Jason Merrill [Mon, 25 Sep 2023 09:15:02 +0000 (10:15 +0100)]
c++: mangle function template constraints
Per https://github.com/itanium-cxx-abi/cxx-abi/issues/24 and
https://github.com/itanium-cxx-abi/cxx-abi/pull/166
We need to mangle constraints to be able to distinguish between function
templates that only differ in constraints. From the latter link, we want to
use the template parameter mangling previously specified for lambdas to also
make explicit the form of a template parameter where the argument is not a
"natural" fit for it, such as when the parameter is constrained or deduced.
I'm concerned about how the latter link changes the mangling for some C++98
and C++11 patterns, so I've limited template_parm_natural_p to avoid two
cases found by running the testsuite with -Wabi forced on:
template <class T, T V> T f() { return V; }
int main() { return f<int,42>(); }
template <int i> int max() { return i; }
template <int i, int j, int... rest> int max()
{
int sub = max<j, rest...>();
return i > sub ? i : sub;
}
int main() { return max<1,2,3>(); }
A third C++11 pattern is changed by this patch:
template <template <typename...> class TT, typename... Ts> TT<Ts...> f();
template <typename> struct A { };
int main() { f<A,int>(); }
I aim to resolve these with the ABI committee before GCC 14.1.
We also need to resolve https://github.com/itanium-cxx-abi/cxx-abi/issues/38
(mangling references to dependent template-ids where the name is fully
resolved) as references to concepts in std:: will consistently run into this
area. This is why mangle-concepts1.C only refers to concepts in the global
namespace so far.
The library changes are to avoid trying to mangle builtins, which fails.
Demangler support and test coverage is not complete yet.
Alexandre Oliva [Fri, 1 Dec 2023 17:31:22 +0000 (14:31 -0300)]
hardcfr: make builtin_return tests more portable [PR112334]
Rework __builtin_return tests to explicitly call __builtin_apply and
use its return value rather than anything else. Also require
untyped_assembly. Avoid the noise out of exceptions escaping the
builtin-applied function, but add a test to cover their effects as
well.
[PR112445][LRA]: Fix "unable to find a register to spill" error
PR112445 is a very complicated bug occurring from interaction of
constraint subpass, inheritance, and hard reg live range splitting.
It is hard to debug this PR only from LRA standard logs. Therefore I
added dumping all func insns at the end of complicated sub-passes
(constraint, inheritance, undoing inheritance, hard reg live range
splitting, and rematerialization). As such output can be quite big,
it is switched only one level 7 of -fira-verbose value. The reason
for the bug is a skip of live-range splitting of hard reg (dx) on the
1st live range splitting subpass. Splitting is done for reload
pseudos around an original insn and its reload insns but the subpass
did not recognize such insn pattern because previous inheritance and
undoing inheritance subpasses extended a bit reload pseudo live range.
Although we undid inheritance in question, the result code was a bit
different from a code before the corresponding inheritance pass. The
following fixes the bug by restoring exact code before the
inheritance.
gcc/ChangeLog:
PR target/112445
* lra.h (lra): Add one more arg.
* lra-int.h (lra_verbose, lra_dump_insns): New externals.
(lra_dump_insns_if_possible): Ditto.
* lra.cc (lra_dump_insns): Dump all insns.
(lra_dump_insns_if_possible): Dump all insns for lra_verbose >= 7.
(lra_verbose): New global.
(lra): Add new arg. Setup lra_verbose from its value.
* lra-assigns.cc (lra_split_hard_reg_for): Dump insns if rtl
was changed.
* lra-remat.cc (lra_remat): Dump insns if rtl was changed.
* lra-constraints.cc (lra_inheritance): Dump insns.
(lra_constraints, lra_undo_inheritance): Dump insns if rtl
was changed.
(remove_inheritance_pseudos): Use restore reg if it is set up.
* ira.cc: (lra): Pass internal_flag_ira_verbose.
gcc/testsuite/ChangeLog:
PR target/112445
* gcc.target/i386/pr112445.c: New test.
Jakub Jelinek [Fri, 1 Dec 2023 15:57:30 +0000 (16:57 +0100)]
extend.texi: Fix up defbuiltin* with spaces in return type
In https://gcc.gnu.org/onlinedocs/gcc/Other-Builtins.html#index-_005f_005fbuiltin_005fstdc_005fbit_005ffloor
I've noticed that while e.g. __builtin_stdc_bit_floor builtin is properly
rendered in bold and bigger size, for the __builtin_stdc_bit_width builtin
it is not the builtin name which is marked like that, but the keyword int
before it. Also, seems such builtins are missing from the index.
I've read the texinfo docs and they seem to suggest in
https://www.gnu.org/software/texinfo/manual/texinfo/html_node/Line-Macros.html
that return types of functions with spaces in the return type should be
wrapped with {}s and we already use that e.g. in
@defbuiltin{{void *} __builtin_thread_pointer (void)}
The following patch adjusts builtins I found which contained one or two
spaces in the return type name (plus two spots which used 2 spaces after
single keyword return type instead of 1 which triggered my search regex as
well).
2023-12-01 Jakub Jelinek <jakub@redhat.com>
* doc/extend.texi (__builtin_addc, __builtin_addcl, __builtin_addcll,
__builtin_subc, __builtin_subcl, __builtin_subcll,
__builtin_stdc_bit_width, __builtin_stdc_count_ones,
__builtin_stdc_count_zeros, __builtin_stdc_first_leading_one,
__builtin_stdc_first_leading_zero, __builtin_stdc_first_trailing_one,
__builtin_stdc_first_trailing_zero, __builtin_stdc_has_single_bit,
__builtin_stdc_leading_ones, __builtin_stdc_leading_zeros,
__builtin_stdc_trailing_ones, __builtin_stdc_trailing_zeros,
__builtin_nvptx_brev, __builtin_nvptx_brevll, __builtin_darn,
__builtin_darn_raw, __builtin_ia32_vec_ext_v2di,
__builtin_ia32_crc32qi, __builtin_ia32_crc32hi,
__builtin_ia32_crc32si, __builtin_ia32_crc32di): Put {}s around
return type with spaces in it.
(__builtin_rx_mvfachi, __builtin_rx_mvfacmi): Remove superfluous
whitespace.
Rainer Orth [Fri, 1 Dec 2023 15:47:28 +0000 (16:47 +0100)]
ada: Fix Ada bootstrap on macOS
The recent warning changes broke Ada bootstrap on macOS:
adaint.c: In function '__gnat_copy_attribs':
adaint.c:3336:10: error: implicit declaration of function 'utimes'; did you
mean 'utime'? [-Wimplicit-function-declaration]
3336 | if (utimes (to, tbuf) == -1) {
| ^~~~~~
| utime
adaint.c: In function '__gnat_kill':
adaint.c:3597:3: error: implicit declaration of function 'kill'
[-Wimplicit-function-declaration]
3597 | kill (pid, sig);
| ^~~~
terminals.c: In function 'allocate_pty_desc':
terminals.c:1196:12: error: implicit declaration of function 'openpty'; did
you mean 'openat'? [-Wimplicit-function-declaration]
1196 | status = openpty (&master_fd, &slave_fd, NULL, NULL, NULL);
| ^~~~~~~
| openat
terminals.c: In function '__gnat_setup_winsize':
terminals.c:1392:6: error: implicit declaration of function 'kill'
[-Wimplicit-function-declaration]
1392 | kill (desc->child_pid, SIGWINCH);
| ^~~~
This patch fixes this by including the necessary headers: <sys/time.h>
for utimes, <signal.h> for kill, and <util.h> for openpty. With those
changes, the build completed on x86_64-apple-darwin2[0-3] (make check
still running).
David Malcolm [Fri, 1 Dec 2023 13:47:41 +0000 (08:47 -0500)]
diagnostics, analyzer: add optional per-diagnostic property bags to SARIF
I've found it useful in debugging the analyzer for the SARIF output to
contain extra analyzer-specific data in each diagnostic.
This patch:
* adds a way for a diagnostic_metadata to populate a property
bag within a SARIF "result" object based on a new vfunc
* reworks how diagnostics are emitted within the analyzer so
that a custom diagnostic_metadata subclass is used, which populates
the property bag with information from the saved_diagnostic, and with
a vfunc hook allowing for per-pending_diagnotic-subclass extra
properties.
Doing so makes it trivial to go from the SARIF output back to
pertinent parts of the analyzer's internals (e.g. the index of
the diagnostic within the ana::diagnostic_manager, the index of
the ana::exploded_node, etc).
It also replaces a lot of boilerplate in the "emit" implementations
in the various pending_diagnostics subclasses. In particular, doing
so fixes missing CVE metadata for -Wanalyzer-fd-phase-mismatch (where
sm-fd.cc's fd_phase_mismatch::emit was failing to use its
diagnostic_metadata instance).
gcc/analyzer/ChangeLog:
* analyzer.h (class saved_diagnostic): New forward decl.
* bounds-checking.cc: Update for changes to
pending_diagnostic::emit.
* call-details.cc: Likewise.
* diagnostic-manager.cc: Include "diagnostic-format-sarif.h".
(saved_diagnostic::maybe_add_sarif_properties): New.
(class pending_diagnostic_metadata): New.
(diagnostic_manager::emit_saved_diagnostic): Create a
pending_diagnostic_metadata and a diagnostic_emission_context.
Pass the latter to the pending_diagnostic::emit vfunc.
* diagnostic-manager.h
(saved_diagnostic::maybe_add_sarif_properties): New decl.
* engine.cc: Update for changes to pending_diagnostic::emit.
* infinite-loop.cc: Likewise.
* infinite-recursion.cc: Likewise.
* kf-analyzer.cc: Likewise.
* kf.cc: Likewise.
* pending-diagnostic.cc
(diagnostic_emission_context::get_pending_diagnostic): New.
(diagnostic_emission_context::warn): New.
(diagnostic_emission_context::inform): New.
* pending-diagnostic.h (class diagnostic_emission_context): New.
(pending_diagnostic::emit): Update params.
(pending_diagnostic::maybe_add_sarif_properties): New vfunc.
* region.cc: Don't include "diagnostic-metadata.h".
* region-model.cc: Include "diagnostic-format-sarif.h". Update
for changes to pending_diagnostic::emit.
(exposure_through_uninit_copy::maybe_add_sarif_properties): New.
* sm-fd.cc: Update for changes to pending_diagnostic::emit.
* sm-file.cc: Likewise.
* sm-malloc.cc: Likewise.
* sm-pattern-test.cc: Likewise.
* sm-sensitive.cc: Likewise.
* sm-signal.cc: Likewise.
* sm-taint.cc: Likewise.
* store.cc: Don't include "diagnostic-metadata.h".
* varargs.cc: Update for changes to pending_diagnostic::emit.
gcc/ChangeLog:
* diagnostic-core.h (emit_diagnostic_valist): New overload decl.
* diagnostic-format-sarif.cc (sarif_builder::make_result_object):
When we have metadata, call its maybe_add_sarif_properties vfunc.
* diagnostic-metadata.h (class sarif_object): Forward decl.
(diagnostic_metadata::~diagnostic_metadata): New.
(diagnostic_metadata::maybe_add_sarif_properties): New vfunc.
* diagnostic.cc (emit_diagnostic_valist): New overload.
gcc/testsuite/ChangeLog:
* gcc.dg/analyzer/fd-accept.c: Update for fix to missing CWE
metadata for -Wanalyzer-fd-phase-mismatch.
* gcc.dg/analyzer/fd-bind.c: Likewise.
* gcc.dg/analyzer/fd-socket-misuse.c: Likewise.
* gcc.dg/plugin/analyzer_cpython_plugin.c: Update for changes to
pending_diagnostic::emit.
* gcc.dg/plugin/analyzer_gil_plugin.c: Likewise.
Signed-off-by: David Malcolm <dmalcolm@redhat.com>
* config/riscv/vector.md: Support highpart overlap for vx/vf.
gcc/testsuite/ChangeLog:
* gcc.target/riscv/rvv/base/pr112431-22.c: New test.
* gcc.target/riscv/rvv/base/pr112431-23.c: New test.
* gcc.target/riscv/rvv/base/pr112431-24.c: New test.
* gcc.target/riscv/rvv/base/pr112431-25.c: New test.
* gcc.target/riscv/rvv/base/pr112431-26.c: New test.
* gcc.target/riscv/rvv/base/pr112431-27.c: New test.
* config/riscv/vector.md: Support highpart overlap for indexed load.
gcc/testsuite/ChangeLog:
* gcc.target/riscv/rvv/base/pr112431-28.c: New test.
* gcc.target/riscv/rvv/base/pr112431-29.c: New test.
* gcc.target/riscv/rvv/base/pr112431-30.c: New test.
* gcc.target/riscv/rvv/base/pr112431-31.c: New test.
* gcc.target/riscv/rvv/base/pr112431-32.c: New test.
* gcc.target/riscv/rvv/base/pr112431-33.c: New test.
Richard Biener [Fri, 1 Dec 2023 10:14:53 +0000 (11:14 +0100)]
Fix ambiguity between vect_get_vec_defs with/without vectype
When querying a single set of vector defs with the overloaded
vect_get_vec_defs API then when you try to use the overload with
the vector type specified the call will be ambiguous with the
variant without the vector type. The following fixes this by
re-ordering the vector type argument to come before the output
def vector argument.
I've changed vectorizable_conversion as that triggered this
so it has coverage showing this works. The motivation is to
reduce the number of (redundant) get_vectype_for_scalar_type
calls.
* tree-vectorizer.h (vect_get_vec_defs): Re-order arguments.
* tree-vect-stmts.cc (vect_get_vec_defs): Likewise.
(vectorizable_condition): Update caller.
(vectorizable_comparison_1): Likewise.
(vectorizable_conversion): Specify the vector type to be
used for invariant/external defs.
* tree-vect-loop.cc (vect_transform_reduction): Update caller.
Jakub Jelinek [Fri, 1 Dec 2023 11:45:58 +0000 (12:45 +0100)]
testsuite: Tweak some further tests for modern C changes
On IRC Richi mentioned some FAILs in gcc.target/x86_64 and in pr83126.c.
The following patch fixes the former ones (they need recent binutils to
be enabled), for pr83126.c because I didn't have graphite configured I've
just verified that the test compiles (didn't without the patch) and that
the gimple dump is identical with one from yesterday's gcc (as it was a
tree-parloops.cc ICE, I guess identical gimple is all we care about
and no need to verify it further).
2023-12-01 Jakub Jelinek <jakub@redhat.com>
* gcc.target/x86_64/abi/avx512fp16/m512h/test_passing_m512.c
(fun_check_passing_m512_8_values, fun_check_passing_m512h_8_values):
Add missing void return type.
* gcc.target/x86_64/abi/avx512fp16/m256h/test_passing_m256.c
(fun_check_passing_m256_8_values, fun_check_passing_m256h_8_values):
Likewise.
* gcc.dg/graphite/pr83126.c (ew): Add missing casts to __INTPTR_TYPE__
and then to int *.
Jakub Jelinek [Fri, 1 Dec 2023 09:56:52 +0000 (10:56 +0100)]
lower-bitint: Fix lowering of middle sized _BitInt operations which can throw [PR112770]
The middle kind _BitInt lowering is mostly done by casting the BITINT_TYPE
operands (if any) to a signed/unsigned integer type which has larger/equal
precision, using such integer type also for the lhs (if BITINT_TYPE) and
and adding a cast after the statement from that new lhs to the old
(BITINT_TYPE) lhs. Note, for middle kind this isn't done for GIMPLE_CALLs.
Most of the time that works nicely, the exception as the following testcase
shows is -fnon-call-exceptions and some operations which can trap. Because
inserting the cast to a new lhs after the statement results in a trapping
statement in the middle of a basic block.
The following patch fixes that by emitting the cast on the fallthru edge
instead.
2023-12-01 Jakub Jelinek <jakub@redhat.com>
PR middle-end/112770
* gimple-lower-bitint.cc (gimple_lower_bitint): When adjusting
lhs of middle _BitInt setter which ends bb, insert cast on
the fallthru edge rather than after stmt.
Jakub Jelinek [Fri, 1 Dec 2023 09:55:49 +0000 (10:55 +0100)]
lower-bitint: Fix up handle_operand_addr for 0 constants [PR112771]
handle_operand_addr for INTEGER_CSTs uses wi::min_precision (UNSIGNED
for non-negative constants, SIGNED for negative ones) and from that
computes mp as minimum number of limbs which can represent that value,
and in some cases creates a test BITINT_TYPE with that precision to
categorize it and decide based on that what types to use on the constant
emitted into memory. For the actual precisions (what will be passed
to libgcc) it actually already uses MAX/MIN to adjust the corner cases:
*prec = MAX (min_prec, 1);
...
*prec = MIN ((int) -min_prec, -2);
but for integer_zerop min_prec will be 0,
mp = CEIL (min_prec, limb_prec) * limb_prec;
will be also 0 and we ICE trying to build unsigned BITINT_TYPE with
0 precision.
Fixed thusly by noting even 0 has to be encoded at least as one limb.
2023-12-01 Jakub Jelinek <jakub@redhat.com>
PR middle-end/112771
* gimple-lower-bitint.cc (bitint_large_huge::handle_operand_addr):
Use mp = 1 if it is zero.
bpf: quote section names whenever necessary in assembly output
In BPF section names are used to encode the kind of BPF program and
other metadata, involving all sort of non alphanumeric characters.
For example:
/* use auto-attach format for section definition. */
SEC("uretprobe//proc/self/exe:trigger_func2")
int handle_uretprobe_byname(struct pt_regs *ctx)
{
uretprobe_byname_res = 6;
return 0;
}
The above requires to quote the section name in the output assembly
file, since otherwise the // in the name would be interpreted by the
assembler lexer as the beginning of a line comment.
This patch makes the BPF backend to emit quoted section names in
.section directives if the name requires to be quoted. Simple section
names whose constituent characters are in the set [0-9a-zA-Z_] are
still emitted unquoted.
Tested in target bpf-unknown-none-gcc and host x86_64-linux-gnu.
gcc/ChangeLog
* config/bpf/bpf.cc (bpf_asm_named_section): New function.
(TARGET_ASM_NAMED_SECTION): Set to bpf_asm_named_section.
gcc/testsuite/ChangeLog
* gcc.target/bpf/section-name-quoting-1.c: New test.
Di Zhao [Fri, 1 Dec 2023 08:16:40 +0000 (16:16 +0800)]
aarch64: modify Ampere CPU tunings on reassociation/FMA
1. Allow reassociation on FP additions.
2. Avoid generating loop-dependant FMA chains. Added a tuning
option 'AARCH64_EXTRA_TUNE_AVOID_CROSS_LOOP_FMA' for this.
gcc/ChangeLog:
* config/aarch64/aarch64-tuning-flags.def
(AARCH64_EXTRA_TUNING_OPTION): New tuning option to avoid
cross-loop FMA.
* config/aarch64/aarch64.cc
(aarch64_override_options_internal): Set
param_avoid_fma_max_bits according to tuning option.
* config/aarch64/tuning_models/ampere1.h (ampere1_tunings):
Modify tunings related with FMA.
* config/aarch64/tuning_models/ampere1a.h (ampere1a_tunings):
Likewise.
* config/aarch64/tuning_models/ampere1b.h (ampere1b_tunings):
Likewise.
SME will add more intrinsics whose expansion code requires
the mode of the function return value. This patch adds an
associated helper routine.
gcc/
* config/aarch64/aarch64-sve-builtins.h
(function_expander::result_mode): New member function.
* config/aarch64/aarch64-sve-builtins-base.cc
(svld234_impl::expand): Use it.
* config/aarch64/aarch64-sve-builtins.cc
(function_expander::get_reg_target): Likewise.
Jakub Jelinek [Fri, 1 Dec 2023 08:26:56 +0000 (09:26 +0100)]
lower-bitint: Fix up maximum addition/subtraction/multiplication result computations
When debugging PR112750, I've noticed some issues in the computation
of precisions and the following patch attempts to fix those.
The pass uses range_to_prec function, which possibly using ranger returns
minimum precision of some operand in the style that libgcc _BitInt
entrypoints expect, i.e. for operands with unsigned types either the
precision of that type or with help of ranger
wi::min_precision (upper_bound, UNSIGNED) (done both if the types
are really unsigned or even when lower_bound is non-negative), while
for operands with signed types either negated precision of that type or
with help of ranger negated value of maximum of SIGNED min_precisions
of lower and upper bound.
Because _BitInt in C only supports unsigned precisions >= 1 and signed
precisions >= 2, the function also ensures that 0 is never returned (returns
1 instead) and should ensure that -1 is never returned (should return -2).
That is the first bug I found though, for the ranger case it ensured that,
but if an operand would be signed 1-bit precision (that means
non-BITINT_TYPE) operand, it could return -1.
Another thing is that both lower_addsub_overflow and lower_mul_overflow
compute from the prec0 and prec1 of the operands (returned by range_to_prec
with the above value meanings) prec2, which is how many bits of the result
we actually need to compute to know the infinite precision result.
This is then used by arith_overflow function together with prec
(TYPE_PRECISION (type)), type (type of the result), prec0 and prec1 to
compute which range of bits should be tested (if any, or that there is never
an overflow) and with which kind (require those bits to be zero vs.
check if all those bits together all all zeros/ones).
The arith_overflow function has one special case, when
prec0 >= 0 && prec1 >= 0 and operation is not .SUB_OVERFLOW; in that case
we treat prec2 as minimum precision to express any infinite precision
unsigned result (the result is never negative in that case), while
in all other cases prec2 is treated as minimum precision to express
any infinite precision signed result (because the result can be also
negative).
The computations of those values were apparently incorrect for all of
.{ADD,SUB}_OVERFLOW (in that case only if one operand was signed and
the other unsigned) and for .MUL_OVERFLOW it was sometimes too large.
It took me a while to get to the right expression how to compute that,
I've started with writing into the comment the possible results for
different prec0 and prec1 values (used just 8/-8/10/-10 as placeholders
for equal or different absolute values of the 2 precisions and cases
with positive and/or negative signs) and then turned into the attached
test program that actually printed what I was writing to make sure
I didn't make mistakes in it and in the end also verified the computation,
this time for all combinations of 1..14 and -2..-14 precisions.
The UNSIGNED vs. SIGNED in the table is what arith_overflow expects
the prec2 to be (see above mentioned exception).
2023-12-01 Jakub Jelinek <jakub@redhat.com>
* gimple-lower-bitint.cc (range_to_prec): Don't return -1 for
signed types.
(bitint_large_huge::lower_addsub_overflow): Fix up computation of
prec2.
(bitint_large_huge::lower_mul_overflow): Likewise.
Jakub Jelinek [Fri, 1 Dec 2023 08:26:24 +0000 (09:26 +0100)]
lower-bitint: Fix ICE on bitint-39.c
torture/bitint-39.c ICEs with -O1; the problem is that the
finish_arith_overflow code in one spot replaces use_stmt with an
assignment or cast, but if unlucky and m_gsi iterator is the same statement,
when the code later
tree clobber = build_clobber (TREE_TYPE (var), CLOBBER_EOL);
g = gimple_build_assign (var, clobber);
gsi_insert_after (&m_gsi, g, GSI_SAME_STMT);
it will insert after iterator which contains already replaced statement and
that causes the gimple chain corruption.
2023-12-01 Jakub Jelinek <jakub@redhat.com>
* gimple-lower-bitint.cc (bitint_large_huge::finish_arith_overflow):
When replacing use_stmt which is gsi_stmt (m_gsi), update m_gsi to
the new statement.
The .{ADD,SUB}_OVERFLOW lowering is implemented by performing normal
addition/subtraction (perhaps extended to even more bits than normally by
continuing with extended values of operands after running of normal bits)
and in addition to that trying to compute if certain sets of bits are either
all zero or all sign extensions of the least significant bit.
That code is in a lot of cases guarded just by a single condition (which
can be idx > startlimb, idx >= startlimb or idx == startlimb) or by
2 conditions - if (idx >= startlimb) { if (idx == startlimb) ... else ... }
Now, if_then_if_then_else when the second argument is NULL works just as
if_then and sets m_gsi to be within the initially empty then block and that is
where we emit code for constant tidx startlimb + (cmp_code == GT_EXPR).
But in the 2 conditions case, m_gsi is set to the initially empty else
block, so using EQ_EXPR for the condition was incorrect (and strangely
nothing in the testsuite caught that), because the code for extracting the
lowest set of bits (i.e. when tidx is startlimb) is then done when idx
is not startlimb rather than when it is.
The following patch fixes that.
Note, when developing the lowering, I was using gcov to make sure all code
is covered by the testsuite with minimum exceptions, so no idea how this
slipped out.
2023-12-01 Jakub Jelinek <jakub@redhat.com>
PR middle-end/112750
* gimple-lower-bitint.cc (bitint_large_huge::lower_addsub_overflow):
Use NE_EXPR rather than EQ_EXPR for g2 if !single_comparison and
adjust probabilities.
* gcc.dg/bitint-41.c: Use -std=c23 rather than -std=c2x.
* gcc.dg/torture/bitint-43.c: Likewise.
* gcc.dg/torture/bitint-44.c: Likewise.
* gcc.dg/torture/bitint-45.c: New test.
Xi Ruoyao [Thu, 30 Nov 2023 09:35:10 +0000 (17:35 +0800)]
doc: Update the status of build directory not fully separated
Recently there are some people building GCC with srcdir == objdir and
the attempts just failed [1]. So stop to say "it should work". OTOH
objdir as a subdirectory of srcdir works: we've built GCC in LFS [2]
and BLFS [3] this way for decades and this is confirmed during the
review of a previous version of this patch [4].
Juergen Christ [Fri, 1 Dec 2023 07:01:36 +0000 (08:01 +0100)]
s390x: Fix s390_md_asm_adjust handling of long doubles [PR112753]
Commit 466b100e5fee808d77598e0f294654deec281150 introduced a bug in
s390_md_asm_adjust if vector extensions are not available. Fix the
control flow of this function to not adjust long double values.
gcc/ChangeLog:
PR target/112753
* config/s390/s390.cc (s390_md_asm_adjust): Return after dealing
with the outputs, if no further processing of long doubles is
required.
Jakub Jelinek [Fri, 1 Dec 2023 07:11:34 +0000 (08:11 +0100)]
s390: Fix builtin-classify-type-1.c on s390 too [PR112725]
s390 suffers from the same issue with __builtin_classify_type vector
arguments, the target hook diagnoses those because it thinks a vector
is passed to an unprototyped function, but in this case it is a type-generic
builtin which can support vector types just fine, by folding into the
appropriate constant.
This fixes
-FAIL: c-c++-common/builtin-classify-type-1.c -Wc++-compat (test for excess errors)
-UNRESOLVED: c-c++-common/builtin-classify-type-1.c -Wc++-compat compilation failed to produce executable
on s390x-linux.
2023-12-01 Jakub Jelinek <jakub@redhat.com>
PR target/112725
* config/s390/s390.cc (s390_invalid_arg_for_unprototyped_fn): Return
NULL for __builtin_classify_type calls with vector arguments.
Florian Weimer [Fri, 1 Dec 2023 07:10:13 +0000 (08:10 +0100)]
c: Turn -Wincompatible-pointer-types into a permerror
The change to build_conditional_expr drops the downgrade
from a pedwarn to warning for builtins for C99 and later
language dialects. It remains a warning in C89 mode (not
a permerror), as the -std=gnu89 -fno-permissive test shows.
PR c/96284
* c-typeck.cc (build_conditional_expr): Upgrade most pointer
type mismatches to a permerror.
(convert_for_assignment): Use permerror_opt and
permerror_init for OPT_Wincompatible_pointer_types warnings.
gcc/testsuite/
* gcc.dg/permerror-default.c (incompatible_pointer_types):
Expect new permerror.
* gcc.dg/permerror-gnu89-nopermissive.c
(incompatible_pointer_types): Likewise.
* gcc.dg/permerror-pedantic.c (incompatible_pointer_types):
Likewise.
* gcc.dg/permerror-system.c: Likewise.
* gcc.dg/Wincompatible-pointer-types-2.c: Compile with
-fpermissive due to expected errors.
* gcc.dg/Wincompatible-pointer-types-5.c: New test. Copied
from gcc.dg/Wincompatible-pointer-types-2.c. Expect errors.
* gcc.dg/anon-struct-11.c: Compile with -fpermissive
due to expected errors.
* gcc.dg/anon-struct-11a.c: New test. Copied from
gcc.dg/anon-struct-11.c. Expect errors.
* gcc.dg/anon-struct-13.c: Compile with -fpermissive
due to expected errors.
* gcc.dg/anon-struct-13a.c: New test. Copied from
gcc.dg/anon-struct-13.c. Expect errors.
* gcc.dg/builtin-arith-overflow-4.c: Compile with -fpermissive
due to expected errors.
* gcc.dg/builtin-arith-overflow-4a.c: New test. Copied from
gcc.dg/builtin-arith-overflow-4.c. Expect errors.
* gcc.dg/c23-qual-4.c: Expect -Wincompatible-pointer-types errors.
* gcc.dg/dfp/composite-type.c: Compile with -fpermissive
due to expected errors.
* gcc.dg/dfp/composite-type-2.c: New test. Copied from
gcc.dg/dfp/composite-type.c. Expect errors.
* gcc.dg/diag-aka-1.c: Compile with -fpermissive
due to expected errors.
* gcc.dg/diag-aka-1a.c: New test. Copied from
gcc.dg/diag-aka-1a.c. Expect errors.
* gcc.dg/enum-compat-1.c: Compile with -fpermissive
due to expected errors.
* gcc.dg/enum-compat-2.c: New test. Copied from
gcc.dg/enum-compat-1.c. Expect errors.
* gcc.dg/func-ptr-conv-1.c: Compile with -fpermissive
due to expected errors.
* gcc.dg/func-ptr-conv-2.c: New test. Copied from
gcc.dg/func-ptr-conv-1.c. Expect errors.
* gcc.dg/init-bad-7.c: Compile with -fpermissive
due to expected errors.
* gcc.dg/init-bad-7a.c: New test. Copied from gcc.dg/init-bad-7.c.
Expect errors.
* gcc.dg/noncompile/incomplete-3.c (foo): Expect
-Wincompatible-pointer-types error.
* gcc.dg/param-type-mismatch-2.c (test8): Likewise.
* gcc.dg/pointer-array-atomic.c: Compile with -fpermissive
due to expected errors.
* gcc.dg/pointer-array-atomic-2.c: New test. Copied from
gcc.dg/pointer-array-atomic.c. Expect errors.
* gcc.dg/pointer-array-quals-1.c (test): Expect
-Wincompatible-pointer-types errors.
* gcc.dg/transparent-union-1.c: Compile with -fpermissive
due to expected errors.
* gcc.dg/transparent-union-1a.c: New test. Copied from
gcc.dg/transparent-union-1.c. Expect errors.
* gcc.target/aarch64/acle/memtag_2a.c
(test_memtag_warning_return_qualifier): Expect additional
errors.
* gcc.target/aarch64/sve/acle/general-c/load_2.c (f1): Likewise.
* gcc.target/aarch64/sve/acle/general-c/load_ext_gather_offset_1.c
(f1): Likewise.
* gcc.target/aarch64/sve/acle/general-c/load_ext_gather_offset_2.c
(f1): Likewise.
* gcc.target/aarch64/sve/acle/general-c/load_ext_gather_offset_3.c
(f1): Likewise.
* gcc.target/aarch64/sve/acle/general-c/load_ext_gather_offset_4.c
(f1): Likewise.
* gcc.target/aarch64/sve/acle/general-c/load_ext_gather_offset_5.c
(f1): Likewise.
* gcc.target/aarch64/sve/acle/general-c/load_ext_gather_offset_restricted_1.c
(f1): Likewise.
* gcc.target/aarch64/sve/acle/general-c/load_ext_gather_offset_restricted_2.c
(f1): Likewise.
* gcc.target/aarch64/sve/acle/general-c/load_ext_gather_offset_restricted_3.c
(f1): Likewise.
* gcc.target/aarch64/sve/acle/general-c/load_ext_gather_offset_restricted_4.c
(f1): Likewise.
* gcc.target/aarch64/sve/acle/general-c/sizeless-1.c (f1): Likewise.
* gcc.target/aarch64/sve/acle/general-c/sizeless-2.c (f1): Likewise.
* gcc.target/aarch64/sve/acle/general-c/store_1.c (f1): Likewise.
* gcc.target/aarch64/sve/acle/general-c/store_2.c (f1): Likewise.
* gcc.target/aarch64/sve/acle/general-c/store_scatter_index_1.c
(f1): Likewise.
* gcc.target/aarch64/sve/acle/general-c/store_scatter_index_restricted_1.c
(f1): Likewise.
* gcc.target/aarch64/sve/acle/general-c/store_scatter_offset_2.c
(f1): Likewise.
* gcc.target/aarch64/sve/acle/general-c/store_scatter_offset_restricted_1.c
(f1): Likewise.
* gcc.target/aarch64/sve/acle/general/attributes_7.c
(f1): Likewise.
* gcc.target/i386/sse2-bfloat16-scalar-typecheck.c (footest):
Expect -Wincompatible-pointer-types errors.
* gcc.target/i386/vect-bfloat16-typecheck_1.c (footest): Likewise.
* gcc.target/i386/vect-bfloat16-typecheck_2.c (footest): Likewise.
Florian Weimer [Fri, 1 Dec 2023 07:10:13 +0000 (08:10 +0100)]
c: Turn -Wreturn-mismatch into a permerror
gcc/
* doc/invoke.texi (Warning Options): Document that
-Wreturn-mismatch is a permerror in C99 and later.
gcc/c/
PR c/96284
* c-typeck.cc (c_finish_return): Use permerrors
for OPT_Wreturn_mismatch diagnostics.
gcc/testsuite/
* gcc.dg/permerror-default.c (return_mismatch_1)
(return_mismatch_2): Expect new permerror.
* gcc.dg/permerror-gnu89-nopermissive.c (return_mismatch_1):
Likewise.
* gcc.dg/permerror-system.c: Likewise.
* gcc.dg/20030906-1.c: Compile with -fpermissive due to
expected -Wreturn-mismatch error.
* gcc.dg/20030906-1a.c: New test. Copied from
gcc.dg/20030906-1.c. Expect the error.
* gcc.dg/20030906-2.c: Compile with -fpermissive due to
expected -Wreturn-mismatch error.
* gcc.dg/20030906-2a.c: New test. Copied from
gcc.dg/20030906-2.c. Expect the error.
* gcc.dg/Wreturn-mismatch-1.c: Compile with -fpermissive due to
expected -Wreturn-mismatch error.
* gcc.dg/Wreturn-mismatch-1a.c: New test. Copied from
gcc.dg/Wreturn-mismatch-1.c. Expect the error.
* gcc.dg/Wreturn-mismatch-2.c: Compile with -fpermissive due to
expected -Wreturn-mismatch error.
* gcc.dg/Wreturn-mismatch-2a.c: New test. Copied from
gcc.dg/Wreturn-mismatch-2.c. Expect the error.
* gcc.dg/diagnostic-range-bad-return.c: Compile with
-fpermissive due to expected -Wreturn-mismatch error.
* gcc.dg/diagnostic-range-bad-return-2.c: New test.
Copied from gcc.dg/diagnostic-range-bad-return.c. Expect the
error.
* gcc.dg/pr105635-2.c: Expect -Wreturn-mismatch error.
* gcc.dg/pr23075.c: Build with -fpermissive due to
expected -Wreturn-mismatch error.
* gcc.dg/pr23075-2.c: New test. Copied from gcc.dg/pr23075.c.
Expect the error.
* gcc.dg/pr29521.c: Compile with -fpermissive due to expected
-Wreturn-mismatch error.
* gcc.dg/pr29521-a.c: New test. Copied from gcc.dg/pr29521.c.
Expect error.
* gcc.dg/pr67730.c: Compile with -fpermissive due to expected
-Wreturn-mismatch error.
* gcc.dg/pr67730-a.c: New test. Copied from
gcc.dg/pr67730-a.c. Expect error.
* gcc.target/powerpc/conditional-return.c: Compile with
-fpermissive due to expected -Wreturn-mismatch error.
Florian Weimer [Fri, 1 Dec 2023 07:10:13 +0000 (08:10 +0100)]
c: Do not ignore some forms of -Wimplicit-int in system headers
Most -Wimplicit-int warnings were unconditionally disabled for system
headers. Only missing types for parameters in old-style function
definitions resulted in warnings. This is inconsistent with the
treatment of other permerrors, which are active in system headers.
gcc/c/
* c-decl.cc (grokdeclarator): Do not skip -Wimplicit-int
warnings or errors in system headers.
gcc/testsuite/
* gcc.dg/permerror-system.c: Expect all -Wimplicit-int
permerrors.
* c-decl.cc (warn_defaults_to): Remove.
(grok_declarator, start_function): Call permerror_opt
instead of warn_defaults_to.
(store_parm_decls_oldstyle): Call permerror_opt for
OPT_Wimplicit_int.
gcc/testsuite/
* gcc.dg/permerror-default.c (implicit_int_1, implicit_int_2)
(implicit_int_3, implicit_int_4): Expect new permerror.
* gcc.dg/permerror-system.c: Expect a single new permerror.
* gcc.dg/Wimplicit-int-1.c: Compile with -fpermissive due to
expected warning.
* gcc.dg/Wimplicit-int-4.c: Likewise.
* gcc.dg/Wimplicit-int-1a.c: New test. Copied from
gcc.dg/Wimplicit-int-1.c, but expect errors.
* gcc.dg/Wimplicit-int-4a.c: New test. Copied from
gcc.dg/Wimplicit-int-4.c, but expect errors.
* gcc.dg/gnu23-attr-syntax-2.c: Compile with -fpermissive
due to expected implicit-int error.
* gcc.dg/gnu23-attr-syntax-3.c: New test. Copied from
gcc.dg/gnu23-attr-syntax-2.c, but expect an error.
* gcc.dg/pr105635.c: Build with -fpermissive due to implicit
int.
* gcc.dg/pr105635-2.c: New test. Copied from
gcc.dg/pr105635.c. Expect implicit int error.
* gcc.dg/noncompile/pr79758.c: Build with -fpermissive due to
implicit int.
* gcc.dg/noncompile/pr79758-2.c: New test. Copied from
gcc.dg/noncompile/pr79758.c. Expect implicit int error.
Florian Weimer [Fri, 1 Dec 2023 07:10:13 +0000 (08:10 +0100)]
c: Turn -Wimplicit-function-declaration into a permerror
In the future, it may make sense to avoid cascading errors from
the implicit declaration, especially its assumed int return type.
This change here only changes the kind of the diagnostic, not
its wording or consequences.
PR c/91092
PR c/96284
* c-decl.cc (implicit_decl_permerror): Rename from
implicit_decl_warning. Call permerror_opt instead of
pedwarn and warning_at.
(implicitly_declare): Adjust callers.
gcc/testsuite/
* gcc.dg/permerror-default.c (implicit_function_declaration):
Expect the new permerror.
* gcc.dg/permerror-system.c: Likewise.
* c-c++-common/spellcheck-reserved.c (test, test_2): Expect
error instead of warning.
(f): Expect error instead of warning.
* gcc.dg/Wimplicit-function-declaration-c99.c: Compile with
-fpermissive due to expected warning.
* gcc.dg/Wimplicit-function-declaration-c99-2.c: New test.
Copied from gcc.dg/Wimplicit-function-declaration-c99.c.
Expect error.
* gcc.dg/missing-header-fixit-1.c: Compile with -fpermissive
due to expect error.
* gcc.dg/missing-header-fixit-1a.c: New test. Copied from
gcc.dg/missing-header-fixit-1.c, but expect error.
* gcc.dg/missing-header-fixit-2.c: Compile with -fpermissive
due to expect error.
* gcc.dg/missing-header-fixit-2a.c: New test. Copied from
gcc.dg/missing-header-fixit-2.c, but expect error.
* gcc.dg/missing-header-fixit-4.c: Compile with -fpermissive
due to expect error.
* gcc.dg/missing-header-fixit-4a.c: New test. Copied from
gcc.dg/missing-header-fixit-4.c, but expect error.
* gcc.dg/missing-header-fixit-5.c: Compile with -fpermissive
due to expect error.
* gcc.dg/missing-header-fixit-5a.c: New test. Copied from
gcc.dg/missing-header-fixit-5.c, but expect error.
* gcc.dg/pr61852.c: Expect implicit-function-declaration
error instead of warning.
* gcc.dg/spellcheck-identifiers-2.c: Compile with
-fpermissive due to expected warnings.
* gcc.dg/spellcheck-identifiers-2a.c: New test. Copied
from gcc.dg/spellcheck-identifiers-2a.c. Expect errors.
* gcc.dg/spellcheck-identifiers-3.c: Compile with
-fpermissive due to expected warnings.
* gcc.dg/spellcheck-identifiers-3a.c: New test. Copied
from gcc.dg/spellcheck-identifiers-2a.c. Expect errors.
* gcc.dg/spellcheck-identifiers-4.c: Compile with
-fpermissive due to expected warnings.
* gcc.dg/spellcheck-identifiers-4a.c: New test. Copied
from gcc.dg/spellcheck-identifiers-2a.c. Expect error.
* gcc.dg/spellcheck-identifiers.c: Compile with
-fpermissive due to expected warnings.
* gcc.dg/spellcheck-identifiers-1a.c: New test. Copied
from gcc.dg/spellcheck-identifiers.c. Expect errors.
* gcc.target/aarch64/sve/acle/general-c/ld1sh_gather_1.c (f1):
Expect error.
* gcc.target/aarch64/sve/acle/general-c/load_ext_gather_index_1.c:
(f1): Likewise.
* gcc.target/aarch64/sve/acle/general-c/load_ext_gather_index_restricted_1.c:
(f1): Likewise.
* gcc.target/aarch64/sve/acle/general-c/load_ext_gather_offset_1.c:
(f1): Likewise.
* gcc.target/aarch64/sve/acle/general-c/load_ext_gather_offset_2.c:
(f1): Likewise.
* gcc.target/aarch64/sve/acle/general-c/load_ext_gather_offset_3.c:
(f1): Likewise.
* gcc.target/aarch64/sve/acle/general-c/load_ext_gather_offset_4.c:
(f1): Likewise.
* gcc.target/aarch64/sve/acle/general-c/load_ext_gather_offset_5.c:
(f1): Likewise.
* gcc.target/aarch64/sve/acle/general-c/load_ext_gather_offset_restricted_1.c:
(f1): Likewise.
* gcc.target/aarch64/sve/acle/general-c/load_ext_gather_offset_restricted_2.c:
(f1): Likewise.
* gcc.target/aarch64/sve/acle/general-c/load_ext_gather_offset_restricted_3.c:
(f1): Likewise.
* gcc.target/aarch64/sve/acle/general-c/load_ext_gather_offset_restricted_4.c:
(f1): Likewise.
PR c/96284
PR c/106416
* c-typeck.cc (build_conditional_expr): Use permerror_opt for
pointer/integer type mismatches, based on -Wint-conversion.
(pedwarn_permerror_init, permerror_init): New function.
(pedwarn_init): Call pedwarn_permerror_init.
(convert_for_assignment): Use permerror_opt and
permerror_init for -Wint-conversion warnings.
gcc/testsuite/
* gcc.dg/permerror-default.c (int_conversion_1)
(int_conversion_2): Expect the new permerrors.
* gcc.dg/permerror-gnu89-nopermissive.c (int_conversion_1)
(int_conversion_2): Likewise.
* gcc.dg/permerror-system.c: Likewise.
* c-c++-common/pr77624-1.c (foo, bar): Expect
error instead of warning.
* gcc.dg/Wint-conversion-2.c: Compile with -fpermissive due
to expected int-conversion warning.
* gcc.dg/Wint-conversion-3.c: Likewise.
* gcc.dg/Wint-conversion-4.c: New test. Based on
gcc.dg/Wint-conversion-3.c. Expect int-conversion errors.
* gcc.dg/assign-warn-1.c: Compile with -fpermissive.
* gcc.dg/assign-warn-4.c: New file. Extracted from
assign-warn1.c. Expect int-conversion errors.
* gcc.dg/diagnostic-types-1.c: Compile with -fpermissive.
* gcc.dg/diagnostic-types-2.c: New file. Extracted from
gcc.dg/diagnostic-types-1.c. Expect some errors instead of
warnings.
* gcc.dg/gomp/pr35738.c: Compile with -fpermissive due to
expected int-conversion error.
* gcc.dg/gomp/pr35738-2.c: New test. Based on
gcc.dg/gomp/pr35738.c. Expect int-converison errors.
* gcc.dg/init-excess-3.c: Expect int-converison errors.
* gcc.dg/overflow-warn-1.c: Likewise.
* gcc.dg/overflow-warn-3.c: Likewise.
* gcc.dg/param-type-mismatch.c: Compile with -fpermissive.
* gcc.dg/param-type-mismatch-2.c: New test. Copied from
gcc.dg/param-type-mismatch.c. Expect errors.
* gcc.dg/pr61162-2.c: Compile with -fpermissive.
* gcc.dg/pr61162-3.c: New test. Extracted from
gcc.dg/pr61162-2.c. Expect int-conversion errors.
* gcc.dg/spec-barrier-3.c: Use -fpermissive due to expected
int-conversion error.
* gcc.dg/spec-barrier-3a.c: New test. Based on
gcc.dg/spec-barrier-3.c. Expect int-conversion errors.
* gcc.target/aarch64/acle/memtag_2.c: Use -fpermissive due to expected
int-conversion error.
* gcc.target/aarch64/acle/memtag_2a.c: New test. Copied from
gcc.target/aarch64/acle/memtag_2.c. Expect error.
* gcc.target/aarch64/sve/acle/general-c/load_3.c (f1): Expect
error.
* gcc.target/aarch64/sve/acle/general-c/store_2.c (f1): Likewise.
* gcc.target/aarch64/sve/acle/general-c/store_scatter_index_1.c
(f1): Likewise.
* gcc.target/aarch64/sve/acle/general-c/store_scatter_index_restricted_1.c
(f1): Likewise.
* gcc.target/aarch64/sve/acle/general-c/store_scatter_offset_2.c
(f1): Likewise.
* gcc.target/aarch64/sve/acle/general-c/store_scatter_offset_restricted_1.c
(f1): Likewise.
Florian Weimer [Fri, 1 Dec 2023 07:10:12 +0000 (08:10 +0100)]
aarch64: Call named function in gcc.target/aarch64/aapcs64/ice_1.c
This test looks like it intends to pass a small struct argument
through both a non-variadic and variadic argument, but due to
the typo, it does not achieve that.
Juzhe-Zhong [Fri, 1 Dec 2023 00:39:57 +0000 (08:39 +0800)]
RISC-V: Fix VSETVL PASS regression
This patch fix 2 regression (one is bug regression, the other is performance regression).
Those 2 regressions are both we are comparing ratio for same AVL in wrong place.
1. BUG regression:
avl_single-84.c:
f0:
li a5,999424
add a1,a1,a5
li a4,299008
add a5,a0,a5
addi a3,a4,992
addi a5,a5,576
addi a1,a1,576
vsetvli a4,zero,e8,m2,ta,ma
add a0,a0,a3
vlm.v v1,0(a5)
vsm.v v1,0(a1)
vl1re64.v v1,0(a0)
beq a2,zero,.L10
li a5,0
vsetvli zero,zero,e64,m1,tu,ma ---> This is totally incorrect since the ratio above is 4, wheras it is demanding ratio = 64 here.
.L3:
fcvt.d.lu fa5,a5
addi a5,a5,1
fadd.d fa5,fa5,fa0
vfmv.s.f v1,fa5
bne a5,a2,.L3
vfmv.f.s fa0,v1
ret
.L10:
vsetvli zero,zero,e64,m1,ta,ma
vfmv.f.s fa0,v1
ret
Jason Merrill [Thu, 30 Nov 2023 02:51:36 +0000 (21:51 -0500)]
c++: lambda capture and explicit object parm
More adjustments to allow for explicit object parameters in lambdas. This
has no practical effect until that patch goes in, but applying this
separately seems reasonable.
Marek Polacek [Tue, 28 Nov 2023 19:44:24 +0000 (14:44 -0500)]
c++: wrong ambiguity in accessing static field [PR112744]
Given
struct A { constexpr static int a = 0; };
struct B : A {};
struct C : A {};
struct D : B, C {};
we give the "'A' is an ambiguous base of 'D'" error for
D{}.A::a;
which seems wrong: 'a' is a static data member so there is only one copy
so it can be unambiguously referred to even if there are multiple A
objects. clang++/MSVC/icx agree.
This patch uses ba_any: [class.access.base] requires conversion to a unique
base subobject for non-static data members, but it does not require that the
base be unique or accessible for static data members.
PR c++/112744
gcc/cp/ChangeLog:
* typeck.cc (finish_class_member_access_expr): When accessing
a static data member, use ba_any for lookup_base.
gcc/testsuite/ChangeLog:
* g++.dg/lookup/scoped11.C: New test.
* g++.dg/lookup/scoped12.C: New test.
* g++.dg/lookup/scoped13.C: New test.
* g++.dg/lookup/scoped14.C: New test.
* g++.dg/lookup/scoped15.C: New test.
__sync_val_compare_and_swap may be used on 128-bit types and either calls the
outline atomic code or uses an inline loop. On AArch64 LDXP is only atomic if
the value is stored successfully using STXP, but the current implementations
do not perform the store if the comparison fails. In this case the value
returned is not read atomically.
gcc/ChangeLog:
PR target/111404
* config/aarch64/aarch64.cc (aarch64_split_compare_and_swap):
For 128-bit store the loaded value and loop if needed.
libgcc/ChangeLog:
PR target/111404
* config/aarch64/lse.S (__aarch64_cas16_acq_rel): Execute STLXP using
either new value or loaded value.
Jonathan Wakely [Wed, 29 Nov 2023 22:26:28 +0000 (22:26 +0000)]
libstdc++: Fix std::ranges::to errors
Fix some errors that Patrick noticed, and remove a #if 0 group that I
didn't mean to leave in the file.
libstdc++-v3/ChangeLog:
* include/std/ranges (__detail::__toable): Fix incorrect use of
_Range instead of _Cont.
(__detail::_ToClosure, __detail::_ToClosure2): Add missing
constexpr specifier on constructors.
* testsuite/std/ranges/conv/1.cc (_Cont, _Cont2, _Cont3): Remove
unnecessary begin() and end() members.
(test_constexpr): New function to check range adaptors are
usable in constant expressions.
After commit r14-5617-gb8592186611, int32x[24]_t types now use
elements of 'long int' type instead of 'int' on arm-eabi (it's still
'int' on arm-linux-gnueabihf). Both are 32-bit types anyway.
This patch adjust the two tests so that they optionnally accept 'long '
before 'int' in the expected error message.
These test cases being 'dg-skip-if [...] { ! amdgcn-*-* }' meant they were
only ever considered for 'amdgcn-*-*' target -- that is, GCN *target*, not
GCN *offload target*, what is intended to be tested here, but for which they
thus always were UNSUPPORTED. Use the same style of 'dg-[...]' directives
as for the nvptx offloading test cases, 'libgomp.c/declare-variant-3*.c'.
Thomas Schwinge [Thu, 16 Nov 2023 22:17:36 +0000 (23:17 +0100)]
GCN: Generally enable the 'gcc.target/gcn/avgpr-[...]' test cases
... added in commit ae0d2c240213c5a7f6959c032bfc9f0703cab787
"amdgcn: Add Accelerator VGPR registers". This way, they're correctly tested
no matter what '-march=[...]' is used with 'make check'.
The cpymemdi/setmemdi implementation doesn't fully support strict alignment.
Block the expansion if the alignment is less than 16 with STRICT_ALIGNMENT.
Clean up the condition when to use MOPS.
gcc/ChangeLog/
PR target/103100
* config/aarch64/aarch64.md (cpymemdi): Remove pattern condition.
(setmemdi): Likewise.
* config/aarch64/aarch64.cc (aarch64_expand_cpymem): Support
strict-align. Cleanup condition for using MOPS.
(aarch64_expand_setmem): Likewise.
Harald Anlauf [Wed, 29 Nov 2023 20:47:24 +0000 (21:47 +0100)]
Fortran: fix TARGET attribute of associating entity in ASSOCIATE [PR112764]
The associating entity in an ASSOCIATE construct has the TARGET attribute
if and only if the selector is a variable and has either the TARGET or
POINTER attribute (e.g. F2018:11.1.3.3).
gcc/fortran/ChangeLog:
PR fortran/112764
* primary.cc (gfc_variable_attr): Set TARGET attribute of associating
entity dependent on TARGET or POINTER attribute of selector.
gcc/testsuite/ChangeLog:
PR fortran/112764
* gfortran.dg/associate_62.f90: New test.
Richard Biener [Thu, 30 Nov 2023 11:38:53 +0000 (12:38 +0100)]
tree-optimization/112767 - spurious diagnostic after sccp/loop-split swap
We are diagnosing an unreachable loop which we only manage to elide
after the copyprop pass after sccp which leaves the code open for
diagnosing by the intermittent ivcanon pass. The following makes sure
to clean things up a bit earlier, propagating constant final values
to uses immediately.
Richard Biener [Thu, 30 Nov 2023 09:58:13 +0000 (10:58 +0100)]
tree-optimization/112766 - improve pruning of uninit diagnostics
Uninit diagnostics has code to prune based on incoming PHI args
that prove the uninit code is never executed. But that only
looks at the first found flag candidate while in the PRs case
only the second candidate would be the one to prune on. The
following patch makes us consider all of the flag candidates
which is cycles well spent IMHO.
PR tree-optimization/112766
* gimple-predicate-analysis.cc (find_var_cmp_const):
Support continuing the iteration and report every candidate.
(uninit_analysis::overlap): Iterate over all flag var
candidates.
* config/riscv/vector.md: Add widening overlap of vf2/vf4.
gcc/testsuite/ChangeLog:
* gcc.target/riscv/rvv/base/pr112431-16.c: New test.
* gcc.target/riscv/rvv/base/pr112431-17.c: New test.
* gcc.target/riscv/rvv/base/pr112431-18.c: New test.
Confirm the codegen above run successfully on both SPIKE/QEMU.
PR target/112431
gcc/ChangeLog:
* config/riscv/vector.md: Remove earlyclobber for wx/wf instructions.
gcc/testsuite/ChangeLog:
* gcc.target/riscv/rvv/base/pr112431-19.c: New test.
* gcc.target/riscv/rvv/base/pr112431-20.c: New test.
* gcc.target/riscv/rvv/base/pr112431-21.c: New test.
Eric Botcazou [Mon, 20 Nov 2023 07:58:52 +0000 (08:58 +0100)]
ada: Rework fix for wrong finalization of qualified aggregate in allocator
The problem is that there is no easy method to insert an action after an
arbitrary node in the tree, so the original fix does not correctly work
when the allocator is nested in another expression.
Therefore this moves the burden of the insertion from Apply_Predicate_Check
to Expand_Allocator_Expression and restricts the new processing to the case
where it is really required.
gcc/ada/
* checks.ads (Apply_Predicate_Check): Add Deref boolean parameter.
* checks.adb (Apply_Predicate_Check): Revert latest change. Use
Loc local variable to hold the source location. Use a common code
path for the generic processing and make a dereference if Deref is
True.
* exp_ch4.adb (Expand_Allocator_Expression): Compute Aggr_In_Place
earlier. If it is true, do not call Apply_Predicate_Check on the
expression on entry but on the temporary on exit with a
dereference.
* sem_res.adb (Resolve_Actuals): Add explicit parameter
association in call to Apply_Predicate_Check.
Steve Baird [Tue, 14 Nov 2023 22:34:47 +0000 (14:34 -0800)]
ada: Support Put_Image for types in user-defined instances of predefined generics.
Predefined units do not generally support the Put_Image attribute.
There are good reasons for this in most cases. But if a user-defined
instantiation of a predefined generic occurs in Ada 2022 code, then
Put_Image can be supported for types declared therein. Add this support.
gcc/ada/
* exp_put_image.adb (Put_Image_Enabled): Return True in more
cases. In particular, when testing to see if a type occurs in a
predefined unit, test the type's code unit
(obtained by calling Get_Code_Unit). In the case of type within a
user-defined instance of a predefined generic, Is_Predefined_Unit
will return True for the type and False for the type's code unit.
Yannick Moy [Fri, 20 Oct 2023 07:39:10 +0000 (09:39 +0200)]
ada: Remove SPARK legality checks
SPARK legality checks apply only to code with SPARK_Mode On, and are
performed again in GNATprove for detecting SPARK-compatible declarations
in code with SPARK_Mode Auto. Remove this duplication, to only perform
SPARK legality checking in GNATprove. After this patch, only a few
special SPARK legality checks are performed in the frontend, which could
be moved to GNATprove later.
gcc/ada/
* contracts.adb (Analyze_Entry_Or_Subprogram_Body_Contract):
Remove checking on volatility. Remove handling of SPARK_Mode, not
needed anymore.
(Analyze_Entry_Or_Subprogram_Contract): Remove checking on
volatility.
(Check_Type_Or_Object_External_Properties): Same.
(Analyze_Object_Contract): Same.
* freeze.adb (Freeze_Record_Type): Same. Also remove checking on
synchronized types and ghost types.
* sem_ch12.adb (Instantiate_Object): Remove checking on
volatility.
(Instantiate_Type): Same.
* sem_ch3.adb (Access_Type_Declaration): Same.
(Derived_Type_Declaration): Remove checking related to untagged
partial view.
(Process_Discriminants): Remove checking on volatility.
* sem_ch5.adb (Analyze_Loop_Parameter_Specification): Same.
* sem_ch6.adb (Analyze_Procedure_Call): Fix use of SPARK_Mode
where GNATprove_Mode was intended.
* sem_disp.adb (Inherited_Subprograms): Protect against Empty
node.
* sem_prag.adb (Analyze_Global_In_Decl_Part): Remove checking on
volatility.
(Analyze_Pragma): Same.
* sem_res.adb (Flag_Effectively_Volatile_Objects): Remove.
(Resolve_Actuals): Remove checking on volatility.
(Resolve_Entity_Name): Same.
* sem_util.adb (Check_Nonvolatile_Function_Profile): Remove.
(Check_Volatility_Compatibility): Remove.
* sem_util.ads: Same.
Sheri Bernstein [Fri, 17 Nov 2023 22:14:48 +0000 (22:14 +0000)]
ada: Remove GNATcheck violations
Remove GNATcheck violations by refactoring code and also using
pragma Annotate to exempt them.
gcc/ada/
* libgnat/i-cstrin.adb (Free): Rewrite code so there is only one
return, to remove Improper_Returns violation.
(Position_Of_Nul): Add pragma to exempt Improper_Returns
violation.
(To_Chars_Ptr): Likewise.
(Value): Likewise
Viljar Indus [Thu, 16 Nov 2023 12:51:13 +0000 (14:51 +0200)]
ada: Ignore defered compile time errors without backend
We defer some compile time warnings and errors until the
backend has added the extra information needed. However
it is not guaranteed that the backend has run by this point.
Avoid checking these errors if the backend has not been activated
and no code has been generated.
gcc/ada/
* sem_prag.adb (Validate_Compile_Time_Warning_Errors): Avoid
checking compile time warnings and errors if backend has not been
activated.
Yannick Moy [Wed, 15 Nov 2023 08:34:35 +0000 (09:34 +0100)]
ada: Fix spelling of functions with(out) "side effects"
Correct spelling does not include an hyphen. Fix comments and one
error message.
Also fix other mispellings of "side-effect" or "side effect" depending
on the case (adjective should have hyphen), and "side-effect-free" with
double hyphen as an adjective.
Javier Miranda [Fri, 17 Nov 2023 17:08:56 +0000 (17:08 +0000)]
ada: Crash initializing component of private record type
The compiler may crash processing the full type declaration of a
private record type that initializes a component with a call to
a function instantiated in the private part of the package.
gcc/ada/
* freeze.adb (Declared_In_Expanded_Body): New subprogram.
(In_Expanded_Body): Minor code cleanup.
(Freeze_Expression): Code cleanup plus factorize in a new function
the code that identifies entities declared in the body of expander
generated subprograms, since such case must be checked also for
other node kinds when climbing the tree to locate the place to
insert the freezing node.
Steve Baird [Fri, 3 Nov 2023 23:00:20 +0000 (16:00 -0700)]
ada: Name resolution in expanded instances
In building the tree for an instance of a generic, expansion sets
entity fields on names that refer to things declared outside of the
instance, but leaves the entity field unset on names that should end
up referring to things declared within the instance. These will instead be
set by analysis - the idea is that if a name resolves a certain way in the
generic, then we should get corresponding results if we resolve the
corresponding name in an instance. For this to work, we have to prevent
unrelated declarations that happen to be visible at the point of the
instantiation from participating in resolution. Add code to filter out such
unwanted name resolution candidates.
gcc/ada/
* sem_ch8.adb (Find_Direct_Name): In the case of a resolving a
name that occurs within an instantiation, add code to detect and
filter out unwanted candidate resolutions. The filtering is
performed via a call to Remove_Interp.
Add a comment in the spec for the default (as opposed to hie) version of
Ada.Real_Time.Timing_Events indicating that it is incompatible with a
a Partition_Elaboration_Policy specification specifying a policy other than
Concurrent.
Steve Baird [Wed, 15 Nov 2023 21:13:04 +0000 (13:13 -0800)]
ada: Too-strict conformance checking for formal discriminated type
The discriminant subtype conformance check for an actual parameter
corresponding to a generic formal discriminated type was too strict and
could incorrectly reject legal instantiations.
gcc/ada/
* sem_ch12.adb (Validate_Discriminated_Formal_Type): Replace
Entity_Id equality test with a call to Subtypes_Match. Distinct
subtypes which are statically matching should pass this test.
(Check_Discriminated_Formal): Replace Entity_Id equality test with
a call to Subtypes_Statically_Match (preceded by a check that the
preconditions for the call are satisfied).
Eric Botcazou [Fri, 17 Nov 2023 10:56:37 +0000 (11:56 +0100)]
ada: Fix predicate check failure in Expand_Allocator_Expression
The For_Special_Return_Object flag needs to be accessed on entry of the
procedure in case the allocator is rewritten during the processing.
gcc/ada/
* exp_ch4.adb (Expand_Allocator_Expression): Add Special_Return
boolean constant to hold the value of For_Special_Return_Object
for the allocator and use it throughout the procedure.
Eric Botcazou [Thu, 16 Nov 2023 18:45:41 +0000 (19:45 +0100)]
ada: Fix wrong finalization for qualified aggregate of limited type in allocator
This happens with -gnata when the limited type has controlled components
and a predicate, because the predicate check generated for the aggregate
causes the creation of a temporary that is used as the expression of the
allocator. Now this combination is illegal for a limited type, so the
compiler does not generate the deep adjustment that would be necessary
for the access value, which ultimately results in a wrong finalization.
gcc/ada/
* checks.adb (Apply_Predicate_Check): Also deal specifically with
an expression that is a qualified aggregate in an allocator.
Steve Baird [Tue, 14 Nov 2023 21:33:31 +0000 (13:33 -0800)]
ada: Constant_Indexing used when context requires a variable
In the case of a call with a formal parameter of mode other than "IN"
where the corresponding actual parameter is a generalized indexing
and the indexable container has both Constant_Indexing and Variable_Indexing
aspects specified, the generalized indexing must be interpreted as a
variable indexing, not as a constant indexing. In some cases involving a
call to a prefixed view of a subprogram, this was not handled correctly.
This error results in spurious compile-time error messages saying that
the actual parameter in the call "must be a variable".
gcc/ada/
* sem_ch4.adb (Constant_Indexing_OK): As a temporary stopgap,
return False in the case of an unanalyzed prefixed-view call.
This patch fixes the other half of the PR, where we were crashing on the
UNSIGNED widest_int modulo of 1 % -128 (where the -128 in UNSIGNED is
131072 bit 0xfffff...fffff80).
The patch actually has 2 independent parts.
The fix for the divmod buffer overflow is in the 2nd-5th and 7th-8th hunks of
the patch. abs (remainder) <= min (abs (dividend), abs (divisor)) and the
wide-int.h callers estimate the remainder size to be the same as the
quotient size, i.e. for UNSIGNED dividend with msb set quotient (same as
remainder) precision based estimation, for SIGNED or dividend with msb
clear estimate based on divident len (+ 1).
For quotient (if any), wi_pack is called as
quotient_len = wi_pack (quotient, b_quotient, m, dividend_prec);
where m is
m = dividend_blocks_needed;
while (m > 1 && b_dividend[m - 1] == 0)
m--;
and
dividend_blocks_needed = 2 * BLOCKS_NEEDED (dividend_prec);
where wi_pack stores to result (in_len + 1) / 2 limbs (or one more
if (in_len & 1) == 0 and in_len / 2 < BLOCKS_NEEDED (dividend_prec)).
In any case, that is never more than BLOCKS_NEEDED (dividend_prec) and
matches caller's estimations (then it is canonicalized and perhaps shrunk
that way).
The problematic case is the remainder, where wi_pack is called as
*remainder_len = wi_pack (remainder, b_remainder, n, dividend_prec);
The last argument is again based on the dividend precision like for
quotient, but n is instead:
n = divisor_blocks_needed;
while (n > 1 && b_divisor[n - 1] == 0)
n--;
so based on the divisor rather than dividend. Now, in the
wi::multiple_of_p (1, -128, UNSIGNED) widest_int precision case,
dividend_prec is very small, the dividend is 1 and while the official
precision is 131072 bits, dividend_len is 1 and
if (sgn == SIGNED || dividend_val[dividend_len - 1] >= 0)
dividend_prec = MIN ((dividend_len + 1) * HOST_BITS_PER_WIDE_INT,
dividend_prec);
shrinks the estimate to 128 bits in this case; but divisor_prec
is huge, because the divisor is negative UNSIGNED, so 131072 bits,
2048 limbs, 4096 half-limbs (so divisor_blocks_needed and n are 4096).
Now, wi_pack relies on the penultimate argument to be smaller or equal
to 2 * BLOCKS_NEEDED of the last argument and that is not the case here,
4096 is certainly larger than 2 * BLOCKS_NEEDED (128) aka 4 and so
wi_pack overflows the array.
Unfortunately looking around, this isn't the only overflow, already
in divmod_internal_2 we have an overflow, though there not overflowing
a buffer during writing, but during reading. When divmod_internal_2
is called with the last 2 arguments like m=1, n=4096 as in this case
(but generally any where m < n), it has a special case for n == 1 (which
won't be the problem, we never have m or n 0), then decides based on
msb of divisor whether there should be some shift canonicalization, then
performs a
for (j = m - n; j >= 0; j--)
main loop and finally initializes b_remainder:
if (s)
for (i = 0; i < n; i++)
b_remainder[i] = (b_dividend[i] >> s)
| (b_dividend[i+1] << (HOST_BITS_PER_HALF_WIDE_INT - s));
else
for (i = 0; i < n; i++)
b_remainder[i] = b_dividend[i];
In the usual case of m >= n, the main loop will iterate at least once,
b_dividend has m + 1 valid elements and the copying to b_remainder
is correct. But for the m < n unusual case, the main loop never iterates
and we want to copy the b_dividend right into b_remainder (possibly with
shifts). Except when it copies n elements, it copies garbage which isn't
there at all for the last n - m elements.
Because the decision whether the main loop should iterate at all or not
and the s computation should be based on the original n, the following
patch sets n to MIN (n, m) only after the main loop, such that we copy
to b_remainder at most what is initialized in b_dividend, and returns
that adjusted value to the caller which then passes it to wi_pack and
so satisfies again the requirement there, rather than trying to
do the n = MIN (n, m) before calling divmod_internal_2 or after it.
I believe the bug is mostly something I've only introduced myself in the
> 576 bit wide_int/widest_int support changes, before that there was
no attempt to decrease the precisions and so I think dividend_prec
and divisor_prec were pretty much always the same (or always?),
and even the copying of garbage wasn't the case because the only time
m or n was decreased from the same precision was if there were 0s in the
upper half-limbs.
The 1st and 6th hunks in the patch is just that while looking at the
code, I've noticed I computed wrong the sizes in the XALLOCAVEC case,
somehow the 4 * in the static arrays led me believe I need to make the
alloced buffers twice (in the multiplication) or 4 times (in the
division/modulo cases) larger than needed.
2023-11-30 Jakub Jelinek <jakub@redhat.com>
PR middle-end/112733
* wide-int.cc (wi::mul_internal): Don't allocate twice as much
space for u, v and r as needed.
(divmod_internal_2): Change return type from void to int, for n == 1
return 1, otherwise before writing b_dividend into b_remainder set
n to MIN (n, m) and at the end return it.
(wi::divmod_internal): Don't allocate 4 times as much space for
b_quotient, b_remainder, b_dividend and b_divisor. Set n to
result of divmod_internal_2.
(wide_int_cc_tests): Add test for unsigned widest_int
wi::multiple_of_p of 1 and -128.
Jakub Jelinek [Thu, 30 Nov 2023 08:09:21 +0000 (09:09 +0100)]
c++: Implement C++26 P2169R4 - Placeholder variables with no name [PR110349]
The following patch implements the C++26 P2169R4 paper.
As written in the PR, the patch expects that:
1) https://eel.is/c++draft/expr.prim.lambda.capture#2
"Ignoring appearances in initializers of init-captures, an identifier
or this shall not appear more than once in a lambda-capture."
is adjusted such that name-independent lambda captures with initializers
can violate this rule (but lambda captures which aren't name-independent
can't appear after name-independent ones)
2) https://eel.is/c++draft/class.mem#general-5
"A member shall not be declared twice in the member-specification,
except that"
having an exception that name-independent non-static data member
declarations can appear multiple times (but again, if there is
a member which isn't name-independent, it can't appear after
name-independent ones)
3) it assumes that any name-independent declarations which weren't
previously valid result in the _ lookups being ambiguous, not just
if there are 2 _ declarations in the same scope, in particular the
https://eel.is/c++draft/basic.scope#block-2 mentioned cases
4) it assumes that _ in static function/block scope structured bindings
is never name-independent like in namespace scope structured bindings;
it matches clang behavior and is consistent with e.g. static type _;
not being name-independent both at namespace scope and at function/block
scope
As you preferred in the PR, for local scope bindings, the ambiguous cases
use a TREE_LIST with the ambiguous cases which can often be directly fed
into print_candidates. For member_vec after sorting/deduping, I chose to use
instead OVERLOAD with a new flag but only internally inside of the
member_vec, get_class_binding_direct turns it into a TREE_LIST. There are
2 reasons for that, in order to keep the member_vec binary search fast, I
think it is better to keep OVL_NAME usable on all elements because having
to special case TREE_LIST would slow everything down, and the callers need
to be able to chain the results anyway and so need an unshared TREE_LIST
they can tweak/destroy anyway.
name-independent declarations (even in older standards) will not have
-Wunused{,-variable,-but-set-variable} or -Wshadow* warnings diagnosed, but
unlike e.g. the clang implementation, this patch does diagnose
-Wunused-parameter for parameters with _ names because they aren't
name-independent and one can just omit their name instead.
2023-11-30 Jakub Jelinek <jakub@redhat.com>
PR c++/110349
gcc/c-family/
* c-cppbuiltin.cc (c_cpp_builtins): Predefine
__cpp_placeholder_variables=202306L for C++26.
gcc/cp/
* cp-tree.h: Implement C++26 P2169R4 - Placeholder variables with no
name.
(OVL_NAME_INDEPENDENT_DECL_P): Define.
(add_capture): Add unsigned * argument.
(name_independent_decl_p): New inline function.
* name-lookup.cc (class name_lookup): Make ambiguous and
add_value members public.
(name_independent_linear_search): New function.
(get_class_binding_direct): Handle member_vec_binary_search
returning OVL_NAME_INDEPENDENT_DECL_P OVERLOAD. Use
name_independent_linear_search rather than fields_linear_search
for linear lookup of _ name if !want_type.
(member_name_cmp): Sort name-independent declarations first.
(member_vec_dedup): Handle name-independent declarations.
(pop_local_binding): Handle binding->value being a TREE_LIST for
ambiguous name-independent declarations.
(supplement_binding): Handle name-independent declarations.
(update_binding): Likewise.
(check_local_shadow): Return tree rather than void, normally
NULL_TREE but old for name-independent declarations which used
to conflict with outer scope declaration. Don't emit -Wshadow*
warnings for name-independent declarations.
(pushdecl): Handle name-independent declarations.
* search.cc (lookup_field_r): Handle nval being a TREE_LIST.
* lambda.cc (build_capture_proxy): Adjust for ___.<number>
names of members.
(add_capture): Add NAME_INDEPENDENT_CNT argument. Use ___.<number>
name rather than ___ for second and following capture with
_ name.
(add_default_capture): Adjust add_capture caller.
* decl.cc (poplevel): Don't warn about name-independent declarations.
(duplicate_decls): If in C++26 a _ named declaration conflicts with
earlier declarations, emit explaining note why the new declaration
is not name-independent.
(reshape_init_class): If field is a TREE_LIST, emit an ambiguity
error with list of candidates rather than error about non-existing
non-static data member.
* parser.cc (cp_parser_lambda_introducer): Adjust add_capture callers.
Allow name-independent capture redeclarations.
(cp_parser_decomposition_declaration): Set decl_specs.storage_class
to sc_static for static structured bindings.
* pt.cc (tsubst_lambda_expr): Adjust add_capture caller.
gcc/testsuite/
* g++.dg/cpp26/name-independent-decl1.C: New test.
* g++.dg/cpp26/name-independent-decl2.C: New test.
* g++.dg/cpp26/name-independent-decl3.C: New test.
* g++.dg/cpp26/name-independent-decl4.C: New test.
* g++.dg/cpp26/name-independent-decl5.C: New test.
* g++.dg/cpp26/name-independent-decl6.C: New test.
* g++.dg/cpp26/feat-cxx26.C: Add __cpp_placeholder_variables test.
* config/i386/sse.md (sdot_prodv64qi): New expander.
(sseunpackmodelower): New mode attr.
(sdot_prod<mode>): Emulate sdot_prodv*qi with sodt_prov*hi
when TARGET_VNNIINT8 is not available.
gcc/testsuite/ChangeLog:
* gcc.target/i386/sdotprodint8_emulate.c: New test.
liuhongt [Tue, 28 Nov 2023 06:46:21 +0000 (14:46 +0800)]
Use vec_extact_lo instead of subreg in reduc_<code>_scal_m.
Loop vectorizer will use vec_perm to select lower part of a vector,
there could be some redundancy when using subreg in
reduc_<code>_scal_m, because rtl cse can't figure out vec_select lower
part is just subreg.
I'm trying to canonicalize vec_select to subreg like aarch64 did, but
there're so many regressions, some are easy to fix, some requires
middle-end adjustment.
So for simplicity, the patch use vec_select instead of subreg in
reduc_<code>_scal_m.
gcc/ChangeLog:
* config/i386/sse.md: (reduc_plus_scal_<mode>): Use
vec_extract_lo instead of subreg.
(reduc_<code>_scal_<mode>): Ditto.
(reduc_<code>_scal_<mode>): Ditto.
(reduc_<code>_scal_<mode>): Ditto.
(reduc_<code>_scal_<mode>): Ditto.
Alexandre Oliva [Thu, 30 Nov 2023 06:40:51 +0000 (03:40 -0300)]
c++: mark short-enums as packed
Unlike C, C++ doesn't mark enums shortened by -fshort-enums as packed.
This makes for undesirable warning differences between C and C++,
e.g. c-c++-common/analyzer/null-deref-pr108251-smp_fetch_ssl_fc_has_early*.c
triggers a warning about a type cast from a pointer to enum that, when
packed, might not be sufficiently aligned.
This change is not enough for that warning to trigger. The tree
expression generated by the C++ front-end is also a little too
complicated for us get to the base pointer. A separate patch takes
care of that.
for gcc/cp/ChangeLog
* decl.cc (finish_enum_value_list): Set TYPE_PACKED if
use_short_enum, and propagate it to variants.
Jason Merrill [Wed, 29 Nov 2023 23:08:23 +0000 (18:08 -0500)]
c++: remove LAMBDA_EXPR_MUTABLE_P
In review of the deducing 'this' patch it came up that LAMBDA_EXPR_MUTABLE_P
doesn't make sense for a lambda with an explicit object parameter. And it
was never necessary, so let's remove it.
gcc/cp/ChangeLog:
* cp-tree.h (LAMBDA_EXPR_MUTABLE_P): Remove.
* cp-tree.def: Remove documentation.
* lambda.cc (build_lambda_expr): Remove reference.
* parser.cc (cp_parser_lambda_declarator_opt): Likewise.
* pt.cc (tsubst_lambda_expr): Likewise.
* ptree.cc (cxx_print_lambda_node): Likewise.
* semantics.cc (capture_decltype): Get the object quals
from the object instead.
Tsukasa OI [Thu, 30 Nov 2023 02:37:54 +0000 (02:37 +0000)]
RISC-V: Fix 'E' extension version to test
Commit 006e90e1 ("RISC-V: Initial RV64E and LP64E support") caused a
regression (test failure) but this is caused by failing to track GCC
changes in that test case (not a true GCC bug).
This commit fixes the test case to track the latest GCC with 'E'
extension version 2.0 (ratified).
gcc/testsuite/ChangeLog:
* gcc.target/riscv/predef-13.c: Fix 'E' extension version to test.
Juzhe-Zhong [Thu, 30 Nov 2023 02:36:30 +0000 (10:36 +0800)]
RISC-V: Support highpart overlap for floating-point widen instructions
This patch leverages the approach of vwcvt/vext.vf2 which has been approved.
Their approaches are totally the same.
Tested no regression and committed.
PR target/112431
gcc/ChangeLog:
* config/riscv/vector.md: Add widenning overlap.
gcc/testsuite/ChangeLog:
* gcc.target/riscv/rvv/base/pr112431-10.c: New test.
* gcc.target/riscv/rvv/base/pr112431-11.c: New test.
* gcc.target/riscv/rvv/base/pr112431-12.c: New test.
* gcc.target/riscv/rvv/base/pr112431-13.c: New test.
* gcc.target/riscv/rvv/base/pr112431-14.c: New test.
* gcc.target/riscv/rvv/base/pr112431-15.c: New test.
* gcc.target/riscv/rvv/base/pr112431-7.c: New test.
* gcc.target/riscv/rvv/base/pr112431-8.c: New test.
* gcc.target/riscv/rvv/base/pr112431-9.c: New test.
Philipp Tomsich [Fri, 20 Oct 2023 20:13:40 +0000 (22:13 +0200)]
aarch64: Add support for Ampere-1B (-mcpu=ampere1b) CPU
This patch adds initial support for Ampere-1B core.
The Ampere-1B core implements ARMv8.7 with the following (compiler
visible) extensions:
- CSSC (Common Short Sequence Compression instructions),
- MTE (Memory Tagging Extension)
- SM3/SM4
gcc/ChangeLog:
* config/aarch64/aarch64-cores.def (AARCH64_CORE): Add ampere-1b
* config/aarch64/aarch64-cost-tables.h: Add ampere1b_extra_costs
* config/aarch64/aarch64-tune.md: Regenerate
* config/aarch64/aarch64.cc: Include ampere1b tuning model
* doc/invoke.texi: Document -mcpu=ampere1b
* config/aarch64/tuning_models/ampere1b.h: New file.
Marek Polacek [Fri, 17 Nov 2023 19:48:44 +0000 (14:48 -0500)]
c++: P2280R4, Using unknown refs in constant expr [PR106650]
This patch is an attempt to implement (part of?) P2280, Using unknown
pointers and references in constant expressions. (Note that R4 seems to
only allow References to unknown/Accesses via this, but not Pointers to
unknown.)
This patch works to the extent that the test case added in [expr.const]
works as expected, as well as the test in
<https://www.open-std.org/jtc1/sc22/wg21/docs/papers/2022/p2280r4.html#the-this-pointer>
Most importantly, the proposal makes this compile:
void check(int const (¶m)[3]) {
constexpr auto s = array_size(param);
static_assert (s == 3);
}
and I think it would be a pity not to have it in GCC 14.
What still doesn't work is the test in $3.2:
struct A2 { constexpr int f() { return 0; } };
struct B2 : virtual A2 {};
void f2(B2 &b) { constexpr int k = b.f(); }
where we say
error: '* & b' is not a constant expression
This will be fixed in the future.
PR c++/106650
gcc/cp/ChangeLog:
* constexpr.cc (cxx_eval_constant_expression) <case PARM_DECL>: Allow
reference to unknown/this as per P2280.
<case VAR_DECL>: Allow reference to unknown as per P2280.
gcc/testsuite/ChangeLog:
* g++.dg/cpp0x/constexpr-array-ptr6.C: Remove dg-error.
* g++.dg/cpp0x/constexpr-ref12.C: Likewise.
* g++.dg/cpp0x/constexpr-ref2.C: Adjust dg-error.
* g++.dg/cpp0x/noexcept34.C: Remove dg-error.
* g++.dg/cpp1y/lambda-generic-const10.C: Likewise.
* g++.dg/cpp0x/constexpr-ref13.C: New test.
* g++.dg/cpp1z/constexpr-ref1.C: New test.
* g++.dg/cpp1z/constexpr-ref2.C: New test.
* g++.dg/cpp2a/constexpr-ref1.C: New test.
Patrick Palka [Wed, 29 Nov 2023 21:42:39 +0000 (16:42 -0500)]
c++: bogus -Wparentheses warning [PR112765]
We need to consistently look through implicit INDIRECT_REF when
setting/checking for -Wparentheses warning suppression. In passing
use the recently introduced STRIP_REFERENCE_REF helper some more.
PR c++/112765
gcc/cp/ChangeLog:
* pt.cc (tsubst_expr) <case MODOP_EXPR>: Look through implicit
INDIRECT_REF when propagating -Wparentheses warning suppression.
* semantics.cc (maybe_warn_unparenthesized_assignment): Replace
REFERENCE_REF_P handling with STRIP_REFERENCE_REF.
(finish_parenthesized_expr): Likewise.
David Faust [Wed, 29 Nov 2023 17:36:02 +0000 (09:36 -0800)]
bpf: change ASM_COMMENT_START to '#'
The BPF "pseudo-C" assembly dialect uses semi-colon (;) to separate
statements, not to begin line comments. The GNU assembler was recently
changed accordingly:
This patch adapts the BPF backend in GCC accordingly, to use a hash (#)
instead of semi-colon (;) for ASM_COMMENT_START. This is supported
already in clang.
gcc/
* config/bpf/bpf.h (ASM_COMMENT_START): Change from ';' to '#'.
gcc/testsuite/
* gcc.target/bpf/core-builtin-enumvalue-opt.c: Change dg-final
scans to not assume a specific comment character.
* gcc.target/bpf/core-builtin-enumvalue.c: Likewise.
* gcc.target/bpf/core-builtin-type-based.c: Likewise.
* gcc.target/bpf/core-builtin-type-id.c: Likewise.
Jakub Jelinek [Wed, 29 Nov 2023 18:19:07 +0000 (19:19 +0100)]
rs6000: Fix up c-c++-common/builtin-classify-type-1.c failure [PR112725]
The rs6000 backend (and s390 one as well) diagnoses passing vector types
to unprototyped functions, which breaks the builtin-classify-type-1.c test.
The builtin isn't really unprototyped, it is just type-generic and accepting
vector types is just fine there, all it does is categorize the vector type.
The following patch makes sure we don't diagnose it for this builtin.
2023-11-29 Jakub Jelinek <jakub@redhat.com>
PR target/112725
* config/rs6000/rs6000.cc (invalid_arg_for_unprototyped_fn): Return
NULL for __builtin_classify_type calls with vector arguments.
Martin Jambor [Wed, 29 Nov 2023 15:24:33 +0000 (16:24 +0100)]
tree-sra: Avoid returns of references to SRA candidates
The enhancement to address PR 109849 contained an importsnt thinko,
and that any reference that is passed to a function and does not
escape, must also not happen to be aliased by the return value of the
function. This has quickly transpired as bugs PR 112711 and PR
112721.
Just as IPA-modref does a good enough job to allow us to rely on the
escaped set of variables, it sems to be doing well also on updating
EAF_NOT_RETURNED_DIRECTLY call argument flag which happens to address
exactly the situation we need to avoid. Of course, if a call
statement ignores any returned value, we also do not need to check the
flag.
Hopefully this does not pessimize things too much, I have verified
that the PR 109849 testcae remains quick and so should also the
benchmark it is derived from.
gcc/ChangeLog:
2023-11-27 Martin Jambor <mjambor@suse.cz>
PR tree-optimization/112711
PR tree-optimization/112721
* tree-sra.cc (build_access_from_call_arg): New parameter
CAN_BE_RETURNED, disqualify any candidate passed by reference if it is
true. Adjust leading comment.
(scan_function): Pass appropriate value to CAN_BE_RETURNED of
build_access_from_call_arg.
Rainer Orth [Wed, 29 Nov 2023 13:52:04 +0000 (14:52 +0100)]
testsuite, i386: Only check for cfi directives if supported [PR112729]
gcc.target/i386/apx-interrupt-1.c and two more tests FAIL on Solaris/x86
with the native assembler. Like Darwin as, it doesn't support cfi
directives. Instead of adding more and more targets in every affected
test, this patch introduces a cfi effective-target keyword to check for
the prerequisite.
Tested on i386-pc-solaris2.11 (as and gas), x86_64-pc-linux-gnu, and
x86_64-apple-darwin23.1.0.
Thomas Schwinge [Tue, 28 Nov 2023 16:52:34 +0000 (17:52 +0100)]
Fix 'g++.dg/cpp26/static_assert1.C' for '-fno-exceptions' configurations
This test case, added in recent commit 6ce952188ab39e303e4f63e474b5cba83b5b12fd
"c++: Implement C++26 P2741R3 - user-generated static_assert messages [PR110348]",
expectedly runs into 'UNSUPPORTED: [...]: exception handling disabled', but
along the way also FAILs a few tests:
UNSUPPORTED: g++.dg/cpp26/static_assert1.C -std=gnu++98
PASS: g++.dg/cpp26/static_assert1.C -std=gnu++11 (test for warnings, line 6)
[...]
PASS: g++.dg/cpp26/static_assert1.C -std=gnu++11 (test for warnings, line 51)
FAIL: g++.dg/cpp26/static_assert1.C -std=gnu++11 (test for errors, line 52)
PASS: g++.dg/cpp26/static_assert1.C -std=gnu++11 (test for warnings, line 56)
PASS: g++.dg/cpp26/static_assert1.C -std=gnu++11 (test for warnings, line 57)
FAIL: g++.dg/cpp26/static_assert1.C -std=gnu++11 at line 58 (test for errors, line 57)
PASS: g++.dg/cpp26/static_assert1.C -std=gnu++11 (test for warnings, line 59)
[...]
PASS: g++.dg/cpp26/static_assert1.C -std=gnu++11 at line 308 (test for errors, line 307)
UNSUPPORTED: g++.dg/cpp26/static_assert1.C -std=gnu++11: exception handling disabled
PASS: g++.dg/cpp26/static_assert1.C -std=gnu++14 (test for warnings, line 6)
[...]
PASS: g++.dg/cpp26/static_assert1.C -std=gnu++14 (test for warnings, line 51)
FAIL: g++.dg/cpp26/static_assert1.C -std=gnu++14 (test for errors, line 52)
PASS: g++.dg/cpp26/static_assert1.C -std=gnu++14 (test for warnings, line 56)
PASS: g++.dg/cpp26/static_assert1.C -std=gnu++14 (test for warnings, line 57)
FAIL: g++.dg/cpp26/static_assert1.C -std=gnu++14 at line 58 (test for errors, line 57)
PASS: g++.dg/cpp26/static_assert1.C -std=gnu++14 (test for warnings, line 59)
[...]
PASS: g++.dg/cpp26/static_assert1.C -std=gnu++14 at line 257 (test for errors, line 256)
FAIL: g++.dg/cpp26/static_assert1.C -std=gnu++14 (test for errors, line 261)
PASS: g++.dg/cpp26/static_assert1.C -std=gnu++14 (test for warnings, line 262)
[...]
PASS: g++.dg/cpp26/static_assert1.C -std=gnu++14 at line 308 (test for errors, line 307)
UNSUPPORTED: g++.dg/cpp26/static_assert1.C -std=gnu++14: exception handling disabled
PASS: g++.dg/cpp26/static_assert1.C -std=gnu++20 (test for warnings, line 6)
[...]
PASS: g++.dg/cpp26/static_assert1.C -std=gnu++20 (test for warnings, line 51)
FAIL: g++.dg/cpp26/static_assert1.C -std=gnu++20 (test for errors, line 52)
PASS: g++.dg/cpp26/static_assert1.C -std=gnu++20 (test for warnings, line 56)
PASS: g++.dg/cpp26/static_assert1.C -std=gnu++20 (test for warnings, line 57)
FAIL: g++.dg/cpp26/static_assert1.C -std=gnu++20 at line 58 (test for errors, line 57)
PASS: g++.dg/cpp26/static_assert1.C -std=gnu++20 (test for warnings, line 59)
[...]
PASS: g++.dg/cpp26/static_assert1.C -std=gnu++20 at line 257 (test for errors, line 256)
FAIL: g++.dg/cpp26/static_assert1.C -std=gnu++20 (test for errors, line 261)
PASS: g++.dg/cpp26/static_assert1.C -std=gnu++20 (test for warnings, line 262)
[...]
PASS: g++.dg/cpp26/static_assert1.C -std=gnu++20 at line 308 (test for errors, line 307)
UNSUPPORTED: g++.dg/cpp26/static_assert1.C -std=gnu++20: exception handling disabled
Use an explicit '-fexceptions' to turn this front end test case all-PASS.
gcc/testsuite/
* g++.dg/cpp26/static_assert1.C: Fix for '-fno-exceptions'
configurations.
Thomas Schwinge [Tue, 28 Nov 2023 16:30:13 +0000 (17:30 +0100)]
Fix '23_containers/span/at.cc' for '-fno-exceptions' configurations
Added in recent commit 1fa85dcf656e2f2c7e483c9ed3c2680bf7db6858
"libstdc++: Add std::span::at for C++26 (P2821R5)", the test case already
does use '#if __cpp_exceptions', but failed to correspondingly guard the
'dg-warning' directives, resulting in:
FAIL: 23_containers/span/at.cc -std=gnu++26 (test for warnings, line 15)
FAIL: 23_containers/span/at.cc -std=gnu++26 (test for warnings, line 26)
PASS: 23_containers/span/at.cc -std=gnu++26 (test for excess errors)
PASS: 23_containers/span/at.cc -std=gnu++26 execution test
libstdc++-v3/
* testsuite/23_containers/span/at.cc: Fix for '-fno-exceptions'
configurations.
git://gcc.gnu.org / gcc.git / log