[Bug 221320] New: ring gfx_0.1.0 timeout, signaled seq=5855035, emitted seq=5855037

[Bug 221320] New: ring gfx_0.1.0 timeout, signaled seq=5855035, emitted seq=5855037

[bugzilla-daemon]

https://bugzilla.kernel.org/show_bug.cgi?id=221320

            Bug ID: 221320
           Summary: ring gfx_0.1.0 timeout, signaled seq=5855035, emitted
                    seq=5855037
           Product: Drivers
           Version: 2.5
          Hardware: Intel
                OS: Linux
            Status: NEW
          Severity: normal
          Priority: P3
         Component: Video(DRI - non Intel)
          Assignee: drivers_video-dri@kernel-bugs.osdl.org
          Reporter: witold.baryluk+kernel@gmail.com
        Regression: No

kernel 7.0.0-rc5

Possibly related to use of LM Studio, with partial GPU and partial CPU usage.
(Started using LM Studio today, and otherwise this kernel and previous ones
were doing fine, for many days on this rc, and weeeks on other kernels).

Using `LM-Studio-0.4.9-1-x64.AppImage` - it is a likely source of issue (had
few hours session with `google/gemma-4-26b-a4b` model with full context window
size), as reset happened just after I pressed enter in the prompt window (but
previous ~50 prompts did not show any issues).

Mesa 26.0.3-1, but appimage and the tooling inside appimage might be using some
other version.

firmware-amd-graphics    20260221-1

Using gnome shell wayland as desktop manager. Also Firefox running in a
background. But nothing special was happening during the gpu timeout (i.e. no
video playback, but some videos maybe were paused).

Linux debian 7.0.0-rc5 #2 SMP PREEMPT Fri Mar 27 00:34:39 UTC 2026 x86_64
GNU/Linux


6900 XT

0c:00.0 VGA compatible controller: Advanced Micro Devices, Inc. [AMD/ATI] Navi
21 [Radeon RX 6800/6800 XT / 6900 XT] (rev c0)

CPU: AMD Ryzen Threadripper 2950X 16-Core Processor

I also do have zram enabled (80GB, no backing_dev), on top of 256GB of ECC RAM.
But probably unrelated.

[673923.157491] amdgpu 0000:0c:00.0: Dumping IP State
[673923.160063] amdgpu 0000:0c:00.0: Dumping IP State Completed
[673923.160138] amdgpu 0000:0c:00.0: [drm] AMDGPU device coredump file has been
created
[673923.160142] amdgpu 0000:0c:00.0: [drm] Check your
/sys/class/drm/card0/device/devcoredump/data
[673923.160146] amdgpu 0000:0c:00.0: ring gfx_0.1.0 timeout, signaled
seq=5855035, emitted seq=5855037
[673923.160150] amdgpu 0000:0c:00.0:  Process gnome-shell pid 9084 thread
gnome-shel:cs0 pid 9106
[673923.160153] amdgpu 0000:0c:00.0: Starting gfx_0.1.0 ring reset
[673923.404103] amdgpu 0000:0c:00.0: Ring gfx_0.1.0 reset failed
[673923.404112] amdgpu 0000:0c:00.0: GPU reset begin!. Source:  1
[673923.704012] amdgpu 0000:0c:00.0: MODE1 reset
[673923.704018] amdgpu 0000:0c:00.0: GPU mode1 reset
[673923.704084] amdgpu 0000:0c:00.0: GPU smu mode1 reset
[673924.205560] amdgpu 0000:0c:00.0: GPU reset succeeded, trying to resume
[673924.205736] amdgpu 0000:0c:00.0: [drm] PCIE GART of 512M enabled (table at
0x0000008000300000).
[673924.205760] amdgpu 0000:0c:00.0: VRAM is lost due to GPU reset!
[673924.205762] amdgpu 0000:0c:00.0: PSP is resuming...
[673924.284210] amdgpu 0000:0c:00.0: reserve 0xa00000 from 0x83fd000000 for PSP
TMR
[673924.421395] amdgpu 0000:0c:00.0: SECUREDISPLAY: optional securedisplay ta
ucode is not available
[673924.421403] amdgpu 0000:0c:00.0: SMU is resuming...
[673924.421409] amdgpu 0000:0c:00.0: smu driver if version = 0x00000040, smu fw
if version = 0x00000041, smu fw program = 0, version = 0x003a5a00 (58.90.0)
[673924.421419] amdgpu 0000:0c:00.0: SMU driver if version not matched
[673924.421450] amdgpu 0000:0c:00.0: use vbios provided pptable
[673924.497199] amdgpu 0000:0c:00.0: SMU is resumed successfully!
[673924.497873] amdgpu 0000:0c:00.0: kiq ring mec 2 pipe 1 q 0
[673924.507153] amdgpu 0000:0c:00.0: [drm] DMUB hardware initialized:
version=0x02020021
[673924.652432] amdgpu 0000:0c:00.0: ring gfx_0.0.0 uses VM inv eng 0 on hub 0
[673924.652437] amdgpu 0000:0c:00.0: ring gfx_0.1.0 uses VM inv eng 1 on hub 0
[673924.652439] amdgpu 0000:0c:00.0: ring comp_1.0.0 uses VM inv eng 4 on hub 0
[673924.652440] amdgpu 0000:0c:00.0: ring comp_1.1.0 uses VM inv eng 5 on hub 0
[673924.652442] amdgpu 0000:0c:00.0: ring comp_1.2.0 uses VM inv eng 6 on hub 0
[673924.652444] amdgpu 0000:0c:00.0: ring comp_1.3.0 uses VM inv eng 7 on hub 0
[673924.652445] amdgpu 0000:0c:00.0: ring comp_1.0.1 uses VM inv eng 8 on hub 0
[673924.652446] amdgpu 0000:0c:00.0: ring comp_1.1.1 uses VM inv eng 9 on hub 0
[673924.652447] amdgpu 0000:0c:00.0: ring comp_1.2.1 uses VM inv eng 10 on hub
0
[673924.652449] amdgpu 0000:0c:00.0: ring comp_1.3.1 uses VM inv eng 11 on hub
0
[673924.652450] amdgpu 0000:0c:00.0: ring kiq_0.2.1.0 uses VM inv eng 12 on hub
0
[673924.652451] amdgpu 0000:0c:00.0: ring sdma0 uses VM inv eng 13 on hub 0
[673924.652452] amdgpu 0000:0c:00.0: ring sdma1 uses VM inv eng 14 on hub 0
[673924.652454] amdgpu 0000:0c:00.0: ring sdma2 uses VM inv eng 15 on hub 0
[673924.652455] amdgpu 0000:0c:00.0: ring sdma3 uses VM inv eng 16 on hub 0
[673924.652456] amdgpu 0000:0c:00.0: ring vcn_dec_0 uses VM inv eng 0 on hub 8
[673924.652457] amdgpu 0000:0c:00.0: ring vcn_enc_0.0 uses VM inv eng 1 on hub
8
[673924.652458] amdgpu 0000:0c:00.0: ring vcn_enc_0.1 uses VM inv eng 4 on hub
8
[673924.652459] amdgpu 0000:0c:00.0: ring vcn_dec_1 uses VM inv eng 5 on hub 8
[673924.652460] amdgpu 0000:0c:00.0: ring vcn_enc_1.0 uses VM inv eng 6 on hub
8
[673924.652461] amdgpu 0000:0c:00.0: ring vcn_enc_1.1 uses VM inv eng 7 on hub
8
[673924.652462] amdgpu 0000:0c:00.0: ring jpeg_dec uses VM inv eng 8 on hub 8
[673924.661748] amdgpu 0000:0c:00.0: GPU reset(1) succeeded!
[673924.661840] amdgpu 0000:0c:00.0: [drm] device wedged, but recovered through
reset
[673924.671004] amdgpu 0000:0c:00.0: [drm] *ERROR* Failed to initialize parser
-125!

After the reset, LM studio operates fine after starting it again.

-- 
You may reply to this email to add a comment.

You are receiving this mail because:
You are watching the assignee of the bug.

[Bug 221320] ring gfx_0.1.0 timeout, signaled seq=5855035, emitted seq=5855037

[bugzilla-daemon]

https://bugzilla.kernel.org/show_bug.cgi?id=221320

--- Comment #1 from Witold Baryluk (witold.baryluk+kernel@gmail.com) ---
Created attachment 309820
  --> https://bugzilla.kernel.org/attachment.cgi?id=309820&action=edit
devcoredump

devcoredump attached

-- 
You may reply to this email to add a comment.

You are receiving this mail because:
You are watching the assignee of the bug.

[Bug 221320] ring gfx_0.1.0 timeout, signaled seq=5855035, emitted seq=5855037

[bugzilla-daemon]

https://bugzilla.kernel.org/show_bug.cgi?id=221320

Witold Baryluk (witold.baryluk+kernel@gmail.com) changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
 Attachment #309820|amdgpu-coredump.issue-22132 |amdgpu-devcoredump.issue-22
           filename|0.data.zst                  |1320.data.zst
 Attachment #309820|devcoredump                 |amdgpu-devcoredump.issue-22
        description|                            |1320.data.zst

-- 
You may reply to this email to add a comment.

You are receiving this mail because:
You are watching the assignee of the bug.

[Bug 221320] ring gfx_0.1.0 timeout, signaled seq=5855035, emitted seq=5855037

[bugzilla-daemon]

https://bugzilla.kernel.org/show_bug.cgi?id=221320

--- Comment #2 from Witold Baryluk (witold.baryluk+kernel@gmail.com) ---
The issue also happens with MATE desktop (X11), and no other programms running
(i.e. no Firefox)

LM studio log:

```
$ tail -f 2026-04-04.1.log 
slot update_slots: id  3 | task 0 | prompt processing progress, n_tokens =
35328, batch.n_tokens = 512, progress = 0.797292
[2026-04-04 20:36:50][DEBUG] slot update_slots: id  3 | task 0 | n_tokens =
35328, memory_seq_rm [35328, end)
[2026-04-04 20:36:50][DEBUG] slot update_slots: id  3 | task 0 | prompt
processing progress, n_tokens = 35840, batch.n_tokens = 512, progress =
0.808847
[2026-04-04 20:36:54][DEBUG] slot update_slots: id  3 | task 0 | n_tokens =
35840, memory_seq_rm [35840, end)
slot update_slots: id  3 | task 0 | prompt processing progress, n_tokens =
36352, batch.n_tokens = 512, progress = 0.820402
[2026-04-04 20:36:57][DEBUG] slot update_slots: id  3 | task 0 | n_tokens =
36352, memory_seq_rm [36352, end)
slot update_slots: id  3 | task 0 | prompt processing progress, n_tokens =
36864, batch.n_tokens = 512, progress = 0.831957
[2026-04-04 20:37:01][DEBUG] slot update_slots: id  3 | task 0 | n_tokens =
36864, memory_seq_rm [36864, end)
[2026-04-04 20:37:01][DEBUG] slot update_slots: id  3 | task 0 | prompt
processing progress, n_tokens = 37376, batch.n_tokens = 512, progress =
0.843512
[2026-04-04 20:37:08][DEBUG] radv/amdgpu: The CS has been cancelled because the
context is lost. This context is innocent.
```

I did verify it is using system vulkan libraries (vulkan loader, and the radeon
mesa driver too).


The issue is reproducible, by loading same model, and replying the same
conversation to the same point.

-- 
You may reply to this email to add a comment.

You are receiving this mail because:
You are watching the assignee of the bug.

[Bug 221320] ring gfx_0.1.0 timeout, signaled seq=5855035, emitted seq=5855037

[bugzilla-daemon]

https://bugzilla.kernel.org/show_bug.cgi?id=221320

--- Comment #3 from Witold Baryluk (witold.baryluk+kernel@gmail.com) ---
Looks to be a likely regression.

6.19.10+deb14-amd64 - works fine, and stable  (firmware-amd-graphics 20260309-1
- newer)
7.0.0-rc5 - timeout (firmware-amd-graphics 20260221-1 - older)
7.0.0-rc6 - timeout (firmware-amd-graphics 20260309-1 - newer)


The 7.0 rc5/rc6 were compiled with different options, compared to stock debian
kernel 6.19.10, so that might be issue too.

-- 
You may reply to this email to add a comment.

You are receiving this mail because:
You are watching the assignee of the bug.

[Bug 221320] ring gfx_0.1.0 timeout, signaled seq=5855035, emitted seq=5855037

[bugzilla-daemon]

https://bugzilla.kernel.org/show_bug.cgi?id=221320

--- Comment #4 from Witold Baryluk (witold.baryluk+kernel@gmail.com) ---
Moving to https://gitlab.freedesktop.org/drm/amd/-/work_items/5146

-- 
You may reply to this email to add a comment.

You are receiving this mail because:
You are watching the assignee of the bug.

[Bug 221320] ring gfx_0.1.0 timeout, signaled seq=5855035, emitted seq=5855037

[bugzilla-daemon]

https://bugzilla.kernel.org/show_bug.cgi?id=221320

Artem S. Tashkinov (aros@gmx.com) changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|---                         |MOVED

-- 
You may reply to this email to add a comment.

You are receiving this mail because:
You are watching the assignee of the bug.