Simon Martin [Wed, 5 Mar 2025 08:08:57 +0000 (09:08 +0100)]
c++: Fix checking assert upon invalid class definition [PR116740]
A checking assert triggers upon the following invalid code since
GCC 11:
=== cut here ===
class { a (struct b;
} struct b
=== cut here ===
The problem is that during error recovery, we call
set_identifier_type_value_with_scope for B in the global namespace, and
the checking assert added via r11-7228-g8f93e1b892850b fails.
This patch relaxes that assert to not fail if we've seen a parser error
(it a generalization of another fix done to that checking assert via r11-7266-g24bf79f1798ad1).
PR c++/116740
gcc/cp/ChangeLog:
* name-lookup.cc (set_identifier_type_value_with_scope): Don't
fail assert with ill-formed input.
Jakub Jelinek [Wed, 5 Mar 2025 06:47:52 +0000 (07:47 +0100)]
openmp, c++: Fix up OpenMP/OpenACC handling in C++ modules [PR119102]
modules.cc has apparently support for extensions and attempts to ensure
that if a module is compiled with those extensions enabled, sources which
use the module are compiled with the same extensions.
The only extension supported is SE_OPENMP right now.
And the use of the extension is keyed on streaming out or in OMP_CLAUSE
tree.
This is undesirable for several reasons.
OMP_CLAUSE is the only tree which can appear in the IL even without
-fopenmp/-fopenmp-simd/-fopenacc (when simd ("notinbranch") or
simd ("inbranch") attributes are used), and it can appear also in all
the 3 modes mentioned above. On the other side, with the exception of
arguments of attributes added e.g. for declare simd where no harm should
be done if -fopenmp/-fopenmp-simd isn't enabled later on, OMP_CLAUSE appears
in OMP_*_CLAUSES of OpenMP/OpenACC construct trees. And those construct
trees often have no clauses at all, so keying the extension on OMP_CLAUSE
doesn't catch many cases that should be caught.
Furthermore, for OpenMP we have 2 modes, -fopenmp-simd which parses some
OpenMP but constructs from that mostly OMP_SIMD and a few other cases,
and -fopenmp which includes that and far more on top of that; and there is
also -fopenacc.
So, this patch stops setting/requesting the extension on OMP_CLAUSE,
introduces 3 extensions rather than one (SE_OPENMP_SIMD, SE_OPENMP and
SE_OPENACC) and keyes those on OpenMP constructs from the -fopenmp-simd
subset, other OpenMP constructs and OpenACC constructs.
2025-03-05 Jakub Jelinek <jakub@redhat.com>
PR c++/119102
gcc/cp/
* module.cc (enum streamed_extensions): Add SE_OPENMP_SIMD
and SE_OPENACC, change value of SE_OPENMP and SE_BITS.
(CASE_OMP_SIMD_CODE, CASE_OMP_CODE, CASE_OACC_CODE): Define.
(trees_out::start): Don't set SE_OPENMP extension for OMP_CLAUSE.
Set SE_OPENMP_SIMD extension for CASE_OMP_SIMD_CODE, SE_OPENMP
for CASE_OMP_CODE and SE_OPENACC for CASE_OACC_CODE.
(trees_in::start): Don't fail for OMP_CLAUSE with missing
SE_OPENMP extension. Do fail for CASE_OMP_SIMD_CODE and missing
SE_OPENMP_SIMD extension, or CASE_OMP_CODE and missing SE_OPENMP
extension, or CASE_OACC_CODE and missing SE_OPENACC extension.
(module_state::write_readme): Write all of SE_OPENMP_SIMD, SE_OPENMP
and SE_OPENACC extensions.
(module_state::read_config): Diagnose missing -fopenmp, -fopenmp-simd
and/or -fopenacc depending on extensions used.
gcc/testsuite/
* g++.dg/modules/pr119102_a.H: New test.
* g++.dg/modules/pr119102_b.C: New test.
* g++.dg/modules/omp-3_a.C: New test.
* g++.dg/modules/omp-3_b.C: New test.
* g++.dg/modules/omp-3_c.C: New test.
* g++.dg/modules/omp-3_d.C: New test.
* g++.dg/modules/oacc-1_a.C: New test.
* g++.dg/modules/oacc-1_b.C: New test.
* g++.dg/modules/oacc-1_c.C: New test.
Jakub Jelinek [Wed, 5 Mar 2025 05:41:00 +0000 (06:41 +0100)]
c++: Apply/diagnose attributes when instatiating ARRAY/POINTER/REFERENCE_TYPE [PR118787]
The following testcase IMO in violation of the P2552R3 paper doesn't
pedwarn on alignas applying to dependent types or alignas with dependent
argument.
tsubst was just ignoring TYPE_ATTRIBUTES.
The following patch fixes it for the POINTER/REFERENCE_TYPE and
ARRAY_TYPE cases, but perhaps we need to do the same also for other
types (INTEGER_TYPE/REAL_TYPE and the like). I guess I'll need to
construct more testcases.
2025-03-05 Jakub Jelinek <jakub@redhat.com>
PR c++/118787
* pt.cc (tsubst) <case ARRAY_TYPE>: Use return t; only if it doesn't
have any TYPE_ATTRIBUTES. Call apply_late_template_attributes.
<case POINTER_TYPE, case REFERENCE_TYPE>: Likewise. Formatting fix.
Xi Ruoyao [Sun, 2 Mar 2025 11:02:50 +0000 (19:02 +0800)]
LoongArch: Fix incorrect reorder of __lsx_vldx and __lasx_xvldx [PR119084]
They could be incorrectly reordered with store instructions like st.b
because the RTL expression does not have a memory_operand or a (mem)
expression. The incorrect reorder has been observed in openh264 LTO
build.
Expand them to a (mem) expression instead of unspec to fix the issue.
Then we need to make loongarch_address_insns return 1 for
ADDRESS_REG_REG because the constraint "R" expects this behavior, or
the vldx instruction will be considered invalid by the register
allocate pass and turned to add.d + vld. Apply the ADDRESS_REG_REG
penalty in loongarch_address_cost instead, loongarch_rtx_costs should
also call loongarch_address_cost instead of loongarch_address_insns
then.
Thomas Koenig [Tue, 4 Mar 2025 19:13:19 +0000 (20:13 +0100)]
C prototypes for external arguments; add warning for mismatch.
The problem was that we were not handling external dummy arguments
with -fc-prototypes-external. In looking at this, I found that we
were not warning about external procedures with different argument
lists. This can actually be legal (see the two test cases) but
creates a problem for the C prototypes: If we have something like
subroutine foo(a,n)
external a
if (n == 1) call a(1)
if (n == 2) call a(2,3)
end subroutine foo
then, pre-C23, we could just have written out the prototype as
void foo_ (void (*a) (), int *n);
but this is illegal in C23. What to do? I finally chose to warn
about the argument mismatch, with a new option. Warn only because the
code above is legal, but include in -Wall because such code seems highly
suspect. This option is also implied in -fc-prototypes-external. I also
put a warning in the generated header file in that case, so users
have a chance to see what is going on (especially since gcc now
defaults to C23).
gcc/fortran/ChangeLog:
PR fortran/119049
PR fortran/119074
* dump-parse-tree.cc (seen_conflict): New static varaible.
(gfc_dump_external_c_prototypes): Initialize it. If it was
set, write out a warning that -std=c23 will not work.
(write_proc): Move the work of actually writing out the
formal arglist to...
(write_formal_arglist): New function. Handle external dummy
parameters and their argument lists. If there were mismatched
arguments, output an empty argument list in pre-C23 style.
* gfortran.h (struct gfc_symbol): Add ext_dummy_arglist_mismatch
flag and formal_at.
* invoke.texi: Document -Wexternal-argument-mismatch.
* lang.opt: Put it in.
* resolve.cc (resolve_function): If warning about external
argument mismatches, build a formal from actual arglist the
first time around, and later compare and warn.
(resolve_call): Likewise
gcc/testsuite/ChangeLog:
PR fortran/119049
PR fortran/119074
* gfortran.dg/interface_55.f90: New test.
* gfortran.dg/interface_56.f90: New test.
real 0m2.422s
user 0m2.421s
sys 0m0.000s
jh@shroud:~> sh bmk.sh blsmsk
real 0m0.973s
user 0m0.973s
sys 0m0.000s
real 0m2.422s
user 0m2.422s
sys 0m0.000s
We already have runable that controls tzcnt together with lzcnt and popcnt.
Since it seems that only tzcnt is affected I added new tunable to control tzcnt
only. I also added splitters for blsi/blsr/blsmsk implemented analogously to
existing splitter for lzcnt.
The patch is neutral on SPEC. We produce blsi and blsr in some internal loops, but
they usually have same destination as source. However it is good to break the
dependency chain to avoid patogolical cases and it is quite cheap overall, so I
think we want to enable this for generic. I will send followup patch for this.
Bootstrapped/regtested x86_64-linux, will commit it shortly.
gcc/ChangeLog:
* config/i386/i386.h (TARGET_AVOID_FALSE_DEP_FOR_TZCNT): New macro.
(TARGET_AVOID_FALSE_DEP_FOR_BLS): New macro.
* config/i386/i386.md (*bmi_blsi_<mode>): Add splitter for false
dependency.
(*bmi_blsi_<mode>_ccno): Add splitter for false dependency.
(*bmi_blsi_<mode>_falsedep): New pattern.
(*bmi_blsmsk_<mode>): Add splitter for false dependency.
(*bmi_blsmsk_<mode>_falsedep): New pattern.
(*bmi_blsr_<mode>): Add splitter for false dependency.
(*bmi_blsr_<mode>_cmp): Add splitter for false dependency
(*bmi_blsr_<mode>_cmp_falsedep): New pattern.
* config/i386/x86-tune.def (X86_TUNE_AVOID_FALSE_DEP_FOR_TZCNT): New tune.
(X86_TUNE_AVOID_FALSE_DEP_FOR_BLS): New tune.
gcc/testsuite/ChangeLog:
* gcc.target/i386/blsi.c: New test.
* gcc.target/i386/blsmsk.c: New test.
* gcc.target/i386/blsr.c: New test.
Jan Hubicka [Mon, 3 Mar 2025 18:12:20 +0000 (19:12 +0100)]
Make ix86_macro_fusion_pair_p and ix86_fuse_mov_alu_p match current CPUs
The current implementation of fussion predicates misses some common
fussion cases on zen and more recent cores. I added knobs for
individual conditionals we test.
1) I split checks for fusing ALU with conditional operands when the ALU
has memory operand. This seems to be supported by zen3+ and by
tigerlake and coperlake (according to Agner Fog's manual)
2) znver4 and 5 supports fussion of ALU and conditional even if ALU has
memory and immediate operands.
This seems to be relatively important enabling 25% more fusions on
gcc bootstrap.
3) no CPU supports fusing when ALU contains IP relative memory
references. I added separate knob so we do not forger about this if
this gets supoorted later.
The patch does not solve the limitation of sched that fuse pairs must be
adjacent on imput and the first operation must be signle-set. Fixing
single-set is easy (I have separate patch for this), for non-adjacent
pairs we need bigger surgery.
To verify what CPU really does I made simpe test script.
jh@ryzen3:~> cat fuse-test.c
int b;
const int z = 0;
const int o = 1;
int
main()
{
int a = 1000000000;
int b;
int z = 0;
int o = 1;
asm volatile ("\n"
".L1234:\n"
"nop\n"
"subl %3, %0\n"
echo ALU with immediate
dotest
echo ALU with memory
dotest -D MEM
echo ALU with IP relative memory
dotest -D MEM -D IPRELATIVE
echo CMP with immediate
dotest -D CMP
echo CMP with memory
dotest -D CMP -D MEM
echo CMP with memory and immediate
dotest -D CMP -D MEMIMM
echo CMP with IP relative memory
dotest -D CMP -D MEM -D IPRELATIVE
echo TEST
dotest -D TEST
On zen5 I get:
ALU with immediate
20,345 ex_ret_fused_instr:u
1,000,020,278 ex_ret_fused_instr:u
ALU with memory
20,367 ex_ret_fused_instr:u
1,000,020,290 ex_ret_fused_instr:u
ALU with IP relative memory
20,395 ex_ret_fused_instr:u
20,403 ex_ret_fused_instr:u
CMP with immediate
20,369 ex_ret_fused_instr:u
1,000,020,301 ex_ret_fused_instr:u
CMP with memory
20,314 ex_ret_fused_instr:u
1,000,020,341 ex_ret_fused_instr:u
CMP with memory and immediate
20,372 ex_ret_fused_instr:u
1,000,020,266 ex_ret_fused_instr:u
CMP with IP relative memory
20,382 ex_ret_fused_instr:u
20,369 ex_ret_fused_instr:u
TEST
20,346 ex_ret_fused_instr:u
1,000,020,301 ex_ret_fused_instr:u
IP relative memory seems to not be documented.
On zen3/4 I get:
ALU with immediate
20,263 ex_ret_fused_instr:u
1,000,020,051 ex_ret_fused_instr:u
ALU with memory
20,255 ex_ret_fused_instr:u
1,000,020,056 ex_ret_fused_instr:u
ALU with IP relative memory
20,253 ex_ret_fused_instr:u
20,266 ex_ret_fused_instr:u
CMP with immediate
20,264 ex_ret_fused_instr:u
1,000,020,052 ex_ret_fused_instr:u
CMP with memory
20,253 ex_ret_fused_instr:u
1,000,019,794 ex_ret_fused_instr:u
CMP with memory and immediate
20,260 ex_ret_fused_instr:u
20,264 ex_ret_fused_instr:u
CMP with IP relative memory
20,258 ex_ret_fused_instr:u
20,256 ex_ret_fused_instr:u
TEST
20,261 ex_ret_fused_instr:u
1,000,020,048 ex_ret_fused_instr:u
zen1 and 2 gets:
ALU with immediate
21,610 ex_ret_fus_brnch_inst:u
21,697 ex_ret_fus_brnch_inst:u
ALU with memory
21,479 ex_ret_fus_brnch_inst:u
21,747 ex_ret_fus_brnch_inst:u
ALU with IP relative memory
21,623 ex_ret_fus_brnch_inst:u
21,684 ex_ret_fus_brnch_inst:u
CMP with immediate
21,708 ex_ret_fus_brnch_inst:u
1,000,021,288 ex_ret_fus_brnch_inst:u
CMP with memory
21,689 ex_ret_fus_brnch_inst:u
1,000,004,270 ex_ret_fus_brnch_inst:u
CMP with memory and immediate
21,604 ex_ret_fus_brnch_inst:u
21,671 ex_ret_fus_brnch_inst:u
CMP with IP relative memory
21,589 ex_ret_fus_brnch_inst:u
21,602 ex_ret_fus_brnch_inst:u
TEST
21,600 ex_ret_fus_brnch_inst:u
1,000,021,233 ex_ret_fus_brnch_inst:u
I tested the patch on zen3 and zen5 and spec2k17 and it seems neutral, however
the number of fussion does go up.
Bootstrapped/regtested x86_64-linux, I plan to commit it tomorrow.
Honza
gcc/ChangeLog:
* config/i386/i386.h (TARGET_FUSE_ALU_AND_BRANCH_MEM): New macro.
(TARGET_FUSE_ALU_AND_BRANCH_MEM_IMM): New macro.
(TARGET_FUSE_ALU_AND_BRANCH_RIP_RELATIVE): New macro.
* config/i386/x86-tune-sched.cc (ix86_fuse_mov_alu_p): Support
non-single-set.
(ix86_macro_fusion_pair_p): Allow ALU which only clobbers;
be more careful about immediates; check TARGET_FUSE_ALU_AND_BRANCH_MEM,
TARGET_FUSE_ALU_AND_BRANCH_MEM_IMM, TARGET_FUSE_ALU_AND_BRANCH_RIP_RELATIVE;
verify that we never use unsigned checks with inc/dec.
* config/i386/x86-tune.def (X86_TUNE_FUSE_ALU_AND_BRANCH): New tune.
(X86_TUNE_FUSE_ALU_AND_BRANCH_MEM): New tune.
(X86_TUNE_FUSE_ALU_AND_BRANCH_MEM_IMM): New tune.
(X86_TUNE_FUSE_ALU_AND_BRANCH_RIP_RELATIVE): New tune.
which output_constructor_regular_field doesn't want to see. This
happens since r9-1483: process_init_constructor_array can now create
a RANGE_EXPR. But the bug isn't in that patch; the problem is that
build_vec_init doesn't handle RANGE_EXPRs.
build_vec_init has a FOR_EACH_CONSTRUCTOR_ELT loop which populates
const_vec. In this case it loops over the elements of
{[0 ... 1]={.low=0, .high=1}}
but assumes that each element initializes one element. So after the
loop num_initialized_elts was 1, and then below:
HOST_WIDE_INT last = tree_to_shwi (maxindex);
if (num_initialized_elts <= last)
{
tree field = size_int (num_initialized_elts);
if (num_initialized_elts != last)
field = build2 (RANGE_EXPR, sizetype, field,
size_int (last));
CONSTRUCTOR_APPEND_ELT (const_vec, field, e);
}
we added the extra initializer.
It seemed convenient to use range_expr_nelts like below.
PR c++/109431
gcc/cp/ChangeLog:
* cp-tree.h (range_expr_nelts): Declare.
* init.cc (build_vec_init): If the CONSTRUCTOR's index is a
RANGE_EXPR, use range_expr_nelts to count how many elements
were initialized.
Tamar Christina [Tue, 4 Mar 2025 11:15:26 +0000 (11:15 +0000)]
aarch64: force operand to fresh register to avoid subreg issues [PR118892]
When the input is already a subreg and we try to make a paradoxical
subreg out of it for copysign this can fail if it violates the subreg
relationship.
Use force_lowpart_subreg instead of lowpart_subreg to then force the
results to a register instead of ICEing.
gcc/ChangeLog:
PR target/118892
* config/aarch64/aarch64.md (copysign<GPF:mode>3): Use
force_lowpart_subreg instead of lowpart_subreg.
gcc/testsuite/ChangeLog:
PR target/118892
* gcc.target/aarch64/copysign-pr118892.c: New test.
Fix folding of BIT_NOT_EXPR for POLY_INT_CST [PR118976]
There was an embarrassing typo in the folding of BIT_NOT_EXPR for
POLY_INT_CSTs: it used - rather than ~ on the poly_int. Not sure
how that happened, but it might have been due to the way that
~x is implemented as -1 - x internally.
gcc/
PR tree-optimization/118976
* fold-const.cc (const_unop): Use ~ rather than - for BIT_NOT_EXPR.
* config/aarch64/aarch64.cc (aarch64_test_sve_folding): New function.
(aarch64_run_selftests): Run it.
The simplify_logical_relational_operation code (in its current form)
was written with arithmetic rather than CC modes in mind. Since CCFP
is a CC mode, it fails the HONOR_NANS check, and so the function assumes
that ge | lt => true.
If one comparison is unsigned then it should be safe to assume that
the other comparison is also unsigned, even for CC modes, since the
optimisation checks that the comparisons are between the same operands.
For the other cases, we can only safely fold comparisons of CC mode
values if the result is always-true (15) or always-false (0).
It turns out that the original testcase for PR117186, which ran at -O,
was relying on the old behaviour for some of the functions. It needs
4-instruction combinations, and so -fexpensive-optimizations, to pass
in its intended form.
gcc/
PR rtl-optimization/119002
* simplify-rtx.cc
(simplify_context::simplify_logical_relational_operation): Handle
comparisons between CC values. If there is no evidence that the
CC values are unsigned, restrict the fold to always-true or
always-false results.
gcc/testsuite/
* gcc.c-torture/execute/ieee/pr119002.c: New test.
* gcc.target/aarch64/pr117186.c: Run at -O2 rather than -O.
When we vectorize a .COND_ADD reduction and apply the single-use-def
cycle optimization we can end up chosing the wrong else value for
subsequent .COND_ADD. The following rectifies this.
PR tree-optimization/119096
* tree-vect-loop.cc (vect_transform_reduction): Use the
correct else value for .COND_fn.
Pan Li [Mon, 3 Mar 2025 06:51:21 +0000 (14:51 +0800)]
RISC-V: Fix the test case bug-3.c failure
The bug-3.c would like to check the slli a[0-9]+, a[0-9]+, 33 for the
big poly int handling. But the underlying insn may change to slli 1
+ slli 32 with sorts of optimization. Thus, update the asm check to
function body check with above slli 1 + slli 32 series.
The below test suites are passed for this patch.
* The rv64gcv fully regression test.
gcc/testsuite/ChangeLog:
* gcc.target/riscv/rvv/autovec/bug-3.c: Update asm check to
function body check.
Harald Anlauf [Sun, 2 Mar 2025 21:20:28 +0000 (22:20 +0100)]
Fortran: reject empty derived type with bind(C) attribute [PR101577]
PR fortran/101577
gcc/fortran/ChangeLog:
* symbol.cc (verify_bind_c_derived_type): Generate error message
for derived type with no components in standard conformance mode,
indicating that this is a GNU extension.
gcc/testsuite/ChangeLog:
* gfortran.dg/empty_derived_type.f90: Adjust dg-options.
* gfortran.dg/empty_derived_type_2.f90: New test.
Andrew Carlotti [Fri, 7 Feb 2025 17:13:36 +0000 (17:13 +0000)]
aarch64: Ignore target pragmas while defining intrinsics
Refactor the switcher classes into two separate classes:
- sve_alignment_switcher takes the alignment switching functionality,
and is used only for ABI correctness when defining sve structure
types.
- aarch64_target_switcher takes the rest of the functionality of
aarch64_simd_switcher and sve_switcher, and gates simd/sve specific
parts upon the specified feature flags.
Additionally, aarch64_target_switcher now adds dependencies of the
specified flags (which adds +fcma and +bf16 to some intrinsic
declarations), and unsets current_target_pragma.
This last change fixes an internal bug where we would sometimes add a
user specified target pragma (stored in current_target_pragma) on top of
an internally specified target architecture while initialising
intrinsics with `#pragma GCC aarch64 "arm_*.h"`. As far as I can tell, this
has no visible impact at the moment. However, the unintended target
feature combinations lead to unwanted behaviour in an under-development
patch.
This also fixes a missing Makefile dependency, which was due to
aarch64-sve-builtins.o incorrectly depending on the undefined $(REG_H).
The correct $(REGS_H) dependency is added to the switcher's new source
location.
gcc/ChangeLog:
* common/config/aarch64/aarch64-common.cc
(struct aarch64_extension_info): Add field.
(aarch64_get_required_features): New.
* config/aarch64/aarch64-builtins.cc
(aarch64_simd_switcher::aarch64_simd_switcher): Rename to...
(aarch64_target_switcher::aarch64_target_switcher): ...this,
and extend to handle sve, nosimd and target pragmas.
(aarch64_simd_switcher::~aarch64_simd_switcher): Rename to...
(aarch64_target_switcher::~aarch64_target_switcher): ...this,
and extend to handle sve, nosimd and target pragmas.
(handle_arm_acle_h): Use aarch64_target_switcher.
(handle_arm_neon_h): Rename switcher and pass explicit flags.
(aarch64_general_init_builtins): Ditto.
* config/aarch64/aarch64-protos.h
(class aarch64_simd_switcher): Rename to...
(class aarch64_target_switcher): ...this, and add new members.
(aarch64_get_required_features): New prototype.
* config/aarch64/aarch64-sve-builtins.cc
(sve_switcher::sve_switcher): Delete
(sve_switcher::~sve_switcher): Delete
(sve_alignment_switcher::sve_alignment_switcher): New
(sve_alignment_switcher::~sve_alignment_switcher): New
(register_builtin_types): Use alignment switcher
(init_builtins): Rename switcher.
(handle_arm_neon_sve_bridge_h): Ditto.
(handle_arm_sme_h): Ditto.
(handle_arm_sve_h): Ditto, and use alignment switcher.
* config/aarch64/aarch64-sve-builtins.h
(class sve_switcher): Delete.
(class sme_switcher): Delete.
(class sve_alignment_switcher): New.
* config/aarch64/t-aarch64 (aarch64-builtins.o): Add $(REGS_H).
(aarch64-sve-builtins.o): Remove $(REG_H).
arm: remove some redundant zero_extend ops on thumb1
The code in gcc.target/unsigned-extend-1.c really should not need an
unsigned extension operations when the optimizers are used. For Arm
and thumb2 that is indeed the case, but for thumb1 code it gets more
complicated as there are too many instructions for combine to look at.
For thumb1 we end up with two redundant zero_extend patterns which are
not removed: the first after the subtract instruction and the second of
the final boolean result.
We can partially fix this (for the second case above) by adding a new
split pattern for LEU and GEU patterns which work because the two
instructions for the [LG]EU pattern plus the redundant extension
instruction are combined into a single insn, which we can then split
using the 3->2 method back into the two insns of the [LG]EU sequence.
Because we're missing the optimization for all thumb1 cases (not just
those architectures with UXTB), I've adjust the testcase to detect all
the idioms that we might use for zero-extending a value, namely:
UXTB
AND ...#255 (in thumb1 this would require a register to hold 255)
LSL ... #24; LSR ... #24
but I've also marked this test as XFAIL for thumb1 because we can't yet
eliminate the first of the two extend instructions.
gcc/
* config/arm/thumb1.md (split patterns for GEU and LEU): New.
gcc/testsuite:
* gcc.target/arm/unsigned-extend-1.c: Expand check for any
insn suggesting a zero-extend. XFAIL for thumb1 code.
--cut here--
/* Otherwise, if this register is used by I3, then this register
now dies here, so we must put a REG_DEAD note here unless there
is one already. */
else if (reg_referenced_p (XEXP (note, 0), PATTERN (i3))
&& ! (REG_P (XEXP (note, 0))
? find_regno_note (i3, REG_DEAD,
REGNO (XEXP (note, 0)))
: find_reg_note (i3, REG_DEAD, XEXP (note, 0))))
{
PUT_REG_NOTE_KIND (note, REG_DEAD);
place = i3;
}
--cut here--
Flags register is used in I3, but there already is a REG_DEAD note in I3.
The above condition doesn't trigger and continues in the "else" part where
REG_DEAD note is put to I2. The proposed solution corrects the above
logic to trigger every time the register is referenced in I3, avoiding the
"else" part.
PR rtl-optimization/118739
gcc/ChangeLog:
* combine.cc (distribute_notes) <case REG_UNUSED>: Correct the
logic when the register is used by I3.
Martin Jambor [Mon, 3 Mar 2025 13:53:03 +0000 (14:53 +0100)]
ipa-vr: Handle non-conversion unary ops separately from conversions (PR 118785)
Since we construct arithmetic jump functions even when there is a
type conversion in between the operation encoded in the jump function
and when it is passed in a call argument, the IPA propagation phase
must also perform the operation and conversion in two steps. IPA-VR
had actually been doing it even before for binary operations but, as
PR 118756 exposes, not in the case on unary operations. This patch
adds the necessary step to rectify that.
Like in the scalar constant case, we depend on
expr_type_first_operand_type_p to determine the type of the result of
the arithmetic operation. On top this, the patch special-cases
ABSU_EXPR because it looks useful an so that the PR testcase exercises
the added code-path. This seems most appropriate for stage 4, long
term we should probably stream the types, probably after also encoding
them with a string of expr_eval_op rather than what we have today.
A check for expr_type_first_operand_type_p was also missing in the
handling of binary ops and the intermediate value_range was
initialized with a wrong type, so I also fixed this.
gcc/ChangeLog:
2025-02-24 Martin Jambor <mjambor@suse.cz>
PR ipa/118785
* ipa-cp.cc (ipa_vr_intersect_with_arith_jfunc): Handle non-conversion
unary operations separately before doing any conversions. Check
expr_type_first_operand_type_p for non-unary operations too. Fix type
of op_res.
gcc/testsuite/ChangeLog:
2025-02-24 Martin Jambor <mjambor@suse.cz>
PR ipa/118785
* g++.dg/lto/pr118785_0.C: New test.
We are detecting a cycle as double reduction where the inner loop
cycle has extra out-of-loop uses. This clashes at least with
assumptions from the SLP discovery code which says the cycle
isn't reachable from another SLP instance. It also was not intended
to support this case, in fact with GCC 14 we seem to generate wrong
code here.
PR tree-optimization/119057
* tree-vect-loop.cc (check_reduction_path): Add argument
specifying whether we're analyzing the inner loop of a
double reduction. Do not allow extra uses outside of the
double reduction cycle in this case.
(vect_is_simple_reduction): Adjust.
Richard Biener [Mon, 3 Mar 2025 08:54:15 +0000 (09:54 +0100)]
ipa/119067 - bogus TYPE_PRECISION check on VECTOR_TYPE
odr_types_equivalent_p can end up using TYPE_PRECISION on vector
types which is a no-go. The following instead uses TYPE_VECTOR_SUBPARTS
for vector types so we also end up comparing the number of vector elements.
PR ipa/119067
* ipa-devirt.cc (odr_types_equivalent_p): Check
TYPE_VECTOR_SUBPARTS for vectors.
* g++.dg/lto/pr119067_0.C: New testcase.
* g++.dg/lto/pr119067_1.C: Likewise.
Fortran: Fix regression on double free on elemental function [PR118747]
Fix a regression were adding a temporary variable inserted a copy of the
argument to the elemental function. That copy was then later used to
free allocated memory, but the freeing was not tracked in the source
array correctly.
PR fortran/118747
gcc/fortran/ChangeLog:
* trans-array.cc (gfc_trans_array_ctor_element): Remove copy to
temporary variable.
* trans-expr.cc (gfc_conv_procedure_call): Use references to
array members instead of copies when freeing after use.
Formatting fix.
gcc/testsuite/ChangeLog:
* gfortran.dg/alloc_comp_auto_array_4.f90: New test.
Jeff Law [Sun, 2 Mar 2025 19:08:34 +0000 (12:08 -0700)]
[RISC-V][PR target/118934] Fix ICE in RISC-V long branch support
I'm not sure if I goof'd this or if I merely upstreamed someone else's goof.
Either way the long branch code isn't working correctly.
We were using 'n' as the output modifier to negate the condition. But 'n' has
a special meaning elsewhere, so when presented with a condition rather than
what was expected, boom, the compiler ICE'd.
Thankfully there's only a few places where we were using %n which I turned into
%r.
The BZ entry includes a good testcase, it just takes a long time to compile as
it's trying to create the out-of-range scenario. I'm not including the
testcase due to how long it takes, but I did test it locally to ensure it's
working properly now.
I'm sure that with a little bit of work I could create at testcase that worked
before and fails with the trunk (by taking advantage of the fuzzyness in length
computations). So I'm going to consider this a regression.
Will push to the trunk after pre-commit testing does its thing.
Sandra Loosemore [Tue, 25 Feb 2025 20:39:31 +0000 (20:39 +0000)]
Fortran: Move "Standard" subheading in documentation [PR47928]
As noted in the issue, the version of the standard an intrinsic was
introduced in is usually not the second-most-important thing a user
needs to know. This patch moves it from near the beginning of each
section towards the end, just ahead of "See also".
Sandra Loosemore [Tue, 25 Feb 2025 23:39:25 +0000 (23:39 +0000)]
Fortran: Rename/move "Syntax" subheading in documentation [PR47928]
As suggested in the issue, it makes more sense to describe the function
call argument syntax before talking about the arguments in the description.
gcc/fortran/ChangeLog
PR fortran/47928
* gfortran.texi: Move all the "Syntax" subheadings ahead of
"Description", and rename to "Synopsis".
* intrinsic.texi: Likewise.
Sandra Loosemore [Tue, 25 Feb 2025 01:03:52 +0000 (01:03 +0000)]
Fortran: Tidy subheadings in Fortran documentation [PR47928]
This is a preparatory patch for the main documentation changes requested
in the issue.
gcc/fortran/ChangeLog
PR fortran/47928
* gfortran.texi: Consistently use "@emph{Notes}:" instead of
other spellings.
* intrinsic.texi: Likewise. Also fix an inconsistent capitalization
and remove a redundant "Standard" entry.
Jakub Jelinek [Sun, 2 Mar 2025 10:30:35 +0000 (11:30 +0100)]
avr: Fix up avr_print_operand diagnostics [PR118991]
As can be seen in gcc/po/gcc.pot:
#: config/avr/avr.cc:2754
#, c-format
msgid "bad I/O address 0x"
msgstr ""
exgettext couldn't retrieve the whole format string in this case,
because it uses a macro in the middle. output_operand_lossage
is c-format function though, so we can't use %wx to print HOST_WIDE_INT,
and HOST_WIDE_INT_PRINT_HEX_PURE is on some hosts %lx, on others %llx
and on others %I64x so isn't really translatable that way.
As Joseph mentioned in the PR, there is no easy way around this
but go through a temporary buffer, which the following patch does.
2025-03-02 Jakub Jelinek <jakub@redhat.com>
PR translation/118991
* config/avr/avr.cc (avr_print_operand): Print ival into
a temporary buffer and use %s in output_operand_lossage to make
the diagnostics translatable.
Filip Kastl [Sun, 2 Mar 2025 05:39:17 +0000 (06:39 +0100)]
gimple: sccopy: Prune removed statements from SCCs [PR117919]
While writing the sccopy pass I didn't realize that 'replace_uses_by ()' can
remove portions of the CFG. This happens when replacing arguments of some
statement results in the removal of an EH edge. Because of this sccopy can
then work with GIMPLE statements that aren't part of the IR anymore. In
PR117919 this triggered an assertion within the pass which assumes that
statements the pass works with are reachable.
This patch tells the pass to notice when a statement isn't in the IR anymore
and remove it from it's worklist.
PR tree-optimization/117919
gcc/ChangeLog:
* gimple-ssa-sccopy.cc (scc_copy_prop::propagate): Prune
statements that 'replace_uses_by ()' removed.
Jakub Jelinek [Sat, 1 Mar 2025 19:48:16 +0000 (20:48 +0100)]
ggc: Fix up ggc_internal_cleared_alloc_no_dtor [PR117047]
Apparently I got one of the !HAVE_ATTRIBUTE_ALIAS fallbacks wrong.
It compiled with a warning:
../../gcc/ggc-common.cc: In function 'void* ggc_internal_cleared_alloc_no_dtor(size_t, void (*)(void*), size_t, size_t)':
../../gcc/ggc-common.cc:154:44: warning: unused parameter 'size' [-Wunused-parameter]
154 | ggc_internal_cleared_alloc_no_dtor (size_t size, void (*f)(void *),
| ~~~~~~~^~~~
and obviously didn't work right (always allocated 0-sized objects).
Fixed thusly.
2025-03-01 Jakub Jelinek <jakub@redhat.com>
PR jit/117047
* ggc-common.cc (ggc_internal_cleared_alloc_no_dtor): Pass size
rather than s as the first argument to ggc_internal_cleared_alloc.
Yuriy Kolerov [Sat, 1 Mar 2025 15:35:55 +0000 (08:35 -0700)]
[PR target/118906] [PATCH v2] RISC-V: Fix a typo in zce to zcf implication
zce must imply zcf but this rule was corrupted after
refactoring in 9e12010b5e724277ea. This may be observed
ater generating an .s file from any source code file with
-mriscv-attribute -march=rv32if_zce -mabi=ilp32 -S
options. A full march will be presented in arch attribute:
As you see, zcf is not presented here though f_zce pair is
passed in -march. According to The RISC-V Instruction
Set Manual:
Specifying Zce on RV32 with F includes Zca, Zcb, Zcmp,
Zcmt and Zcf.
PR target/118906
gcc/ChangeLog:
* common/config/riscv/riscv-common.cc: fix zce to zcf
implication.
gcc/testsuite/ChangeLog:
* gcc.target/riscv/attribute-zce-1.c: New test.
* gcc.target/riscv/attribute-zce-2.c: New test.
* gcc.target/riscv/attribute-zce-3.c: New test.
* gcc.target/riscv/attribute-zce-4.c: New test.
Jan Dubiec [Sat, 1 Mar 2025 15:21:16 +0000 (08:21 -0700)]
[PATCH] H8/300, libgcc: PR target/114222 For HImode call internal ffs() implementation instead of an external one
When INT_TYPE_SIZE < BITS_PER_WORD gcc emits a call to an external ffs()
implementation instead of a call to "__builtin_ffs()" – see function
init_optabs() in <SRCROOT>/gcc/optabs-libfuncs.cc. External ffs()
(which is usually the one from newlib) in turn calls __builtin_ffs()
what causes infinite recursion and stack overflow. This patch overrides
default gcc bahaviour for H8/300H (and newer) and provides a generic
ffs() implementation for HImode.
PR target/114222
gcc/ChangeLog:
* config/h8300/h8300.cc (h8300_init_libfuncs): For HImode override
calls to external ffs() (from newlib) with calls to __ffshi2() from
libgcc. The implementation of ffs() in newlib calls __builtin_ffs()
what causes infinite recursion and finally a stack overflow.
libgcc/ChangeLog:
* config/h8300/t-h8300: Add __ffshi2().
* config/h8300/ffshi2.c: New file.
Jakub Jelinek [Sat, 1 Mar 2025 15:09:07 +0000 (16:09 +0100)]
input: Fix UB during self-tests [PR119052]
As the comment in check_line says:
/* get_buffer is not null terminated, but the sscanf stops after a number. */
the buffer is not null terminated, there is line.length () to determine
the size of the line. But unlike what the comment says, sscanf actually
still requires null terminated string argument, anything else is UB.
E.g. glibc when initializing the temporary FILE stream for the string does
if (size == 0)
end = strchr (ptr, '\0');
and this strchr/rawmemchr is what shows up in valgrind report on cc1/cc1plus
doing self-tests.
The function is used only in a test with 1000 lines, each containg its
number, so numbers from 1 to 1000 inclusive (each time with '\n' separator,
but that isn't included in line.length ()).
So the function just uses a temporary buffer which can fit numbers from 1 to
1000 as strings with terminating '\0' and runs sscanf on that (why not
strtoul?).
Furthermore, the caller allocated number of lines * 15 bytes for the
string, but 1000\n is 5 bytes, so I think * 5 is more than enough.
2025-03-01 Jakub Jelinek <jakub@redhat.com>
PR other/119052
* input.cc (check_line): Don't call sscanf on non-null terminated
buffer, instead copy line.length () bytes from line.get_buffer ()
to a local buffer, null terminate it and call sscanf on that.
Formatting fix.
(test_replacement): Just allocate maxline * 5 rather than maxline * 15
bytes for the file. Formatting fix.
Jakub Jelinek [Sat, 1 Mar 2025 10:22:27 +0000 (11:22 +0100)]
ggc: Avoid using ATTRIBUTE_MALLOC for allocations that need finalization [PR117047]
As analyzed by Andrew/David/Richi/Sam in the PR, the reason for the
libgccjit ICE is that there are GC allocations with finalizers and we
still mark ggc_internal_{,cleared_}alloc with ATTRIBUTE_MALLOC, which
to the optimizers hints that nothing will actually read the state
of the objects when they get out of lifetime. The finalizer actually
inspects those though. What actually happens in the testcases is that on
tree expr_size = TYPE_SIZE (expr->get_type ()->as_tree ());
we see that expr->get_type () was allocated using something with malloc
attribute but it doesn't escape and only the type size from it is queried,
so there is no need to store other members of it. Except that it does escape
in the GC internals. Normal GC allocations are fine, they don't look at the
data in the allocated objects on "free", but the ones with finalizers actually
call a function on that object and expect the data to be in there.
So that we don't lose ATTRIBUTE_MALLOC for the common case when no
finalization is needed, the following patch uses the approach used e.g.
for glibc error function which can sometimes be noreturn but at other
times just return normally.
If possible, it uses __attribute__((alias ("..."))) to add an alias
to the function, where one is without ATTRIBUTE_MALLOC and one
(with _no_dtor suffix) is with ATTRIBUTE_MALLOC (note, as this is
C++ and I didn't want to hardcode particular mangling I used an
extern "C" function with 2 aliases to it), and otherwise adds a wrapper
(for the ggc-page/ggc-common case with noinline attribute if possible,
for ggc-none that doesn't matter because ggc-none doesn't support
finalizers).
The *_no_dtor aliases/wrappers are then used in inline functions which
pass unconditional NULL, 0 as the f/s pair.
2025-03-01 Jakub Jelinek <jakub@redhat.com>
PR jit/117047
* acinclude.m4 (gcc_CHECK_ATTRIBUTE_ALIAS): New.
* configure.ac: Add gcc_CHECK_ATTRIBUTE_ALIAS.
* ggc.h (ggc_internal_alloc): Remove ATTRIBUTE_MALLOC from
overload with finalizer pointer. Call ggc_internal_alloc_no_dtor
in inline overload without finalizer pointer.
(ggc_internal_alloc_no_dtor): Declare.
(ggc_internal_cleared_alloc): Remove ATTRIBUTE_MALLOC from
overload with finalizer pointer. Call
ggc_internal_cleared_alloc_no_dtor in inline overload without
finalizer pointer.
(ggc_internal_cleared_alloc_no_dtor): Declare.
(ggc_alloc): Call ggc_internal_alloc_no_dtor if no finalization
is needed.
(ggc_alloc_no_dtor): Call ggc_internal_alloc_no_dtor.
(ggc_cleared_alloc): Call ggc_internal_cleared_alloc_no_dtor if no
finalization is needed.
(ggc_vec_alloc): Call ggc_internal_alloc_no_dtor if no finalization
is needed.
(ggc_cleared_vec_alloc): Call ggc_internal_cleared_alloc_no_dtor if no
finalization is needed.
* ggc-page.cc (ggc_internal_alloc): If HAVE_ATTRIBUTE_ALIAS, turn
overload with finalizer into alias to ggc_internal_alloc_ and
rename it to ...
(ggc_internal_alloc_): ... this, make it extern "C".
(ggc_internal_alloc_no_dtor): New alias if HAVE_ATTRIBUTE_ALIAS,
otherwise new noinline wrapper.
* ggc-common.cc (ggc_internal_cleared_alloc): If HAVE_ATTRIBUTE_ALIAS,
turn overload with finalizer into alias to ggc_internal_alloc_ and
rename it to ...
(ggc_internal_cleared_alloc_): ... this, make it extern "C".
(ggc_internal_cleared_alloc_no_dtor): New alias if
HAVE_ATTRIBUTE_ALIAS, otherwise new noinline wrapper.
* ggc-none.cc (ggc_internal_alloc): If HAVE_ATTRIBUTE_ALIAS, turn
overload with finalizer into alias to ggc_internal_alloc_ and
rename it to ...
(ggc_internal_alloc_): ... this, make it extern "C".
(ggc_internal_alloc_no_dtor): New alias if HAVE_ATTRIBUTE_ALIAS,
otherwise new wrapper.
(ggc_internal_cleared_alloc): If HAVE_ATTRIBUTE_ALIAS, turn overload
with finalizer into alias to ggc_internal_alloc_ and rename it to ...
(ggc_internal_cleared_alloc_): ... this, make it extern "C".
(ggc_internal_cleared_alloc_no_dtor): New alias if
HAVE_ATTRIBUTE_ALIAS, otherwise new wrapper.
* genmatch.cc (ggc_internal_cleared_alloc, ggc_free): Formatting fix.
(ggc_internal_cleared_alloc_no_dtor): Define.
* config.in: Regenerate.
* configure: Regenerate.
Jakub Jelinek [Sat, 1 Mar 2025 08:15:57 +0000 (09:15 +0100)]
openmp: Fix up simd clone mask argument creation on x86 [PR115871]
The following testcase ICEs since r14-5057.
The Intel vector ABI says that in the ZMM case the masks is passed
in unsigned int or unsigned long long arguments and how many bits in
them and how many of those arguments are is determined by the characteristic
data type of the function. In the testcase simdlen is 32 and characteristic
data type is double, so return as well as first argument is passed in 4
V8DFmode arguments and the mask is supposed to be passed in 4 unsigned int
arguments (8 bits in each).
Before the r14-5057 change there was
sc->args[i].orig_type = parm_type;
...
case SIMD_CLONE_ARG_TYPE_LINEAR_VAL_CONSTANT_STEP:
case SIMD_CLONE_ARG_TYPE_LINEAR_VAL_VARIABLE_STEP:
case SIMD_CLONE_ARG_TYPE_VECTOR:
if (INTEGRAL_TYPE_P (parm_type) || POINTER_TYPE_P (parm_type))
veclen = sc->vecsize_int;
else
veclen = sc->vecsize_float;
if (known_eq (veclen, 0U))
veclen = sc->simdlen;
else
veclen
= exact_div (veclen,
GET_MODE_BITSIZE (SCALAR_TYPE_MODE (parm_type)));
for the argument handling and
if (sc->inbranch)
{
tree base_type = simd_clone_compute_base_data_type (sc->origin, sc);
...
if (INTEGRAL_TYPE_P (base_type) || POINTER_TYPE_P (base_type))
veclen = sc->vecsize_int;
else
veclen = sc->vecsize_float;
if (known_eq (veclen, 0U))
veclen = sc->simdlen;
else
veclen = exact_div (veclen,
GET_MODE_BITSIZE (SCALAR_TYPE_MODE (base_type)));
for the mask handling. r14-5057 moved this argument creation later and
unified that:
case SIMD_CLONE_ARG_TYPE_MASK:
case SIMD_CLONE_ARG_TYPE_LINEAR_VAL_CONSTANT_STEP:
case SIMD_CLONE_ARG_TYPE_LINEAR_VAL_VARIABLE_STEP:
case SIMD_CLONE_ARG_TYPE_VECTOR:
if (sc->args[i].arg_type == SIMD_CLONE_ARG_TYPE_MASK
&& sc->mask_mode != VOIDmode)
elem_type = boolean_type_node;
else
elem_type = TREE_TYPE (sc->args[i].vector_type);
if (INTEGRAL_TYPE_P (elem_type) || POINTER_TYPE_P (elem_type))
veclen = sc->vecsize_int;
else
veclen = sc->vecsize_float;
if (known_eq (veclen, 0U))
veclen = sc->simdlen;
else
veclen
= exact_div (veclen,
GET_MODE_BITSIZE (SCALAR_TYPE_MODE (elem_type)));
This is correct for the argument cases (so linear or vector) (though
POINTER_TYPE_P will never appear as TREE_TYPE of a vector), but the
boolean_type_node in there is completely bogus, when using AVX512 integer
masks as I wrote above we need the characteristic data type, not bool,
and bool is strange in that it has bitsize of 8 (or 32 on darwin), while
the masks are 1 bit per lane anyway.
Fixed thusly.
2025-03-01 Jakub Jelinek <jakub@redhat.com>
PR middle-end/115871
* omp-simd-clone.cc (simd_clone_adjust): For SIMD_CLONE_ARG_TYPE_MASK
and sc->mask_mode not VOIDmode, set elem_type to the characteristic
type rather than boolean_type_node.
Jan Dubiec [Sat, 1 Mar 2025 05:01:42 +0000 (22:01 -0700)]
[PATCH] H8/300: PR target/109189 Silence -Wformat warnings on Windows
This patch fixes annoying -Wformat warnings when gcc is built
on Windows/MinGW64. Instead of %ld it uses HOST_WIDE_INT_PRINT_DEC
macro, just like many other targets do.
PR target/109189
gcc/ChangeLog:
* config/h8300/h8300.cc (h8300_print_operand): Replace %ld format
strings with HOST_WIDE_INT_PRINT_DEC macro in order to silence
-Wformat warnings when building on Windows/MinGW64.
Jonathan Wakely [Fri, 28 Feb 2025 21:44:41 +0000 (21:44 +0000)]
libstdc++: Fix ranges::iter_move handling of rvalues [PR106612]
The specification for std::ranges::iter_move apparently requires us to
handle types which do not satisfy std::indirectly_readable, for example
with overloaded operator* which behaves differently for different value
categories.
libstdc++-v3/ChangeLog:
PR libstdc++/106612
* include/bits/iterator_concepts.h (_IterMove::__iter_ref_t):
New alias template.
(_IterMove::__result): Use __iter_ref_t instead of
std::iter_reference_t.
(_IterMove::__type): Remove incorrect __dereferenceable
constraint.
(_IterMove::operator()): Likewise. Add correct constraints. Use
__iter_ref_t instead of std::iter_reference_t. Forward parameter
as correct value category.
(iter_swap): Add comments.
* testsuite/24_iterators/customization_points/iter_move.cc: Test
that iter_move is found by ADL and that rvalue arguments are
handled correctly.
Jonathan Wakely [Thu, 27 Feb 2025 13:27:17 +0000 (13:27 +0000)]
libstdc++: Fix ranges::move and ranges::move_backward to use iter_move [PR105609]
The ranges::move and ranges::move_backward algorithms are supposed to
use ranges::iter_move(iter) instead of std::move(*iter), which matters
for an iterator type with an iter_move overload findable by ADL.
Currently those algorithms use std::__assign_one which uses std::move,
so define a new ranges::__detail::__assign_one helper function that uses
ranges::iter_move.
libstdc++-v3/ChangeLog:
PR libstdc++/105609
* include/bits/ranges_algobase.h (__detail::__assign_one): New
helper function.
(__copy_or_move, __copy_or_move_backward): Use new function
instead of std::__assign_one.
* testsuite/25_algorithms/move/constrained.cc: Check that
ADL iter_move is used in preference to std::move.
* testsuite/25_algorithms/move_backward/constrained.cc:
Likewise.
Jonathan Wakely [Thu, 27 Feb 2025 15:48:49 +0000 (15:48 +0000)]
libstdc++: Add static_assertions to ranges::to adaptor factory [PR112803]
The standard requires that we reject attempts to create a ranges::to
adaptor for cv-qualified types and non-class types. Currently we only
diagnose it once the adaptor is used in a pipeline.
This adds static assertions to diagnose it immediately.
libstdc++-v3/ChangeLog:
PR libstdc++/112803
* include/std/ranges (ranges::to): Add static assertions to
enforce Mandates conditions.
* testsuite/std/ranges/conv/112803.cc: New test.
Iain Buclaw [Fri, 28 Feb 2025 18:22:36 +0000 (19:22 +0100)]
d: Fix comparing uninitialized memory in dstruct.d [PR116961]
Floating-point emulation in the D front-end is done via a type named
`struct longdouble`, which in GDC is a small interface around the
real_value type. Because the D code cannot include gcc/real.h directly,
a big enough buffer is used for the data instead.
On x86_64, this buffer is actually bigger than real_value itself, so
when a new longdouble object is created with
there is uninitialized padding at the end of `r`. This was never a
problem when D was implemented in C++ (until GCC 12) as comparing two
longdouble objects with `==' would be forwarded to the relevant
operator== overload that extracted the underlying real_value.
However when the front-end was translated to D, such conditions were
instead rewritten into identity comparisons
return exp.toReal() is CTFloat.zero
The `is` operator gets lowered as a call to `memcmp() == 0', which is
where the read of uninitialized memory occurs, as seen by valgrind.
==26778== Conditional jump or move depends on uninitialised value(s)
==26778== at 0x911F41: dmd.dstruct._isZeroInit(dmd.expression.Expression) (dstruct.d:635)
==26778== by 0x9123BE: StructDeclaration::finalizeSize() (dstruct.d:373)
==26778== by 0x86747C: dmd.aggregate.AggregateDeclaration.determineSize(ref const(dmd.location.Loc)) (aggregate.d:226)
[...]
To avoid accidentally reading uninitialized data, explicitly initialize
all `longdouble` variables with an empty constructor on C++ side of the
implementation before initializing underlying real_value type it holds.
The initial &<retval>._M_local_buf was not constant, but since
HelloWorld is a static VAR_DECL, the resulting &HelloWorld._M_local_buf
should have been marked as TREE_CONSTANT. And since we're taking
its address, the whole thing should be TREE_ADDRESSABLE.
PR c++/114913
PR c++/110822
gcc/cp/ChangeLog:
* constexpr.cc (replace_decl_r): If we've replaced something
inside of an ADDR_EXPR, call cxx_mark_addressable and
recompute_tree_invariant_for_addr_expr on the resulting ADDR_EXPR.
gcc/testsuite/ChangeLog:
* g++.dg/cpp0x/constexpr-nsdmi4.C: New test.
* g++.dg/cpp0x/constexpr-nsdmi5.C: New test.
so when we call maybe_constant_init, the object we're initializing is D.2701,
and the init is the expr_stmt. We unwrap the EXPR_STMT/INIT_EXPR/TARGET_EXPR
in maybe_constant_init_1 and so end up evaluating the f1 call. But f1 returns
c2 whereas the type of D.2701 is ._anon_0 -- the closure.
cxx_eval_outermost_constant_expr is already ready for the types to be
different, in which case the result isn't constant. But replace_decl
is called before that check.
I'm leaving the assert in replace_decl on purpose, maybe we'll find
another use for it.
PR c++/118986
gcc/cp/ChangeLog:
* constexpr.cc (cxx_eval_call_expression): Check that the types match
before calling replace_decl, if not, set *non_constant_p.
(maybe_constant_init_1): Don't strip INIT_EXPR if it would change the
type of the expression.
Martin Jambor [Fri, 28 Feb 2025 16:34:10 +0000 (17:34 +0100)]
ipa-sra: Avoid clashes with ipa-cp when pulling accesses across calls (PR 118243)
Among other things, IPA-SRA checks whether splitting out a bit of an
aggregate or something passed by reference would lead into a clash
with an already known IPA-CP constant a way which would cause problems
later on. Unfortunately the test is done only in
adjust_parameter_descriptions and is missing when accesses are
propagated from callees to callers, which leads to miscompilation
reported as PR 118243 (where the callee is a function created by
ipa-split).
The matter is then further complicated by the fact that we consider
complex numbers as scalars even though they can be modified piecemeal
(IPA-CP can detect and propagate the pieces separately too) which then
confuses the parameter manipulation machinery furter.
This patch simply adds the missing check to avoid the IPA-SRA
transform in these cases too, which should be suitable for backporting
to all affected release branches. It is a bit of a shame as in the PR
testcase we do propagate both components of the complex number in
question and the transformation phase could recover. I have some
prototype patches in this direction but that is something for (a)
stage 1.
gcc/ChangeLog:
2025-02-10 Martin Jambor <mjambor@suse.cz>
PR ipa/118243
* ipa-sra.cc (pull_accesses_from_callee): New parameters
caller_ipcp_ts and param_idx. Check that scalar pulled accesses would
not clash with a known IPA-CP aggregate constant.
(param_splitting_across_edge): Pass IPA-CP transformation summary and
caller parameter index to pull_accesses_from_callee.
When a generic lambda calls an overload set containing an iobj member
function we speculatively capture 'this'. We need to do the same
for an xobj member function.
PR c++/119038
gcc/cp/ChangeLog:
* lambda.cc (maybe_generic_this_capture): Consider xobj
member functions as well, not just iobj. Update function
comment.
It turns out the reason the behavior of this testcase changed after CWG
2369 is because validity of the substituted return type is now checked
later, after constraints. So a more reliable workaround for this issue
is to add a constraint to check the validity of the return type earlier,
matching the pre-CWG 2369 semantics.
PR libstdc++/104606
libstdc++-v3/ChangeLog:
* include/std/optional (operator<=>): Revert r14-9771 change.
Add constraint checking the validity of the return type
compare_three_way_result_t before the three_way_comparable_with
constraint.
Patrick Palka [Fri, 28 Feb 2025 14:39:57 +0000 (09:39 -0500)]
libstdc++: Fix constraint recursion in basic_const_iterator relops [PR112490]
Here for
using RCI = reverse_iterator<basic_const_iterator<vector<int>::iterator>>
static_assert(std::totally_ordered<RCI>);
we effectively need to check the requirement
requires (RCI x) { x RELOP x; } for each RELOP in {<, >, <=, >=}
which we expect to be straightforwardly satisfied by reverse_iterator's
namespace-scope relops. But due to ADL we find ourselves also
considering the basic_const_iterator relop friends, which before CWG
2369 would be quickly discarded since RCI clearly isn't convertible to
basic_const_iterator. After CWG 2369 though we must first check these
relops' constraints (with _It = vector<int>::iterator and _It2 = RCI),
which entails checking totally_ordered<RCI> recursively.
This patch fixes this by turning the problematic non-dependent function
parameters of type basic_const_iterator<_It> into dependent ones of
type basic_const_iterator<_It3> where _It3 is constrained to match _It.
Thus the basic_const_iterator relop friends now get quickly discarded
during deduction and before the constraint check if the second operand
isn't a specialization of basic_const_iterator (or derived from one)
like before CWG 2369.
PR libstdc++/112490
libstdc++-v3/ChangeLog:
* include/bits/stl_iterator.h (basic_const_iterator::operator<):
Replace non-dependent basic_const_iterator function parameter with
a dependent one of type basic_const_iterator<_It3> where _It3
matches _It.
(basic_const_iterator::operator>): Likewise.
(basic_const_iterator::operator<=): Likewise.
(basic_const_iterator::operator>=): Likewise.
* testsuite/24_iterators/const_iterator/112490.cc: New test.
I've added the asserts that probe == target because {REAL,IMAG}PART_EXPR
always implies a scalar type and so applying ARRAY_REF/COMPONENT_REF
etc. on it further doesn't make sense and the later code relies on it
to be the last one in refs array. But as the following testcase shows,
we can fail those assertions in case there is a reference or pointer
to the __real__ or __imag__ part, in that case we just evaluate the
constant expression and so probe won't be the same as target.
That case doesn't push anything into the refs array though.
The following patch changes those asserts to verify that refs is still
empty, which fixes it.
2025-02-28 Jakub Jelinek <jakub@redhat.com>
PR c++/119045
* constexpr.cc (cxx_eval_store_expression) <case REALPART_EXPR>:
Assert that refs->is_empty () rather than probe == target.
(cxx_eval_store_expression) <case IMAGPART_EXPR>: Likewise.
Jakub Jelinek [Fri, 28 Feb 2025 14:17:37 +0000 (15:17 +0100)]
c++: Adjust #embed support for P1967R14
Now that the #embed paper has been voted in, the following patch
removes the pedwarn for C++26 on it (and adjusts pedwarn warning for
older C++ versions) and predefines __cpp_pp_embed FTM.
Also, the patch changes cpp_error to cpp_pedwarning with for C++
-Wc++26-extensions guarding, and for C add -Wc11-c23-compat warning
about #embed.
I believe we otherwise implement everything in the paper already,
except I'm really confused by the
[Example:
— end example]
part. My reading of both C23 and C++ with the P1967R14 paper in
is that the first case (#embed with __has_include or __has_embed in its
clauses) is what is clearly invalid and so the ill-formed note should be
for #embed. And the __has_include/__has_embed in __has_embed is actually
questionable.
Both C and C++ have something like
"The identifiers __has_include, __has_embed, and __has_c_attribute
shall not appear in any context not mentioned in this subclause."
or
"The identifiers __has_include and __has_cpp_attribute shall not appear
in any context not mentioned in this subclause."
(into which P1967R14 adds __has_embed) in the conditional inclusion
subclause. #embed is defined in a different one, so using those in there
is invalid (unless "using the rules specified for conditional inclusion"
wording e.g. in limit clause overrides that).
The reason why I think it is fuzzy for __has_embed is that __has_embed
is actually defined in the Conditional inclusion subclause (so that
would mean one can use __has_include, __has_embed and __has_*attribute
in there) but its clauses are described in a different one.
GCC currently accepts
#embed __FILE__ limit (__has_include (<stdarg.h>))
#if __has_embed (__FILE__ limit (__has_include (<stdarg.h>)))
#endif
#embed __FILE__ limit (__has_embed (__FILE__))
#if __has_embed (__FILE__ limit (__has_embed (__FILE__)))
#endif
Note, it isn't just about limit clause, but also about
prefix/suffix/if_empty, except that in those cases the "using the rules
specified for conditional inclusion" doesn't apply.
In any case, I'd hope that can be dealt with incrementally (and should
be handled the same for both C and C++).
2025-02-28 Jakub Jelinek <jakub@redhat.com>
libcpp/
* include/cpplib.h (enum cpp_warning_reason): Add
CPP_W_CXX26_EXTENSIONS enumerator.
* init.cc (lang_defaults): Set embed for GNUCXX26 and CXX26.
* directives.cc (do_embed): Adjust pedwarn wording for embed in C++,
use cpp_pedwarning instead of cpp_error and add CPP_W_C11_C23_COMPAT
warning of cpp_pedwarning hasn't diagnosed anything.
gcc/c-family/
* c.opt (Wc++26-extensions): Add CppReason(CPP_W_CXX26_EXTENSIONS).
* c-cppbuiltin.cc (c_cpp_builtins): Predefine __cpp_pp_embed=202502
for C++26.
gcc/testsuite/
* g++.dg/cpp/embed-1.C: Adjust for pedwarn wording change and don't
expect any error for C++26.
* g++.dg/cpp/embed-2.C: Adjust for pedwarn wording change and don't
expect any warning for C++26.
* g++.dg/cpp26/feat-cxx26.C: Test __cpp_pp_embed value.
* gcc.dg/cpp/embed-17.c: New test.
Richard Biener [Fri, 28 Feb 2025 13:09:29 +0000 (14:09 +0100)]
lto/91299 - weak definition inlined with LTO
The following fixes a thinko in the handling of interposed weak
definitions which confused the interposition check in
get_availability by setting DECL_EXTERNAL too early.
PR lto/91299
gcc/lto/
* lto-symtab.cc (lto_symtab_merge_symbols): Set DECL_EXTERNAL
only after calling get_availability.
gcc/testsuite/
* gcc.dg/lto/pr91299_0.c: New testcase.
* gcc.dg/lto/pr91299_1.c: Likewise.
Richard Biener [Fri, 28 Feb 2025 10:44:26 +0000 (11:44 +0100)]
ipa/111245 - bogus modref analysis for store in call that might throw
We currently record a kill for
*x_4(D) = always_throws ();
because we consider the store always executing since the appropriate
check for whether the stmt could throw is guarded by
!cfun->can_throw_non_call_exceptions.
PR ipa/111245
* ipa-modref.cc (modref_access_analysis::analyze_store): Do
not guard the check of whether the stmt could throw by
cfun->can_throw_non_call_exceptions.
Jakub Jelinek [Fri, 28 Feb 2025 11:42:27 +0000 (12:42 +0100)]
ifcvt: Fix ICE with (fix:SI (fix:DF (reg:DF))) [PR117712]
As documented in the manual, FIX/UNSIGNED_FIX from floating point
mode to integral mode has unspecified rounding and FIX from floating point
mode to the same floating point mode is expressing rounding toward zero.
So, some targets (arc, arm, csky, m68k, mmix, nds32, pdp11, sparc and
visium) use
(fix:SI (fix:SF (match_operand:SF 1 "..._operand")))
etc. to express the rounding toward zero during conversion to integer.
For some reason other targets don't use that.
Anyway, the 2 FIXes (or inner FIX with outer UNSIGNED_FIX) cause problems
since the r15-2890 which removed some strict checks in ifcvt.cc on what
SET_SRC can be actually conditionalized (I must say I'm still worried
about the change, don't know why one can't get e.g. inline asm or
something with UNSPEC or some complex backend specific RTLs that
force_operand can't handle), force_operand just ICEs on it, it can only
handle (through expand_fix) conversions from floating point to integral.
The following patch fixes this by detecting this case and just pretend
the inner FIX isn't there, i.e. call expand_fix with the inner FIX's
operand instead, which works and on targets like arm it will just
create the nested FIXes again.
2025-02-28 Jakub Jelinek <jakub@redhat.com>
PR rtl-optimization/117712
* expr.cc (force_operand): Handle {,UNSIGNED_}FIX with
FIX operand using expand_fix on the inner FIX operand.
Richard Biener [Fri, 28 Feb 2025 09:36:11 +0000 (10:36 +0100)]
tree-optimization/87984 - hard register assignments not preserved
The following disables redundant store elimination to hard register
variables which isn't valid.
PR tree-optimization/87984
* tree-ssa-dom.cc (dom_opt_dom_walker::optimize_stmt): Do
not perform redundant store elimination to hard register
variables.
* tree-ssa-sccvn.cc (eliminate_dom_walker::eliminate_stmt):
Likewise.
When the C++ frontend clones a CTOR we do not copy ASM_EXPR constraints
fully as walk_tree does not recurse to TREE_PURPOSE of TREE_LIST nodes.
At this point doing that seems too dangerous so the following instead
avoids gimplification of ASM_EXPRs to clobber the shared constraints
and unshares it there, like it also unshares TREE_VALUE when it
re-writes a "+" output constraint to separate "=" output and matching
input constraint.
PR middle-end/66279
* gimplify.cc (gimplify_asm_expr): Copy TREE_PURPOSE before
rewriting it for "+" processing.
Jakub Jelinek [Fri, 28 Feb 2025 09:13:57 +0000 (10:13 +0100)]
testsuite: Remove -m32 from another i386/ test
I found another test which uses -m32 in gcc.target/i386/ . Similarly
to the previously posted test, the test ought to be tested during i686-linux
testing or x86_64-linux test with --target_board=unix\{-m32,-m64\}
There is nothing ia32 specific on the test, so I've just dropped the -m32.
2025-02-28 Jakub Jelinek <jakub@redhat.com>
* gcc.target/i386/strub-pr118006.c: Remove -m32 from dg-options.
Jakub Jelinek [Fri, 28 Feb 2025 09:12:14 +0000 (10:12 +0100)]
testsuite: Fix up gcc.target/i386/pr118940.c test [PR118940]
The testcase uses -m32 in dg-options, something we try hard not to do,
if something should be tested only for -m32, it is { target ia32 } test,
if it can be tested for -m64/-mx32 too, just some extra options are
needed for ia32, it should have dg-additional-options with ia32 target.
Also, the test wasn't reduced, so I've reduced it using cvise and manual
tweaks and verified the test still FAILs before r15-7700 and succeeds
with current trunk.
2025-02-28 Jakub Jelinek <jakub@redhat.com>
PR target/118940
* gcc.target/i386/pr118940.c: Drop -w, -g and -m32 from dg-options, move
-march=i386 -mregparm=3 to dg-additional-options for ia32 and -fno-pie
to dg-additional-options for pie. Reduce the test.
Pan Li [Sat, 22 Feb 2025 11:34:52 +0000 (19:34 +0800)]
RISC-V: Fix bug for expand_const_vector interleave [PR118931]
This patch would like to fix one bug when expanding const vector for the
interleave case. For example, we have:
base1 = 151
step = 121
For vec_series, we will generate vector in format of v[i] = base + i * step.
Then the vec_series will have below result for HImode, and we can find
that the result overflow to the highest 8 bits of HImode.
Unfortunately, the base1 + i * step1 in HImode may overflow to the high
8 bits, and the high 8 bits will pollute the v2 and result in incorrect
value in const_vector.
This patch would like to perform the overflow to smode check before the
optimized interleave code generation. If overflow or VLA, it will fall
back to the default merge approach.
The below test suites are passed for this patch.
* The rv64gcv fully regression test.
PR target/118931
gcc/ChangeLog:
* config/riscv/riscv-v.cc (expand_const_vector): Add overflow to
smode check and clean up highest bits if overflow.
gcc/testsuite/ChangeLog:
* gcc.target/riscv/rvv/base/pr118931-run-1.c: New test.
Iain Buclaw [Thu, 27 Feb 2025 22:37:21 +0000 (23:37 +0100)]
libphobos: Run unittest tests with dg-runtest.
Use `dg-runtest' test driver rather than `dg-test' to run the libphobos
unittest testsuite, same as all other libphobos tests. This prevents
the tests from being ran multiple times when parallelized.
Set `libphobos_test_name' as well so that all tests get a unique name.
libphobos/ChangeLog:
* testsuite/libphobos.unittest/unittest.exp: Use `dg-runtest' rather
than `dg-test'. Set `libphobos_test_name'.
Jakub Jelinek [Thu, 27 Feb 2025 21:11:51 +0000 (22:11 +0100)]
gimple-fold: Fix a pasto in fold_truth_andor_for_ifcombine [PR119030]
The following testcase is miscompiled since r15-7597.
The left comparison is unsigned (x & 0x8000U) != 0) while the
right one is signed (x >> 16) >= 0 and is actually a signbit test,
so rsignbit is 64.
After debugging this and reading the r15-7597 change, I believe there
is just a pasto, the if (lsignbit) and if (rsignbit) blocks are pretty
much identical with just the first l on all variables starting with l
replaced with r (the only difference is that if (lsignbit) has a comment
explaining the sign <<= 1; stuff, while it isn't repeated in the second one.
Except the second one was using ll_unsignedp instead of rl_unsignedp
in one spot. I think it should use the latter, the signedness of the left
comparison doesn't affect the other one, they are basically independent
with the exception that we check that after transformations they are both
EQ or both NE and later on we try to merge them together.
Jakub Jelinek [Thu, 27 Feb 2025 21:10:46 +0000 (22:10 +0100)]
input: Fix up ICEs with --param=file-cache-files=N for N > 16 [PR118860]
The following testcase ICEs, because we first construct file_cache object
inside of *global_dc, then process options and then call file_cache::tune.
The earlier construction allocates the m_file_slots array (using new)
based on the static data member file_cache::num_file_slots, but then tune
changes it, without actually reallocating all m_file_slots arrays in already
constructed file_cache objects.
I think it is just weird to have the count be a static data member and
the pointer be non-static data member, that is just asking for issues like
this.
So, this patch changes num_file_slots into m_num_file_slots and turns tune
into a non-static member function and changes toplev.cc to call it on the
global_gc->get_file_cache () object. And let's the tune just delete the
array and allocate it freshly if there is a change in the number of slots
or lines.
Note, file_cache_slot has similar problem, but because there are many, I
haven't moved the count into those objects; I just hope that when tune
is called there is exactly one file_cache constructed and all the
file_cache_slot objects constructed are pointed by its m_file_slots member,
so also on lines change it just deletes it and allocates again. I think
it should be unlikely that the cache actually has any used slots by the time
it is called.
2025-02-27 Jakub Jelinek <jakub@redhat.com>
PR middle-end/118860
* input.h (file_cache::tune): No longer static. Rename argument
from num_file_slots_ to num_file_slots. Formatting fix.
(file_cache::num_file_slots): Renamed to ...
(file_cache::m_num_file_slots): ... this. No longer static.
* input.cc (file_cache_slot::tune): Change return type from void to
size_t, return previous file_cache_slot::line_record_size value.
Formatting fixes.
(file_cache::tune): Rename argument from num_file_slots_ to
num_file_slots. Set m_num_file_slots rather than num_file_slots.
If m_num_file_slots or file_cache_slot::line_record_size changes,
delete[] m_file_slots and new it again.
(file_cache::num_file_slots): Remove definition.
(file_cache::lookup_file): Use m_num_file_slots rather than
num_file_slots.
(file_cache::evicted_cache_tab_entry): Likewise.
(file_cache::file_cache): Likewise. Initialize m_num_file_slots
to 16.
(file_cache::dump): Use m_num_file_slots rather than num_file_slots.
(file_cache_slot::get_next_line): Formatting fixes.
(file_cache_slot::read_line_num): Likewise.
(get_source_text_between): Likewise.
* toplev.cc (toplev::main): Call global_dc->get_file_cache ().tune
rather than file_cache::tune.
Thomas Schwinge [Wed, 26 Feb 2025 14:39:37 +0000 (15:39 +0100)]
nvptx: '#define MAX_FIXED_MODE_SIZE 128'
... instead of 64 via 'gcc/defaults.h':
MAX_FIXED_MODE_SIZE GET_MODE_BITSIZE (DImode)
This fixes ICEs:
[-FAIL: c-c++-common/pr111309-1.c -Wc++-compat (internal compiler error: in expand_fn_using_insn, at internal-fn.cc:268)-]
[-FAIL:-]{+PASS:+} c-c++-common/pr111309-1.c -Wc++-compat (test for excess errors)
[-UNRESOLVED:-]{+PASS:+} c-c++-common/pr111309-1.c -Wc++-compat [-compilation failed to produce executable-]{+execution test+}
[-FAIL: c-c++-common/pr111309-1.c -std=gnu++17 (internal compiler error: in expand_fn_using_insn, at internal-fn.cc:268)-]
[-FAIL:-]{+PASS:+} c-c++-common/pr111309-1.c -std=gnu++17 (test for excess errors)
[-UNRESOLVED:-]{+PASS:+} c-c++-common/pr111309-1.c -std=gnu++17 [-compilation failed to produce executable-]{+execution test+}
[-FAIL: c-c++-common/pr111309-1.c -std=gnu++26 (internal compiler error: in expand_fn_using_insn, at internal-fn.cc:268)-]
[-FAIL:-]{+PASS:+} c-c++-common/pr111309-1.c -std=gnu++26 (test for excess errors)
[-UNRESOLVED:-]{+PASS:+} c-c++-common/pr111309-1.c -std=gnu++26 [-compilation failed to produce executable-]{+execution test+}
[-FAIL: c-c++-common/pr111309-1.c -std=gnu++98 (internal compiler error: in expand_fn_using_insn, at internal-fn.cc:268)-]
[-FAIL:-]{+PASS:+} c-c++-common/pr111309-1.c -std=gnu++98 (test for excess errors)
[-UNRESOLVED:-]{+PASS:+} c-c++-common/pr111309-1.c -std=gnu++98 [-compilation failed to produce executable-]{+execution test+}
[-FAIL: gcc.dg/torture/pr116480-1.c -O0 (internal compiler error: in expand_fn_using_insn, at internal-fn.cc:268)-]
[-FAIL:-]{+PASS:+} gcc.dg/torture/pr116480-1.c -O0 (test for excess errors)
[-FAIL: gcc.dg/torture/pr116480-1.c -O1 (internal compiler error: in expand_fn_using_insn, at internal-fn.cc:268)-]
[-FAIL:-]{+PASS:+} gcc.dg/torture/pr116480-1.c -O1 (test for excess errors)
PASS: gcc.dg/torture/pr116480-1.c -O2 (test for excess errors)
PASS: gcc.dg/torture/pr116480-1.c -O3 -g (test for excess errors)
PASS: gcc.dg/torture/pr116480-1.c -Os (test for excess errors)
..., where we ran into 'gcc_assert (icode != CODE_FOR_nothing);' in
'gcc/internal-fn.cc:expand_fn_using_insn' for '__int128' '__builtin_clzg' etc.:
during RTL pass: expand
[...]/c-c++-common/pr111309-1.c: In function 'clzI':
[...]/c-c++-common/pr111309-1.c:69:10: internal compiler error: in expand_fn_using_insn, at internal-fn.cc:268
0x120ec2cf internal_error(char const*, ...)
[...]/gcc/diagnostic-global-context.cc:517
0x102c7c5b fancy_abort(char const*, int, char const*)
[...]/gcc/diagnostic.cc:1722
0x109708eb expand_fn_using_insn
[...]/gcc/internal-fn.cc:268
0x1098114f expand_internal_call(internal_fn, gcall*)
[...]/gcc/internal-fn.cc:5273
0x1098114f expand_internal_call(gcall*)
[...]/gcc/internal-fn.cc:5281
0x10594fc7 expand_call_stmt
[...]/gcc/cfgexpand.cc:3049
[...]
Likewise, as of commit e8ad697a75b0870a833366daf687668a57cabb6e
"libstdc++: Use new type-generic built-ins in <bit> [PR118855]",
the libstdc++ target library build ICEd in the same way.
Additionally, this change fixes:
[-FAIL:-]{+PASS:+} gcc.dg/pr105094.c (test for excess errors)
..., which was:
[...]/gcc.dg/pr105094.c: In function 'foo':
[...]/gcc.dg/pr105094.c:11:12: error: size of variable 's' is too large
And, finally, regarding 'gcc.target/nvptx/stack_frame-1.c'. Before, in
'gcc/cfgexpand.cc': 'expand_used_vars' -> 'expand_used_vars_for_block' ->
'expand_one_var' for 'ww' -> 'gcc/function.cc:use_register_for_decl' due to
'DECL_MODE (decl) == BLKmode' did 'return false;', thus -> 'add_stack_var'
(even if 'ww' wasn't then actually living on the stack). Now, 'ww' has
'TImode' and 'use_register_for_decl' does 'return true;', thus ->
'expand_one_register_var', and therefore no unused stack frame emitted.
Thomas Schwinge [Tue, 25 Feb 2025 21:31:25 +0000 (22:31 +0100)]
nvptx: Build libgfortran with '-mfake-ptx-alloca' [PR107635]
As of recent commit 8bf0ee8d62b8a08e808344d31354ab713157e15d
"Fortran: Add transfer_between_remotes [PR107635]", we've got 'alloca' usage
in 'libgfortran/caf/single.c:_gfortran_caf_transfer_between_remotes', and
the libgfortran target library fails to build for legacy configurations where
PTX 'alloca' is not available:
With '-mfake-ptx-alloca', libgfortran again succeeds to build, and compared
to before, we've got only a small number of regressions due to nvptx 'ld'
complaining about 'unresolved symbol __GCC_nvptx__PTX_alloca_not_supported':
[-PASS:-]{+FAIL:+} gfortran.dg/coarray/codimension_2.f90 -fcoarray=lib -O2 -lcaf_single (test for excess errors)
[-PASS:-]{+FAIL:+} gfortran.dg/coarray/event_4.f08 -fcoarray=lib -O2 -lcaf_single (test for excess errors)
[-PASS:-]{+UNRESOLVED:+} gfortran.dg/coarray/event_4.f08 -fcoarray=lib -O2 -lcaf_single [-execution test-]{+compilation failed to produce executable+}
[-PASS:-]{+FAIL:+} gfortran.dg/coarray/fail_image_2.f08 -fcoarray=lib -O2 -lcaf_single (test for excess errors)
[-PASS:-]{+UNRESOLVED:+} gfortran.dg/coarray/fail_image_2.f08 -fcoarray=lib -O2 -lcaf_single [-execution test-]{+compilation failed to produce executable+}
[-PASS:-]{+FAIL:+} gfortran.dg/coarray/proc_pointer_assign_1.f90 -fcoarray=lib -O2 -lcaf_single (test for excess errors)
[-PASS:-]{+UNRESOLVED:+} gfortran.dg/coarray/proc_pointer_assign_1.f90 -fcoarray=lib -O2 -lcaf_single [-execution test-]{+compilation failed to produce executable+}
[-PASS:-]{+FAIL:+} gfortran.dg/coarray_43.f90 -O (test for excess errors)
That's acceptable for such legacy PTX configurations.
Thomas Schwinge [Mon, 24 Feb 2025 15:13:11 +0000 (16:13 +0100)]
nvptx: Support '-mfake-ptx-alloca'
With '-mfake-ptx-alloca' enabled, the user-visible behavior changes only
for configurations where PTX 'alloca' is not available. Rather than a
compile-time 'sorry, unimplemented: dynamic stack allocation not supported'
in presence of dynamic stack allocation, compilation and assembly then
succeeds. However, attempting to link in such '*.o' files then fails due
to unresolved symbol '__GCC_nvptx__PTX_alloca_not_supported'.
This is meant to be used in scenarios where large volumes of code are
compiled, a small fraction of which runs into dynamic stack allocation, but
these parts are not important for specific use cases, and we'd thus like the
build to succeed, and error out just upon actual, very rare use of the
offending '*.o' files.
Thomas Schwinge [Mon, 24 Feb 2025 09:09:11 +0000 (10:09 +0100)]
nvptx: Delay 'sorry, unimplemented: dynamic stack allocation not supported' from expansion time to code generation
This gives the back end a chance to clean out a few more unnecessary instances
of dynamic stack allocation. This progresses:
PASS: gcc.dg/pr78902.c (test for warnings, line 7)
PASS: gcc.dg/pr78902.c (test for warnings, line 8)
PASS: gcc.dg/pr78902.c (test for warnings, line 9)
PASS: gcc.dg/pr78902.c (test for warnings, line 10)
PASS: gcc.dg/pr78902.c (test for warnings, line 11)
PASS: gcc.dg/pr78902.c (test for warnings, line 12)
PASS: gcc.dg/pr78902.c (test for warnings, line 13)
PASS: gcc.dg/pr78902.c strndup excessive bound at line 14 (test for warnings, line 13)
[-UNSUPPORTED: gcc.dg/pr78902.c: dynamic stack allocation not supported-]
{+PASS: gcc.dg/pr78902.c (test for excess errors)+}
UNSUPPORTED: gcc.dg/torture/pr71901.c -O0 : dynamic stack allocation not supported
[-UNSUPPORTED:-]{+PASS:+} gcc.dg/torture/pr71901.c -O1 [-: dynamic stack allocation not supported-]{+(test for excess errors)+}
UNSUPPORTED: gcc.dg/torture/pr71901.c -O2 : dynamic stack allocation not supported
UNSUPPORTED: gcc.dg/torture/pr71901.c -O3 -fomit-frame-pointer -funroll-loops -fpeel-loops -ftracer -finline-functions : dynamic stack allocation not supported
UNSUPPORTED: gcc.dg/torture/pr71901.c -O3 -g : dynamic stack allocation not supported
[-UNSUPPORTED:-]{+PASS:+} gcc.dg/torture/pr71901.c -Os [-: dynamic stack allocation not supported-]{+(test for excess errors)+}
UNSUPPORTED: gcc.dg/torture/pr78742.c -O0 : dynamic stack allocation not supported
[-UNSUPPORTED:-]{+PASS:+} gcc.dg/torture/pr78742.c -O1 [-: dynamic stack allocation not supported-]{+(test for excess errors)+}
[-UNSUPPORTED:-]{+PASS:+} gcc.dg/torture/pr78742.c -O2 [-: dynamic stack allocation not supported-]{+(test for excess errors)+}
[-UNSUPPORTED:-]{+PASS:+} gcc.dg/torture/pr78742.c -O3 -fomit-frame-pointer -funroll-loops -fpeel-loops -ftracer -finline-functions [-: dynamic stack allocation not supported-]{+(test for excess errors)+}
[-UNSUPPORTED:-]{+PASS:+} gcc.dg/torture/pr78742.c -O3 -g [-: dynamic stack allocation not supported-]{+(test for excess errors)+}
UNSUPPORTED: gcc.dg/torture/pr78742.c -Os : dynamic stack allocation not supported
[-UNSUPPORTED:-]{+PASS:+} gfortran.dg/pr101267.f90 -O [-: dynamic stack allocation not supported-]{+(test for excess errors)+}
[-UNSUPPORTED:-]{+PASS:+} gfortran.dg/pr112404.f90 -O [-: dynamic stack allocation not supported-]{+(test for excess errors)+}
Jerry DeLisle [Thu, 27 Feb 2025 01:26:26 +0000 (17:26 -0800)]
GCC: Documentation of -x option
This change updates information about the -x option to clarify
that it does not ensure standards compliance. Sparked by
discussions in the following PR.
PR fortran/108369
gcc/ChangeLog:
* doc/invoke.texi: Add a note to clarify. Adjust some wording.
Marek Polacek [Wed, 19 Feb 2025 19:06:33 +0000 (14:06 -0500)]
c++: ICE with GOTO_EXPR [PR118928]
In this PR we crash in cxx_eval_constant_expression/GOTO_EXPR on:
gcc_assert (cxx_dialect >= cxx23);
The code obviously doesn't expect to see a goto pre-C++23. But we can
get here with the new prvalue optimization. In this test we found
ourselves in synthesize_method for X::X(). This function calls:
a) finish_function, which does cp_genericize -> ... -> genericize_c_loops,
which creates the GOTO_EXPR;
b) expand_or_defer_fn -> maybe_clone_body -> ... -> cp_fold_function
where we reach the new maybe_constant_init call and crash on the
goto.
Since we can validly get to that assert, I think we should just remove
it. I don't see other similar asserts like this one.
PR c++/118928
gcc/cp/ChangeLog:
* constexpr.cc (cxx_eval_constant_expression) <case GOTO_EXPR>: Remove
an assert.
Marek Polacek [Thu, 20 Feb 2025 19:35:25 +0000 (14:35 -0500)]
c++: too many errors with sneaky template [PR118516]
Since C++20 P0846, a name followed by a < can be treated as a template-name
even though name lookup did not find a template-name. That happens
in this test with "i < foo ()":
for (int id = 0; i < foo(); ++id);
and results in a raft of errors about non-constant foo(). The problem
is that the require_potential_constant_expression call in
cp_parser_template_argument emits errors even when we're parsing
tentatively. So we repeat the error when we're trying to parse
as a nested-name-specifier, type-name, etc.
Guarding the call with !cp_parser_uncommitted_to_tentative_parse_p would
mean that require_potential_constant_expression never gets called. But
we don't need the call at all as far as I can tell. Stuff like
template<int N> struct S { };
int foo () { return 4; }
void
g ()
{
S<foo()> s;
}
gets diagnosed in convert_nontype_argument. In fact, with this patch,
we only emit "call to non-constexpr function" once. (That is, in C++17
only; C++14 uses a different path.)
Richard Earnshaw [Thu, 27 Feb 2025 15:11:47 +0000 (15:11 +0000)]
testsuite: arm: Avoid incremental link warnings in pr61123-enum-size
This test uses incremental linking, but that can generate warnings if
the LTO step contains a mix of LTO and non-LTO object files (this can
happen when there's a testglue file that is normally included during
linking).
We don't care about the testglue, though, so just tell the LTO
optimizer to generate nolto-rel output, which is what it is falling
back to anyway.
gcc/testsuite:
* gcc.target/arm/lto/pr61123-enum-size_0.c: (dg-lto-options) Move
linker related options to ...
(dg-extra-ld-options): ... here. Add -flinker-output=nolto-rel.
projects / thirdparty / gcc.git / log
