illwieckz wrote:
Here is a script to reproduce the bug:
```bash
#! /usr/bin/env bash
set -x -u -e -o pipefail
version="${1:-18}"
CMAKE_BUILD_PARALLEL_LEVEL="$(nproc)"
export CMAKE_BUILD_PARALLEL_LEVEL="${CMAKE_BUILD_PARALLEL_LEVEL:-4}"
workspace="llvm-bug95484-${version}"
rm -rf "${workspac
illwieckz wrote:
```$
$ rg HSA_AMD_AGENT_INFO_TIMESTAMP_FREQUENCY
libc/utils/gpu/loader/amdgpu/Loader.cpp
521: HSA_AMD_AGENT_INFO_TIMESTAMP_FREQUENCY),
openmp/libomptarget/plugins-nextgen/amdgpu/dynamic_hsa/hsa_ext_amd.h
74: HSA_AMD_AGENT_INFO_TIMESTAMP_FREQUENCY
illwieckz wrote:
> We made a change recently that made the dynamic_hsa version the default. The
> error you're seeing is from an old HSA, so if you're overriding the default
> to use an old library that's probably not worth working around.
The error I see comes from the fact there is no old HS
https://github.com/jhuber6 edited
https://github.com/llvm/llvm-project/pull/95484
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
Author: Schrodinger ZHU Yifan
Date: 2024-06-13T20:22:21-07:00
New Revision: 4076c3004f09e95d1fcd299452843f99235ff422
URL:
https://github.com/llvm/llvm-project/commit/4076c3004f09e95d1fcd299452843f99235ff422
DIFF:
https://github.com/llvm/llvm-project/commit/4076c3004f09e95d1fcd299452843f99235ff4
https://github.com/jhuber6 commented:
We made a change recently that made the dynamic_hsa version the default. The
error you're seeing is from an old HSA, so if you're overriding the default to
use an old library that's probably not worth working around.
https://github.com/llvm/llvm-project/pu
@@ -354,6 +354,23 @@ Given that ``signedPointer`` matches the layout for signed
pointers signed with
the given key, extract the raw pointer from it. This operation does not trap
and cannot fail, even if the pointer is not validly signed.
+``ptrauth_sign_constant``
+^
@@ -354,6 +354,23 @@ Given that ``signedPointer`` matches the layout for signed
pointers signed with
the given key, extract the raw pointer from it. This operation does not trap
and cannot fail, even if the pointer is not validly signed.
+``ptrauth_sign_constant``
+^
@@ -58,6 +58,35 @@ void test_string_discriminator(const char *str) {
}
+void test_sign_constant(int *dp, int (*fp)(int)) {
+ __builtin_ptrauth_sign_constant(&dv, VALID_DATA_KEY); // expected-error
{{too few arguments}}
+ __builtin_ptrauth_sign_constant(&dv, VALID_DATA_KEY,
@@ -2061,6 +2071,58 @@ ConstantLValueEmitter::VisitCallExpr(const CallExpr *E) {
}
}
+ConstantLValue
+ConstantLValueEmitter::emitPointerAuthSignConstant(const CallExpr *E) {
+ llvm::Constant *UnsignedPointer = emitPointerAuthPointer(E->getArg(0));
+ unsigned Key = emitPoin
illwieckz wrote:
@pranav-sivaraman try this patch:
```diff
diff --git a/openmp/libomptarget/plugins/amdgpu/CMakeLists.txt
b/openmp/libomptarget/plugins/amdgpu/CMakeLists.txt
index 92523c23f68b..92bcd94edb7a 100644
--- a/openmp/libomptarget/plugins/amdgpu/CMakeLists.txt
+++ b/openmp/libomptarget
illwieckz wrote:
The 14 branch seems to be very old, espially the file you link is in `plugins/`
directory, while the files I modify are in `plugins-nextgen/` directory, witht
the `plugins/` directory not existing anymore. So I strongly doubt the patch is
useful for LLVM 14, but your problem p
llvmbot wrote:
@llvm/pr-subscribers-llvm-transforms
Author: shaw young (shawbyoung)
Changes
Test Plan: tbd
---
Full diff: https://github.com/llvm/llvm-project/pull/95486.diff
2 Files Affected:
- (modified) bolt/lib/Profile/StaleProfileMatching.cpp (+29-8)
- (modified) llvm/include/
https://github.com/shawbyoung closed
https://github.com/llvm/llvm-project/pull/95486
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
https://github.com/shawbyoung created
https://github.com/llvm/llvm-project/pull/95486
Test Plan: tbd
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
pranav-sivaraman wrote:
This is different from this
[file](https://github.com/llvm/llvm-project/blob/release/14.x/openmp/libomptarget/plugins/amdgpu/impl/hsa_api.h)
right? I'm trying to fix an issue when building LLVM 14 with a newer ROCm
releases which fails to find the newer `hsa/hsa.h` head
illwieckz wrote:
I first noticed the issue when building the chipStar fork of LLVM 17:
https://github.com/CHIP-SPV/llvm-project (branch `chipStar-llvm-17`), but the
code being the same in LLVM 18, it is expected to fail in LLVM 18 too.
The whole folder disappeared in `main` so I made this patc
llvmbot wrote:
@llvm/pr-subscribers-backend-amdgpu
Author: Thomas Debesse (illwieckz)
Changes
The `dynamic_hsa/` include directory is required by both optional
`dynamic_hsa/hsa.cpp` and non-optional `src/rtl.cpp`.
It should then always be included or the build will fail if only `src/rtl.
https://github.com/illwieckz created
https://github.com/llvm/llvm-project/pull/95484
The `dynamic_hsa/` include directory is required by both optional
`dynamic_hsa/hsa.cpp` and non-optional `src/rtl.cpp`.
It should then always be included or the build will fail if only `src/rtl.cpp`
is built.
https://github.com/ahmedbougacha updated
https://github.com/llvm/llvm-project/pull/94394
>From 1e9a3fde97d907c3cd6be33db91d1c18c7236ffb Mon Sep 17 00:00:00 2001
From: Ahmed Bougacha
Date: Tue, 4 Jun 2024 12:41:47 -0700
Subject: [PATCH 1/7] [Support] Reformat SipHash.cpp to match libSupport.
Wh
ahmedbougacha wrote:
[37c84b9](https://github.com/llvm/llvm-project/pull/94394/commits/37c84b9dce70f40db8a7c27b7de8232c4d10f78f)
shows what I had in mind, let me know what you all think. I added:
```
void getSipHash_2_4_64(const uint8_t *In, uint64_t InLen,
const uint8_t
https://github.com/ahmedbougacha updated
https://github.com/llvm/llvm-project/pull/94394
>From 1e9a3fde97d907c3cd6be33db91d1c18c7236ffb Mon Sep 17 00:00:00 2001
From: Ahmed Bougacha
Date: Tue, 4 Jun 2024 12:41:47 -0700
Subject: [PATCH 1/6] [Support] Reformat SipHash.cpp to match libSupport.
Wh
krzysz00 wrote:
On the other hand, it's a lot easier to handle ugly types down in instruction
selection, where you get to play much more fast and loose with types.
And there are buffer uses that don't fit into the fat pointer use use case
where we'd still want them to work.
For example, both `
@@ -5745,6 +5745,14 @@ IntrinsicLibrary::genReduce(mlir::Type resultType,
int rank = arrayTmp.rank();
assert(rank >= 1);
+ // Arguements to the reduction operation are passed by reference or value?
+ bool argByRef = true;
+ if (auto embox =
+ mlir::dyn_cast_or_
https://github.com/AtariDreams reopened
https://github.com/llvm/llvm-project/pull/92478
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
llvmbot wrote:
@uweigand What do you think about merging this PR to the release branch?
https://github.com/llvm/llvm-project/pull/95463
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listi
https://github.com/llvmbot milestoned
https://github.com/llvm/llvm-project/pull/95463
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
https://github.com/llvmbot created
https://github.com/llvm/llvm-project/pull/95463
Backport 7e4c6e98fa05f5c3bf14f96365ae74a8d12c6257
Requested by: @nikic
>From 016c200faf4bcf1a531dabd4411a2ec4d0a23068 Mon Sep 17 00:00:00 2001
From: Jonas Paulsson
Date: Mon, 15 Apr 2024 16:32:14 +0200
Subject:
llvmbot wrote:
@llvm/pr-subscribers-testing-tools
Author: Tom Stellard (tstellar)
Changes
---
Full diff: https://github.com/llvm/llvm-project/pull/95458.diff
2 Files Affected:
- (modified) llvm/CMakeLists.txt (+1-1)
- (modified) llvm/utils/lit/lit/__init__.py (+1-1)
``diff
https://github.com/tstellar created
https://github.com/llvm/llvm-project/pull/95458
None
>From 2edf6218b7e74cc76035e4e1efa8166b1c22312d Mon Sep 17 00:00:00 2001
From: Tom Stellard
Date: Thu, 13 Jun 2024 12:33:39 -0700
Subject: [PATCH] Bump version to 18.1.8
---
llvm/CMakeLists.txt
https://github.com/gbMattN converted_to_draft
https://github.com/llvm/llvm-project/pull/95387
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
kbeyls wrote:
> Yes, this doesn't have tests by itself because there's no exposed interface.
> It's certainly trivial to add one (which would allow using the reference test
> vectors).
>
> I don't have strong arguments either way, but I figured the conservative
> option is to force hypothetic
Author: Florian Mayer
Date: 2024-06-13T09:55:29-07:00
New Revision: 7fe862d0a1f6dfa67c236f5af32ad15546797404
URL:
https://github.com/llvm/llvm-project/commit/7fe862d0a1f6dfa67c236f5af32ad15546797404
DIFF:
https://github.com/llvm/llvm-project/commit/7fe862d0a1f6dfa67c236f5af32ad15546797404.diff
https://github.com/jeanPerier approved this pull request.
LGTM
https://github.com/llvm/llvm-project/pull/95353
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
@@ -592,10 +599,15 @@ void preprocessUnreachableBlocks(FlowFunction &Func) {
/// Decide if stale profile matching can be applied for a given function.
/// Currently we skip inference for (very) large instances and for instances
/// having "unexpected" control flow (e.g., having
krzysz00 wrote:
Yeah, makes sense.
... what prevents a match-bitwidth operator from existing?
Context from where I'm standing is that you should be able to
`raw.buffer.load/store` any (non-aggregate, let's say, since that could be
better handled in `addrspace(7)` handling) type you could `loa
@@ -5745,6 +5745,14 @@ IntrinsicLibrary::genReduce(mlir::Type resultType,
int rank = arrayTmp.rank();
assert(rank >= 1);
+ // Arguements to the reduction operation are passed by reference or value?
+ bool argByRef = true;
+ if (auto embox =
+ mlir::dyn_cast_or_
arsenm wrote:
I don't think we should be trying to handle the unreasonable illegal types in
the intrinsics themselves. Theoretically the intrinsic should correspond to
direct support.
We would handle the ugly types in the fat pointer lowering in terms of the
intrinsics.
https://github.com/l
gbMattN wrote:
This may be a side effect of a different bug tracking global variables. I think
fixing that bug first, and then applying this change if the problem persists is
a better idea. Because of this, I'm switching this to a draft for now.
Discourse link is
https://discourse.llvm.org/t/
Author: Schrodinger ZHU Yifan
Date: 2024-06-13T07:54:57-07:00
New Revision: 93e7f145bc38c7c47d797e652d891695eb44fcfa
URL:
https://github.com/llvm/llvm-project/commit/93e7f145bc38c7c47d797e652d891695eb44fcfa
DIFF:
https://github.com/llvm/llvm-project/commit/93e7f145bc38c7c47d797e652d891695eb44fc
Author: Schrodinger ZHU Yifan
Date: 2024-06-13T08:38:05-07:00
New Revision: 91323a6ea8f32a9fe2cec7051e8a99b87157133e
URL:
https://github.com/llvm/llvm-project/commit/91323a6ea8f32a9fe2cec7051e8a99b87157133e
DIFF:
https://github.com/llvm/llvm-project/commit/91323a6ea8f32a9fe2cec7051e8a99b8715713
arsenm wrote:
That's what we've traditionally done and I think we should stop. We currently
skip inserting the casts if the type is legal.
It introduces extra bitcasts, which have a cost and increase pattern match
complexity. We have a bunch of patterns that don't bother to look through the
krzysz00 wrote:
So, general question on this patch series:
Wouldn't it be more reasonable to, instead of having separate handling for all
the possible register types, always do loads as `i8`, `i16`, `i32` `<2 x i32>`,
`<3 x i32>, or `<4 x i32>` and then `bitcast`/`merge_values`/... the results
@@ -117,13 +117,44 @@ void test_update_dpp(global int* out, int arg1, int arg2)
}
// CHECK-LABEL: @test_ds_fadd
-// CHECK: {{.*}}call{{.*}} float @llvm.amdgcn.ds.fadd.f32(ptr addrspace(3)
%out, float %src, i32 0, i32 0, i1 false)
+// CHECK: atomicrmw fadd ptr addrspace(3) %ou
gbMattN wrote:
@fhahn
https://github.com/llvm/llvm-project/pull/95387
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
@@ -2331,40 +2337,74 @@ static Value *upgradeARMIntrinsicCall(StringRef Name,
CallBase *CI, Function *F,
llvm_unreachable("Unknown function for ARM CallBase upgrade.");
}
+// These are expected to have have the arguments:
cdevadas wrote:
```suggestion
// T
https://github.com/gbMattN updated
https://github.com/llvm/llvm-project/pull/95387
>From 432f994b1bc21e4db0778fff9cc1425f788f8168 Mon Sep 17 00:00:00 2001
From: Matthew Nagy
Date: Thu, 13 Jun 2024 09:54:04 +
Subject: [PATCH] [TySan] Fixed false positive when accessing offset member
variabl
@@ -1421,27 +1421,21 @@ let OtherPredicates = [HasPackedD16VMem] in {
defm : MUBUF_LoadIntrinsicPat;
} // End HasPackedD16VMem.
-defm : MUBUF_LoadIntrinsicPat;
-defm : MUBUF_LoadIntrinsicPat;
-defm : MUBUF_LoadIntrinsicPat;
-defm : MUBUF_LoadIntrinsicPat;
-defm : MUBUF_LoadI
https://github.com/arsenm updated
https://github.com/llvm/llvm-project/pull/95379
>From 14695322d92821374dd6599d8f0f76d212e50169 Mon Sep 17 00:00:00 2001
From: Matt Arsenault
Date: Wed, 12 Jun 2024 10:10:20 +0200
Subject: [PATCH] AMDGPU: Fix buffer load/store of pointers
Make sure we test all
https://github.com/arsenm updated
https://github.com/llvm/llvm-project/pull/95378
>From 1dfcc0961e82bbe656faded0c38e694da0d76c9b Mon Sep 17 00:00:00 2001
From: Matt Arsenault
Date: Sun, 9 Jun 2024 23:12:31 +0200
Subject: [PATCH] AMDGPU: Cleanup selection patterns for buffer loads
We should jus
Author: Younan Zhang
Date: 2024-06-13T18:53:46+08:00
New Revision: 0e8c9bca863137f14aea2cee0e05d4270b33e0e8
URL:
https://github.com/llvm/llvm-project/commit/0e8c9bca863137f14aea2cee0e05d4270b33e0e8
DIFF:
https://github.com/llvm/llvm-project/commit/0e8c9bca863137f14aea2cee0e05d4270b33e0e8.diff
llvmbot wrote:
@llvm/pr-subscribers-compiler-rt-sanitizer
Author: None (gbMattN)
Changes
This patch fixes a bug the current TySan implementation has. Currently if you
access a member variable other than the first, TySan reports an error. TySan
believes you are accessing the struct type w
github-actions[bot] wrote:
Thank you for submitting a Pull Request (PR) to the LLVM Project!
This PR will be automatically labeled and the relevant teams will be
notified.
If you wish to, you can add reviewers by using the "Reviewers" section on this
page.
If this is not working for you, it
https://github.com/jayfoad approved this pull request.
https://github.com/llvm/llvm-project/pull/95377
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
https://github.com/arsenm ready_for_review
https://github.com/llvm/llvm-project/pull/95377
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
llvmbot wrote:
@llvm/pr-subscribers-backend-amdgpu
Author: Matt Arsenault (arsenm)
Changes
We should just support these for all register types.
---
Full diff: https://github.com/llvm/llvm-project/pull/95378.diff
2 Files Affected:
- (modified) llvm/lib/Target/AMDGPU/BUFInstructions.td (
llvmbot wrote:
@llvm/pr-subscribers-backend-amdgpu
Author: Matt Arsenault (arsenm)
Changes
Make sure we test all the address spaces since this support isn't
free in gisel.
---
Patch is 38.37 KiB, truncated to 20.00 KiB below, full version:
https://github.com/llvm/llvm-project/pull/95379
https://github.com/arsenm ready_for_review
https://github.com/llvm/llvm-project/pull/95379
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
https://github.com/arsenm ready_for_review
https://github.com/llvm/llvm-project/pull/95378
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
llvmbot wrote:
@llvm/pr-subscribers-backend-amdgpu
Author: Matt Arsenault (arsenm)
Changes
---
Full diff: https://github.com/llvm/llvm-project/pull/95377.diff
2 Files Affected:
- (modified) llvm/lib/Target/AMDGPU/SIISelLowering.cpp (+2-2)
- (modified) llvm/test/CodeGen/AMDGPU/llvm.am
arsenm wrote:
> [!WARNING]
> This pull request is not mergeable via GitHub because a downstack PR is
> open. Once all requirements are satisfied, merge this PR as a stack href="https://app.graphite.dev/github/pr/llvm/llvm-project/95378?utm_source=stack-comment-downstack-mergeability-warning";
>
arsenm wrote:
> [!WARNING]
> This pull request is not mergeable via GitHub because a downstack PR is
> open. Once all requirements are satisfied, merge this PR as a stack href="https://app.graphite.dev/github/pr/llvm/llvm-project/95377?utm_source=stack-comment-downstack-mergeability-warning";
>
arsenm wrote:
> [!WARNING]
> This pull request is not mergeable via GitHub because a downstack PR is
> open. Once all requirements are satisfied, merge this PR as a stack href="https://app.graphite.dev/github/pr/llvm/llvm-project/95379?utm_source=stack-comment-downstack-mergeability-warning";
>
https://github.com/arsenm created
https://github.com/llvm/llvm-project/pull/95379
Make sure we test all the address spaces since this support isn't
free in gisel.
>From b05179ed684e289ce31f7aee8b57939c7bf2809c Mon Sep 17 00:00:00 2001
From: Matt Arsenault
Date: Wed, 12 Jun 2024 10:10:20 +0200
https://github.com/arsenm created
https://github.com/llvm/llvm-project/pull/95378
We should just support these for all register types.
>From 46c7f8b4529827204e5273472ea5b642ecb7266e Mon Sep 17 00:00:00 2001
From: Matt Arsenault
Date: Sun, 9 Jun 2024 23:12:31 +0200
Subject: [PATCH] AMDGPU: Clea
https://github.com/arsenm created
https://github.com/llvm/llvm-project/pull/95377
None
>From 520d91d73339d8bea65f2e30e2a4d7fd0eb3d92b Mon Sep 17 00:00:00 2001
From: Matt Arsenault
Date: Sun, 9 Jun 2024 22:54:35 +0200
Subject: [PATCH] AMDGPU: Fix buffer intrinsic store of bfloat
---
llvm/lib/
66 matches
Mail list logo
