO_Fortran_binding_6.f90 : New test
* gfortran.dg/ISO_Fortran_binding_6.c : Support previous test.
* gfortran.dg/pr32599.f03 : Set standard to F2008.
2019-02-23 Paul Thomas <pault@gcc.gnu.org>
PR fortran/89385
PR fortran/89366
* runtime/ISO_Fortran_binding.c (cfi_desc_to_gfc_desc) : In the
interchange between character and derived, the character type
was being set incorrectly.
(gfc_desc_to_cfi_desc) : Eliminate the interchange of types in
this function. Do not add the kind and length information to
the type field of structures. Lbounds were incorrectly being
set to zero for allocatable and pointer descriptors. Should
have been non-pointer, non-allocatables that received this
treatment.
Jonathan Wakely [Sat, 23 Feb 2019 03:01:59 +0000 (03:01 +0000)]
PR libstdc++/89446 fix null pointer dereference in char_traits
PR libstdc++/89446
* include/bits/char_traits.h (__constant_char_array): Check index is
in range before dereferencing.
(char_traits<char>::compare, char_traits<char>::find)
(char_traits<char8_t>::compare, char_traits<char8_t>::find): Return
immediately if n is zero.
(char_traits<wchar_t>::compare, char_traits<wchar_t>::find): Likewise.
Remove workarounds for PR 67026.
* testsuite/21_strings/basic_string_view/operators/char/89446.cc:
New test.
* testsuite/21_strings/basic_string_view/operators/wchar_t/89446.cc:
New test.
David Malcolm [Sat, 23 Feb 2019 01:19:38 +0000 (01:19 +0000)]
Capture source location of dtors (PR c++/89390)
gcc/cp/ChangeLog:
PR c++/89390
* parser.c (cp_parser_unqualified_id): Capture and use locations
for destructors.
gcc/testsuite/ChangeLog:
PR c++/89390
* g++.dg/diagnostic/pr89390.C: Update expected location of error,
renaming to a multicharacter name, so that start != finish. Add
tests for dtor locations.
David Malcolm [Sat, 23 Feb 2019 00:14:52 +0000 (00:14 +0000)]
re PR tree-optimization/88074 (g++ hangs on math expression)
PR middle-end/88074
* simplify.c (norm2_do_sqrt, gfc_simplify_norm2): Use
mpfr_number_p && !mpfr_zero_p instead of mpfr_regular_p.
(norm2_add_squared): Likewise. Use mp_exp_t rather than mpfr_exp_t.
Co-Authored-By: Jakub Jelinek <jakub@redhat.com>
From-SVN: r269139
Marek Polacek [Fri, 22 Feb 2019 19:24:37 +0000 (19:24 +0000)]
PR c++/89420 - ICE with CAST_EXPR in explicit-specifier.
* decl.c (build_explicit_specifier): Don't check
processing_template_decl. Call instantiation_dependent_expression_p
instead of value_dependent_expression_p. Call
instantiate_non_dependent_expr_sfinae before
build_converted_constant_expr instead of calling
instantiate_non_dependent_expr after it. Add
processing_template_decl_sentinel.
Martin Sebor [Fri, 22 Feb 2019 17:38:11 +0000 (10:38 -0700)]
PR tree-optimization/88993 - GCC 9 -Wformat-overflow=2 should reflect real libc limits
PR tree-optimization/88993 - GCC 9 -Wformat-overflow=2 should reflect real libc limits
PR tree-optimization/88835 - overly aggressive -Werror=format-overflow for printf
gcc/ChangeLog:
PR tree-optimization/88993
PR tree-optimization/88853
* gimple-ssa-sprintf.c (sprintf_dom_walker::call_info::is_file_func):
New helper.
(sprintf_dom_walker::call_info::is_string_func): New helper.
(format_directive): Only issue "may exceed" 4095/INT_MAX warnings
for formatted string functions.
(sprintf_dom_walker::handle_gimple_call): Fix a typo in a comment.
gcc/testsuite/ChangeLog:
PR tree-optimization/88993
PR tree-optimization/88853
* gcc.dg/tree-ssa/builtin-fprintf-warn-2.c: New test.
* gcc.dg/tree-ssa/builtin-printf-warn-2.c: New test.
* gcc.dg/tree-ssa/builtin-snprintf-warn-3.c: Adjust.
* gcc.dg/tree-ssa/builtin-sprintf-warn-18.c: Same.
Jeff Law [Fri, 22 Feb 2019 16:38:43 +0000 (09:38 -0700)]
re PR rtl-optimization/87761 ([MIPS] New FAIL: gcc.target/mips/fix-r4000-10.c -O1 start with r265398)
PR rtl-optimization/87761
* config/mips/mips.md: Add new combiner pattern to recognize
a bitfield extraction using (ashiftrt (truncate (ashift (...)))).
In general the stack pointer was not handled for many SUBS/ADDS patterns in
aarch64.md.
Both the "extended register" and "immediate" forms allow the stack pointer to be
used as the source register, while no form allows the stack pointer for the
destination register.
The define_insn patterns generating ADDS/SUBS did not allow the stack pointer
for any operand, while the define_peephole2 patterns that generated RTX to be
matched by these patterns allowed the stack pointer for any operand.
The patterns are fixed by adding the 'k' constraint for the first source operand
to all define_insns that generate the ADDS/SUBS "extended register" and
"immediate" forms (but not the "shifted register" form).
In peephole optimizations, constraint strings are ignored (see "(gccint) C
Constraint Interface" info node in the documentation), so the decision to act or
not is based solely on the predicate and condition.
This patch introduces a new predicate "aarch64_general_reg" to be used in
define_peephole2 patterns where only GENERAL_REGS registers are acceptable and
uses that predicate in the peepholes that generate patterns for ADDS/SUBS.
Full bootstrap and regtest done on aarch64-none-linux-gnu.
Regression tests done on aarch64-none-linux-gnu and aarch64-none-elf cross
compiler.
OK for trunk?
gcc/ChangeLog:
2019-02-22 Matthew Malcomson <matthew.malcomson@arm.com>
PR target/89324
* config/aarch64/aarch64.md: Use aarch64_general_reg predicate on
destination register in peepholes generating patterns for ADDS/SUBS.
(add<mode>3_compare0,
*addsi3_compare0_uxtw, add<mode>3_compareC,
add<mode>3_compareV_imm, add<mode>3_compareV,
*adds_<optab><ALLX:mode>_<GPI:mode>,
*subs_<optab><ALLX:mode>_<GPI:mode>,
*adds_<optab><ALLX:mode>_shift_<GPI:mode>,
*subs_<optab><ALLX:mode>_shift_<GPI:mode>,
*adds_<optab><mode>_multp2, *subs_<optab><mode>_multp2,
*sub<mode>3_compare0, *subsi3_compare0_uxtw,
sub<mode>3_compare1): Allow stack pointer for source register.
* config/aarch64/predicates.md (aarch64_general_reg): New predicate.
gcc/testsuite/ChangeLog:
2019-02-22 Matthew Malcomson <matthew.malcomson@arm.com>
PR target/89324
* gcc.dg/rtl/aarch64/subs_adds_sp.c: New test.
* gfortran.fortran-torture/compile/pr89324.f90: New test.
to avoid partial XMM register stall. This patch adds a pass to generate
a single
vxorps %xmmN, %xmmN, %xmmN
at entry of the nearest dominator for basic blocks with SF/DF conversions,
which is in the fake loop that contains the whole function, instead of
generating one
vxorp[ds] %xmmN, %xmmN, %xmmN
for each SF/DF conversion.
NB: The LCM algorithm isn't appropriate here since it may place a vxorps
inside the loop. Simple testcase show this:
$ cat badcase.c
extern float f;
extern double d;
void
foo (int n, int k)
{
for (int j = 0; j != n; j++)
if (j < k)
d = f;
}
Sudakshina Das [Fri, 22 Feb 2019 10:55:22 +0000 (10:55 +0000)]
[GCC, AArch64] Fix a couple of bugs in BTI
While doing more testing I found a couple of issues with my BTI patches.
This patch fixes them:
1) Remove a reference to return address key. The original patch was written
based on a different not yet committed patch ([PATCH 3/3][GCC][AARCH64] Add
support for pointer authentication B key) and I missed out on cleaning this
up. This is hidden behind the configuration option and thus went unnoticed.
2) Add a missed case for adding the BTI instruction in thunk functions.
*** gcc/ChangeLog ***
2019-02-22 Sudakshina Das <sudi.das@arm.com>
* config/aarch64/aarch64.c (aarch64_output_mi_thunk): Add bti
instruction if enabled.
(aarch64_override_options): Remove reference to return address key.
Thomas Schwinge [Fri, 22 Feb 2019 10:50:18 +0000 (11:50 +0100)]
Revert earlier OpenACC 'routine' changes
..., which were committed as part of, but logically don't belong into r261813
"Update OpenACC data clause semantics to the 2.5 behavior", and which confuse
later changes.
Thomas Schwinge [Fri, 22 Feb 2019 10:49:43 +0000 (11:49 +0100)]
[C, C++] Use correct location information for OpenACC shape and simple clauses
gcc/c/
* c-parser.c (c_parser_oacc_shape_clause): Add loc formal
parameter. Adjust all users.
(c_parser_oacc_simple_clause): Replace parser with loc formal
parameter. Adjust all users.
gcc/cp/
* parser.c (cp_parser_oacc_simple_clause): Remove parser formal
parameter, move loc formal parameter to the front. Adjust all
users.
(cp_parser_oacc_shape_clause): Add loc formal parameter. Adjust
all users.
Kyrylo Tkachov [Fri, 22 Feb 2019 08:59:20 +0000 (08:59 +0000)]
[arm] Add support for Neoverse N1
This patch adds support for the Neoverse N1 [1]. This CPU was previously supported through the Ares codename.
-mcpu=ares is retained as an alias of the new -mcpu=neoverse-n1.
Bootstrapped and tested on arm-none-linux-gnueabihf.
Kyrylo Tkachov [Fri, 22 Feb 2019 08:58:04 +0000 (08:58 +0000)]
[AArch64] Add support for Neoverse E1
This patch adds -mcpu and -mtune support for the Neoverse E1 CPU [1].
The new option is -mcpu=neoverse-e1.
Bootstrapped and tested on aarch64-none-linux-gnu.
Kyrylo Tkachov [Fri, 22 Feb 2019 08:56:50 +0000 (08:56 +0000)]
[AArch64] Add support for Neoverse N1
This patch adds support for the Neoverse N1 CPU [1]. This was supported in GCC earlier through the codename Ares,
which it now replaces. -mcpu=ares is still accepted as there's been a binutils release supporting it,
but the internal structures are renamed to use Neoverse N1-related identifiers.
Bootstrapped and tested on aarch64-none-linux-gnu.
* config/aarch64/aarch64.c (ares_tunings): Rename to...
(neoversen1_tunings): ... This.
* config/aarch64/aarch64-cores.def (ares): Change tuning to the above.
(neoverse-n1): New CPU.
* config/aarch64/aarch64-tune.md: Regenerate.
* doc/invoke.txt (AArch64 Options): Document neoverse-n1.
Richard Biener [Fri, 22 Feb 2019 08:38:14 +0000 (08:38 +0000)]
re PR tree-optimization/87609 (miscompilation with restrict and loop)
2019-02-22 Richard Biener <rguenther@suse.de>
PR middle-end/87609
* cfghooks.h (dependence_hash): New typedef.
(struct copy_bb_data): New type.
(cfg_hooks::duplicate_block): Adjust to take a copy_bb_data argument.
(duplicate_block): Likewise.
* cfghooks.c (duplicate_block): Pass down copy_bb_data.
(copy_bbs): Create and pass down copy_bb_data.
* cfgrtl.c (cfg_layout_duplicate_bb): Adjust.
(rtl_duplicate_bb): Likewise.
* tree-cfg.c (gimple_duplicate_bb): If the copy_bb_data arg is not NULL
remap dependence info.
Richard Biener [Fri, 22 Feb 2019 08:36:30 +0000 (08:36 +0000)]
re PR tree-optimization/87609 (miscompilation with restrict and loop)
2019-02-22 Richard Biener <rguenther@suse.de>
PR tree-optimization/87609
* tree-core.h (tree_base): Document special clique values.
* tree-inline.c (remap_dependence_clique): Do not use the
special clique value of one.
(maybe_set_dependence_info): Use clique one.
(clear_dependence_clique): New callback.
(compute_dependence_clique): Clear clique one from all refs
before assigning it (again).
Eric Botcazou [Fri, 22 Feb 2019 07:25:52 +0000 (07:25 +0000)]
configure.host (abi_baseline_pair): Adjust for SPARC64/Linux.
* configure.host (abi_baseline_pair): Adjust for SPARC64/Linux.
* config/abi/post/sparc64-linux-gnu: New directory.
* config/abi/post/sparc64-linux-gnu/baseline_symbols.txt: New file.
* config/abi/post/sparc64-linux-gnu/32: New directory.
* config/abi/post/sparc64-linux-gnu/32/baseline_symbols.txt: New file.
The standard says that in a generic lambda we should speculatively capture
'this' if we see a call to an overload set that contains a non-static member
function, but it seems wrong to reject the program if we can't capture,
since it might not actually be needed.
* lambda.c (lambda_expr_this_capture): Change add_capture_p to int.
(maybe_generic_this_capture): Pass -1.
Jason Merrill [Fri, 22 Feb 2019 02:47:33 +0000 (21:47 -0500)]
PR c++/88394 - ICE with VLA init-capture.
We mostly use is_normal_capture_proxy to decide whether or not to use
DECL_CAPTURED_VARIABLE; we could just check whether it's set. VLA capture
is still mostly broken, but this fixes this ICE.
Jason Merrill [Fri, 22 Feb 2019 02:08:05 +0000 (21:08 -0500)]
PR c++/88869 - C++17 ICE with CTAD and explicit specialization.
The members of an explicit specialization of a class template don't have the
template parameters of that class template, so we shouldn't try to provide
arguments for them. Only set outer_args when the class is an instantiation.
* pt.c (do_class_deduction): Don't include explicit specialization
args in outer_args.
Tom Honermann [Fri, 22 Feb 2019 01:16:08 +0000 (01:16 +0000)]
P0482R5 char8_t: Updates to existing standard library tests
This patch augments existing tests to validate behavior for char8_t. In
all cases, added test cases are cloned from existing tests for wchar_t
or char16_t.
A few tests required updates to line numbers for diagnostic messages.
2019-02-22 Tom Honermann <tom@honermann.net>
* testsuite/18_support/byte/ops.cc: Validate
std::to_integer<char8_t>, std::to_integer<char16_t>, and
std::to_integer<char32_t>.
* testsuite/18_support/numeric_limits/dr559.cc: Validate
std::numeric_limits<char8_t>.
* testsuite/18_support/numeric_limits/lowest.cc: Validate
std::numeric_limits<char8_t>::lowest().
* testsuite/18_support/numeric_limits/max_digits10.cc: Validate
std::numeric_limits<char8_t>::max_digits10.
* testsuite/18_support/type_info/fundamental.cc: Validate
typeinfo for char8_t.
* testsuite/20_util/from_chars/1_c++20_neg.cc: New test, validating
std::from_chars with char8_t.
* testsuite/20_util/hash/requirements/explicit_instantiation.cc:
Validate explicit instantiation of std::hash<char8_t>.
* testsuite/20_util/is_integral/value.cc: Validate
std::is_integral<char8_t>.
* testsuite/20_util/make_signed/requirements/typedefs-4.cc:
Validate std::make_signed<char8_t>.
* testsuite/21_strings/basic_string/cons/char/deduction.cc:
Validate u8string construction from char8_t sources.
* testsuite/21_strings/basic_string/types/pmr_typedefs.cc: Validate
std::pmr::u8string.
* testsuite/21_strings/basic_string_view/operations/compare/
char/70483.cc: Validate substr operations on u8string_view.
* testsuite/21_strings/basic_string_view/typedefs.cc: Validate that
the u8string_view typedef is defined.
* testsuite/21_strings/char_traits/requirements/
constexpr_functions.cc: Validate char_traits<char8_t> constexpr
member functions.
* testsuite/21_strings/char_traits/requirements/
constexpr_functions_c++17.cc: Validate char_traits<char8_t> C++17
constexpr member functions.
* testsuite/21_strings/headers/string/types_std_c++0x.cc: Validate
that the u8string typedef is defined.
* testsuite/22_locale/locale/cons/unicode.cc: Validate the presence
of the std::codecvt<char16_t, char8_t, std::mbstate_t> and
std::codecvt<char32_t, char8_t, std::mbstate_t> facets.
* testsuite/29_atomics/atomic/cons/assign_neg.cc: Update line
numbers.
* testsuite/29_atomics/atomic/cons/copy_neg.cc: Likewise.
* testsuite/29_atomics/atomic_integral/cons/assign_neg.cc:
Likewise.
* testsuite/29_atomics/atomic_integral/cons/copy_neg.cc: Likewise.
* testsuite/29_atomics/atomic_integral/is_always_lock_free.cc:
Validate std::atomic<char8_t>::is_always_lock_free
* testsuite/29_atomics/atomic_integral/operators/bitwise_neg.cc:
Update line numbers.
* testsuite/29_atomics/atomic_integral/operators/decrement_neg.cc:
Likewise.
* testsuite/29_atomics/atomic_integral/operators/increment_neg.cc:
Likewise.
* testsuite/experimental/polymorphic_allocator/pmr_typedefs_string.cc:
Validate std::experimental::pmr::u8string.
* testsuite/experimental/string_view/typedefs.cc: Validate that the
u8string_view typedef is defined.
* testsuite/util/testsuite_common_types.h: Add char8_t, char16_t and
char32_t to the typelists.
This patch corrects ambiguous partial specializations of
typelist::detail::append_. Previously, neither append_<chain<Hd, Tl>,
Typelist_Chain> nor append_<Typelist_Chain, null_type> was a better
match for append_<chain<Hd, Tl>, null_type>.
2019-02-22 Tom Honermann <tom@honermann.net>
* include/ext/typelist.h: Constrain a partial specialization of
typelist::detail::append_ to only match chain<T1,T2>.
Ian Lance Taylor [Fri, 22 Feb 2019 00:41:25 +0000 (00:41 +0000)]
cmd/go: remove work directory on usage error
Backport https://golang.org/cl/163237 from the master library:
Ensure that cmd/go consistently calls base.Exit rather than os.Exit,
so that we don't incorrectly leave the work directory around on exit.
Test this by modifying the testsuite to run all the tests with TMPDIR
set to a temporary directory, and then check that no files are left
behind in that temporary directory. Adjust a couple of tests to make
this approach work.
Jason Merrill [Thu, 21 Feb 2019 23:07:47 +0000 (18:07 -0500)]
PR c++/89422 - ICE with -g and lambda in default arg in template.
Here, we were trying to instantiate the default argument before setting
DECL_FRIEND_CONTEXT, so that the instantiated lambda ended up being treated
as part of the S template, which confused dwarf2out.