Bugzilla – Bug 1216194
i915 drm: drivers/gpu/drm/drm_vblank.c:1310 drm_wait_one_vblank+0x1cb/0x1f0 (Intermittent complete freezes of the system)
Last modified: 2023-11-17 08:44:55 UTC
Hi, I've experienced this with my old notebook from time to time: The complete system freezes and there is no other option than resetting it. On the old notebook that happened every or every second day. With the new one it happens twice or three times a day. There are various reports on the internet, e.g. https://www.reddit.com/r/openSUSE/comments/12dup7j/opensuse_freezes_completely_i_cant_stand_it/?rdt=35354 or https://forums.opensuse.org/t/where-to-look-at-to-diagnose-for-random-freeze-crash/132293 Some facts: - It is not related to any tumbleweed snapshot as I am following them as they are released - Both, old and new, are Dell notebooks - When it happens nothing can be done at all, not even starting a non-desktop console session - KDE desktop - X, not wayland Here is the boot log I think is relevant: https://paste.opensuse.org/pastes/e4f2a769a9ba I am kind of stuck as I don't know how to debug this. Let me know what I can do to investigate this further. Regards Daniel
Please try to get at least some crash dump or contributing factor, otherwise this is not actionable in any way.
Thank you Andreas. I've enabled crash dumps and will upload once it re-produces. Best regards Daniel
Created attachment 870143 [details] journalctl -fn output while the system is not responding
So it happened again but did not create a crash dump, as it actually did not crash but just became unresponsive. I've waited a bit longer and was able to login in a non graphical session. I've attached the journalctl capture for that time. Is this of any help? I've unplugged the external monitor (connected via hdmi) and the system came back to normal operations.
Created attachment 870174 [details] journactl -fn until unresponsiveness happens This is a "journalctl -fn" which was running in the background until the issue happens
Now properly indented. > i915 0000:00:02.0: [drm] *ERROR* [CRTC:131:pipe B] flip_done timed out > i915 0000:00:02.0: [drm] *ERROR* flip_done timed out > i915 0000:00:02.0: [drm] *ERROR* [CRTC:131:pipe B] commit wait timed out > i915 0000:00:02.0: [drm] *ERROR* [CRTC:131:pipe B] flip_done timed out > i915 0000:00:02.0: [drm] *ERROR* flip_done timed out > i915 0000:00:02.0: [drm] *ERROR* [CRTC:131:pipe B] commit wait timed out > i915 0000:00:02.0: [drm] *ERROR* [CRTC:131:pipe B] flip_done timed out > i915 0000:00:02.0: [drm] *ERROR* flip_done timed out > i915 0000:00:02.0: [drm] *ERROR* [CRTC:131:pipe B] commit wait timed out > i915 0000:00:02.0: [drm] *ERROR* flip_done timed out > i915 0000:00:02.0: [drm] *ERROR* [PLANE:82:plane 1B] commit wait timed out > ------------[ cut here ]------------ > i915 0000:00:02.0: vblank wait timed out on crtc 1 > WARNING: CPU: 8 PID: 1538 at drivers/gpu/drm/drm_vblank.c:1310 drm_wait_one_vblank+0x1cb/0x1f0 > Modules linked in: usblp vhost_net ... > CPU: 8 PID: 1538 Comm: Xorg.bin Kdump: loaded Tainted: G W 6.5.6-1-default #1 openSUSE Tumbleweed 70271d4c483b2ae14bf13038e00d865831f85bbc > Hardware name: Dell Inc. Latitude 5540/0M5RRN, BIOS 1.7.1 09/12/2023 > RIP: 0010:drm_wait_one_vblank+0x1cb/0x1f0 > Code: c9 fe ff ff 48 8b 7b 08 4c 8b 67 50 4d 85 e4 74 25 e8 a9 2a 02 00 89 e9 4c 89 e2 48 c7 c7 70 b8 86 95 48 89 c6 e8 c5 08 7e ff <0f> 0b e9 9c fe ff ff 48 8b 1f eb 9a 4c 8b 27 eb d6 e8 9f dd 39 00 > RSP: 0000:ffff9e078104b9b8 EFLAGS: 00010286 > RAX: 0000000000000000 RBX: ffff900602c00000 RCX: 0000000000000027 > RDX: ffff900d5a0274c8 RSI: 0000000000000001 RDI: ffff900d5a0274c0 > RBP: 0000000000000001 R08: 0000000000000000 R09: ffff9e078104b860 > R10: 0000000000000003 R11: ffff900d59afffe8 R12: ffff900601972d10 > R13: 00000000001776d4 R14: ffff9006086ed160 R15: 0000000010000000 > FS: 00007f99b56cc980(0000) GS:ffff900d5a000000(0000) knlGS:0000000000000000 > CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 > CR2: 000014dc0fb8bc50 CR3: 0000000106508000 CR4: 0000000000f52ee0 > PKRU: 55555554 > Call Trace: > <TASK> > hsw_audio_codec_disable+0xe5/0x150 [i915 155b0d0d67e4ec00151b49d39e823b5e9f2b3fe4] > intel_audio_codec_disable+0xa8/0x170 [i915 155b0d0d67e4ec00151b49d39e823b5e9f2b3fe4] > intel_disable_ddi+0x76/0x210 [i915 155b0d0d67e4ec00151b49d39e823b5e9f2b3fe4] > intel_encoders_disable+0x8a/0xb0 [i915 155b0d0d67e4ec00151b49d39e823b5e9f2b3fe4] > hsw_crtc_disable+0x129/0x140 [i915 155b0d0d67e4ec00151b49d39e823b5e9f2b3fe4] > intel_atomic_commit_tail+0x3f2/0xe00 [i915 155b0d0d67e4ec00151b49d39e823b5e9f2b3fe4] > intel_atomic_commit+0x341/0x390 [i915 155b0d0d67e4ec00151b49d39e823b5e9f2b3fe4] > drm_atomic_commit+0x97/0xd0 > drm_atomic_connector_commit_dpms+0xce/0xf0 > drm_mode_obj_set_property_ioctl+0x189/0x3a0 > drm_connector_property_set_ioctl+0x3b/0x60 > drm_ioctl_kernel+0xc2/0x170 > drm_ioctl+0x256/0x490
Created attachment 870270 [details] Here is a dmesg from another freeze
Is there anything else I can provide to help fixing this?
No, removing needinfo flag.
Could you give the hardware details? At least hwinfo output. Then boot with nomodeset boot option. This will disable the use of the native graphics driver, hence the vblank problem must be avoided.
Created attachment 870335 [details] hwinfo
Small update: I've applied the settings mentioned here: https://community.frame.work/t/tracking-hard-freezing-on-fedora-36-with-the-new-12th-gen-system/20675/146 This seems to avoid freezes without an external monitor, but freezes still happen when an external monitor is connected. It doesn't matter if HDMI or USB-C. I'll try the nomodeset boot option now and check if that helps. Regards Daniel
I've tested the nomodeset boot mode option, but immediately reverted as this disabled the detection of the external monitor
Anything else I can do?
Was this somehow fixed? I did not have a single freeze the last days
It is possibly fixed by some of these commits: $ git log --oneline v6.5.6..v6.5.9 drivers/gpu/drm/i915/ 935df6cfa78a drm/i915/cx0: Only clear/set the Pipe Reset bit of the PHY Lanes Owned 2c9e90773fab drm/i915: Retry gtt fault when out of fence registers 1d61ce0a321f drm/i915: Register engines early to avoid type confusion 2f92524f5a7f drm/i915: Don't set PIPE_CONTROL_FLUSH_L3 for aux inval
Cool, thanks for the update
Small update: This is still happening from time to time