On Thu, Aug 14, 2025 at 3:43 PM Jeff Law wrote:
>
> While working to remove mvconst_internal I stumbled over a regression in
> the code to handle signed division by a power of two.
>
> In that sequence we want to select between 0, 2^n-1 by pairing a sign
> bit splat with a subsequent logical right
On Thu, Aug 14, 2025 at 12:08 AM wrote:
>
> From: Pan Li
>
> This patch would like to try to match the the unsigned
> SAT_MUL form 3, aka below:
>
> #define DEF_SAT_U_MUL_FMT_3(NT, WT) \
> NT __attribute__((noinline))\
> sat_u_mul_##NT##_from_##WT##_fmt_3 (NT
On Tue, Aug 19, 2025 at 6:09 AM Andrew Pinski
wrote:
>
> When expanding malloc like functions, we copy the return register into a
> temporary
> and then mark that temporary register with a noalias regnote and the
> alignment.
> This works fine unless you are calling the function with a return ty
On Mon, 18 Aug 2025, Richard Biener wrote:
> SRA handles outermost VIEW_CONVERT_EXPRs but it wrongly ignores
> those when building an access which leads to the wrong size
> used when the VIEW_CONVERT_EXPR does not have the same size as
> its operand which is valid GENERIC and is used by Ada upcast
On Tue, Aug 19, 2025 at 10:51 AM H.J. Lu wrote:
>
> We can't place a TLS call before a conditional jump in a basic block like
>
> (code_label 13 11 14 4 2 (nil) [1 uses])
> (note 14 13 16 4 [bb 4] NOTE_INSN_BASIC_BLOCK)
> (jump_insn 16 14 17 4 (set (pc)
> (if_then_else (le (reg:CCNO 17 fla
Am Montag, dem 18.08.2025 um 23:19 + schrieb Joseph Myers:
> On Fri, 15 Aug 2025, Alejandro Colomar wrote:
>
> > Hi Joseph,
> >
> > On Thu, Aug 14, 2025 at 10:03:00PM +, Joseph Myers wrote:
> > > On Thu, 14 Aug 2025, Alejandro Colomar wrote:
> > >
> > > > And I'm proposing it as a GNU ex
This adds scaffolding for supporting narrowing IFNs inside the vectorizer in a
similar way as how widening is supported. However because narrowing operations
always have the same number of elements as the input and output we need to be
able to combine the results. One way this could have been don
On Mon, Aug 18, 2025 at 9:25 PM Tamar Christina
wrote:
> commit g:fb59c5719c17a04ecfd58b5e566eccd6d2ac583a stops passing the scalar
> type
> (confusingly named vectype) to the costing hook when doing scalar costing.
>
> As a result, we could no longer distinguish between FPR and GPR scalar
> stmt
From: Pan Li
This patch would like to introduce the combine of vec_dup + vmacc.vv
into vmacc.vx on the cost value of GR2VR. The late-combine will take
place if the cost of GR2VR is zero, or reject the combine if non-zero
like 1, 2, 15 in test.
From:
| ...
| vmv.v.x
| L1:
| vmacc.vv
|
On Tue, Aug 19, 2025 at 10:55 AM H.J. Lu wrote:
>
> On Mon, Aug 18, 2025 at 6:56 PM Hongtao Liu wrote:
> >
> > On Tue, Aug 19, 2025 at 4:40 AM H.J. Lu wrote:
> > >
> > > On Mon, Aug 18, 2025 at 12:59 AM Hongtao Liu wrote:
> > > >
> > > > On Mon, Aug 18, 2025 at 4:12 PM Hongtao Liu wrote:
> > >
I'd accidentally had a custom GXX_TESTSUITE_STDS leftover in my
environment, so I'd missed testing as C++14.
Tested on x86_64-pc-linux-gnu, pushing as obvious.
-- >8 --
This testcase (added in r16-3233-g7921bb4afcb7a3) mistakenly only
required C++14, but auto template paramaters are a C++17 feat
Given a sequence such as
int foo ()
{
#pragma GCC unroll 4
for (int i = 0; i < N; i++)
if (a[i] == 124)
return 1;
return 0;
}
where a[i] is long long, we will unroll the loop and use an OR reduction for
early break on Adv. SIMD. Afterwards the sequence is followed by a compression
This implements the new vector optabs vec_addh_narrow_hi_,
vec_addh_narrow_lo_, vec_addh_narrow adding support for
in-vectorizer recognition of addhn.
The existing codegen tests will now recognize the instructions through the
optabs
rather than combine.
Bootstrapped Regtested on aarch64-none-lin
If the user has requested loop unrolling through pragma GCC unroll then at the
moment we only set LOOP_VINFO_USER_UNROLL if the vectorizer has not overrode the
unroll factor (through backend costing) or if the VF made the requested unroll
factor be 1.
But of these events are costing related, and s
commit g:1786be14e94bf1a7806b9dc09186f021737f0227 stops storing in
STMT_VINFO_VECTYPE the vectype of the current stmt being vectorized and instead
requires the use of SLP_TREE_VECTYPE for everything but data-refs.
This means that STMT_VINFO_VECTYPE (stmt_info) will always be NULL and so
aarch64_bo
This adds support for detectioon of the ADDHN pattern in the vectorizer.
Concretely try to detect
_1 = (W)a
_2 = (W)b
_3 = _1 + _2
_4 = _3 >> (precision(a) / 2)
_5 = (N)_4
where
W = precision (a) * 2
N = precision (a) / 2
Bootstrapped Regtested on aarch64-none-linux-gnu,
arm-none-l
commit g:1786be14e94bf1a7806b9dc09186f021737f0227 stops storing in
STMT_VINFO_VECTYPE the vectype of the current stmt being vectorized and instead
requires the use of SLP_TREE_VECTYPE for everything but data-refs.
However contrary to what the commit says not all usages of STMT_VINFO_VECTYPE
have b
commit g:fb59c5719c17a04ecfd58b5e566eccd6d2ac583a stops passing the scalar type
(confusingly named vectype) to the costing hook when doing scalar costing.
As a result, we could no longer distinguish between FPR and GPR scalar stmts.
A later commit also removed STMT_VINFO_VECTYPE from stmt_info.
T
When expanding malloc like functions, we copy the return register into a
temporary
and then mark that temporary register with a noalias regnote and the alignment.
This works fine unless you are calling the function with a return type of void.
At this point then the valreg will be null and a crash
Looks like I mispelled "Merrill". Here is the corrected patch.
-Ben
On Mon, Aug 18, 2025 at 7:31 PM Ben Wu wrote:
> Sorry about that, I'm not sure why the formatting issues are there.
> Gmail seems to remove tabs from the patch as well. I attached the
> revised patch in case this patch still ha
From: Pan Li
This patch would like to combine the vec_duplicate + vmacc.vv to the
vmacc.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 th
From: Pan Li
After enable the vmacc.vx by introducing the define_insn, the
below asm need to adjust for this change.
gcc/testsuite/ChangeLog:
* gcc.target/riscv/rvv/base/ternop_vx_constraint-4.c: Adjust
asm check for vx.
* gcc.target/riscv/rvv/base/ternop_vx_constraint-5
From: Pan Li
Add asm dump check and run test for vec_duplicate + vmacc.vvm
combine to vmacc.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 vx combine.
* gcc.target/riscv/rvv/autovec/vx_v
From: Pan Li
Add asm dump check and run test for vec_duplicate + vmacc.vvm
combine to vmacc.vx, with the GR2VR cost is 0, 2 and 15.
gcc/testsuite/ChangeLog:
* gcc.target/riscv/rvv/autovec/vx_vf/vx-1-i16.c: Add asm check
for vx combine.
* gcc.target/riscv/rvv/autovec/vx_v
On Mon, Aug 18, 2025 at 6:56 PM Hongtao Liu wrote:
>
> On Tue, Aug 19, 2025 at 4:40 AM H.J. Lu wrote:
> >
> > On Mon, Aug 18, 2025 at 12:59 AM Hongtao Liu wrote:
> > >
> > > On Mon, Aug 18, 2025 at 4:12 PM Hongtao Liu wrote:
> > > >
> > > > On Mon, Aug 18, 2025 at 4:50 AM H.J. Lu wrote:
> > >
We can't place a TLS call before a conditional jump in a basic block like
(code_label 13 11 14 4 2 (nil) [1 uses])
(note 14 13 16 4 [bb 4] NOTE_INSN_BASIC_BLOCK)
(jump_insn 16 14 17 4 (set (pc)
(if_then_else (le (reg:CCNO 17 flags)
(const_int 0 [0]))
(label_ref
Sorry about that, I'm not sure why the formatting issues are there.
Gmail seems to remove tabs from the patch as well. I attached the
revised patch in case this patch still has some formatting problems like
the indentation for error_at.
I have adjusted the testcases to use dg-prune-output and boot
Changes in v9:
* Split out implementations from public pre-26 and 26 APIs.
* Add remaining support for non-radix2 float types in C++26.
* Handle pow() and floor() compatibly with non-built-in types, and Clang.
* Add testing for a non-radix2 float type.
* Iron out gratuitous differences between
On Tue, Aug 19, 2025 at 4:40 AM H.J. Lu wrote:
>
> On Mon, Aug 18, 2025 at 12:59 AM Hongtao Liu wrote:
> >
> > On Mon, Aug 18, 2025 at 4:12 PM Hongtao Liu wrote:
> > >
> > > On Mon, Aug 18, 2025 at 4:50 AM H.J. Lu wrote:
> > > >
> > > > We can't place a TLS call before a conditional jump in a b
在 2025/8/18 下午7:28, Xi Ruoyao 写道:
On Mon, 2025-08-18 at 11:10 +0800, Xi Ruoyao wrote:
On Mon, 2025-08-18 at 09:15 +0800, Lulu Cheng wrote:
Pushed to r16-3247 ... r16-3264.
Sorry it took so long to merge.
LoongArch: Implement 16-byte CAS with sc.q
Sorry but it seems we need to revert th
On Fri, 15 Aug 2025, Alejandro Colomar wrote:
> Hi Joseph,
>
> On Thu, Aug 14, 2025 at 10:03:00PM +, Joseph Myers wrote:
> > On Thu, 14 Aug 2025, Alejandro Colomar wrote:
> >
> > > And I'm proposing it as a GNU extension, which means we don't even need
> > > to care about what ISO C says abo
On Mon, Aug 18, 2025 at 12:59 AM Hongtao Liu wrote:
>
> On Mon, Aug 18, 2025 at 4:12 PM Hongtao Liu wrote:
> >
> > On Mon, Aug 18, 2025 at 4:50 AM H.J. Lu wrote:
> > >
> > > We can't place a TLS call before a conditional jump in a basic block like
> > >
> > > (code_label 13 11 14 4 2 (nil) [1 us
On 8/16/25 2:02 PM, Ben Wu wrote:
This testcase caused an ICE when mangling the invalid type-constraint in
write_requirement since write_type_constraint expects a TEMPLATE_TYPE_PARM.
Setting the trailing return type to NULL_TREE when a
return-type-requirement is found in place of a type-constrai
On 8/16/25 11:06 PM, Nathaniel Shead wrote:
Bootstrapped and regtested on x86_64-pc-linux-gnu, OK for trunk?
OK.
-- >8 --
We have logic to adjust a function decl if it gets re-declared as a
using-decl with different purviewness, but we also need to do the same
if it gets redeclared with diff
On 8/16/25 11:13 PM, Nathaniel Shead wrote:
Tested on x86_64-pc-linux-gnu and verified with a powerpc64-linux-gnu
cross. OK for trunk and 15?
OK (obvious)
-- >8 --
I added a testcase for the (temporary) warning that we don't currently
support the 'gnu::optimize' or 'gnu::target' attributes
On 8/18/25 11:36 AM, Patrick Palka wrote:
Bootstrapped anh regtested on x86_64-pc-linux-gnu, does this look
OK for trunk and 15?
OK.
-- >8 --
When a using refers to a member from a partial specialization, we
need to substitute the arguments relative to partial specialization
into the constra
This example used to work (with C) in GCC 14 before the
warning for different pointer types without a cast was changed
to an error.
The fix is to make the q variable `int*` rather than the current `char*`.
This also fixes the example for C++ too.
Pushed as obvious after doing a `make html`.
gcc/C
On Mon, 18 Aug 2025 at 19:24, Christophe Lyon
wrote:
>
> A few arm effective-targets call check_effective_target_arm32 even
> though they would force a -march=XXX flag which supports Arm and/or
> Thumb-2, thus making the arm32 check useless. This has an impact when
> the toolchain is configured w
We get lots of error messages when compiling arm_neon.h under
e.g. -mcpu=cortex-m55, because Neon builtins are enabled only when
!TARGET_HAVE_MVE. This has been the case since MVE support was
introduced.
This patch uses an approach similar to what we do on aarch64, but only
partially since Neon i
A few arm effective-targets call check_effective_target_arm32 even
though they would force a -march=XXX flag which supports Arm and/or
Thumb-2, thus making the arm32 check useless. This has an impact when
the toolchain is configured with a default -march or -mcpu which
supports Thumb-1 only: in su
Like we do in other effective-targets, add
"-mcpu=unset -march=armv8-a"
directly when setting et_arm_v8_neon_flags in arm_v8_neon_ok_nocache,
to avoid having to add these two flags in all users of arm_v8_neon_ok.
This avoids duplication and possible typos / oversights.
gcc/testsuite/ChangeLog:
From: Dhruv Chawla
This patch folds the following patterns:
- umax (a, add (a, b)) -> [sum, ovf] = adds (a, b); !ovf ? sum : a
- umin (a, add (a, b)) -> [sum, ovf] = adds (a, b); !ovf ? a : sum
- umax (a, sub (a, b)) -> [diff, ovf] = subs (a, b); ovf ? diff : a
- umin (a, sub (a, b)) -> [diff,
Ping!
please review.
Thanks & Regards
Kishan
On 05/06/25 12:36 pm, Kishan Parmar wrote:
> Hi All,
>
> The following patch has been bootstrapped and regtested on powerpc64le-linux.
>
> After r12-5752-gd08236359eb229, a new bif infrastructure was introduced
> which stopped using opaque vector typ
Bootstrapped anh regtested on x86_64-pc-linux-gnu, does this look
OK for trunk and 15?
-- >8 --
When a using refers to a member from a partial specialization, we
need to substitute the arguments relative to partial specialization
into the constraints, not those relative to the primary template.
O
>> + { NULL, 0ULL, 0UL, false, false }
>
> sanitize_code_type (0) (or just 0) instead of 0ULL. Same for the other
> uses of the same constants.
>
> OK with that change, thanks.
>
> Richard
>
Patch pushed including indicated mods.
Thank you,
Claudiu
>> + (const_int MEMTAG_TAG_MASK)) ;; 0xf0ff...
>
> I think we can drop the comments now, since the name of the constant
> makes it clear enough. Same for the uses below.
>
> OK with that change. Thanks for doing this.
>
> Richard
Patch pushed with the additional mods.
Thank you,
On Mon, 18 Aug 2025 at 15:40, Jonathan Wakely wrote:
>
> On Thu, 24 Jul 2025 at 08:33, Tomasz Kamiński wrote:
> >
> > Removed the wrong_stuff() function, which was effectively empty for
> > actual test runs. Replaced the manual failure counter with the VERIFY
> > macro to simplify identifying fai
On Thu, 24 Jul 2025 at 08:33, Tomasz Kamiński wrote:
>
> Removed the wrong_stuff() function, which was effectively empty for
> actual test runs. Replaced the manual failure counter with the VERIFY
> macro to simplify identifying failures.
>
> PR libstdc++/104874
>
> libstdc++-v3/ChangeLog:
SRA handles outermost VIEW_CONVERT_EXPRs but it wrongly ignores
those when building an access which leads to the wrong size
used when the VIEW_CONVERT_EXPR does not have the same size as
its operand which is valid GENERIC and is used by Ada upcasting.
Bootstrapped on x86_64-unknown-linux-gnu with
The FUNCTION_VALUE and LIBCALL_VALUE macros are deprecated in favor of
the TARGET_FUNCTION_VALUE and TARGET_LIBCALL_VALUE target hooks. This
patch replaces the macro definitions with proper target hook implementations.
This change is also a preparatory step for VLS calling convention support,
whi
On Mon, 2025-08-18 at 11:10 +0800, Xi Ruoyao wrote:
> On Mon, 2025-08-18 at 09:15 +0800, Lulu Cheng wrote:
> > Pushed to r16-3247 ... r16-3264.
> >
> > Sorry it took so long to merge.
>
> > > LoongArch: Implement 16-byte CAS with sc.q
>
> Sorry but it seems we need to revert this one particul
> -Original Message-
> From: Richard Sandiford
> Sent: Monday, August 18, 2025 11:30 AM
> To: gcc-patches@gcc.gnu.org
> Cc: Alex Coplan ; Alice Carlotti
> ;
> pins...@gmail.com; ktkac...@nvidia.com; Richard Earnshaw
> ; Tamar Christina ;
> Wilco Dijkstra
> Subject: [GCC 15] aarch64: F
> -Original Message-
> From: Prathamesh Kulkarni
> Sent: 10 August 2025 20:04
> To: Matthew Malcomson ; gcc-patches@gcc.gnu.org
> Cc: Joseph Myers ; Thomas Schwinge
> ; Sam James
> Subject: RE: [v2] PR81358: Enable automatic linking of libatomic
>
> External email: Use caution opening
This PR is about a case where we used aarch64_expand_sve_const_pred_trn
to combine two predicates, one of which was constructing using
aarch64_sve_move_pred_via_while. The former requires the inputs
to have mode VNx16BI, but the latter returned VNx8BI for a .H
WHILELO.
The proper fix, used on tru
Richard Biener writes:
> On Mon, Aug 18, 2025 at 10:51 AM Richard Sandiford
> wrote:
>>
>> This patch fixes an internal disagreement in gcse about how to
>> handle partial clobbers. Like many passes, gcse doesn't track
>> the modes of live values, so if a call clobbers only part of
>> a register
On Mon, 18 Aug 2025 at 10:14, Tomasz Kamiński wrote:
>
> This patch adds the [[nodiscard]] attribute to the operator() of ranges
> algorithm function objects if their std counterpart has it.
>
> Furthermore, we [[nodiscard]] the operator() of the following ranges
> algorithms that lack a std counte
This patch adds the [[nodiscard]] attribute to the operator() of ranges
algorithm function objects if their std counterpart has it.
Furthermore, we [[nodiscard]] the operator() of the following ranges
algorithms that lack a std counterpart:
* find_last, find_last_if, find_last_if_not (to match oth
On Mon, Aug 18, 2025 at 10:51 AM Richard Sandiford
wrote:
>
> This patch fixes an internal disagreement in gcse about how to
> handle partial clobbers. Like many passes, gcse doesn't track
> the modes of live values, so if a call clobbers only part of
> a register, the pass has to make conservati
This patch fixes an internal disagreement in gcse about how to
handle partial clobbers. Like many passes, gcse doesn't track
the modes of live values, so if a call clobbers only part of
a register, the pass has to make conservative assumptions.
As the comment in the patch says, this means:
(1) ig
> * moved the 'defer' keyword behind a -fdefer-ts flag, and added texinfo
> docs plus related failure test.
it was mentioned to me that having both -std=c2y and -fdefer-ts would
be excessive. i'll wait a bit for more feedback, then send a v3 dropping
1/4 and adjusting the following commits to remo
For __n == 0, the elements were self move-assigned by
std::move_backward(__ins, __old_finish - __n, __old_finish).
PR libstdc++/121313
libstdc++-v3/ChangeLog:
* include/bits/vector.tcc (vector::insert_range): Add check for
empty size.
* testsuite/23_containers/vec
On Mon, Aug 11, 2025 at 8:57 PM Richard Biener wrote:
>
> On Sun, 10 Aug 2025, liuhongt wrote:
>
> > >
> > > The comment doesn't match the bool type.
> > >
> > Fixed.
> >
> > >
> > > is_gimple_assign (stmt_info->stmt)
> > >
> > Changed.
> >
> > > There's also SAD_EXPR? The vectorizer has lane_red
From: Dhruv Chawla
This patch is a respin of the RFC originally posted at
https://gcc.gnu.org/pipermail/gcc-patches/2025-June/686835.html.
The patch reads the file names from the GCOV file and compares them
against DECL_SOURCE_FILE for symbols to decide which profile count to
annotate them with.
The overlapping range, will lead to move from moved-from object, which I
think is generally expected to work.
My preference is to land this fix for us misbehaving for correct inputs,
and consider any additional assertions separately.
On Fri, Aug 15, 2025 at 12:27 PM Jonathan Wakely
wrote:
> Mayb
On Mon, Aug 18, 2025 at 4:12 PM Hongtao Liu wrote:
>
> On Mon, Aug 18, 2025 at 4:50 AM H.J. Lu wrote:
> >
> > We can't place a TLS call before a conditional jump in a basic block like
> >
> > (code_label 13 11 14 4 2 (nil) [1 uses])
> > (note 14 13 16 4 [bb 4] NOTE_INSN_BASIC_BLOCK)
> > (jump_ins
On Mon, Aug 18, 2025 at 4:50 AM H.J. Lu wrote:
>
> We can't place a TLS call before a conditional jump in a basic block like
>
> (code_label 13 11 14 4 2 (nil) [1 uses])
> (note 14 13 16 4 [bb 4] NOTE_INSN_BASIC_BLOCK)
> (jump_insn 16 14 17 4 (set (pc)
> (if_then_else (le (reg:CCNO 17 flag
66 matches
Mail list logo
