Hi,
This fixes an old issue whereby we generate wrong code in Ada for an indexed
component in an array with a ludicrously large index type instead of raising
Storage_Error. We would probably need the counterpart of int_const_binop for
unop in the general case, but that's not worth the hassle s
Currently, the shuffle in which LoongArch selects two vectors at
corresponding positions is implemented through the [x]vshuf instruction,
but this will introduce additional index copies. In this case, the
[x]vbitsel.v instruction can be used for optimization.
gcc/ChangeLog:
* config/loong
In LoongArch, when the permutation idx comes from different vectors and
idx is not repeated, for V8SI/V8SF/V4DI/V4DF type vectors, we can use
two xvperm.w + one xvbitsel.v instructions or two xvpermi.d + one
xvbitsel.v instructions for shuffle optimization.
gcc/ChangeLog:
* config/loongar
In LoongArch, we have xvshuf.{b/h/w/d} instructions which can dealt the
situation that all low 128-bit elements of the target vector are shuffled
by concatenating the low 128-bit elements of the two input vectors, and
all high 128-bit elements of the target vector are similarly shuffled.
Therefore,
r16-3760-g9ff5cadac4579f generates more condition move for
-march=cascadelake, and causes new failures as below
FAIL: gcc.target/i386/pr116896.c scan-assembler-times \tjp\t 2
FAIL: g++.target/i386/pr116896-1.C -std=gnu++20 scan-assembler-times \tjp\t 1
FAIL: g++.target/i386/pr116896-1.C -std=gn
Hi,
I can reproduce this on x86_64. The issue is that bb-slp-41.c expects BB
vevtorization to not happen. Look like the original test case had other checks
too that has since gone.
I have added -fno-tree-loop-im and tested to keep the behaviour the same. Is
this OK?
Thanks,
Kugan
> On 23 No
changes v2: more tests, and the one-argument version of
__iter_concept_or_category is now constexpr instead of
consteval
-- >8 --
>From the paper's abstract:
Change the iterator requirements for non-Ranges algorithms. For
forward iterators and above that are constant iterators, instead of
> > It can be just put within BFmode should be fine, I mean
> > load/store/move/insert/extract could just use normal vector
> > instruction with SEW=16,
> > Only tha casting from/to other data types need real BF16 instruction.
> > ('real' might not be the right term to describe, but I guess you
> >
Add missing dg-require-effective-target
---
gcc/testsuite/c-c++-common/Wdangling-pointer-9.c | 1 +
gcc/testsuite/c-c++-common/analyzer/computed-goto-1.c | 1 +
gcc/testsuite/c-c++-common/analyzer/infinite-recursion-5.c | 1 +
gcc/testsuite/c-c++-common/analyzer/pr102695.c
Make floating point test compile on platforms without FE_UPWARD
---
gcc/testsuite/gcc.dg/torture/fp-double-convert-float-1.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/gcc/testsuite/gcc.dg/torture/fp-double-convert-float-1.c
b/gcc/testsuite/gcc.dg/torture/fp-double-conv
Add scan-weak function to testsuite, which would work like scan-hidden,
and allow for format-specific regexes
---
gcc/testsuite/gcc.dg/weak/weak-1.c | 16 +-
gcc/testsuite/gcc.dg/weak/weak-10.c | 2 +-
gcc/testsuite/gcc.dg/weak/weak-11.c | 2 +-
gcc/testsuite/gcc.dg/weak/weak-12.c | 2
Move tests for signaling floats to ieee directory
---
.../ieee/bfloat16-builtin-issignaling-1.c | 11 ++
.../ieee/bfloat16-builtin-issignaling-1.x | 11 ++
.../execute/ieee}/builtin-issignaling-1.c | 6 --
.../execute/ieee/builtin-issignaling-1.x | 20
While hacking on GCC, I found some tests that don't properly set their
requirements, or just plain don't compile on some systems. These patches
are intended to address that, as well as do some cleanups.
feedable (4):
testsuite: Clean up the weak declaration scanning
testsuite: Move float tests
From: Pan Li
Add asm dump check and run test for vec_duplicate + vmsltu.vv
combine to vmsltu.vx, with the GR2VR cost is 0, 2 and 15.
gcc/testsuite/ChangeLog:
* gcc.target/riscv/rvv/autovec/vx_vf/vx-1-u16.c: Add asm check
for vmsltu.vx.
* gcc.target/riscv/rvv/autovec/vx_v
From: Pan Li
This patch would like to introduce the combine of vec_dup + vmsltu.wv
into vmsltu.wx on the cost value of GR2VR. The late-combine will
take place if the cost of GR2VRlike 1, 2, 15 in test.
From:
| ...
| vmv.v.x
| L1:
| vmsltu.vv
| J L1
| ...
To:
| ...
| L1:
|
From: Pan Li
This patch would like to combine the vec_duplicate + vmsltu.wv to the
vmsltu.vx. From example as below code. The related pattern will depend
on the cost of vec_duplicate from GR2VR. Then the late-combine will
take action if the cost of GR2VR is zero, and reject the combination
if
在 2025/11/22 下午10:37, Xi Ruoyao 写道:
On Sat, 2025-11-22 at 22:11 +0800, Xi Ruoyao wrote:
On Mon, 2025-11-17 at 14:38 +0800, zhaozhou wrote:
Add new template muldi3_ui12 for mul(reg, ui12), which mode is
DImode
and op2 is const_uns_arith_operand. And the template is matched when
the mul operati
在 2025/11/22 下午10:37, Xi Ruoyao 写道:
On Sat, 2025-11-22 at 22:11 +0800, Xi Ruoyao wrote:
On Mon, 2025-11-17 at 14:38 +0800, zhaozhou wrote:
Add new template muldi3_ui12 for mul(reg, ui12), which mode is
DImode
and op2 is const_uns_arith_operand. And the template is matched when
the mul operati
On Thu, Nov 20, 2025 at 9:12 AM FX Coudert wrote:
>
> Provide a sensible default value for native_system_header_dir, namely
> /mingw/include, on aarch64-mingw. This is in line with the expectations
> for mingw file locations, and is already set on both x86- and
> x86_64-mingw.
>
> gcc/ChangeLog:
>
On Sat, Nov 22, 2025 at 5:47 PM Richard Biener wrote:
>
>
>
> > Am 22.11.2025 um 00:29 schrieb H.J. Lu :
> >
> > On Sat, Nov 22, 2025 at 5:33 AM Florian Weimer wrote:
> >>
> >> * Richard Biener:
> >>
> >>> +/* When the glibc we build against does not have dl_find_object
> >>> tentatively
> >>>
From: Lúcio Boari Fleury
gcc/testsuite/ChangeLog:
* rust/compile/macros/mbe/macro-issue3608.rs: New Test. The test skips
an issue at line 11
Signed-off-by: Lúcio Boari Fleury
---
This change was merged into the gccrs repository and is posted here for
upstream visibility and potential
From: Lúcio Boari Fleury
gcc/rust/ChangeLog:
* parse/rust-parse-impl.h: Add early exit condition to parsing loop.
Signed-off-by: Lúcio Boari Fleury
---
This change was merged into the gccrs repository and is posted here for
upstream visibility and potential drive-by review, as requeste
Hi!
The testcase in the PR is miscompiled on aarch64 with
--param=ggc-min-expand=0 --param=ggc-min-heapsize=0 -O2
(not including it in the testsuite because it is too much of
a lottery).
Anyway, the problem is that the testcase only uses unsigned _BitInt(66)
and never uses _BitInt(66), get_alias_
Small step toward getting 64-bit glibc working on hppa.
Dave
---
hppa: Define GLIBC_DYNAMIC_LINKER in pa64-linux.h
2025-11-23 John David Anglin
gcc/ChangeLog:
* config/pa/pa64-linux.h (GLIBC_DYNAMIC_LINKER): Define.
diff --git a/gcc/config/pa/pa64-linux.h b/gcc/config/pa/pa64-linux
Committed to trunk.
Dave
---
hppa: Enable LRA as default
2025-11-23 John David Anglin
gcc/ChangeLog:
PR target/113932
PR target/113933
* config/pa/pa.opt (mlra): Default to LRA instead of reload.
diff --git a/gcc/config/pa/pa.opt b/gcc/config/pa/pa.opt
index fae4247
Tested on hppa64-hp-hpux11.11. Committed to trunk.
Dave
---
hppa: Fix scaled and unscaled index support on targets with non-equivalent
space registers
HP-UX targets have non-equivalent space registers. The base register
in most loads and stores selects the space register used to calculate
the
This moves the `(pointer_diff (pointer_plus @0 @2) (pointer_plus @1 @2))`
pattern
to right below the `(pointer_diff (pointer_plus @0 @1) (pointer_plus @0 @2))`
pattern
to make easier to see both versions are supported.
Bootstrapped and tested on x86_64-linux-gnu.
gcc/ChangeLog:
* match
> Am 21.11.2025 um 19:44 schrieb Andrew Pinski :
>
> This moves 2 patterns, `1/x` and `m1 CMP m2) * d` patterns to
> use `fold_before_rtl_expansion_p()` instead of `!canonicalize_math_p ()`
> it was before.
>
> These 2 were checking !canonicalize_math_p() before because there was no
> other p
This patch bumps the libgm2 soname to 21:0:0 ready for the release of GCC 16.
libgm2/ChangeLog:
PR modula2/122801
* configure: Regenerate.
* configure.ac (libtool_VERSION): Assign to 21:0:0.
Signed-off-by: Gaius Mulley
---
libgm2/configure| 2 +-
libgm2/configure.ac
Bootstrapped and regtested on x86_64-pc-linux-gnu, OK for trunk/15?
-- >8 --
We don't generally stream the TREE_CHAIN of a DECL, as this would cause
us to unnecessarily walk into the next member in its scope chain any
time it was referenced by an expression.
Unfortunately, REQUIRES_EXPR_PARMS is
> On Sat, Nov 22, 2025 at 3:41 AM Jose E. Marchesi
> wrote:
>>
>> [Changes from V5:
>> - Rebased to today's master.
>> - We have added a modules system to support separated compilation,
>>based on the MR design, as well as many tests. Compilation units,
>>or packets, are either partic
On Sat, Nov 22, 2025 at 3:41 AM Jose E. Marchesi
wrote:
>
> [Changes from V5:
> - Rebased to today's master.
> - We have added a modules system to support separated compilation,
>based on the MR design, as well as many tests. Compilation units,
>or packets, are either particular program
On Fri, Nov 21, 2025 at 2:41 PM Li, Pan2 wrote:
>
> Thanks Richard.
>
> > Please use (nop_convert (bitop...
>
> It complains predicate cannot be outermost expr with blow change, is there
> something missing from my side?
> I searched match.pd with ^\s(nop_con but failed to find anything I can ref
On Linux/x86_64,
ed1911b9f1f1ab0c1b631f0b6427b798c7056200 is the first bad commit
commit ed1911b9f1f1ab0c1b631f0b6427b798c7056200
Author: Kugan Vivekanandarajah
Date: Sun Nov 23 15:27:10 2025 +1100
[tree-optimization] Allow LICM to hoist loads in "self write" patterns
caused
FAIL: gcc.dg
34 matches
Mail list logo
