On Tue, Jul 22, 2025 at 5:21 PM Robin Dapp wrote:
>
> > Note if-conversion emits IFN_MASK_LOAD/STORE, only the vectorizer later
> > emits the LEN variants. So this is about whether there are (might) be
> > uarchs that have vector aligned loads (aka target alignment is
> > sizeof(vector))
> > and
Am Montag, dem 21.07.2025 um 16:03 -0700 schrieb Bill Wendling:
> On Mon, Jul 21, 2025 at 2:19 PM Martin Uecker wrote:
> > Am Montag, dem 21.07.2025 um 04:29 -0700 schrieb Bill Wendling:
> > > On Sun, Jul 20, 2025 at 4:41 AM Martin Uecker wrote:
..
>
> > > Re original context: The 'struct c_par
On Tue, 15 Jul 2025, H.J. Lu wrote:
>> This feels a bit complex to parse. How about something like
>>
>> + The new --enable-x86-64-mfentry configure option
>> + makes -mfentry use __fentry__ instead
>> + of mcount for profiling on x86-64. This option is
>> + enabled by default for
IMHO there are enough reasons to reject delayed parsing
as bad design for C. We should work towards proper
language features that cleanly fit into the language,
instead of pushing the boundaries with macros.
But this requires true collaboration, which can not
exist when one side is not able to
Since r16-372-g064cac730f88dc fn1 is now inlined into main
which meant the scan dump was failing since it was looking
for it only once. Marking fn1 as noinline gets us back to
the old behavior and no longer dependent on the inliner.
Pushed as obvious after a quick test.
PR testsuite/12010
From: Pan Li
Add asm check to make sure vx combine of vaaddu.vx will not pollute
the vxrm.
gcc/testsuite/ChangeLog:
* gcc.target/riscv/rvv/autovec/vx_vf/vx-fixed-vxrm-1-u16.c: New test.
* gcc.target/riscv/rvv/autovec/vx_vf/vx-fixed-vxrm-1-u32.c: New test.
* gcc.target/ri
From: Pan Li
The vaaddu.vx combine almost comes from avg_floor, it will
requires the vxrm to be RDN. But not all vaaddu.vx should
depends on the RDN. The vaaddu.vx combine should leverage
the VXRM value as is instead of pollute them all to RDN.
This patch would like to fix this and set it as i
From: Pan Li
The RVV fixed point insn VX combine should focus on the insn itself,
instead of any standard name like avg_floor, the vxrm should be
the value of insn as is.
The below test suites are passed for this patch series.
* The rv64gcv fully regression test.
Pan Li (2):
RISC-V: Avoid vaa
[Explicitly CCing Joseph since this thread[1] is about the C front-end]
On Tue, Jul 22, 2025 at 09:01:35PM +, Qing Zhao wrote:
> However, given the current situation, I also think that provide two different
> interfaces from CLANG
> and GCC might be the practical solution to this endless disc
On 7/22/25 3:03 AM, Robin Dapp wrote:
Hi,
During the last weeks it became clear that our current broadcast
handling needs an overhaul in order to improve maintainability.
PR121073 showed that my intermediate fix wasn't enough and caused
regressions.
This patch now goes a first step towards u
On 2025-07-21 04:52, Siddhesh Poyarekar wrote:
On 2025-07-14 03:02, Richard Biener wrote:
On Fri, Jul 11, 2025 at 8:27 PM Siddhesh Poyarekar wrote:
The test ends up writing a byte beyond bounds of the buffer, which
On Fri, 18 Jul 2025, Pietro Monteiro wrote:
> I think you need to update the description of MAX_FIXED_MODE_SIZE in
> gcc/doc/tm.texi.in with the new default.
Oops, absolutely, thanks for pointing that out.
Here's v2, with info and dvi results inspected for sanity.
Ok to commit now? (Note: sciss
On 7/20/25 6:54 PM, Sam James wrote:
This is vanilla zlib-1.3.1 imported over the existing zlib/ dir with:
* README adjusted to add the GCC note at the top;
* GCC's ChangeLog merged with the upstream one, as before;
* Deleted upstream Makefile as has been done before (we use an autoconf-
ge
On Fri, Apr 18, 2025 at 5:08 PM Andrew Pinski wrote:
>
> This fixes a long standing (since GCC 5) issue where the malloc+memset->calloc
> optimization would happen even if the memset was not always executed.
> This is a varient of Nathan's patch:
> https://inbox.sourceware.org/gcc-patches/f4b5d10
[Apologies if I've missed some context in my reading since I'm coming
back to this after a big break]
On 2025-07-22 12:33, Qing Zhao wrote:
Why it's wrong to pass the VALUE of the original pointer as the first argument
to
the call to .ACCESS_WITH_SIZE
For a pointer field with counted_by attri
On Mon, Jul 21, 2025 at 8:09 PM Hongtao Liu wrote:
>
> On Tue, Jul 22, 2025 at 4:47 AM H.J. Lu wrote:
> >
> > For TLS calls:
> >
> > 1. UNSPEC_TLS_GD:
> >
> > (parallel [
> > (set (reg:DI 0 ax)
> > (call:DI (mem:QI (symbol_ref:DI ("__tls_get_addr")))
> > (const_in
For TLS calls:
1. UNSPEC_TLS_GD:
(parallel [
(set (reg:DI 0 ax)
(call:DI (mem:QI (symbol_ref:DI ("__tls_get_addr")))
(const_int 0 [0])))
(unspec:DI [(symbol_ref:DI ("e") [flags 0x50])
(reg/f:DI 7 sp)] UNSPEC_TLS_GD)
(clobber (reg:DI 5 di))]
This patch would add a new middle-end representation for matching the
x264 narrowing clip idiom:
inline
U_NT clip_uint8 (S_WT x)
{
return x & (~((U_NT)-1)) ? (-x) >> 31 : x;
}
which would be accessible through the define_expand clip2
optabs.
For example, truncating int32_t to uint8_t would pro
On 7/22/25 11:28 AM, Harald Anlauf wrote:
Dear all,
the attached patch fixes a long-standing issue with legacy code,
where an assumed-length character function is passed as actual
argument to a procedure, and when there is no explicit interface.
The solution is to do the same as in the case when
Tested x86_64-pc-linux-gnu, applying to trunk.
-- 8< --
The note and example in [class.union] p6 think that placement new can be
used to change the active member of a union, but we didn't support that for
array members in constant-evaluation even after implementing P1330 and
P2747.
First I tried
On Sat, 19 Jul 2025, 1nfocalypse wrote:
> Implements Philox Engine (P2075R6) and associated tests.
>
> Implements additional feedback from v3 from Patrick Palka.
>
> I went ahead and qualified the STL functions to avoid ADL ambiguity,
> as recommended. However, if this was erroneous, I'd be happ
Hi, Bill,
Thanks a lot for your effort to try to provide a consistent interface of
counted_by attribute to the users.
Really appreciate for this.
However, given the current situation, I also think that provide two different
interfaces from CLANG
and GCC might be the practical solution to this
On 7/22/25 4:18 PM, Patrick Palka wrote:
Bootstrapped and rgetested on x86_64-pc-linux-gnu, does this look OK for
trunk?
In order to properly handle all rewritten operator expressions, I
suppose we could teach build_min_non_dep_op_overload to carefully look
through 'non_dep' and note when argume
On 7/21/25 6:28 PM, Marek Polacek wrote:
On Mon, Jul 21, 2025 at 11:26:08AM -0400, Jason Merrill wrote:
On 7/18/25 5:11 PM, Marek Polacek wrote:
On Thu, Jul 17, 2025 at 06:44:08PM -0400, Jason Merrill wrote:
On 7/17/25 6:25 PM, Jakub Jelinek wrote:
On Thu, Jul 17, 2025 at 05:31:27PM -0400, Ja
On 7/22/25 4:18 PM, Patrick Palka wrote:
Bootstrapped and regtested on x86_64-pc-linux-gnu, does this look OK for trunk?
OK, and I think go ahead and apply it to the 15 branch as well.
-- >8 --
Our implementation of the INVOKE spec ([func.require]) was incorrectly
treating reference_wrapper:
Bootstrapped and regtested on x86_64-pc-linux-gnu, does this look OK for trunk?
-- >8 --
Our implementation of the INVOKE spec ([func.require]) was incorrectly
treating reference_wrapper::get() as returning T instead of T&, which
notably makes a difference when invoking a ref-qualified memfn poin
Bootstrapped and rgetested on x86_64-pc-linux-gnu, does this look OK for
trunk?
In order to properly handle all rewritten operator expressions, I
suppose we could teach build_min_non_dep_op_overload to carefully look
through 'non_dep' and note when arguments are reversed etc. But rather
than that
On Tue, Jul 22, 2025 at 12:50 PM Andrew Pinski wrote:
>
> On Thu, Jul 17, 2025 at 10:35 PM wrote:
> >
> > From: Soumya AR
> >
> > This patch adds functionality to dump AArch64 CPU tuning parameters to a
> > JSON
> > file. The new '-fdump-tuning-model=' flag allows users to export the current
>
On Thu, Jul 17, 2025 at 10:35 PM wrote:
>
> From: Soumya AR
>
> This patch adds functionality to dump AArch64 CPU tuning parameters to a JSON
> file. The new '-fdump-tuning-model=' flag allows users to export the current
> tuning model configuration to a JSON file.
>
> This patch was bootstrapped
On 26/06/25 9:20 pm, Segher Boessenkool wrote:
> Hi!
>
> [ Please don't say "patch v1", it's just clutter. ]
>
> The point of the patch is to improve some code that evokes undefined
> behaviour. The sanitizer for that is how these problems were found,
> you can remark that somewhere later in the
+josmy...@redhat.com
On Sun, Jul 20, 2025 at 4:20 AM wrote:
>
> From: Bill Wendling
>
> Also, this code doesn't go further than parsing. I.e., it doesn't generate the
> internal gimple code that accesses the struct fields. The code is meant to
> show
> that it *is* possible to perform a delayed
On 7/22/25 17:22, Iain Sandoe wrote:
Hi Andre
2. If I try linking a trivial Fortran code thus:
gfortran /src-local/test/fortran/hello.f90 -o h -lcaf_shmem -fcoarray=lib
the exe then hangs - it looks like it might be deadlocked in startup code:
Call graph:
2166 Thread_408707810 Disp
On Tue, Jul 22, 2025, 11:05 AM Remi Machet wrote:
>
> On 7/22/25 13:32, Andrew Pinski wrote:
> > External email: Use caution opening links or attachments
> >
> >
> > Regrename can fail in some case and `insn_rr[INSN_UID (insn)].op_info`
> > will be null. The FMA steering code was not expecting th
Dear all,
the attached patch fixes a long-standing issue with legacy code,
where an assumed-length character function is passed as actual
argument to a procedure, and when there is no explicit interface.
The solution is to do the same as in the case when there is an
actual interface: take the cha
On 7/22/25 13:32, Andrew Pinski wrote:
> External email: Use caution opening links or attachments
>
>
> Regrename can fail in some case and `insn_rr[INSN_UID (insn)].op_info`
> will be null. The FMA steering code was not expecting the failure to happen.
> This started to happen after early RA was
The switch conversion code will generate an array with VLA vector constants in
it
in some cases but this does not work as the length of the vector type is not
known
at compile time; only the min size is known.
I tried to reject this in initializer_constant_valid_p but code from the C++
front-en
Regrename can fail in some case and `insn_rr[INSN_UID (insn)].op_info`
will be null. The FMA steering code was not expecting the failure to happen.
This started to happen after early RA was added but it has been a latent bug
before that.
Build and tested for aarch64-linux-gnu.
PR target/1
On Wed, Jul 16, 2025 at 1:13 AM Takayuki 'January June' Suwa
wrote:
>
> This patch fixes the following defects in the function:
>
>- The cost of move instructions larger than the natural word width,
> specifically "movd[if]_internal", cannot be estimated correctly
>- Floating-point or
Hi,
More thought on this over the weekend. -:)
I summarized my thought in the following small writeup, please provide any
comment
or suggestion.
Thanks a lot!
Qing
===
Why it's wrong to pass the VALUE of the original pointer as the first argument
to
the call
> Am 22.07.2025 um 17:05 schrieb Martin Jambor :
>
> Hi,
>
> we currently use the types encountered in the function body and not in
> type declaration to perform total scalarization. Bug PR 119085
> uncovered that we miss a check that when the same data is accessed
> with aggregate types tha
Gentle ping.
On Mon, Jul 14, 2025 at 09:43:29PM +, Spencer Abson wrote:
> Hi all,
>
> Currently, using '-moverride' in the testsuite will reset the baseline tuning
> model from 'generic' to the default tuning for the architecture level. Since
> 'generic' is the tuning model relied upon by mo
Hi Andre
It would be very helpful if there were such a thing as a “hello-coarrays.fxx”
code for
those of us who do not have existing installations of mpi &c. aand limited time
to
so tests…
… experiments so far - below
thanks
Iain
> >> > > We like everyone to test the new caf_shmem library an
Note if-conversion emits IFN_MASK_LOAD/STORE, only the vectorizer later
emits the LEN variants. So this is about whether there are (might) be
uarchs that have vector aligned loads (aka target alignment is sizeof(vector))
and in addition to that have support for misaligned loads but those with sti
Hi, an update on this feature:
For a more complicated testing case like the following:
#define MAX_LENGTH 10
int a[MAX_LENGTH];
void __attribute__ ((noinline)) foo (int i, bool is_day, bool is_dollar)
{
if (i < MAX_LENGTH)
{
if (is_day)
__builtin_printf ("day");
else
Hi,
we currently use the types encountered in the function body and not in
type declaration to perform total scalarization. Bug PR 119085
uncovered that we miss a check that when the same data is accessed
with aggregate types that those are actually compatible. Without it,
we can base total scal
On Tue, 22 Jul 2025, Tomasz Kamiński wrote:
> Adds negative tests for preconditions on inserting into a full
> inplace_vector and erasing non-existent elementsi at compile-time.
> This ensures coverage for the inplace_vector specialization.
>
> Also extends element access tests to cover front() a
The following makes dataref analysis not set STMT_VINFO_VECTYPE.
This leaves the field ready for taking when vectorizable_*
transition to use SLP_TREE_VECTYPE. It also delays determining
the vectorization mode a little bit - SLP build will first do
that then. I have kept a check that there is any
The following makes us use the SLP graph and its recorded vector
types to perform alignment analysis even for loops. This avoids
accessing STMT_VINFO_VECTYPE we want to get rid of.
v2 adjusts alignment peeling to ignore DRs that are dead because
of conditional store pattern recognition.
On Tue, Jul 22, 2025, 6:09 AM Richard Biener wrote:
> With a patch still in development we get NULL STMT_VINFO_VECTYPE.
> One side-effect is that during scalar stmt testing we no longer
> pass a vectype. The following adjusts aarch64_vector_costs::add_stmt_cost
> to check for a non-NULL vectype
Adds negative tests for preconditions on inserting into a full
inplace_vector and erasing non-existent elementsi at compile-time.
This ensures coverage for the inplace_vector specialization.
Also extends element access tests to cover front() and back()
methods, and const and mutable overloads for
Adds negative tests for preconditions on inserting into a full
inplace_vector and erasing non-existent elementsi at compile-time.
This ensures coverage for the inplace_vector specialization.
Also extends element access tests to cover front() and back()
methods, and const and mutable overloads for
The following removes the STMT_VINFO_VECTYPE usage from
vect_check_gather_scatter and instead gets the vector type passed
in from the caller. For the two early callers, the one from
dataref analysis and the one from vect_mark_stmts_to_be_vectorized
we add a mode that works with a NULL vector type
With a patch still in development we get NULL STMT_VINFO_VECTYPE.
One side-effect is that during scalar stmt testing we no longer
pass a vectype. The following adjusts aarch64_vector_costs::add_stmt_cost
to check for a non-NULL vectype before accessing it, like all the
code surrounding it. The ot
On Tue, 22 Jul 2025, Tomasz Kamiński wrote:
> All functions in testsuite_iterators.h are now marked constexpr,
> targeting the earliest possible standard. Most functions use C++14 due
> to multi-statement bodies, with exceptions:
>
> * BoundsContainer and some constructors are C++11 compatible.
>
Hi Tomasz,
there's no time-pressure or rush from my side. We've finished the
C++23 parts of mdspan, I think that's the important milestone. Hence,
I feel it makes a lot of sense if you deprioritize reviewing mdspan
related patches for a while.
Personally, I think it would be least work for every
constant_byte_string fails to consider the string type might be VLA
when initialized by an empty string CTOR.
Bootstrap and regtest running on x86_64-unknown-linux-gnu.
PR middle-end/121216
* expr.cc (constant_byte_string): Check the string type
size fits an uhwi before co
On 7/22/25 3:04 AM, Robin Dapp wrote:
Hi,
This patch fixes the vf_vfmacc-run-1-f16.c test failures on rv32
by adding zvfh requirements as well as options to the test and
the target harness.
Regtested on rv64gcv_zvl512b and rv32gcv_zvl512b. Going to commit
as obvious if the CI agrees that it
Hi All,
Apologies, that previous patch was clearly hastily made and included some
rubbish temp files.
Cleaned up patch attached
-- >8 --
This commit introduces support for the target_version attribute in the c
frontend, following the behavior defined in the Arm C Language Extension.
Key change
Hello,
Quick update for this patch as I found a minor bug with it.
Changed the decl merging logic slightly and added a new test.
Reg tested on Aarch64 and x68_64.
Kind regards,
Alfie
-- >8 --
This commit introduces support for the target_version attribute in the c
frontend, following the beha
With a patch still in development we get NULL STMT_VINFO_VECTYPE.
One side-effect is that during scalar stmt testing we no longer
pass a vectype. The following adjusts aarch64_vector_costs::add_stmt_cost
to check for a non-NULL vectype before accessing it, like all the
code surrounding it. The ot
namaskaaram
Revised patch is attached. This patch no longer changes the default option
value for
xl-barrel-shift
On Mon, Jul 14, 2025 at 10:29 AM Gopi Kumar Bulusu wrote:
>
>
> On Fri, Jul 11, 2025 at 8:12 PM Michael Eager wrote:
>
>> On 7/10/25 4:41 AM, Gopi Kumar Bulusu wrote:
>> >
>> > nam
When we have a vector shift with a scalar the shift operand can be
external - in that case we should not use the shift operand def
as hint where to place the vector shift instruction. The ICE
in the PR is because stmt dominance queries only work inside of
the vector region. But we should also nev
On Tue, Jul 22, 2025 at 11:28 AM Tomasz Kaminski
wrote:
>
>
> On Tue, Jul 22, 2025 at 9:22 AM Nathan Myers wrote:
>
>> This should be close to ready. However, std::is_invocable
>> and noexcept still fail oddly applied to the not_fp result.
>> The remaining failing test cases in */nttp.cc, commen
On Tue, Jul 22, 2025 at 12:02 PM Robin Dapp wrote:
>
> > So this is the only part I think is odd - there is a dataref, it just
> > has only DR_REF as relevant data. I would have expected we can
> > adjust vect_supportable_dr_alignment to deal with the scatter/gather
> > case. I'm OK with doing i
On Tue, 22 Jul 2025, Robin Dapp wrote:
> The more I look at our heuristic the more it appears due for a rewrite.
> But that's really not in my plans right now. I just sent a riscv patch
> that does the necessary preparations so you can basically
>
> s/STMT_VINFO_TYPE (stmt_info)/SLP_TREE_TYPE (n
On Mon, Jul 21, 2025 at 01:54:55PM -0700, H.J. Lu wrote:
> On Mon, Jul 21, 2025 at 4:16 AM Stefan Schulze Frielinghaus
> wrote:
> >
> > On Sat, Jul 19, 2025 at 08:26:22AM -0600, Jeff Law wrote:
> > >
> > >
> > > On 7/17/25 2:24 AM, Stefan Schulze Frielinghaus wrote:
> > > > On Wed, Jul 09, 2025 at
While scanning the instructions and upon reaching an instruction that
doesn't satisfy the constraints that we have set, we were removing the
already detected stores, but we were continuing adding stores from that
point onward. This was causing issues when the address ranges from later
stores overla
Hi,
These commits mostly LGTM, as mentioned, but as I have mentioned in other
emails,
They may stay as is for some time.
However, we could land a (test rewrite) patch soon, as it does not
introduce any functional changes,
so you could send a v2 of that, and then patch series with only functional
p
On Tue, Jul 22, 2025 at 11:58 AM Tomasz Kaminski
wrote:
>
>
> On Mon, Jul 21, 2025 at 10:19 AM Luc Grosheintz
> wrote:
>
>> All test code of default_accessor can be reused. This commit moves
>> the reuseable code into a file generic.cc and prepares the tests for
>> reuse with aligned_accessor.
>
On Mon, Jul 21, 2025 at 10:25 AM Luc Grosheintz
wrote:
> This commit completes the implementation of P2897R7 by implementing and
> testing the template class aligned_accessor.
>
> PR libstdc++/120994
>
> libstdc++-v3/ChangeLog:
>
> * include/bits/version.def (aligned_accessor): Ad
So this is the only part I think is odd - there is a dataref, it just
has only DR_REF as relevant data. I would have expected we can
adjust vect_supportable_dr_alignment to deal with the scatter/gather
case. I'm OK with doing it how you did it here, but seeing
the
/* For now assume all condit
On Mon, Jul 21, 2025 at 10:19 AM Luc Grosheintz
wrote:
> All test code of default_accessor can be reused. This commit moves
> the reuseable code into a file generic.cc and prepares the tests for
> reuse with aligned_accessor.
>
> The AllocatorTrait creates a unified interface for creating both
>
The more I look at our heuristic the more it appears due for a rewrite.
But that's really not in my plans right now. I just sent a riscv patch
that does the necessary preparations so you can basically
s/STMT_VINFO_TYPE (stmt_info)/SLP_TREE_TYPE (node)/
once it lands.
I regtested with your patc
Hi,
This patch prepares the dynamic LMUL vector costing to use the coming
SLP_TREE_TYPE instead of the (to-be-removed) STMT_VINFO_TYPE.
Even though the whole approach should be reviewed and adjusted at some
point, the patch chooses the path of least resistance and uses a hash
map for the stmt_in
This patch optimizes out calls to __dynamic_cast when the type being cast to is
final (or its destructor), replacing them with a simple comparison of the
types' vtable addresses. This is already implemented and done by default in
clang (https://reviews.llvm.org/D154658), but can be turned off wi
Hi,
IPA split is now running before auto-profile annotation. Changing it
after annotation as it is done for PGO.
gcc/ChangeLog:
2025-07-21 Kugan Vivekanandarajah
* ipa-split.cc (pass_split_functions::gate): Do not run when
flag_auto_profile.
(pass_feedback_split_funct
On Tue, Jul 22, 2025 at 9:22 AM Nathan Myers wrote:
> This should be close to ready. However, std::is_invocable
> and noexcept still fail oddly applied to the not_fp result.
> The remaining failing test cases in */nttp.cc, commented
> out, need careful examination to see whether they should
> be
The following makes us use the SLP graph and its recorded vector
types to perform alignment analysis even for loops. This avoids
accessing STMT_VINFO_VECTYPE we want to get rid of.
v2 adjusts alignment peeling to ignore DRs that are dead because
of conditional store pattern recognition.
Bootstap
On Mon, Jul 21, 2025 at 6:28 PM Stefan Schulze Frielinghaus
wrote:
>
> Bootstrapped successfully on s390x and tests dg.exp=asm-hard-reg-\*.c
> and s390.exp=asm-hard-reg-\*.c are successful, too. Ok for mainline?
OK.
Richard.
> -- >8 --
>
> The GNU extension rawmemchr cannot be used. Therefore
> Bootstrapped on sparc-sun-solaris2.11 and sparcv9-sun-solaris2.11 with
> as/ld, gas/ld, and gas/gld configurations.
>
> There are currently two regressions exposed by this patch (PRs 121191
> and 121192), which are only present in gcc 16 resp. 15/16.
>
> There's one small caveat: while Solaris
From: Gary Dismukes
The compiler mishandles nested use_type_clauses in the case where the
outer one is a normal use_type_clause and the inner one has "all".
Upon leaving the scope of the inner use_type_clause, the outer one
is effectively disabled, because it's not considered redundant (and
in fa
From: Piotr Trojanek
Fix missing checks for prefixes of array attributes in GNATprove mode.
gcc/ada/ChangeLog:
* sem_attr.adb (Eval_Attribute): Only fold array attributes when prefix
is static or at least safe to evaluate
Tested on x86_64-pc-linux-gnu, committed on master.
---
From: Ronan Desplanques
gcc/ada/ChangeLog:
* einfo.ads (Is_Controlled_Active): Fix pasto in comment.
* sem_util.ads (Propagate_Controlled_Flags): Update comment for
Destructor aspect.
Tested on x86_64-pc-linux-gnu, committed on master.
---
gcc/ada/einfo.ads| 2 +-
From: Ronan Desplanques
Before this patch, Make_Init_Call and Make_Adjust_Call made the
assumption that if the type they were called with was untagged and a
derived type, it was the untagged private view of a tagged type. That
assumption made it possible to inspect the root type's primitives to
h
Hi,
In afdo_propagate_edge, annotate edge from the connected
basic_block if they have single predcessor/succcessor.
Tested with spec2017 intrate/intspeed and shows slight improvement.
gcc/ChangeLog:
2025-07-21 Kugan Vivekanandarajah
* auto-profile.cc (afdo_propagate_edge): Propagate
From: Denis Mazzucato
This patch makes sure that we return the same decision for all aliased
types when checking if the BIP task extra actuals are needed.
gcc/ada/ChangeLog:
* sem_ch6.adb (Might_Need_BIP_Task_Actuals): Before retrieving the
original corresponding
operation we r
From: Javier Miranda
The compiler generates wrong code in a dispatching call on result
when the call is performed under dependent conditional expressions
or case-expressions.
gcc/ada/ChangeLog:
* sinfo.ads (Is_Expanded_Dispatching_Call): New flag.
(Tag_Propagated): New flag.
Hi,
This patch fixes the vf_vfmacc-run-1-f16.c test failures on rv32
by adding zvfh requirements as well as options to the test and
the target harness.
Regtested on rv64gcv_zvl512b and rv32gcv_zvl512b. Going to commit
as obvious if the CI agrees that it's obvious ;)
Regards
Robin
gcc/testsuit
Hi,
During the last weeks it became clear that our current broadcast
handling needs an overhaul in order to improve maintainability.
PR121073 showed that my intermediate fix wasn't enough and caused
regressions.
This patch now goes a first step towards untangling broadcast
(vmv.v.x), "set first"
From: Gary Dismukes
This change test an additional condition as part of the criteria used
for deciding whether to generate a call to a container type's Length
function (for passing to the Empty function) when determining the
size of the object to allocate for a bounded container aggregate
with a
From: Piotr Trojanek
When checking restriction No_Relative_Delay and detecting calls to
Ada.Real_Time.Timing_Events.Set_Handler with a Time_Span parameter,
we looked at the exact type of the actual parameter, while we should
look at its base type.
This patch looks at the type of actual parameter
From: Steve Baird
Fix bugs related to mutably tagged types in streaming operations, Put_Image
attributes, aggregates, composite equality comparisons with mutably-tagged
components, and other issues.
gcc/ada/ChangeLog:
* exp_aggr.adb (Build_Record_Aggr_Code.Gen_Assign): In the case of
All functions in testsuite_iterators.h are now marked constexpr,
targeting the earliest possible standard. Most functions use C++14 due
to multi-statement bodies, with exceptions:
* BoundsContainer and some constructors are C++11 compatible.
* OutputContainer is C++20 due to operator new/delete us
From: Ronan Desplanques
This patch adds a GNAT-specific extension which enables "destructors".
Destructors are an optional replacement for Ada.Finalization where some
aspects of the interaction with type derivation are different.
gcc/ada/ChangeLog:
* doc/gnat_rm/gnat_language_extensions
On Mon, Jul 21, 2025 at 10:18 AM Luc Grosheintz
wrote:
> This commit implements and tests the function is_sufficiently_aligned
> from P2897R7.
>
> PR libstdc++/120994
>
> libstdc++-v3/ChangeLog:
>
> * include/bits/align.h (is_sufficiently_aligned): New function.
> * includ
From: Gary Dismukes
Minor change to satisfy GNAT SAS checker.
gcc/ada/ChangeLog:
* exp_aggr.adb (Build_Size_Expr): Change test of "not Present (...)"
to "No (...)".
Tested on x86_64-pc-linux-gnu, committed on master.
---
gcc/ada/exp_aggr.adb | 3 +--
1 file changed, 1 inserti
Prompted by the discussions around a recent clang bug, I realized that
gcc still defaults to -mcpu=v9 on Solaris/SPARC.
This is an oversight since the Oracle Studio 12.6 cc, released in 2017,
already defaults to -xarch=sparcvis2, the equivalent of
-mcpu=ultrasparc3. Besides, both the 32 and 64-bi
From: Eric Botcazou
The assertion is:
pragma Assert (Side_Effect_Free (L));
in Make_Tag_Ctrl_Assignment and demonstrates that the sequence:
Remove_Side_Effects (L);
pragma Assert (Side_Effect_Free (L));
does not hold in this case.
What happens is that Remove_Side_Effects uses a ren
On Mon, Jul 21, 2025 at 5:51 PM Luc Grosheintz
wrote:
> This commit implements the C++26 feature std::dims described in P2389R2.
> It sets the feature testing macro to 202406 and adds tests.
>
> Also fixes the test mdspan/version.cc
>
> libstdc++-v3/ChangeLog:
>
> * include/bits/version.d
From: Eric Botcazou
That's a kludge added to work around the limitations of the stack checking
mechanism used in the early days.
gcc/ada/ChangeLog:
* exp_util.ads (May_Generate_Large_Temp): Delete.
* exp_util.adb (May_Generate_Large_Temp): Likewise.
(Safe_Unchecked_Type_
1 - 100 of 114 matches
Mail list logo
