On 01/11/2024 19:18, Torbjorn SVENSSON wrote:
>
>
> On 2024-11-01 19:40, Richard Earnshaw (lists) wrote:
>> On 24/10/2024 09:50, Torbjörn SVENSSON wrote:
>>> Ok for trunk and releases/gcc-14?
>>>
>>> --
>>>
>>> As these tests are set to execute and require neon hardware to do so,
>>> add the mis
"Yuta Mukai (Fujitsu)" writes:
> Thank you for the reviews! I attached a patch that fixes the problems.
>
>>> On 31 Oct 2024, at 11:50, Richard Sandiford
>>> wrote:
>>>
>>> "Yuta Mukai (Fujitsu)" writes:
Hello,
This patch adds initial support for FUJITSU-MONAKA CPU, which we ar
On 2024-11-04 12:44, Richard Earnshaw (lists) wrote:
On 01/11/2024 19:18, Torbjorn SVENSSON wrote:
On 2024-11-01 19:40, Richard Earnshaw (lists) wrote:
On 24/10/2024 09:50, Torbjörn SVENSSON wrote:
Ok for trunk and releases/gcc-14?
--
As these tests are set to execute and require neon h
As Yuta Mukai pointed out, the manual wrongly said that LS64 is
enabled by default for Armv8.7-A and above, and for Armv9.2-A
and above. LS64 is not mandatory at any architecture level
(and the code correctly implemented that).
I think this was a leftover from an early version of the spec.
gcc/
On Sun, Nov 3, 2024 at 11:23 PM Lewis Hyatt wrote:
>
> Hello-
>
> There is no shortage of PRs complaining about things that go wrong when the
> line_maps data structure in libcpp starts to run into its limits. Being
> restricted to a 32-bit location_t to cover all source locations means it has
> t
On Sat, Nov 02, 2024 at 08:25:45AM +0100, Jakub Jelinek wrote:
> So, I think the main question to decide is if we want to keep existing
> standard constraints work as is even in toplevel asm and add new constraints
> which will allow more stuff (like "i" which will just allow SYMBOL_REFs
> even wit
> On 4 Nov 2024, at 13:55, Richard Biener wrote:
>
> On Thu, Oct 31, 2024 at 4:30 PM Jeff Law wrote:
>>
>>
>>
>> On 10/27/24 10:21 AM, Kyrylo Tkachov wrote:
>>> Hi all,
>>>
>>> simplify-rtx can transform (X << C1) | (X >> C2) into ROTATE (X, C1) when
>>> C1 + C2 == mode-width. But the tra
Patch 1-5 of v1 have already been pushed. This is v2 of patch 6 and 7
of that series.
Changes since v1:
RISC-V: Make vectorized memset handle more cases
* Removed vector memset loop generation.
RISC-V: Disable by pieces for vector setmem length > UNITS_PER_WORD
* No changes.
gcc/config/riscv/ri
For fast unaligned access targets, by pieces uses up to UNITS_PER_WORD
size pieces resulting in more store instructions than needed. For
example gcc.target/riscv/rvv/base/setmem-2.c:f1 built with
`-O3 -march=rv64gcv -mtune=thead-c906`:
```
f1:
vsetivlizero,8,e8,mf2,ta,ma
vm
`expand_vec_setmem` only generated vectorized memset if it fitted into a
single vector store of at least (TARGET_MIN_VLEN / 8) bytes. Also,
without dynamic LMUL the operation was always TARGET_MAX_LMUL even if it
would have fitted a smaller LMUL.
Allow vectorized memset to be generated for smalle
> On 4 Nov 2024, at 15:20, Jakub Jelinek wrote:
>
> On Mon, Nov 04, 2024 at 02:31:29PM +0100, Jakub Jelinek wrote:
>> On Mon, Nov 04, 2024 at 01:07:33PM +, Kyrylo Tkachov wrote:
This seems to have broken bootstrap on multiple targets and is causing
the arm CI to hang, miscompiling
Hi,
Joseph pointed out a pretty major flaw that this previously
did not merge fmv declarations and definitons or check properly
for duplicate definitons.
I rewrote the fmv logic and moved it to pushdecl resulting in a
patch that I much prefer.
The logic in pushdecl now detects the FMV case and a
Hi Jason,
On 1 Nov 2024, at 16:31, Jason Merrill wrote:
> On 11/1/24 5:07 AM, Simon Martin wrote:
>> Since r10-3793-g1a37b6d9a7e57c, we ICE upon the following valid code
>> with -std=c++17 and above
>>
>> === cut here ===
>> struct Base {
>>unsigned int *intarray;
>> };
>> template struct Su
> Am 04.11.2024 um 16:01 schrieb Andrew MacLeod :
>
> The invert() range operation is not supported on values of either VARYING or
> UNDEFINED. Primarily this is because UNDEFINED has no type, which makes it
> impossible to perform invert() twice on a value, and produce that same value.
>
On 11/4/24 10:27, Richard Biener wrote:
Am 04.11.2024 um 16:01 schrieb Andrew MacLeod :
The invert() range operation is not supported on values of either VARYING or
UNDEFINED. Primarily this is because UNDEFINED has no type, which makes it
impossible to perform invert() twice on a value,
> Am 04.11.2024 um 16:32 schrieb Andrew MacLeod :
>
>
>> On 11/4/24 10:27, Richard Biener wrote:
>>
Am 04.11.2024 um 16:01 schrieb Andrew MacLeod :
>>>
>>> The invert() range operation is not supported on values of either VARYING
>>> or UNDEFINED. Primarily this is because UNDEFINED
Samuel Thibault writes:
> I reworked the patch to factorize the bsd signal definitions.
> I have split off the system definitions because the priority range of
> GNU/Mach has diverged from the original BSD kernels.
> gcc/ada/ChangeLog | 17 ++
Hello,
The changes lo
> On 4 Nov 2024, at 16:03, Kyrylo Tkachov wrote:
>
>
>
>> On 4 Nov 2024, at 15:20, Jakub Jelinek wrote:
>>
>> On Mon, Nov 04, 2024 at 02:31:29PM +0100, Jakub Jelinek wrote:
>>> On Mon, Nov 04, 2024 at 01:07:33PM +, Kyrylo Tkachov wrote:
> This seems to have broken bootstrap on multiple
On Mon, Nov 04, 2024 at 03:43:29PM +, Kyrylo Tkachov wrote:
> Here’s a rework of the original patch which now doesn’t call simplify_rtx on
> its operands in the new simplify_rotate_op.
> It wasn’t necessary to do the necessary optimization and it was blowing up
> the recursion depth needlessl
On 20/10/2024 16:51, Torbjörn SVENSSON wrote:
> Ok for trunk and releases/gcc-14?
>
> --
>
> The tests assume that a neon fpu is avialable and fails it not, so
> explicitly require it.
> diff --git a/gcc/testsuite/gcc.target/arm/attr-neon2.c
> b/gcc/testsuite/gcc.target/arm/attr-neon2.c
> index
On 31/10/2024 18:24, Torbjörn SVENSSON wrote:
> Ok for trunk and releases/gcc-14?
>
> --
>
> The test case is written in a way that it should be using hard float
> ABI, but the use of -mfloat-abi=hard could be overriden by
> dg-add-options arm_neon. Ensure that -mfloat-abi=hard is always after.
>
Am 04.11.2024 um 16:32 schrieb Andrew MacLeod :
On 11/4/24 10:27, Richard Biener wrote:
Am 04.11.2024 um 16:01 schrieb Andrew MacLeod :
The invert() range operation is not supported on values of either VARYING or
UNDEFINED. Primarily this is because UNDEFINED has no type, which make
On 31/10/2024 18:26, Torbjörn SVENSSON wrote:
> Ok for trunk and releases/gcc-14?
>
> --
>
> Tests uses neon, so add effective-target arm_neon.
>
> gcc/testsuite/ChangeLog:
>
> * gcc.target/arm/pr68620.c: Use effective-target arm_neon.
> * gcc.target/arm/pr78041.c: Likewise.
>
> Si
From: Piotr Trojanek
Code cleanup; semantics is unaffected.
gcc/ada/ChangeLog:
* sem_res.adb (In_Decl): Rename and move local variable where
it is used.
Tested on x86_64-pc-linux-gnu, committed on master.
---
gcc/ada/sem_res.adb | 16
1 file changed, 8 insert
From: Javier Miranda
When a derived tagged type implements interface types in addition
to deriving from its parent type, and a primitive inherited from
its parent type corresponds to an inherited primitive that has
class-wide preconditions, then the generated code fails to check
the class-wide pr
From: Piotr Trojanek
Resolution of binary and unary intrinsic operators differed when
expansion was inactive. In particular, this affected GNATprove
handling of Ada.Real_Time."abs" operator. This patch makes unary
resolution behave like binary resolution.
gcc/ada/ChangeLog:
* sem_res.ad
From: Piotr Trojanek
Code cleanup.
gcc/ada/ChangeLog:
* exp_ch4.adb (Expand_N_Op_Multiply): Remove extra whitespace.
Tested on x86_64-pc-linux-gnu, committed on master.
---
gcc/ada/exp_ch4.adb | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/gcc/ada/exp_ch4.a
From: Piotr Trojanek
By declaring access parameter with non-null qualifier, the compiler
should avoid generating run-time checks in debug builds, resulting in
a tiny performance improvement.
Code cleanup; semantics is unaffected.
gcc/ada/ChangeLog:
* sem_res.adb (Type_In_P): Add non-nu
From: Claire Dross
The formal containers have been part of the SPARK library for some
time now. However, some units used only by these containers are still
part of the gnat repository. Move them to the SPARK library.
gcc/ada/ChangeLog:
* Makefile.rtl: Remove references to moved units.
From: Ronan Desplanques
The primary motivation for this change is making the taskset command
line tool work as expected for tasking programs that don't use features
from section D.16 of the Ada reference manual. A couple of components
are added to the ATCB record to make it possible to tell value
From: Piotr Trojanek
Code cleanup and tiny performance improvement; semantics is unaffected.
gcc/ada/ChangeLog:
* exp_ch4.adb (Expand_N_Op_Subtract): Replace numeric literal
with universal integer constant, just like it is done in
expansion of addition operator.
Tested
From: Eric Botcazou
They are related to the special support for text encoding on Windows.
gcc/ada/ChangeLog:
* adaint.c: Replace initialize.c with rtinit.c in comment.
* sysdep.c (__gnat_set_mode): Fix reference in comment.
* libgnat/i-cstrea.ads (Content_Encoding): Adju
From: Viljar Indus
Add an invocation node to the SARIF report that contains the
command line use to activate gnat and whether the execution was
successful or not.
gcc/ada/ChangeLog:
* diagnostics-sarif_emitter.adb (Print_Runs): Add printing for
the invocation node that consists
From: Daniel King
Exception declarations require elaboration on the full run-time to
register the exceptions. The package Interfaces.CHERI, however, is
used on bare-metal targets during early initialization, before
elaboration and is therefore marked No_Elaboration_Code_All.
Refactoring the excep
From: Eric Botcazou
The front-end raises Program_Error on an alignment clause for a type in a
generic unit that references the alignment of another type in the unit.
gcc/ada/ChangeLog:
PR ada/117051
* freeze.adb (Freeze_Entity): Call the layout procedure on subtypes
decl
From: Piotr Trojanek
When uninstalling private package declarations we must mark Taft
amendment types hidden, just like we mark other types.
Looking at previous revisions of this code, it is quite clear that this
bug comes from a code evolution and marking types should happen in all
ELSE branche
From: Eric Botcazou
The existing comment is a bit too vague.
gcc/ada/ChangeLog:
* exp_aggr.ads (Is_Two_Pass_Aggregate): Beef up comment.
Tested on x86_64-pc-linux-gnu, committed on master.
---
gcc/ada/exp_aggr.ads | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/
From: Javier Miranda
When the type imposed by the context for an interpolated string is
constrained, the compiler silently omits adding a runtime check.
gcc/ada/ChangeLog:
* exp_ch2.adb (Expand_N_Interpolated_String_Literal): Use the
base type of the type imposed by the context
From: Daniel King
SIGPROT is a new signal on CheriBSD that signals a CHERI protection violation.
The full runtime converts these to the appropriate Ada exception declared in
Interfaces.CHERI.Exceptions.
gcc/ada/ChangeLog:
* Makefile.rtl: Build support for Morello CheriBSD.
* lib
From: Viljar Indus
Imporve the wording to explicitly state which options are turned on
by -gnatVa and that -gnatVd is enabled by default.
It can be somewhat hard to decifer that information from the old
wording. Especially when compared to -gnatWxx switches where there
is an elaborate scheme for
From: Ronan Desplanques
Before this patch, nothing was reported to the user when an exception
was raised during generation of a minimal reproducer. This patch fixes
this.
gcc/ada/ChangeLog:
* comperr.adb (Compiler_Abort): Display message in exception handler.
Tested on x86_64-pc-linux-
From: Raphaël AMIARD
gcc/ada/ChangeLog:
* doc/gnat_rm/gnat_language_extensions.rst: Adjust documentation.
* gnat_rm.texi: Regenerate.
* gnat_ugn.texi: Regenerate.
Tested on x86_64-pc-linux-gnu, committed on master.
---
.../doc/gnat_rm/gnat_language_extensions.rst | 14
From: Eric Botcazou
It has gone unnoticed for decades because it changes nothing in practice.
gcc/ada/ChangeLog:
* gcc-interface/trans.cc (process_decls): Remove tests on Nkind that
contain a typo and would be redundant if written correctly.
Tested on x86_64-pc-linux-gnu, commi
From: Eric Botcazou
The problem is that the freeze node for the type of the element ends up in
the component list of the record type declared with the default value.
gcc/ada/ChangeLog:
PR ada/113036
* freeze.adb (Freeze_Expression): Deal with freezing actions coming
from
From: Raphaël AMIARD
gcc/ada/ChangeLog:
* sem_prag.adb (Analyze_Pragma): Fix format for second line of
warning (should be a continuation line)
Tested on x86_64-pc-linux-gnu, committed on master.
---
gcc/ada/sem_prag.adb | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
di
From: Viljar Indus
gcc/ada/ChangeLog:
* diagnostics-sarif_emitter.adb (Print_SARIF_Report): Add a
Schema field to the SARIF report.
Tested on x86_64-pc-linux-gnu, committed on master.
---
gcc/ada/diagnostics-sarif_emitter.adb | 14 +-
1 file changed, 13 insertions(
From: Piotr Trojanek
Intrinsic operators are resolved by rewriting into a corresponding
operator from the Standard package. Traversing homonyms just to find the
corresponding operator was not particularly efficient; also, for the
binary "-" it was finding the unary "-".
There appears to be no di
From: Bob Duff
The Library_Unit field was used for all sorts of different purposes,
which led to confusing code.
This patch splits Library_Unit into much more specific wrapper
subprograms that should be called instead of [Set_]Library_Unit.
Predicates and pragmas Assert are used to catch misuses
From: Daniel King
gcc/ada/ChangeLog:
* Makefile.rtl: Use s-secsta__cheri.adb on Morello CheriBSD.
* libgnat/s-secsta__cheri.adb: New file.
Tested on x86_64-pc-linux-gnu, committed on master.
---
gcc/ada/Makefile.rtl|3 +-
gcc/ada/libgnat/s-secsta__cheri.adb
From: Eric Botcazou
The resolution status of Resolve_Aggr_Expr is lost when the routine is
invoked indirectly from Resolve_Iterated_Component_Association.
gcc/ada/ChangeLog:
* sem_aggr.adb (Resolve_Iterated_Component_Association): Change to
function returning Boolean and return
From: Bob Duff
Follow-on to previous change "Disable self-referential with_clauses",
which caused some regressions. Remove useless use clauses referring
to useless self-referential with'ed packages. This is necessary
because in some cases, such use clauses cause the compiler to
crash or give spur
From: Nicolas Roche
Improve performance of iteration using Element function.
Improve performance of Append.
gcc/ada/ChangeLog:
* libgnat/a-strunb__shared.adb: Restructure code to inline only
the most common cases. Remove whenever possible runtime checks.
* libgnat/a-stru
From: Piotr Trojanek
Previously the lower bound of string literals indexed by non-static
integer types was artificially set to 1 in the frontend. This was to
avoid an overflow in calculation of a null string size by the GCC
backend, which was causing an excessively large binary object file.
Howe
From: Eric Botcazou
It dates back to when analysis was performed on a copy of the expression.
gcc/ada/ChangeLog:
* sem_aggr.adb (Resolve_Iterated_Component_Association): Move up
declaration of Expr and remove dead code from older processing.
Tested on x86_64-pc-linux-gnu, commi
From: Piotr Trojanek
Apparently we no longer need to ignore string literal subtypes case
when validating size of a type entity.
Code cleanup; behavior appears to be unaffected.
gcc/ada/ChangeLog:
* gcc-interface/decl.cc (gnat_to_gnu_entity): Remove special
case for string liter
On Tue, Oct 22, 2024 at 07:18:55PM +0200, Torbjorn SVENSSON wrote:
>
>
> On 2024-10-22 13:36, Richard Earnshaw (lists) wrote:
> > On 20/10/2024 16:48, Torbj�rn SVENSSON wrote:
> > > Ok for trunk and releases/gcc-14?
> > >
> > > --
> > >
> > > Converted the tests to use check-function-bodies
On 01/11/2024 18:40, Richard Earnshaw (lists) wrote:
> On 24/10/2024 09:50, Torbjörn SVENSSON wrote:
>> Ok for trunk and releases/gcc-14?
>>
>> --
>>
>> As these tests are set to execute and require neon hardware to do so,
>> add the missing dg-require-effective-target arm_neon_hw.
>>
>> gcc/testsu
I've back ported and committed the 3 patch set for assume p;rocessing to
GCC14. The new file is almost the same, just a couple of minor
differences related to
1) value_range in gcc15 was Value_Range in gcc14, and
2) there was no gori() component to a range_query, so I had to pass a
ranger
The invert() range operation is not supported on values of either
VARYING or UNDEFINED. Primarily this is because UNDEFINED has no type,
which makes it impossible to perform invert() twice on a value, and
produce that same value. There were also times when it wasn't precisely
clear what the c
When we end up expanding a SSA name copy with BLKmode regs which can
happen for vectors, possibly wrapped in a NOP-conversion or
a PAREN_EXPR and we are not optimizing we can end up with two
BLKmode MEMs that expand_gimple_stmt_1 doesn't properly handle
when expanding, trying to emit_move_insn them
Hi Richard,
On Fri, 18 Oct 2024 at 13:44, Richard Biener wrote:
>
> The following makes -fwrapv explicit.
>
> * gcc.dg/vect/vect.exp: Remove special-casing of tests
> named wrapv-*
It looks like this does a bit more than that.
As can be seen in
https://gcc.gnu.org/pipermail/gcc-r
On Mon, Nov 04, 2024 at 11:48:16AM +0100, Richard Biener wrote:
> When we end up expanding a SSA name copy with BLKmode regs which can
> happen for vectors, possibly wrapped in a NOP-conversion or
> a PAREN_EXPR and we are not optimizing we can end up with two
> BLKmode MEMs that expand_gimple_stmt
On Mon, Nov 04, 2024 at 02:31:29PM +0100, Jakub Jelinek wrote:
> On Mon, Nov 04, 2024 at 01:07:33PM +, Kyrylo Tkachov wrote:
> > > This seems to have broken bootstrap on multiple targets and is causing
> > > the arm CI to hang, miscompiling stage2 - can
> > > you please revert for now?
> >
> >
On 10/31/24 6:59 PM, H.J. Lu wrote:
On Fri, Nov 1, 2024 at 8:32 AM Sam James wrote:
"H.J. Lu" writes:
When gdb defaults to use debuginfod, gdb warns guality tests:
Spawning: gdb -nx -nw -quiet -batch -x pr36728-2.gdb ./pr36728-2.exe
spawn gdb -nx -nw -quiet -batch -x pr36728-2.gdb ./pr3
From: Viljar Indus
When the Extend_System pragma is used then we are supposed
to check the extended system for referenced entities. Otherwise
we would get an incorrect unused with warning.
This was previously done on body files but it should also be
done specs as well.
gcc/ada/ChangeLog:
Hi all,
The gcc-patches list recently saw a couple of bounces of emails that
didn't pass dmarc policy (with p=reject) with spf and dkim failing.
All people who got unsubscribed because of this have been resubscribed
again.
To prevent getting email on the list that don't pass a strict dmarc
polic
On 21/10/2024 06:56, François Dumont wrote:
On 17/10/2024 23:11, Jonathan Wakely wrote:
On Thu, 17 Oct 2024 at 21:39, Jonathan Wakely
wrote:
On Thu, 17 Oct 2024 at 20:52, François Dumont
wrote:
Here is an updated version that compiles, I think, all your
feedb
On Oct 20, 2024, Alexandre Oliva wrote:
> the x86 hits are newer, and
> the line numbers didn't help me find the errors, so I don't even know
> whether it's macros that we're talking about. These new warnings were
> encouraging, for they showed that the ifcombine incarnation of the patch
> cover
On 11/4/24 8:13 PM, H.J. Lu wrote:
On Tue, Nov 5, 2024 at 10:57 AM Jeff Law wrote:
On 11/4/24 7:52 PM, H.J. Lu wrote:
On Tue, Nov 5, 2024 at 8:48 AM Jeff Law wrote:
On 11/4/24 5:42 PM, H.J. Lu wrote:
On Tue, Nov 5, 2024 at 8:07 AM Jeff Law wrote:
On 11/1/24 4:32 PM, H.J. Lu wr
Ping and also CC Jeff and Kito. This is also useful for RISC-V since
RISC-V uses target_version, which defines
TARGET_HAS_FMV_TARGET_ATTRIBUTE to zero.
On 10/28/24 00:00, Yangyu Chen wrote:
Following the implementation of commit b8ce8129a5 ("Redirect call
within specific target attribute among
On Tue, Nov 5, 2024 at 2:41 PM Hu, Lin1 wrote:
>
> > -Original Message-
> > From: Hu, Lin1
> > Sent: Tuesday, November 5, 2024 1:34 PM
> > To: gcc-patches@gcc.gnu.org
> > Cc: Liu, Hongtao ; ubiz...@gmail.com
> > Subject: [PATCH v2] i386: Handling exception input of
> > __builtin_ia32_pref
On 02/11/2024 12:58, Robin Dapp wrote:
From: Robin Dapp
This patch adds an undefined else operand to the masked loads.
gcc/ChangeLog:
* config/gcn/predicates.md (maskload_else_operand): New
predicate.
* config/gcn/gcn-valu.md: Use new predicate.
---
gcc/config/gcn/gc
This patch merges two isomorphic vector sequences by using the
redundancy in the lane utilization in these sequences.
This redundancy in lane utilization comes from the way how specific
scalar statements end up vectorized: two VEC_PERMs on top, binary operations
on both of them, and a final VEC_PER
On Wed, Oct 30, 2024 at 11:26 AM Christoph Müllner
wrote:
>
> On Fri, Oct 18, 2024 at 1:08 PM Richard Biener wrote:
> >
> > On Fri, 18 Oct 2024, Tamar Christina wrote:
> >
> > > > -Original Message-
> > > > From: Richard Biener
> > > > Sent: Friday, October 18, 2024 11:03 AM
> > > > To:
Hi Joseph,
Thank you for the review and nice to meet you.
@@ -3191,6 +3240,14 @@ duplicate_decls (tree newdecl, tree olddecl)
return false;
}
+ /* If both new and old are Function Multi Versioned functions then they are
+ not duplicates. */
+ if (TREE_CODE (newdecl) == FUNCTION_
This patch adds stmt_vec_info to TARGET_SIMD_CLONE_USABLE to make sure the
target can reject a simd_clone based on the vector mode it is using.
This is needed because for VLS SVE vectorization the vectorizer accepts
Advanced SIMD simd clones when vectorizing using SVE types because the simdlens
mig
On 10/21/24 3:18 AM, Richard Biener wrote:
Bootstrapped and tested on x86_64-linux-gnu.
PR tree-optimization/112418
gcc/ChangeLog:
* tree-ssa-phiopt.cc (is_factor_profitable): New function.
(factor_out_conditional_operation): Add merge argument. Remove
arg0/arg
On 11/1/24 10:44 AM, Uros Bizjak wrote:
On Fri, Nov 1, 2024 at 3:51 PM Jeff Law wrote:
On 11/1/24 8:45 AM, Uros Bizjak wrote:
On Fri, Nov 1, 2024 at 2:18 PM Jeff Law wrote:
On 11/1/24 5:34 AM, Uros Bizjak wrote:
Stack pointer modifications in asm are currently not flagged in
crtl->sp
This patch adds stmt_vec_info to TARGET_SIMD_CLONE_USABLE to make sure the
target can reject a simd_clone based on the vector mode it is using.
This is needed because for VLS SVE vectorization the vectorizer accepts
Advanced SIMD simd clones when vectorizing using SVE types because the simdlens
mig
On 11/4/24 10:19 AM, Simon Martin wrote:
Hi Jason,
On 1 Nov 2024, at 16:31, Jason Merrill wrote:
On 11/1/24 5:07 AM, Simon Martin wrote:
Since r10-3793-g1a37b6d9a7e57c, we ICE upon the following valid code
with -std=c++17 and above
=== cut here ===
struct Base {
unsigned int *intarray;
}
On 11/1/24 3:17 PM, Jakub Jelinek wrote:
On Fri, Nov 01, 2024 at 12:28:52PM -0400, Jason Merrill wrote:
Can this block be
if (aligned_allocation_fn_p (decl))
args = TREE_CHAIN (args);
?
Yes.
+ && TREE_CODE (t) == RECORD_TYPE
+ && DECL_NAMESPACE_ST
On 10/31/24 9:50 AM, Kyrylo Tkachov wrote:
Is there a reason why we don't have the target reject the vector rotation cases
it can't natively handle and the expander code would then try the rotate via
permuation?
It seems like your patch defers everything to the target which calls back int
The current codegen code to support VF's that are multiples of a simdclone
simdlen rely on BIT_FIELD_REF to create multiple input vectors. This does not
work for non-constant simdclones, so we should disable using such clones when
the VF is a multiple of the non-constant simdlen until we change th
On 10/31/24 1:35 PM, Vineet Gupta wrote:
An INSN can have multiple dependencies/predecessor nodes, some of them
being true dependency REG_DEP_TRUE meaning the predecessor register output
is a must have for the INSN to be scheduled.
e.g.
In the sched1 dump below, insn 70 has multiple deps, bu
On 11/1/24 1:22 PM, Simon Martin wrote:
We currently crash upon the following invalid code (notice the "void
void**" parameter)
=== cut here ===
using size_t = decltype(sizeof(int));
void *operator new(size_t, void void **p) noexcept { return p; }
int x;
void f() {
int y;
new (&y) int(
On 11/4/24 5:42 PM, H.J. Lu wrote:
On Tue, Nov 5, 2024 at 8:07 AM Jeff Law wrote:
On 11/1/24 4:32 PM, H.J. Lu wrote:
For targets, like x86, which define TARGET_PROMOTE_PROTOTYPES to return
true, all integer arguments smaller than int are passed as int:
[hjl@gnu-tgl-3 pr14907]$ cat x.c
e
On 11/1/24 2:19 AM, KuanLin Chen wrote:
Hi Jeff,
I'm really sorry for the regression failure.
I missed one patch to fix these issues.
Thanks for your review.
The GTY skip makes GGC clean the registered functions wrongly in lto.
Example:
riscv64-unknown-elf-gcc -flto gcc/testsuite/gcc.target
On 11/1/24 4:32 PM, H.J. Lu wrote:
For targets, like x86, which define TARGET_PROMOTE_PROTOTYPES to return
true, all integer arguments smaller than int are passed as int:
[hjl@gnu-tgl-3 pr14907]$ cat x.c
extern int baz (char c1);
int
foo (char c1)
{
return baz (c1);
}
[hjl@gnu-tgl-3 pr149
Following a few bugfixes in the if-convert pass which were previously
found to degrade performance in the proposed SVE libmvec
autovectorization of conditional calls to math functions, this
patch-series carries on the work initially presented by Andre Vieira
and last discussed at:
- https://patch
On 10/30/24 7:06 PM, Jeff Law wrote:
So this patch is a very conservative approach to eliminate more vsetvl
instructions.
As we know, scheduling can scramble the instruction stream based on a
variety of factors and can easily result in an instruction sequence
where we ping-pong between dif
This patch finalizes adding support for the generation of SVE simd clones when
no simdlen is provided, following the ABI rules where the widest data type
determines the minimum amount of elements in a length agnostic vector.
gcc/ChangeLog:
* config/aarch64/aarch64-protos.h (add_sve_type_a
On 11/2/24 7:05 AM, Nathaniel Shead wrote:
On Fri, Nov 01, 2024 at 11:23:41AM -0400, Jason Merrill wrote:
On 8/22/24 8:02 PM, Nathaniel Shead wrote:
On Thu, Aug 22, 2024 at 02:20:14PM -0400, Patrick Palka wrote:
On Mon, 12 Aug 2024, Nathaniel Shead wrote:
Bootstrapped and regtested on x86_64
gcc/testsuite/ChangeLog:
* c-c++-common/gomp/declare-variant-14.c: Make i?86 and x86_64 target
only test.
* gfortran.dg/gomp/declare-variant-14.f90: Likewise.
---
gcc/testsuite/c-c++-common/gomp/declare-variant-14.c | 12 +---
.../gfortran.dg/gomp/declare-variant-1
cc'ing Jakub due to email address typo in original patch submission.
Apologies,
Victor
Victor Do Nascimento writes:
> gcc/testsuite/ChangeLog:
>
> * c-c++-common/gomp/declare-variant-14.c: Make i?86 and x86_64 target
> only test.
> * gfortran.dg/gomp/declare-variant-14.f90:
On 11/4/24 5:35 PM, Nathaniel Shead wrote:
This patch is missing a testcase because I wasn't able to easily
construct one that reliably exhausts all location_t values and then
causes an error.
Bootstrapped and regtested on x86_64-pc-linux-gnu, OK for trunk?
OK.
-- >8 --
The 'location_t' typ
On Tue, Nov 5, 2024 at 8:07 AM Jeff Law wrote:
>
>
>
> On 11/1/24 4:32 PM, H.J. Lu wrote:
> > For targets, like x86, which define TARGET_PROMOTE_PROTOTYPES to return
> > true, all integer arguments smaller than int are passed as int:
> >
> > [hjl@gnu-tgl-3 pr14907]$ cat x.c
> > extern int baz (cha
On 10/31/24 4:40 PM, Andi Kleen wrote:
From: Andi Kleen
autofdo looks up inline stacks and tries to match them with the profile
data using their symbol name. Make sure all decls that can be in a inline stack
have a valid assembler name.
This fixes a bootstrap problem with autoprofiledbootstrap
Hi Jeff:
This should work with the second patch in the patch set, but I think that's
not ready yet...
Jeff Law 於 2024年11月5日 週二 08:52 寫道：
>
>
> On 11/1/24 2:19 AM, KuanLin Chen wrote:
> > Hi Jeff,
> >
> > I'm really sorry for the regression failure.
> > I missed one patch to fix these issues.
>
On Fri, Nov 1, 2024 at 11:24 AM Haochen Jiang wrote:
>
> Hi all,
>
> I have just landed new ISA patches on trunk. The next step will
> be the arch support for ISE055 mentioned CPUs.
>
> There are two changes in ISE055 on CPUs:
>
> - A new model number is added for Arrow Lake.
> - Diamond Rapid
This is version 2 of the patch to add functions to allow high performance
libraries to write their code so that they can do streaming math functions
using the Power10 load vector pair and store vector pair instructions.
This patch adds 2 functions to swap the even and odd elements. This is from
f
1 - 100 of 185 matches
Mail list logo
