llvm-project
llvm-project copied to clipboard
RadeonOpenCompute
[Issue]: Atomic optimizer reorder causes memory access fault in Blender
Problem Description
Following https://projects.blender.org/blender/blender/issues/112084, I've bisected the rocm-6.0.x branch and found that commit 30a3adf50e2d49dfc97c1b614d9b93638eba672d caused any Blender render (using HIP, of course) to crash with message along the lines of "Memory access fault by GPU node-1 (Agent handle: 0x7f1db8337e00) on address 0x7f1bf177e000. Reason: Page not present or supervisor privilege."
Operating System
Solus 4.5 Resilience
CPU
AMD Ryzen 7 5800H with Radeon Graphics
GPU
AMD Instinct MI250, AMD Radeon VII
ROCm Version
ROCm 6.0.0
ROCm Component
llvm-project
Steps to Reproduce
-
Build this project at the commit mentioned.
-
Download the Blender 4.1 release binaries:
curl -O https://download.blender.org/release/Blender4.1/blender-4.1.0-linux-x64.tar.xz,tar xf blender-4.1.0-linux-x64.tar.xz. You should now have a folderblender-4.1.0-linux-x64. -
Clone Blender. Just cloning the
v4.1.0tag is enough:git clone https://projects.blender.org/blender/blender.git --depth 1 --branch v4.1.0. -
In the Blender repo, compile the HIP fatbin used to run Blender render:
hipcc --offload-arch=$arch --genco intern/cycles/kernel/device/hip/kernel.cpp -D CCL_NAMESPACE_BEGIN= -D CCL_NAMESPACE_END= -D HIPCC -I intern/cycles/kernel/.. -I intern/cycles/kernel/device/hip -ffast-math -o kernel_$arch.fatbin. AdjustHIP_ROCCLR_HOME,HIP_CLANG_PATHas necessary to point to the Clang you just compiled. Replace$archwith the GPU architecture to run on, e.g.gfx900orgfx1030. Don't add extra attributes like:xnack-.If you want to run on multiple architectures, repeat step 4 and 5 for each architecture.
-
Put this file into
blender-4.1.0-linux-x64/4.1/scripts/addons/cycles/lib/kernel_$arch.fatbin. -
Get the BMW27 Blender demo file.
curl -O https://download.blender.org/demo/test/BMW27.blend.zip,unzip BMW27.blend.zip. You should have a fileBMW27.blend. -
Now run Blender render.
blender-4.1.0-linux-x64/blender -b <path-to-BMW27.blend> -f 0 -- --cycles-device HIP. By default it runs on GPU with device ID 0, so adjustHIP_VISIBLE_DEVICESas necessary to run on the desired GPU.You should almost immediately see Blender crash with an error message similar to "Memory access fault by GPU node-1 (Agent handle: 0x7f1db8337e00) on address 0x7f1bf177e000. Reason: Page not present or supervisor privilege."
-
Now, build LLVM at 1 commit prior, e.g.
git switch --detach 30a3adf50e2d49dfc97c1b614d9b93638eba672d~1. Repeat step 4-7, and Blender should render normally.
All of this is on ROCm 6.0.0. If you get a hang instead of a crash when running Blender (likely your on an APU), Ctrl+C and run again with environment variable HSA_ENABLE_SMDA=0.
(Optional for Linux users) Output of /opt/rocm/bin/rocminfo --support
rocminfo --support
ROCk module is loaded
=====================
HSA System Attributes
=====================
Runtime Version: 1.1
System Timestamp Freq.: 1000.000000MHz
Sig. Max Wait Duration: 18446744073709551615 (0xFFFFFFFFFFFFFFFF) (timestamp count)
Machine Model: LARGE
System Endianness: LITTLE
Mwaitx: DISABLED
DMAbuf Support: YES
==========
HSA Agents
==========
*******
Agent 1
*******
Name: AMD Ryzen 7 5800H with Radeon Graphics
Uuid: CPU-XX
Marketing Name: AMD Ryzen 7 5800H with Radeon Graphics
Vendor Name: CPU
Feature: None specified
Profile: FULL_PROFILE
Float Round Mode: NEAR
Max Queue Number: 0(0x0)
Queue Min Size: 0(0x0)
Queue Max Size: 0(0x0)
Queue Type: MULTI
Node: 0
Device Type: CPU
Cache Info:
L1: 32768(0x8000) KB
Chip ID: 0(0x0)
ASIC Revision: 0(0x0)
Cacheline Size: 64(0x40)
Max Clock Freq. (MHz): 2200
BDFID: 0
Internal Node ID: 0
Compute Unit: 16
SIMDs per CU: 0
Shader Engines: 0
Shader Arrs. per Eng.: 0
WatchPts on Addr. Ranges:1
Features: None
Pool Info:
Pool 1
Segment: GLOBAL; FLAGS: FINE GRAINED
Size: 61576816(0x3ab9670) KB
Allocatable: TRUE
Alloc Granule: 4KB
Alloc Alignment: 4KB
Accessible by all: TRUE
Pool 2
Segment: GLOBAL; FLAGS: KERNARG, FINE GRAINED
Size: 61576816(0x3ab9670) KB
Allocatable: TRUE
Alloc Granule: 4KB
Alloc Alignment: 4KB
Accessible by all: TRUE
Pool 3
Segment: GLOBAL; FLAGS: COARSE GRAINED
Size: 61576816(0x3ab9670) KB
Allocatable: TRUE
Alloc Granule: 4KB
Alloc Alignment: 4KB
Accessible by all: TRUE
ISA Info:
*******
Agent 2
*******
Name: gfx1032
Uuid: GPU-XX
Marketing Name: AMD Radeon RX 6600M
Vendor Name: AMD
Feature: KERNEL_DISPATCH
Profile: BASE_PROFILE
Float Round Mode: NEAR
Max Queue Number: 128(0x80)
Queue Min Size: 64(0x40)
Queue Max Size: 131072(0x20000)
Queue Type: MULTI
Node: 1
Device Type: GPU
Cache Info:
L1: 16(0x10) KB
L2: 2048(0x800) KB
L3: 32768(0x8000) KB
Chip ID: 29695(0x73ff)
ASIC Revision: 0(0x0)
Cacheline Size: 64(0x40)
Max Clock Freq. (MHz): 2720
BDFID: 768
Internal Node ID: 1
Compute Unit: 28
SIMDs per CU: 2
Shader Engines: 2
Shader Arrs. per Eng.: 2
WatchPts on Addr. Ranges:4
Coherent Host Access: FALSE
Features: KERNEL_DISPATCH
Fast F16 Operation: TRUE
Wavefront Size: 32(0x20)
Workgroup Max Size: 1024(0x400)
Workgroup Max Size per Dimension:
x 1024(0x400)
y 1024(0x400)
z 1024(0x400)
Max Waves Per CU: 32(0x20)
Max Work-item Per CU: 1024(0x400)
Grid Max Size: 4294967295(0xffffffff)
Grid Max Size per Dimension:
x 4294967295(0xffffffff)
y 4294967295(0xffffffff)
z 4294967295(0xffffffff)
Max fbarriers/Workgrp: 32
Packet Processor uCode:: 116
SDMA engine uCode:: 76
IOMMU Support:: None
Pool Info:
Pool 1
Segment: GLOBAL; FLAGS: COARSE GRAINED
Size: 8372224(0x7fc000) KB
Allocatable: TRUE
Alloc Granule: 4KB
Alloc Alignment: 4KB
Accessible by all: FALSE
Pool 2
Segment: GLOBAL; FLAGS: EXTENDED FINE GRAINED
Size: 8372224(0x7fc000) KB
Allocatable: TRUE
Alloc Granule: 4KB
Alloc Alignment: 4KB
Accessible by all: FALSE
Pool 3
Segment: GROUP
Size: 64(0x40) KB
Allocatable: FALSE
Alloc Granule: 0KB
Alloc Alignment: 0KB
Accessible by all: FALSE
ISA Info:
ISA 1
Name: amdgcn-amd-amdhsa--gfx1032
Machine Models: HSA_MACHINE_MODEL_LARGE
Profiles: HSA_PROFILE_BASE
Default Rounding Mode: NEAR
Default Rounding Mode: NEAR
Fast f16: TRUE
Workgroup Max Size: 1024(0x400)
Workgroup Max Size per Dimension:
x 1024(0x400)
y 1024(0x400)
z 1024(0x400)
Grid Max Size: 4294967295(0xffffffff)
Grid Max Size per Dimension:
x 4294967295(0xffffffff)
y 4294967295(0xffffffff)
z 4294967295(0xffffffff)
FBarrier Max Size: 32
*******
Agent 3
*******
Name: gfx90c
Uuid: GPU-XX
Marketing Name: AMD Radeon Graphics
Vendor Name: AMD
Feature: KERNEL_DISPATCH
Profile: BASE_PROFILE
Float Round Mode: NEAR
Max Queue Number: 128(0x80)
Queue Min Size: 64(0x40)
Queue Max Size: 131072(0x20000)
Queue Type: MULTI
Node: 2
Device Type: GPU
Cache Info:
L1: 16(0x10) KB
L2: 1024(0x400) KB
Chip ID: 5688(0x1638)
ASIC Revision: 0(0x0)
Cacheline Size: 64(0x40)
Max Clock Freq. (MHz): 2000
BDFID: 2048
Internal Node ID: 2
Compute Unit: 8
SIMDs per CU: 4
Shader Engines: 1
Shader Arrs. per Eng.: 1
WatchPts on Addr. Ranges:4
Coherent Host Access: FALSE
Features: KERNEL_DISPATCH
Fast F16 Operation: TRUE
Wavefront Size: 64(0x40)
Workgroup Max Size: 1024(0x400)
Workgroup Max Size per Dimension:
x 1024(0x400)
y 1024(0x400)
z 1024(0x400)
Max Waves Per CU: 40(0x28)
Max Work-item Per CU: 2560(0xa00)
Grid Max Size: 4294967295(0xffffffff)
Grid Max Size per Dimension:
x 4294967295(0xffffffff)
y 4294967295(0xffffffff)
z 4294967295(0xffffffff)
Max fbarriers/Workgrp: 32
Packet Processor uCode:: 471
SDMA engine uCode:: 40
IOMMU Support:: None
Pool Info:
Pool 1
Segment: GLOBAL; FLAGS: COARSE GRAINED
Size: 4194304(0x400000) KB
Allocatable: TRUE
Alloc Granule: 4KB
Alloc Alignment: 4KB
Accessible by all: FALSE
Pool 2
Segment: GLOBAL; FLAGS: EXTENDED FINE GRAINED
Size: 4194304(0x400000) KB
Allocatable: TRUE
Alloc Granule: 4KB
Alloc Alignment: 4KB
Accessible by all: FALSE
Pool 3
Segment: GROUP
Size: 64(0x40) KB
Allocatable: FALSE
Alloc Granule: 0KB
Alloc Alignment: 0KB
Accessible by all: FALSE
ISA Info:
ISA 1
Name: amdgcn-amd-amdhsa--gfx90c:xnack-
Machine Models: HSA_MACHINE_MODEL_LARGE
Profiles: HSA_PROFILE_BASE
Default Rounding Mode: NEAR
Default Rounding Mode: NEAR
Fast f16: TRUE
Workgroup Max Size: 1024(0x400)
Workgroup Max Size per Dimension:
x 1024(0x400)
y 1024(0x400)
z 1024(0x400)
Grid Max Size: 4294967295(0xffffffff)
Grid Max Size per Dimension:
x 4294967295(0xffffffff)
y 4294967295(0xffffffff)
z 4294967295(0xffffffff)
FBarrier Max Size: 32
*** Done ***
Additional Information
This behavior has been reproduced on MI250, RX6600M, Vega 10, and Ryzen 7 5800H. (gfx90a, gfx1032, gfx900, gfx90c, respectively)
Kernel version: 6.6.22-281.current, with https://github.com/torvalds/linux/commit/96c211f1f9ef82183493f4ceed4e347b52849149 reverted (ref: https://lists.freedesktop.org/archives/amd-gfx/2023-October/100298.html and https://github.com/ROCm/ROCm/issues/2596#issuecomment-1874371163)
I can confirm that reverting that commit (https://github.com/ROCm/llvm-project/commit/30a3adf50e2d49dfc97c1b614d9b93638eba672d) fixes the crash with other scenes like classroom but not with blender-3.2 (https://cloud.blender.org/p/gallery/629f23f908e12d4ff15241d3) which still crashed with the similar error (happens only with rocm-6.x):
Compiling HIP kernel ...
hipcc -Wno-parentheses-equality -Wno-unused-value --hipcc-func-supp -O3 -ffast-math --amdgpu-target=gfx1031 -I /usr/share/blender/4.1/scripts/addons/cycles/source --genco /usr/share/blender/4.1/scripts/addons/cycles/source/kernel/device/hip/kernel.cpp -o "/home/lakostis/.cache/cycles/kernels/cycles_kernel_gfx1031_F68CBA054A76B5B26A931C269B2CF458"
Warning: The --hipcc-func-supp option has been deprecated and will be removed in the future.
Warning: The --amdgpu-target option has been deprecated and will be removed in the future. Use --offload-arch instead.
Kernel compilation finished in 50.97s.
Read blend: "/home/lakostis/Downloads/Blender 3.blend"
Memory access fault by GPU node-1 (Agent handle: 0x7fd727132600) on address 0x7fd7932ca000. Reason: Page not present or supervisor privilege.
Aborted
@LAKostis this is weird, because the blender-3.2 splashscreen renders fine for me on all the devices I've tested. Are you sure your Blender is not using your cycles cache? Maybe try clearing ~/.cache/cycles before every run and install the fatbin files into /usr/share/blender/4.1/scripts/addons/cycles/lib for consistency?
I also see that you seem to be using the Blender provided by your distro. For consistency, can you download the official binaries from Blender and try testing with that instead?
@LAKostis this is weird, because the blender-3.2 splashscreen renders fine for me on all the devices I've tested. Are you sure your Blender is not using your cycles cache? Maybe try clearing
~/.cache/cyclesbefore every run and install the fatbin files into/usr/share/blender/4.1/scripts/addons/cycles/libfor consistency?I also see that you seem to be using the Blender provided by your distro. For consistency, can you download the official binaries from Blender and try testing with that instead?
Yes, I've specially tested this before rocm-6.0.x migration and this demo started failing only with rocm-6.0.x. Cache clearing doesn't help. It can be device specific issue - this demo crashes on my RX 6700 XT (gfx1031) but works on gfx900 (with rendering artifacts and only after setting HSA_ENABLE_SMDA=0)
Regarding the blender build - I'm the blender package maintainer in this distro, so I'm sure what build options where used there :) For the sake of clarity official blender builds crash exactly the same way. I can provide any additional information or logs if you need any.
Thanks for the quick response!
It can be device specific issue - this demo crashes on my RX 6700 XT (gfx1031) but works on gfx900 (with rendering artifacts and only after setting HSA_ENABLE_SMDA=0)
Just curious, does compiling against gfx1030 and running Blender through HSA_OVERRIDE_GFX_VERSION on your RX 6700XT change anything? Also what kernel version are you on?
HSA_OVERRIDE_GFX_VERSION
No, with HSA_OVERRIDE_GFX_VERSION=gfx1030 it doesn't start with error HIP hipInit: Invalid device
But interesting, it works with lowering -O level:
Read prefs: "/home/lakostis/.config/blender/4.1/config/userpref.blend"
Read blend: "/home/lakostis/Downloads/Blender 3.blend"
Compiling HIP kernel ...
hipcc -Wno-parentheses-equality -Wno-unused-value --hipcc-func-supp -O3 -ffast-math --amdgpu-target=gfx1031 -I /usr/share/blender/4.1/scripts/addons/cycles/source --genco /usr/share/blender/4.1/scripts/addons/cycles/source/kernel/device/hip/kernel.cpp -o "/home/lakostis/.cache/cycles/kernels/cycles_kernel_gfx1031_F68CBA054A76B5B26A931C269B2CF458"
Warning: The --hipcc-func-supp option has been deprecated and will be removed in the future.
Warning: The --amdgpu-target option has been deprecated and will be removed in the future. Use --offload-arch instead.
Kernel compilation finished in 50.93s.
Memory access fault by GPU node-1 (Agent handle: 0x7fa997041200) on address 0x7fa9f1b5f000. Reason: Page not present or supervisor privilege.
Aborted
...
❯ blender
Read prefs: "/home/lakostis/.config/blender/4.1/config/userpref.blend"
Read blend: "/home/lakostis/Downloads/Blender 3.blend"
❯ hipcc -Wno-parentheses-equality -Wno-unused-value --hipcc-func-supp -O1 -ffast-math --amdgpu-target=gfx1031 -I /usr/share/blender/4.1/scripts/addons/cycles/source --genco /usr/share/blender/4.1/scripts/addons/cycles/source/kernel/device/hip/kernel.cpp -o "/home/lakostis/.cache/cycles/kernels/cycles_kernel_gfx1031_F68CBA054A76B5B26A931C269B2CF458"
Warning: The --hipcc-func-supp option has been deprecated and will be removed in the future.
Warning: The --amdgpu-target option has been deprecated and will be removed in the future. Use --offload-arch instead.
...
with -O1 rendering works. And it start crashing with >=-O2. So something is not right with optimization.
Also what kernel version are you on?
I'm using 6.6.25 kernel + patches up to v6.5-2638-gbf901afac5d5f from amd-staging-drm-next
UPDATE: more funny things with compiler:
hipcc -Wno-parentheses-equality -Wno-unused-value --hipcc-func-supp -O1 -ffast-math --amdgpu-target=gfx1031 -I /usr/share/blender/4.1/scripts/addons/cycles/source --genco /usr/share/blender/4.1/scripts/addons/cycles/source/kernel/device/hip/kernel.cpp -o "/home/lakostis/.cache/cycles/kernels/cycles_kernel_gfx1031_F68CBA054A76B5B26A931C269B2CF458"
This command produces workable kernel despite of warnings about deprecated commands. But this command:
hipcc -Wno-parentheses-equality -Wno-unused-value -O1 -ffast-math --offload-arch=gfx1031 -I /usr/share/blender/4.1/scripts/addons/cycles/source --genco /usr/share/blender/4.1/scripts/addons/cycles/source/kernel/device/hip/kernel.cpp -o "/home/lakostis/.cache/cycles/kernels/cycles_kernel_gfx1031_F68CBA054A76B5B26A931C269B2CF458"
Produces kernel which crashes. And those kernels are not equal:
-rw-r--r-- 1 lakostis lakostis 4286552 Apr 7 13:10 cycles_kernel_gfx1031_F68CBA054A76B5B26A931C269B2CF458.crash_O1
-rw-r--r-- 1 lakostis lakostis 3213824 Apr 7 13:09 cycles_kernel_gfx1031_F68CBA054A76B5B26A931C269B2CF458.works_O1
UPDATE: checked with recent rocm and blender 4.2.0
kernel 6.9.10
❯ rpm -qa|fgrep 6.1.2-alt0
fgrep: warning: fgrep is obsolescent; using grep -F
rocm-comgr-devel-6.1.2-alt0.2.x86_64
llvm-rocm-6.1.2-alt0.2.x86_64
clang-rocm-6.1.2-alt0.2.x86_64
hip-devel-6.1.2-alt0.2.x86_64
librocm-smi1-6.1.2-alt0.2.x86_64
rocminfo-6.1.2-alt0.1.x86_64
clang-rocm-tools-6.1.2-alt0.2.x86_64
rocm-smi-6.1.2-alt0.2.x86_64
rocm-opencl-runtime-6.1.2-alt0.2.x86_64
llvm-rocm-filesystem-6.1.2-alt0.2.x86_64
libhsakmt1-6.1.2-alt0.1.x86_64
clang-rocm-libs-support-6.1.2-alt0.2.x86_64
libhsa-runtime1-6.1.2-alt0.1.x86_64
hip-runtime-amd-6.1.2-alt0.2.x86_64
clang-rocm-libs-6.1.2-alt0.2.x86_64
lld-rocm-6.1.2-alt0.2.x86_64
rocm-device-libs-6.1.2-alt0.2.x86_64
libamd_comgr2-6.1.2-alt0.2.x86_64
hipcc-6.1.2-alt0.2.x86_64
❯ rpm -q blender
blender-4.2.0-alt0.1.x86_64
If I compile gfx1031 with previous workaround (--hipcc-func-supp -O1) blender crashes on every rendering with errors:
❯ blender
register_class(...):
Info: Registering key-config preferences class: 'Prefs', bl_idname 'Blender' has been registered before, unregistering previous
register_class(...):
Info: Registering key-config preferences class: 'Prefs', bl_idname 'Blender' has been registered before, unregistering previous
Read blend: "/home/lakostis/Downloads/bmw27_gpu.blend"
Warning: region type 4 missing in space type "Info" (id: 7) - removing region
:0:rocdevice.cpp :2895: 420951542923 us: [pid:4113852 tid:0x7f5723200000] Callback: Queue 0x7f5600700000 aborting with error : HSA_STATUS_ERROR_MEMORY_APERTURE_VIOLATION: The agent attempted to access memory beyond the largest legal address. code: 0x29
Aborted
If I compile kernel with default upstream options (-O3) it renders most scenes (bmw/classroom) but still crashes on Blender 3 scene:
Read blend: "/home/lakostis/Downloads/bmw27_gpu.blend"
Warning: region type 4 missing in space type "Info" (id: 7) - removing region
Compiling HIP kernel ...
hipcc -Wno-parentheses-equality -Wno-unused-value -O3 -ffast-math --offload-arch=gfx1031 -I /usr/share/blender/4.2/scripts/addons_core/cycles/source --genco /usr/share/blender/4.2/scripts/addons_core/cycles/source/kernel/device/hip/kernel.cpp -o "/home/lakostis/.cache/cycles/kernel
s/cycles_kernel_gfx1031_210C856BB7ABA617B857E9D03ED272C1"
Kernel compilation finished in 110.52s.
...
Read blend: "/home/lakostis/Downloads/Blender 3.blend"
Memory access fault by GPU node-1 (Agent handle: 0x7f115fbc1800) on address 0x7f0ffe842000. Reason: Page not present or supervisor privilege.
Aborted
If I compile with -O2, still the same behavior, Blender 3 crashes.
But everything works with -O1:
❯ hipcc -Wno-parentheses-equality -Wno-unused-value -O1 -ffast-math --offload-arch=gfx1031 -I /usr/share/blender/4.2/scripts/addons_core/cycles/source --genco /usr/share/blender/4.2/scripts/addons_core/cycles/source/kernel/device/hip/kernel.cpp -o "/home/lakostis/.cache/cycles/kernels/cycles_kernel_gfx1031_210C856BB7ABA617B857E9D03ED272C1"
❯ blender
Read blend: "/home/lakostis/Downloads/Blender 3.blend"
Saved session recovery to "/tmp/.private/lakostis/quit.blend"
Writing userprefs: "/home/lakostis/.config/blender/4.2/config/userpref.blend" ok
Info: Preferences saved
Blender quit
Does reverting 30a3adf50e2d49dfc97c1b614d9b93638eba672d still help? I'm considering just reverting this patch for Solus's ROCm 6.1.2.
I've tested with reverted commit, can try again without reverting.
On Mon, 22 Jul 2024, 19:28 Gavin Zhao, @.***> wrote:
Does reverting 30a3adf https://github.com/ROCm/llvm-project/commit/30a3adf50e2d49dfc97c1b614d9b93638eba672d still help? I'm considering just reverting this patch for Solus's ROCm 6.1.2.
— Reply to this email directly, view it on GitHub https://github.com/ROCm/llvm-project/issues/58#issuecomment-2243462619, or unsubscribe https://github.com/notifications/unsubscribe-auth/ACMG6QPSN2R6CGG5PFJCJZTZNU6LZAVCNFSM6AAAAABFUT43JCVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDENBTGQ3DENRRHE . You are receiving this because you were mentioned.Message ID: @.***>
Can try after applying following patches: https://github.com/llvm/llvm-project/commit/c86a1e6903e9935b808c1406f480c769279b69fa https://github.com/llvm/llvm-project/commit/9ff71814cb5d71e907feaa0b3165e866b882f9aa https://github.com/llvm/llvm-project/commit/56af0e913ce7ec29690cc7295d75fc5573153bbf
I've tested with reverted commit, can try again without reverting.
Without reverting that commit everything works with -O1 but fails if O >=2.
Can try after applying following patches: llvm@c86a1e6 llvm@9ff7181 llvm@56af0e9
Hey! Those patches are already applied in rocm-llvm somewhere in this bulk commit (https://github.com/ROCm/llvm-project/commit/1ce2523b021e0bfb9b5c5d0d42bcd09e220c1380)
❯ patch -p1 --dry-run < ../c86a1e6903e9935b808c1406f480c769279b69fa.patch
checking file llvm/lib/Transforms/Scalar/GVN.cpp
Hunk #1 succeeded at 487 (offset 14 lines).
Hunk #2 FAILED at 2789.
1 out of 2 hunks FAILED
checking file llvm/lib/Transforms/Scalar/NewGVN.cpp
Reversed (or previously applied) patch detected! Assume -R? [n]
Apply anyway? [n]
Skipping patch.
1 out of 1 hunk ignored
checking file llvm/test/Transforms/GVN/convergent.ll
Reversed (or previously applied) patch detected! Assume -R? [n]
Apply anyway? [n]
Skipping patch.
1 out of 1 hunk ignored
checking file llvm/test/Transforms/NewGVN/convergent.ll
Reversed (or previously applied) patch detected! Assume -R? [n]
Apply anyway? [n]
Skipping patch.
1 out of 1 hunk ignored
❯ patch -p1 --dry-run < ../9ff71814cb5d71e907feaa0b3165e866b882f9aa.patch
checking file llvm/lib/Transforms/Scalar/EarlyCSE.cpp
Reversed (or previously applied) patch detected! Assume -R? [n]
Apply anyway? [n]
Skipping patch.
3 out of 3 hunks ignored
checking file llvm/test/Transforms/EarlyCSE/AMDGPU/convergent-call.ll
Reversed (or previously applied) patch detected! Assume -R? [n]
Apply anyway? [n]
Skipping patch.
2 out of 2 hunks ignored
❯ patch -p1 --dry-run < ../56af0e913ce7ec29690cc7295d75fc5573153bbf.patch
checking file llvm/lib/Transforms/Scalar/EarlyCSE.cpp
Hunk #1 succeeded at 336 with fuzz 2 (offset 18 lines).
Hunk #2 FAILED at 352.
1 out of 2 hunks FAILED
checking file llvm/test/CodeGen/AMDGPU/cse-convergent.ll
Reversed (or previously applied) patch detected! Assume -R? [n]
Apply anyway? [n]
Skipping patch.
1 out of 1 hunk ignored
ROCm 6.0.0 might have this issue. Looks like those patches (fixes) landed in next rocm-release immediately. AFAIU, this should be fixed in the latest rocm release. Is this exists on rocm-6.2?