Open Bug 1812265 Opened 1 year ago Updated 1 year ago

Firefox+Thunderbird+other apps/AMD/X11+Wayland: Triple-buffering loop with amdgpu_dm_commit_planes Waiting for fences timed out, Google Docs and AMD Radeon RX 6400

Categories

(Core :: Graphics, defect)

Product:
Core
Component:
Graphics
Platform:
Desktop
Linux
Type:
defect

Tracking

()

People

(Reporter: hsivonen, Unassigned)

References

(Blocks 1 open bug)

Details

Configuration

  • Firefox Nightly
  • Ubuntu 22.04 with all updates applied
  • AMD Radeon RX 6400
  • Gnome Wayland session
  • WebRender compositing
  • amdgpu kernel driver, about:support saying:
    • Description AMD BEIGE_GOBY (LLVM 13.0.1, DRM 3.42, 5.15.0-58-generic)
    • Vendor ID 0x1002
    • Device ID 0x743f
    • Driver Vendor mesa/radeonsi
    • Driver Version 22.0.5.0

Problem

From time to time, when the frontmost window is showing a wordprocessing GDoc (but, ironically, also once when the frontmost window was showing about:support):

  1. The screen goes black with the DisplayPort link dropping for a moment (the screen shows the input source symbol that it shows when the link comes back).
  2. The picture comes back
  3. Mouse cursor can be moved and changes according to position based on what's in the Firefox window
  4. In the Firefox window, the pixels show a loop of three (not 2, but not sure whether more than 3) most recently shown pixel states. It's like triple-buffering buffers getting reshown in a loop.
  5. Attempting to interact with the Gnome Dock does not cause any new non-mouse-cursor pixels on the screen.
  6. It seems that new windows can't be created even without being shown: pressing ctrl-alt-t to hopefully create a terminal and then typing something like echo foo > thetest.txt does not leave a file on the disk.
  7. Connecting to the computer via ssh or switching to a text-based local virtual console works.
  8. When switching to a text-based virtual console, after logging in, the following line is dumped to the console two or three times: [timestamp] [drm:amdgpu_cs_ioctl [amdgpu]] *ERROR* Failed to initialize parser -125!, where [timestamp] is a kernel log timestamp since boot. (This line does not appear if switching to a text-based console before the bug has manifested in the GUI session.)

Additional info

I have recently switched from a computer that had Radeon RX 460 to a computer that has Radeon RX 6400 (both running Ubuntu 22.04 and Gnome Wayland session). This bug didn't appear on the old computer, so this is either a new bug introduced in the last 5 or so weeks or this is specific to the newer microarchitecture in RX 6400 relative to RX 460.

From kernel.log:

Jan 25 10:02:10 thread kernel: [ 4481.192402] amdgpu 0000:53:00.0: [drm] *ERROR* mstb 00000000552db651 port 8: DPCD read on addr 0x60 for 1 bytes NAKed
Jan 25 10:05:44 thread kernel: [ 4690.661755] [drm:amdgpu_dm_commit_planes [amdgpu]] *ERROR* Waiting for fences timed out!
Jan 25 10:05:44 thread kernel: [ 4695.791621] [drm:amdgpu_job_timedout [amdgpu]] *ERROR* ring gfx_0.0.0 timeout, signaled seq=474675, emitted seq=474677
Jan 25 10:05:44 thread kernel: [ 4695.791867] [drm:amdgpu_job_timedout [amdgpu]] *ERROR* Process information: process firefox-bin pid 3827 thread firefox:cs0 pid 3896
Jan 25 10:05:44 thread kernel: [ 4695.792080] amdgpu 0000:53:00.0: amdgpu: GPU reset begin!
Jan 25 10:05:45 thread kernel: [ 4696.347045] amdgpu 0000:53:00.0: [drm:amdgpu_ring_test_helper [amdgpu]] *ERROR* ring kiq_2.1.0 test failed (-110)
Jan 25 10:05:45 thread kernel: [ 4696.347129] [drm:gfx_v10_0_hw_fini [amdgpu]] *ERROR* KGQ disable failed
Jan 25 10:05:45 thread kernel: [ 4696.619532] amdgpu 0000:53:00.0: [drm:amdgpu_ring_test_helper [amdgpu]] *ERROR* ring kiq_2.1.0 test failed (-110)
Jan 25 10:05:45 thread kernel: [ 4696.619605] [drm:gfx_v10_0_hw_fini [amdgpu]] *ERROR* KCQ disable failed
Jan 25 10:05:45 thread kernel: [ 4696.892055] [drm:gfx_v10_0_cp_gfx_enable.isra.0 [amdgpu]] *ERROR* failed to halt cp gfx
Jan 25 10:05:45 thread kernel: [ 4696.916295] [drm] free PSP TMR buffer
Jan 25 10:05:45 thread kernel: [ 4696.957785] amdgpu 0000:53:00.0: AMD-Vi: Event logged [IO_PAGE_FAULT domain=0x0021 address=0xf7e676ef900 flags=0x0020]
Jan 25 10:05:45 thread kernel: [ 4696.957795] amdgpu 0000:53:00.0: AMD-Vi: Event logged [IO_PAGE_FAULT domain=0x0021 address=0xf7e43424000 flags=0x0000]
Jan 25 10:05:45 thread kernel: [ 4696.957797] amdgpu 0000:53:00.0: amdgpu: MODE1 reset
Jan 25 10:05:45 thread kernel: [ 4696.957800] amdgpu 0000:53:00.0: amdgpu: GPU mode1 reset
Jan 25 10:05:45 thread kernel: [ 4696.957802] amdgpu 0000:53:00.0: AMD-Vi: Event logged [IO_PAGE_FAULT domain=0x0021 address=0xf7e43426500 flags=0x0000]
Jan 25 10:05:45 thread kernel: [ 4696.957808] amdgpu 0000:53:00.0: AMD-Vi: Event logged [IO_PAGE_FAULT domain=0x0021 address=0xf7e4342fc00 flags=0x0000]
Jan 25 10:05:45 thread kernel: [ 4696.957814] amdgpu 0000:53:00.0: AMD-Vi: Event logged [IO_PAGE_FAULT domain=0x0021 address=0xf7e676eff00 flags=0x0020]
Jan 25 10:05:45 thread kernel: [ 4696.957818] amdgpu 0000:53:00.0: AMD-Vi: Event logged [IO_PAGE_FAULT domain=0x0021 address=0xf7e43427500 flags=0x0000]
Jan 25 10:05:45 thread kernel: [ 4696.957823] amdgpu 0000:53:00.0: AMD-Vi: Event logged [IO_PAGE_FAULT domain=0x0021 address=0xf7e676f9a00 flags=0x0020]
Jan 25 10:05:45 thread kernel: [ 4696.957827] amdgpu 0000:53:00.0: AMD-Vi: Event logged [IO_PAGE_FAULT domain=0x0021 address=0xf7e676fbb00 flags=0x0020]
Jan 25 10:05:45 thread kernel: [ 4696.957830] amdgpu 0000:53:00.0: AMD-Vi: Event logged [IO_PAGE_FAULT domain=0x0021 address=0xf7e43427300 flags=0x0000]
Jan 25 10:05:45 thread kernel: [ 4696.957832] amdgpu 0000:53:00.0: AMD-Vi: Event logged [IO_PAGE_FAULT domain=0x0021 address=0xf7e4342f400 flags=0x0000]
Jan 25 10:05:45 thread kernel: [ 4696.957866] amdgpu 0000:53:00.0: amdgpu: GPU smu mode1 reset
Jan 25 10:05:45 thread kernel: [ 4696.958606] AMD-Vi: IOMMU event log overflow
Jan 25 10:05:46 thread kernel: [ 4697.477860] amdgpu 0000:53:00.0: amdgpu: GPU reset succeeded, trying to resume
Jan 25 10:05:46 thread kernel: [ 4697.478184] [drm] PCIE GART of 512M enabled (table at 0x0000008000E10000).
Jan 25 10:05:46 thread kernel: [ 4697.478208] [drm] VRAM is lost due to GPU reset!
Jan 25 10:05:46 thread kernel: [ 4697.479036] [drm] PSP is resuming...
Jan 25 10:05:46 thread kernel: [ 4697.572163] [drm] reserve 0xa00000 from 0x80fe000000 for PSP TMR
Jan 25 10:05:46 thread kernel: [ 4697.663842] amdgpu 0000:53:00.0: amdgpu: RAS: optional ras ta ucode is not available
Jan 25 10:05:46 thread kernel: [ 4697.678613] amdgpu 0000:53:00.0: amdgpu: SECUREDISPLAY: securedisplay ta ucode is not available
Jan 25 10:05:46 thread kernel: [ 4697.678616] amdgpu 0000:53:00.0: amdgpu: SMU is resuming...
Jan 25 10:05:46 thread kernel: [ 4697.678619] amdgpu 0000:53:00.0: amdgpu: smu driver if version = 0x0000000d, smu fw if version = 0x0000000f, smu fw version = 0x00491a00 (73.26.0)
Jan 25 10:05:46 thread kernel: [ 4697.678622] amdgpu 0000:53:00.0: amdgpu: SMU driver if version not matched
Jan 25 10:05:46 thread kernel: [ 4697.720949] amdgpu 0000:53:00.0: amdgpu: SMU is resumed successfully!
Jan 25 10:05:46 thread kernel: [ 4697.721488] [drm] kiq ring mec 2 pipe 1 q 0
Jan 25 10:05:46 thread kernel: [ 4697.722860] [drm] DMUB hardware initialized: version=0x0202000C
Jan 25 10:05:46 thread kernel: [ 4698.030403] [drm] VCN decode and encode initialized successfully(under DPG Mode).
Jan 25 10:05:46 thread kernel: [ 4698.030418] amdgpu 0000:53:00.0: amdgpu: ring gfx_0.0.0 uses VM inv eng 0 on hub 0
Jan 25 10:05:46 thread kernel: [ 4698.030421] amdgpu 0000:53:00.0: amdgpu: ring comp_1.0.0 uses VM inv eng 1 on hub 0
Jan 25 10:05:46 thread kernel: [ 4698.030422] amdgpu 0000:53:00.0: amdgpu: ring comp_1.1.0 uses VM inv eng 4 on hub 0
Jan 25 10:05:46 thread kernel: [ 4698.030422] amdgpu 0000:53:00.0: amdgpu: ring comp_1.2.0 uses VM inv eng 5 on hub 0
Jan 25 10:05:46 thread kernel: [ 4698.030423] amdgpu 0000:53:00.0: amdgpu: ring comp_1.3.0 uses VM inv eng 6 on hub 0
Jan 25 10:05:46 thread kernel: [ 4698.030424] amdgpu 0000:53:00.0: amdgpu: ring comp_1.0.1 uses VM inv eng 7 on hub 0
Jan 25 10:05:46 thread kernel: [ 4698.030424] amdgpu 0000:53:00.0: amdgpu: ring comp_1.1.1 uses VM inv eng 8 on hub 0
Jan 25 10:05:46 thread kernel: [ 4698.030425] amdgpu 0000:53:00.0: amdgpu: ring comp_1.2.1 uses VM inv eng 9 on hub 0
Jan 25 10:05:46 thread kernel: [ 4698.030425] amdgpu 0000:53:00.0: amdgpu: ring comp_1.3.1 uses VM inv eng 10 on hub 0
Jan 25 10:05:46 thread kernel: [ 4698.030426] amdgpu 0000:53:00.0: amdgpu: ring kiq_2.1.0 uses VM inv eng 11 on hub 0
Jan 25 10:05:46 thread kernel: [ 4698.030427] amdgpu 0000:53:00.0: amdgpu: ring sdma0 uses VM inv eng 12 on hub 0
Jan 25 10:05:46 thread kernel: [ 4698.030427] amdgpu 0000:53:00.0: amdgpu: ring vcn_dec_0 uses VM inv eng 0 on hub 1
Jan 25 10:05:46 thread kernel: [ 4698.033257] amdgpu 0000:53:00.0: amdgpu: recover vram bo from shadow start
Jan 25 10:05:46 thread kernel: [ 4698.033456] amdgpu 0000:53:00.0: amdgpu: recover vram bo from shadow done
Jan 25 10:05:46 thread kernel: [ 4698.033458] [drm] Skip scheduling IBs!
Jan 25 10:05:46 thread kernel: [ 4698.033459] [drm] Skip scheduling IBs!
Jan 25 10:05:46 thread kernel: [ 4698.033490] amdgpu 0000:53:00.0: amdgpu: GPU reset(2) succeeded!
Jan 25 10:05:46 thread kernel: [ 4698.033500] [drm] Skip scheduling IBs!

The last line repeats, then the previously-mentioned error starts repeating:

Jan 25 10:05:46 thread kernel: [ 4698.033730] [drm:amdgpu_cs_ioctl [amdgpu]] *ERROR* Failed to initialize parser -125!

With lines like:

Jan 25 10:06:00 thread kernel: [ 4711.417825] amdgpu_cs_ioctl: 19 callbacks suppressed

interleaved from time to time with the number of callbacks differing.

Looking at multiple log sequences of this searching for "GPU reset begin", it always starts with [drm:amdgpu_dm_commit_planes [amdgpu]] *ERROR* Waiting for fences timed out!.

Upstream report with a GPU microarchitecture between RX 460 and RX 6400 with Gecko-based processes as trigger under X session:
https://gitlab.freedesktop.org/drm/amd/-/issues/934

Upstream report with a non-Gnome X window manager as the trigger:
https://gitlab.freedesktop.org/drm/amd/-/issues/1766

Summary: Gnome Shell or WebRender triple-buffering loop with Google Docs and AMD Radeon RX 6400 → Triple-buffering loop with amdgpu_dm_commit_planes Waiting for fences timed out, Google Docs and AMD Radeon RX 6400

The last line repeats, then the previously-mentioned error starts repeating:

The repeating line being [drm] Skip scheduling IBs!. Bugzilla doesn't make it obvious that the preformatted section is scrollable.

(In reply to Henri Sivonen (:hsivonen) from comment #3)

Upstream report with a non-Gnome X window manager as the trigger:
https://gitlab.freedesktop.org/drm/amd/-/issues/1766

This also mentions scrolling in Firefox (under X, the manifestation on about:support for me happened when scrolling; the GDoc instances didn't involve scrolling but typing text).

Blocks: wr-linux
OS: Unspecified → Linux
Hardware: Unspecified → Desktop
Summary: Triple-buffering loop with amdgpu_dm_commit_planes Waiting for fences timed out, Google Docs and AMD Radeon RX 6400 → Firefox+Thunderbird+other apps/AMD/X11+Wayland: Triple-buffering loop with amdgpu_dm_commit_planes Waiting for fences timed out, Google Docs and AMD Radeon RX 6400

Arch Wiki associates what looks like this with dynamic power management that seems to be enabled by default for RX 6400 but perhaps not enabled by default for RX 460.

I'm now testing with the radeon.dpm=0 kernel parameter suggested by Arch Wiki, which should disable dynamic power management.

(In reply to Henri Sivonen (:hsivonen) from comment #6)

I'm now testing with the radeon.dpm=0 kernel parameter suggested by Arch Wiki, which should disable dynamic power management.

This kernel parameter did not make the problem go away.

(In reply to Henri Sivonen (:hsivonen) from comment #7)

(In reply to Henri Sivonen (:hsivonen) from comment #6)

I'm now testing with the radeon.dpm=0 kernel parameter suggested by Arch Wiki, which should disable dynamic power management.

This kernel parameter did not make the problem go away.

Hmm. Since the kernel driver is "amdgpu" and not "radeon", perhaps I need a kernel parameter starting with "amdgpu."

For now, I'm trying to force the power/performance mode to "low" in order to get rid of power level changes that way.

amdgpu.dpm=0 makes the screen stay black at boot.

Starting Firefox with LIBGL_ALWAYS_SOFTWARE=1 seems to avoid the problem.

Found a news item about a fix for a bug involving the same error message.

Now trying 6.2.0rc6 mainline kernel.

For now, I'm trying to force the power/performance mode to "low" in order to get rid of power level changes that way.

This didn't help.

Making a note of the kernel patches so that I don't lose the link.

With the 6.2.0rc6 kernel, the error about fences goes away, but the GPU restart followed by the triple(?) buffering loop does not.

In the logs, this had happened and had been recoverable:

Jan 30 16:01:46 thread kernel: [19949.305134] amdgpu 0000:53:00.0: amdgpu: [gfxhub] page fault (src_id:0 ring:24 vmid:4 pasid:32770, for
 process firefox-bin pid 3433 thread firefox:cs0 pid 3502)
Jan 30 16:01:46 thread kernel: [19949.305141] amdgpu 0000:53:00.0: amdgpu:   in page starting at address 0x000080031d20c000 from client 
0x1b (UTCL2)
Jan 30 16:01:46 thread kernel: [19949.305144] amdgpu 0000:53:00.0: amdgpu: GCVM_L2_PROTECTION_FAULT_STATUS:0x00401030
Jan 30 16:01:46 thread kernel: [19949.305145] amdgpu 0000:53:00.0: amdgpu:       Faulty UTCL2 client ID: TCP (0x8)
Jan 30 16:01:46 thread kernel: [19949.305146] amdgpu 0000:53:00.0: amdgpu:       MORE_FAULTS: 0x0
Jan 30 16:01:46 thread kernel: [19949.305147] amdgpu 0000:53:00.0: amdgpu:       WALKER_ERROR: 0x0
Jan 30 16:01:46 thread kernel: [19949.305148] amdgpu 0000:53:00.0: amdgpu:       PERMISSION_FAULTS: 0x3
Jan 30 16:01:46 thread kernel: [19949.305148] amdgpu 0000:53:00.0: amdgpu:       MAPPING_ERROR: 0x0
Jan 30 16:01:46 thread kernel: [19949.305149] amdgpu 0000:53:00.0: amdgpu:       RW: 0x0
Jan 30 16:01:56 thread kernel: [19959.364522] [drm:amdgpu_job_timedout [amdgpu]] *ERROR* ring gfx_0.0.0 timeout, but soft recovered

Then later, the GPU reset leading to the triple buffering loop looked like this in the logs (no log lines between the above and the below copypaste!):

Jan 30 17:16:01 thread kernel: [24404.045796] amdgpu 0000:53:00.0: amdgpu: [gfxhub] page fault (src_id:0 ring:24 vmid:4 pasid:32770, for
 process firefox-bin pid 3433 thread firefox:cs0 pid 3502)
Jan 30 17:16:01 thread kernel: [24404.045806] amdgpu 0000:53:00.0: amdgpu:   in page starting at address 0x00008002f134e000 from client 
0x1b (UTCL2)
Jan 30 17:16:01 thread kernel: [24404.045811] amdgpu 0000:53:00.0: amdgpu: GCVM_L2_PROTECTION_FAULT_STATUS:0x00401030
Jan 30 17:16:01 thread kernel: [24404.045813] amdgpu 0000:53:00.0: amdgpu:       Faulty UTCL2 client ID: TCP (0x8)
Jan 30 17:16:01 thread kernel: [24404.045816] amdgpu 0000:53:00.0: amdgpu:       MORE_FAULTS: 0x0
Jan 30 17:16:01 thread kernel: [24404.045817] amdgpu 0000:53:00.0: amdgpu:       WALKER_ERROR: 0x0
Jan 30 17:16:01 thread kernel: [24404.045819] amdgpu 0000:53:00.0: amdgpu:       PERMISSION_FAULTS: 0x3
Jan 30 17:16:01 thread kernel: [24404.045820] amdgpu 0000:53:00.0: amdgpu:       MAPPING_ERROR: 0x0
Jan 30 17:16:01 thread kernel: [24404.045822] amdgpu 0000:53:00.0: amdgpu:       RW: 0x0
Jan 30 17:16:11 thread kernel: [24414.083593] [drm:amdgpu_job_timedout [amdgpu]] *ERROR* ring gfx_0.0.0 timeout, but soft recovered
Jan 30 17:16:16 thread kernel: [24418.930649] usb 1-1.2: input irq status -75 received
Jan 30 17:16:21 thread kernel: [24424.333131] [drm:amdgpu_job_timedout [amdgpu]] *ERROR* ring gfx_0.0.0 timeout, signaled seq=779410, em
itted seq=779413
Jan 30 17:16:21 thread kernel: [24424.333539] [drm:amdgpu_job_timedout [amdgpu]] *ERROR* Process information: process firefox-bin pid 34
33 thread firefox:cs0 pid 3502
Jan 30 17:16:21 thread kernel: [24424.333910] amdgpu 0000:53:00.0: amdgpu: GPU reset begin!
Jan 30 17:16:21 thread kernel: [24424.696596] amdgpu 0000:53:00.0: amdgpu: MODE1 reset
Jan 30 17:16:21 thread kernel: [24424.696602] amdgpu 0000:53:00.0: amdgpu: GPU mode1 reset
Jan 30 17:16:21 thread kernel: [24424.696669] amdgpu 0000:53:00.0: amdgpu: GPU smu mode1 reset
Jan 30 17:16:22 thread kernel: [24425.219386] amdgpu 0000:53:00.0: amdgpu: GPU reset succeeded, trying to resume
Jan 30 17:16:22 thread kernel: [24425.219755] [drm] PCIE GART of 512M enabled (table at 0x0000008000F00000).
Jan 30 17:16:22 thread kernel: [24425.219835] [drm] VRAM is lost due to GPU reset!
Jan 30 17:16:22 thread kernel: [24425.219837] [drm] PSP is resuming...
Jan 30 17:16:22 thread kernel: [24425.296000] [drm] reserve 0xa00000 from 0x80fd000000 for PSP TMR
Jan 30 17:16:22 thread kernel: [24425.386492] amdgpu 0000:53:00.0: amdgpu: RAS: optional ras ta ucode is not available
Jan 30 17:16:22 thread kernel: [24425.401368] amdgpu 0000:53:00.0: amdgpu: SECUREDISPLAY: securedisplay ta ucode is not available
Jan 30 17:16:22 thread kernel: [24425.401371] amdgpu 0000:53:00.0: amdgpu: SMU is resuming...
Jan 30 17:16:22 thread kernel: [24425.401375] amdgpu 0000:53:00.0: amdgpu: smu driver if version = 0x0000000d, smu fw if version = 0x0000000f, smu fw program = 0, version = 0x00491a00 (73.26.0)
Jan 30 17:16:22 thread kernel: [24425.401378] amdgpu 0000:53:00.0: amdgpu: SMU driver if version not matched
Jan 30 17:16:22 thread kernel: [24425.401410] amdgpu 0000:53:00.0: amdgpu: use vbios provided pptable
Jan 30 17:16:22 thread kernel: [24425.445423] amdgpu 0000:53:00.0: amdgpu: SMU is resumed successfully!
Jan 30 17:16:22 thread kernel: [24425.446380] [drm] DMUB hardware initialized: version=0x0202000C
Jan 30 17:16:22 thread kernel: [24425.764052] [drm] kiq ring mec 2 pipe 1 q 0
Jan 30 17:16:22 thread kernel: [24425.766993] [drm] VCN decode and encode initialized successfully(under DPG Mode).
Jan 30 17:16:22 thread kernel: [24425.767007] amdgpu 0000:53:00.0: amdgpu: ring gfx_0.0.0 uses VM inv eng 0 on hub 0
Jan 30 17:16:22 thread kernel: [24425.767009] amdgpu 0000:53:00.0: amdgpu: ring comp_1.0.0 uses VM inv eng 1 on hub 0
Jan 30 17:16:22 thread kernel: [24425.767010] amdgpu 0000:53:00.0: amdgpu: ring comp_1.1.0 uses VM inv eng 4 on hub 0
Jan 30 17:16:22 thread kernel: [24425.767011] amdgpu 0000:53:00.0: amdgpu: ring comp_1.2.0 uses VM inv eng 5 on hub 0
Jan 30 17:16:22 thread kernel: [24425.767012] amdgpu 0000:53:00.0: amdgpu: ring comp_1.3.0 uses VM inv eng 6 on hub 0
Jan 30 17:16:22 thread kernel: [24425.767012] amdgpu 0000:53:00.0: amdgpu: ring comp_1.0.1 uses VM inv eng 7 on hub 0
Jan 30 17:16:22 thread kernel: [24425.767013] amdgpu 0000:53:00.0: amdgpu: ring comp_1.1.1 uses VM inv eng 8 on hub 0
Jan 30 17:16:22 thread kernel: [24425.767014] amdgpu 0000:53:00.0: amdgpu: ring comp_1.2.1 uses VM inv eng 9 on hub 0
Jan 30 17:16:22 thread kernel: [24425.767014] amdgpu 0000:53:00.0: amdgpu: ring comp_1.3.1 uses VM inv eng 10 on hub 0
Jan 30 17:16:22 thread kernel: [24425.767015] amdgpu 0000:53:00.0: amdgpu: ring kiq_2.1.0 uses VM inv eng 11 on hub 0
Jan 30 17:16:22 thread kernel: [24425.767015] amdgpu 0000:53:00.0: amdgpu: ring sdma0 uses VM inv eng 12 on hub 0
Jan 30 17:16:22 thread kernel: [24425.767016] amdgpu 0000:53:00.0: amdgpu: ring vcn_dec_0 uses VM inv eng 0 on hub 1
Jan 30 17:16:22 thread kernel: [24425.769407] amdgpu 0000:53:00.0: amdgpu: recover vram bo from shadow start
Jan 30 17:16:22 thread kernel: [24425.769709] amdgpu 0000:53:00.0: amdgpu: recover vram bo from shadow done
Jan 30 17:16:22 thread kernel: [24425.769716] amdgpu 0000:53:00.0: amdgpu: GPU reset(4) succeeded!
Jan 30 17:16:22 thread kernel: [24425.769763] [drm] Skip scheduling IBs!
Jan 30 17:16:22 thread kernel: [24425.769780] [drm] Skip scheduling IBs!
Jan 30 17:16:22 thread kernel: [24425.769789] [drm] Skip scheduling IBs!
Jan 30 17:16:22 thread kernel: [24425.769796] [drm] Skip scheduling IBs!
Jan 30 17:16:22 thread kernel: [24425.769800] [drm] Skip scheduling IBs!
Jan 30 17:16:22 thread kernel: [24425.769810] [drm] Skip scheduling IBs!
Jan 30 17:16:22 thread kernel: [24425.769810] [drm] Skip scheduling IBs!

(The last line was repeated some more times.)

(The last line was repeated some more times.)

The last line being [drm] Skip scheduling IBs!. Bugzilla, again, does not make it obvious that the block of text is scrollable.

Severity: -- → S3
You need to log in before you can comment on or make changes to this bug.