labrinea wrote:
Ping
https://github.com/llvm/llvm-project/pull/110297
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/labrinea updated
https://github.com/llvm/llvm-project/pull/110297
>From 82047dc021979fe58001b15c740649c3d44dd23b Mon Sep 17 00:00:00 2001
From: Alexandros Lamprineas
Date: Fri, 27 Sep 2024 16:37:26 +0100
Subject: [PATCH] [FMV][AArch64] Unify features ssbs and ssbs2.
Accordin
@@ -1125,31 +1125,32 @@ int caller(void) { return
used_def_without_default_decl() + used_decl_without_de
// CHECK: attributes #[[ATTR10]] = { noinline nounwind optnone
"no-trapping-math"="true" "stack-protector-buffer-size"="8"
"target-features"="+fp-armv8,+ls64,+neon" }
// C
@@ -88,9 +88,9 @@ def : FMVExtension<"sve-bf16", "FEAT_SVE_BF16",
"+sve,+bf16,+fullfp16,+fp-armv8,
def : FMVExtension<"sve-ebf16", "FEAT_SVE_EBF16",
"+sve,+bf16,+fullfp16,+fp-armv8,+neon", 330>;
def : FMVExtension<"sve-i8mm", "FEAT_SVE_I8MM",
"+sve,+i8mm,+fullfp16,+fp-armv8,+
@@ -1117,39 +1117,41 @@ int caller(void) { return
used_def_without_default_decl() + used_decl_without_de
// CHECK: attributes #[[ATTR2]] = { noinline nounwind optnone
"no-trapping-math"="true" "stack-protector-buffer-size"="8"
"target-features"="+fp-armv8,+lse,+neon,+sha2" }
labrinea wrote:
Adding the GCC folks @andrewcarlotti and @Wilco1 for visibility.
https://github.com/llvm/llvm-project/pull/110816
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/labrinea updated
https://github.com/llvm/llvm-project/pull/110297
>From 82047dc021979fe58001b15c740649c3d44dd23b Mon Sep 17 00:00:00 2001
From: Alexandros Lamprineas
Date: Fri, 27 Sep 2024 16:37:26 +0100
Subject: [PATCH] [FMV][AArch64] Unify features ssbs and ssbs2.
Accordin
https://github.com/labrinea created
https://github.com/llvm/llvm-project/pull/110297
According to https://developer.arm.com/documentation/102105/latest Arm
Architecture Reference Manual for A-profile architecture: Known issues
2.206 D22789
In section C5.2.25 "SSBS, Speculative Store Bypass Saf
labrinea wrote:
Fix here -> https://github.com/llvm/llvm-project/pull/110291
https://github.com/llvm/llvm-project/pull/108383
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/labrinea closed
https://github.com/llvm/llvm-project/pull/101712
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/labrinea closed
https://github.com/llvm/llvm-project/pull/108383
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
labrinea wrote:
> I don't quite get the code, but it seems that +mte or +mte2 may be emitted in
> disassembly and/or supported by the assembler?
They are supported in assembler for the sole purpose of providing more
informative error messages.
https://github.com/llvm/llvm-project/pull/109299
labrinea wrote:
> From the patch it is clear all CPUs that implement MTE only support MTE2 -
> and it seems unlikely this will change.
Not quite true. I didn't change those CPU entries to keep the behavior NFC,
which may have been a mistake actually. The reference manual suggests that
Neovers
labrinea wrote:
ping
https://github.com/llvm/llvm-project/pull/108383
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/labrinea closed
https://github.com/llvm/llvm-project/pull/108024
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/labrinea updated
https://github.com/llvm/llvm-project/pull/108024
>From b6abb4f0fcf50d7498a6f0d0ef8b75808c8b1eb1 Mon Sep 17 00:00:00 2001
From: Alexandros Lamprineas
Date: Tue, 10 Sep 2024 14:02:17 +0100
Subject: [PATCH 1/2] [FMV][AArch64] Unify ls64, ls64_v and ls64_accdata.
labrinea wrote:
Just letting you know that I have reopened
https://github.com/llvm/llvm-project/pull/101712 as a draft just to demonstrate
that backwards compatibility can be preserved if we split the features
(previously I thought it wasn't possible, but I found a way). However I
couldn't co
labrinea wrote:
Adding the GCC folks @andrewcarlotti and @Wilco1 for visibility.
https://github.com/llvm/llvm-project/pull/109299
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/labrinea created
https://github.com/llvm/llvm-project/pull/109299
These extensions depend on each other as follows: mte -> mte2
Currently in LLVM FeatureMTE models both FEAT_MTE and FEAT_MTE2 lumped
together. However the architecture does not mandate that both need to be
im
https://github.com/labrinea updated
https://github.com/llvm/llvm-project/pull/108024
>From b6abb4f0fcf50d7498a6f0d0ef8b75808c8b1eb1 Mon Sep 17 00:00:00 2001
From: Alexandros Lamprineas
Date: Tue, 10 Sep 2024 14:02:17 +0100
Subject: [PATCH 1/2] [FMV][AArch64] Unify ls64, ls64_v and ls64_accdata.
https://github.com/labrinea updated
https://github.com/llvm/llvm-project/pull/108383
>From 50820f4f3ded438162a33eabca84c5077fd2a5aa Mon Sep 17 00:00:00 2001
From: Alexandros Lamprineas
Date: Thu, 12 Sep 2024 13:55:24 +0100
Subject: [PATCH] [FMV][AArch64] Remove feature sha1 from FMV.
Sha1 has
https://github.com/labrinea edited
https://github.com/llvm/llvm-project/pull/108024
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/labrinea updated
https://github.com/llvm/llvm-project/pull/108024
>From b6abb4f0fcf50d7498a6f0d0ef8b75808c8b1eb1 Mon Sep 17 00:00:00 2001
From: Alexandros Lamprineas
Date: Tue, 10 Sep 2024 14:02:17 +0100
Subject: [PATCH 1/2] [FMV][AArch64] Unify ls64, ls64_v and ls64_accdata.
https://github.com/labrinea updated
https://github.com/llvm/llvm-project/pull/108024
>From b6abb4f0fcf50d7498a6f0d0ef8b75808c8b1eb1 Mon Sep 17 00:00:00 2001
From: Alexandros Lamprineas
Date: Tue, 10 Sep 2024 14:02:17 +0100
Subject: [PATCH] [FMV][AArch64] Unify ls64, ls64_v and ls64_accdata.
Or
https://github.com/labrinea updated
https://github.com/llvm/llvm-project/pull/101712
>From 2fe8ef63846989952a1c72b83114aabe6f36637a Mon Sep 17 00:00:00 2001
From: Alexandros Lamprineas
Date: Fri, 2 Aug 2024 17:32:23 +0100
Subject: [PATCH 1/2] [AArch64] Split FeatureLS64 to LS64_ACCDATA and LS64
https://github.com/labrinea updated
https://github.com/llvm/llvm-project/pull/101712
>From 2fe8ef63846989952a1c72b83114aabe6f36637a Mon Sep 17 00:00:00 2001
From: Alexandros Lamprineas
Date: Fri, 2 Aug 2024 17:32:23 +0100
Subject: [PATCH] [AArch64] Split FeatureLS64 to LS64_ACCDATA and LS64_V.
@@ -3682,6 +3682,8 @@ static const struct Extension {
{"sve2-bitperm", {AArch64::FeatureSVE2BitPerm}},
{"sve2p1", {AArch64::FeatureSVE2p1}},
{"ls64", {AArch64::FeatureLS64}},
+{"ls64_v", {AArch64::FeatureLS64_V}},
+{"ls64_accdata", {AArch64::FeatureLS64_ACCD
https://github.com/labrinea converted_to_draft
https://github.com/llvm/llvm-project/pull/101712
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/labrinea updated
https://github.com/llvm/llvm-project/pull/101712
>From 2fe8ef63846989952a1c72b83114aabe6f36637a Mon Sep 17 00:00:00 2001
From: Alexandros Lamprineas
Date: Fri, 2 Aug 2024 17:32:23 +0100
Subject: [PATCH] [AArch64] Split FeatureLS64 to LS64_ACCDATA and LS64_V.
https://github.com/labrinea reopened
https://github.com/llvm/llvm-project/pull/101712
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
labrinea wrote:
This may sound like a bad idea in terms of future maintenance, but how about we
guarded the enum entries I am removing with
`#if __FUNCTION_MULTI_VERSIONING_SUPPORT_LEVEL < __ARM_ACLE_VERSION(2024, 3,
0)` ?
https://github.com/llvm/llvm-project/pull/108024
__
https://github.com/labrinea closed
https://github.com/llvm/llvm-project/pull/108857
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
labrinea wrote:
Thanks for reviewing!
https://github.com/llvm/llvm-project/pull/108857
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/labrinea updated
https://github.com/llvm/llvm-project/pull/108857
>From cc3f72918aadab7df7397c25de280e713d1257f2 Mon Sep 17 00:00:00 2001
From: Alexandros Lamprineas
Date: Mon, 16 Sep 2024 17:39:07 +0100
Subject: [PATCH 1/2] [AArch64] Add missing ACLE predefined macros and up
https://github.com/labrinea created
https://github.com/llvm/llvm-project/pull/108857
Adds __ARM_ACLE_VERSION and __FUNCTION_MULTI_VERSIONING_SUPPORT_LEVEL as
defined here https://github.com/ARM-software/acle/pull/301 and here
https://github.com/ARM-software/acle/pull/302.
Also bumps __ARM_ACL
@@ -73,8 +73,6 @@ enum CPUFeatures {
FEAT_SSBS,
FEAT_SSBS2,
FEAT_BTI,
- FEAT_LS64,
- FEAT_LS64_V,
labrinea wrote:
Thanks for flagging this up. I am puzzled though since I was under the
impression that FMV is still not production ready. When did the en
@@ -33,7 +33,6 @@ enum CPUFeatures {
FEAT_FP,
FEAT_SIMD,
FEAT_CRC,
- FEAT_SHA1,
labrinea wrote:
It was raised on another pr, so perhaps it is.
https://github.com/llvm/llvm-project/pull/108383
___
cfe-commits
https://github.com/labrinea edited
https://github.com/llvm/llvm-project/pull/108383
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/labrinea created
https://github.com/llvm/llvm-project/pull/108383
Sha1 has been unified with sha2 in the ACLE spec
(see https://github.com/ARM-software/acle/pull/347)
so I am changing the compiler to adhere to it.
>From 50820f4f3ded438162a33eabca84c5077fd2a5aa Mon Sep 17 00:0
https://github.com/labrinea created
https://github.com/llvm/llvm-project/pull/108024
Originally I tried spliting these features in the compiler with
https://github.com/llvm/llvm-project/pull/101712, but I realized that there's
no way to preserve backwards compatibility, therefore we decided to
https://github.com/labrinea closed
https://github.com/llvm/llvm-project/pull/99816
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
labrinea wrote:
I am abandoning this PR in favor of
https://github.com/ARM-software/acle/pull/346
https://github.com/llvm/llvm-project/pull/99816
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/c
https://github.com/labrinea closed
https://github.com/llvm/llvm-project/pull/101712
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
labrinea wrote:
I am abandoning this PR in favor of
https://github.com/ARM-software/acle/pull/346
https://github.com/llvm/llvm-project/pull/101712
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/
@@ -3682,6 +3682,8 @@ static const struct Extension {
{"sve2-bitperm", {AArch64::FeatureSVE2BitPerm}},
{"sve2p1", {AArch64::FeatureSVE2p1}},
{"ls64", {AArch64::FeatureLS64}},
+{"ls64_v", {AArch64::FeatureLS64_V}},
+{"ls64_accdata", {AArch64::FeatureLS64_ACCD
https://github.com/labrinea updated
https://github.com/llvm/llvm-project/pull/101712
>From 2fe8ef63846989952a1c72b83114aabe6f36637a Mon Sep 17 00:00:00 2001
From: Alexandros Lamprineas
Date: Fri, 2 Aug 2024 17:32:23 +0100
Subject: [PATCH] [AArch64] Split FeatureLS64 to LS64_ACCDATA and LS64_V.
@@ -3682,6 +3682,8 @@ static const struct Extension {
{"sve2-bitperm", {AArch64::FeatureSVE2BitPerm}},
{"sve2p1", {AArch64::FeatureSVE2p1}},
{"ls64", {AArch64::FeatureLS64}},
+{"ls64_v", {AArch64::FeatureLS64_V}},
+{"ls64_accdata", {AArch64::FeatureLS64_ACCD
@@ -3682,6 +3682,8 @@ static const struct Extension {
{"sve2-bitperm", {AArch64::FeatureSVE2BitPerm}},
{"sve2p1", {AArch64::FeatureSVE2p1}},
{"ls64", {AArch64::FeatureLS64}},
+{"ls64_v", {AArch64::FeatureLS64_V}},
+{"ls64_accdata", {AArch64::FeatureLS64_ACCD
@@ -305,9 +305,17 @@ def FeatureWFxT : ExtensionWithMArch<"wfxt", "WFxT",
"FEAT_WFxT",
def FeatureHCX : Extension<"hcx", "HCX", "FEAT_HCX",
"Enable Armv8.7-A HCRX_EL2 system register">;
-def FeatureLS64 : ExtensionWithMArch<"ls64", "LS64",
- "FEAT_LS64, FEAT_LS64_V, FEAT_L
@@ -305,9 +305,17 @@ def FeatureWFxT : ExtensionWithMArch<"wfxt", "WFxT",
"FEAT_WFxT",
def FeatureHCX : Extension<"hcx", "HCX", "FEAT_HCX",
"Enable Armv8.7-A HCRX_EL2 system register">;
-def FeatureLS64 : ExtensionWithMArch<"ls64", "LS64",
- "FEAT_LS64, FEAT_LS64_V, FEAT_L
labrinea wrote:
ping
https://github.com/llvm/llvm-project/pull/101712
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/labrinea updated
https://github.com/llvm/llvm-project/pull/101712
>From 2fe8ef63846989952a1c72b83114aabe6f36637a Mon Sep 17 00:00:00 2001
From: Alexandros Lamprineas
Date: Fri, 2 Aug 2024 17:32:23 +0100
Subject: [PATCH] [AArch64] Split FeatureLS64 to LS64_ACCDATA and LS64_V.
labrinea wrote:
> @labrinea @jroelofs Is there anything that still needs to be done here or can
> this be merged?
Currently the dependent PR #87939 is on hold. Unless it moves forward this PR
is not needed.
https://github.com/llvm/llvm-project/pull/87942
__
labrinea wrote:
Just letting you know that I will be on holiday for a couple of weeks before I
have a chance to respond to review comments.
https://github.com/llvm/llvm-project/pull/101712
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https:
https://github.com/labrinea created
https://github.com/llvm/llvm-project/pull/101712
These extensions depend on each other as follows:
ls64 -> ls64_v -> ls64_accdata
Currently in LLVM FeatureLS64 models all three lumped together. However the
architecture does not mandate that all three need to
labrinea wrote:
Created a PR for the backport https://github.com/llvm/llvm-project/pull/101007
https://github.com/llvm/llvm-project/pull/100181
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-
labrinea wrote:
/cherry-pick
[f8ae128](https://github.com/llvm/llvm-project/commit/f8ae128755777424cf4133e4e8e819c0bc08d2b1)
https://github.com/llvm/llvm-project/pull/100181
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.or
https://github.com/labrinea closed
https://github.com/llvm/llvm-project/pull/100181
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/labrinea updated
https://github.com/llvm/llvm-project/pull/100181
>From 6799eb869ed1f87eb65dae84b11e3dd3907b2157 Mon Sep 17 00:00:00 2001
From: Alexandros Lamprineas
Date: Tue, 23 Jul 2024 19:24:41 +0100
Subject: [PATCH 1/5] [clang][FMV][AArch64] Improve streaming mode
compa
https://github.com/labrinea edited
https://github.com/llvm/llvm-project/pull/100181
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/labrinea updated
https://github.com/llvm/llvm-project/pull/100181
>From 6799eb869ed1f87eb65dae84b11e3dd3907b2157 Mon Sep 17 00:00:00 2001
From: Alexandros Lamprineas
Date: Tue, 23 Jul 2024 19:24:41 +0100
Subject: [PATCH 1/4] [clang][FMV][AArch64] Improve streaming mode
compa
@@ -3183,6 +3183,25 @@ class ASTContext : public RefCountedBase {
const FunctionDecl *FD,
llvm::function_ref Pred) const;
+ bool areFMVCompatible(const FunctionDecl *FD1,
+const FunctionDecl *FD2) const {
+if (!hasSameType(FD1->getRetur
https://github.com/labrinea edited
https://github.com/llvm/llvm-project/pull/100181
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/labrinea updated
https://github.com/llvm/llvm-project/pull/100181
>From 6799eb869ed1f87eb65dae84b11e3dd3907b2157 Mon Sep 17 00:00:00 2001
From: Alexandros Lamprineas
Date: Tue, 23 Jul 2024 19:24:41 +0100
Subject: [PATCH 1/3] [clang][FMV][AArch64] Improve streaming mode
compa
@@ -11145,7 +11148,24 @@ bool Sema::areMultiversionVariantFunctionsCompatible(
FunctionType::ExtInfo OldTypeInfo = OldType->getExtInfo();
FunctionType::ExtInfo NewTypeInfo = NewType->getExtInfo();
-if (OldTypeInfo.getCC() != NewTypeInfo.getCC())
+const auto *Ol
@@ -11145,7 +11148,24 @@ bool Sema::areMultiversionVariantFunctionsCompatible(
FunctionType::ExtInfo OldTypeInfo = OldType->getExtInfo();
FunctionType::ExtInfo NewTypeInfo = NewType->getExtInfo();
-if (OldTypeInfo.getCC() != NewTypeInfo.getCC())
+const auto *Ol
https://github.com/labrinea updated
https://github.com/llvm/llvm-project/pull/100181
>From 6799eb869ed1f87eb65dae84b11e3dd3907b2157 Mon Sep 17 00:00:00 2001
From: Alexandros Lamprineas
Date: Tue, 23 Jul 2024 19:24:41 +0100
Subject: [PATCH 1/2] [clang][FMV][AArch64] Improve streaming mode
compa
https://github.com/labrinea created
https://github.com/llvm/llvm-project/pull/100181
* Allow arm-streaming if all the functions versions adhere to it.
* Allow arm-streaming-compatible if all the functions versions adhere to it.
* Allow arm-locally-streaming regardless of the other functions vers
https://github.com/labrinea converted_to_draft
https://github.com/llvm/llvm-project/pull/99816
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/labrinea closed
https://github.com/llvm/llvm-project/pull/99522
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/labrinea created
https://github.com/llvm/llvm-project/pull/99522
When generating the body of the ifunc resolver, clang skips runtime checks for
features that are implied from the command line. We bend this rule for certain
features (memtag, bti, dgh), but this happens quite
https://github.com/labrinea closed
https://github.com/llvm/llvm-project/pull/97761
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
@@ -261,9 +261,9 @@ __attribute__((target_version("jscvt"))) int
default_def_with_version_decls(void
// CHECK: attributes #[[ATTR3]] = { noinline nounwind optnone
"no-trapping-math"="true" "stack-protector-buffer-size"="8"
"target-features"="+lse,-v9.5a" }
// CHECK: attribute
labrinea wrote:
> It's really hard to tell what is changing here because the existing tests are
> so non-specific.
I appreciate that. The tests would benefit from some tidying up, we should
prioritize this at some point. In case it helps most of the test changes are
due to symbols being gener
@@ -4224,10 +4204,8 @@ void CodeGenModule::emitMultiVersionFunctions() {
llvm::Function *Func = createFunction(CurFD);
Options.emplace_back(Func, TA->getArchitecture(), Feats);
} else if (const auto *TVA = CurFD->getAttr()) {
-bool
@@ -4210,9 +4192,7 @@ void CodeGenModule::emitMultiVersionFunctions() {
return cast(Func);
};
-bool HasDefaultDecl = !FD->isTargetVersionMultiVersion();
-bool ShouldEmitResolver =
-!getContext().getTargetInfo().getTriple().isAArch64();
+bool Shoul
@@ -59,15 +59,22 @@ int bar() {
return m.goo(1) + foo(1) + foo();
}
+// Example to demonstrate that at the point of use we haven't yet seen the
default.
+// At that point a declaration for the unmangled symbol is emitted, which is
later
+// replaced by the ifunc symbol (on
@@ -11,7 +11,7 @@ int __attribute__((target_version("fp+aes"))) fmv(void) {
return 6; }
int __attribute__((target_version("crc+ls64_v"))) fmv(void) { return 7; }
int __attribute__((target_version("bti"))) fmv(void) { return 8; }
int __attribute__((target_version("sme2"))) fmv(
@@ -261,9 +261,9 @@ __attribute__((target_version("jscvt"))) int
default_def_with_version_decls(void
// CHECK: attributes #[[ATTR3]] = { noinline nounwind optnone
"no-trapping-math"="true" "stack-protector-buffer-size"="8"
"target-features"="+lse,-v9.5a" }
// CHECK: attribute
@@ -59,15 +59,22 @@ int bar() {
return m.goo(1) + foo(1) + foo();
}
+// Example to demonstrate that at the point of use we haven't yet seen the
default.
+// At that point a declaration for the unmangled symbol is emitted, which is
later
+// replaced by the ifunc symbol (on
https://github.com/labrinea edited
https://github.com/llvm/llvm-project/pull/97761
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/labrinea edited
https://github.com/llvm/llvm-project/pull/97761
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/labrinea closed
https://github.com/llvm/llvm-project/pull/96628
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/labrinea updated
https://github.com/llvm/llvm-project/pull/96628
>From ff4635208e9cd83c6735c95ebf12125ca737029a Mon Sep 17 00:00:00 2001
From: Alexandros Lamprineas
Date: Tue, 25 Jun 2024 00:27:45 +0100
Subject: [PATCH 1/3] [clang][FMV] Do not omit explicit default target_ver
https://github.com/labrinea updated
https://github.com/llvm/llvm-project/pull/96628
>From ff4635208e9cd83c6735c95ebf12125ca737029a Mon Sep 17 00:00:00 2001
From: Alexandros Lamprineas
Date: Tue, 25 Jun 2024 00:27:45 +0100
Subject: [PATCH 1/3] [clang][FMV] Do not omit explicit default target_ver
https://github.com/labrinea updated
https://github.com/llvm/llvm-project/pull/96628
>From ff4635208e9cd83c6735c95ebf12125ca737029a Mon Sep 17 00:00:00 2001
From: Alexandros Lamprineas
Date: Tue, 25 Jun 2024 00:27:45 +0100
Subject: [PATCH 1/3] [clang][FMV] Do not omit explicit default target_ver
https://github.com/labrinea edited
https://github.com/llvm/llvm-project/pull/96628
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/labrinea edited
https://github.com/llvm/llvm-project/pull/96628
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/labrinea updated
https://github.com/llvm/llvm-project/pull/96628
>From ff4635208e9cd83c6735c95ebf12125ca737029a Mon Sep 17 00:00:00 2001
From: Alexandros Lamprineas
Date: Tue, 25 Jun 2024 00:27:45 +0100
Subject: [PATCH 1/3] [clang][FMV] Do not omit explicit default target_ver
@@ -11465,6 +11465,10 @@ static bool CheckMultiVersionFirstFunction(Sema &S,
FunctionDecl *FD) {
// otherwise it is treated as a normal function.
if (TA && !TA->isDefaultVersion())
return false;
+ // The target_version attribute only causes Multiversioning if this
+
https://github.com/labrinea edited
https://github.com/llvm/llvm-project/pull/96628
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
@@ -11576,22 +11584,6 @@ static bool CheckTargetCausesMultiVersioning(Sema &S,
FunctionDecl *OldFD,
}
}
- if (NewTVA) {
labrinea wrote:
The assertion should be conditional when NewTVA is not null because the
function is used for the target attribute
https://github.com/labrinea edited
https://github.com/llvm/llvm-project/pull/96628
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
@@ -11523,10 +11525,17 @@ static bool CheckTargetCausesMultiVersioning(Sema &S,
FunctionDecl *OldFD,
const auto *OldTVA = OldFD->getAttr();
// If the old decl is NOT MultiVersioned yet, and we don't cause that
// to change, this is a simple redeclaration.
- if ((NewTA &
@@ -11465,6 +11465,10 @@ static bool CheckMultiVersionFirstFunction(Sema &S,
FunctionDecl *FD) {
// otherwise it is treated as a normal function.
if (TA && !TA->isDefaultVersion())
return false;
+ // The target_version attribute only causes Multiversioning if this
+
https://github.com/labrinea edited
https://github.com/llvm/llvm-project/pull/96628
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/labrinea edited
https://github.com/llvm/llvm-project/pull/96628
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
@@ -11576,22 +11584,6 @@ static bool CheckTargetCausesMultiVersioning(Sema &S,
FunctionDecl *OldFD,
}
}
- if (NewTVA) {
labrinea wrote:
This is checking things like
```
//expected-note@+1 {{previous declaration is here}}
void __attribute__((target_ver
https://github.com/labrinea updated
https://github.com/llvm/llvm-project/pull/96628
>From ff4635208e9cd83c6735c95ebf12125ca737029a Mon Sep 17 00:00:00 2001
From: Alexandros Lamprineas
Date: Tue, 25 Jun 2024 00:27:45 +0100
Subject: [PATCH 1/2] [clang][FMV] Do not omit explicit default target_ver
@@ -11523,10 +11525,16 @@ static bool CheckTargetCausesMultiVersioning(Sema &S,
FunctionDecl *OldFD,
const auto *OldTVA = OldFD->getAttr();
// If the old decl is NOT MultiVersioned yet, and we don't cause that
// to change, this is a simple redeclaration.
- if ((NewTA &
1 - 100 of 367 matches
Mail list logo
