=== Eickmeyer is now known as Erich
=== Erich is now known as Eickmeyer
[10:15] <KGB-X> clutter Jeremy Bicha 435423 * commented merge request !1 * https://deb.li/80Eg
[14:50] <KGB-X> gedit signed tags d50b8b6 Amin Bandali upstream/46.1 * Upstream version 46.1 * https://deb.li/XuNe
[14:51] <KGB-X> gedit upstream/latest 20a82b3 Amin Bandali * pushed 216 commits (first 5 follow) * https://deb.li/3mu9y
[14:51] <KGB-X> gedit upstream/latest a45bce6 Sébastien Wilmet (via Amin Bandali) gedit/gedit-multi-notebook.c * MultiNotebook: fix compilation warning [-Wold-style-definition] * https://deb.li/3U8SG
[14:51] <KGB-X> gedit upstream/latest 9e49806 Sébastien Wilmet (via Amin Bandali) plugins/modelines/modeline-parser.c * modelines: fix compilation warning [-Wold-style-definition] * https://deb.li/itwXM
[14:51] <KGB-X> gedit upstream/latest d83c0cc Sébastien Wilmet (via Amin Bandali) gedit/gedit-documents-panel.c * DocumentsPanel: fix compil warning [-Wdeclaration-after-statement] * https://deb.li/3Yfy
[14:52] <KGB-X> gedit upstream/latest 31a688f Sébastien Wilmet (via Amin Bandali) plugins/filebrowser/gedit-file-browser-widget.c * filebrowser: fix compil warning [-Wdeclaration-after-statement] * https://deb.li/32ii4
[14:52] <KGB-X> gedit upstream/latest ebf60f5 Sébastien Wilmet (via Amin Bandali) plugins/modelines/modeline-parser.c * modelines: fix compil warning [-Wdeclaration-after-statement] * https://deb.li/z5BM
[14:52] <KGB-X> gedit pristine-tar fec8830 Amin Bandali gedit_46.1.orig.tar.xz.delta gedit_46.1.orig.tar.xz.id * pristine-tar data for gedit_46.1.orig.tar.xz * https://deb.li/387hj
[16:04] <Eickmeyer> jbicha: Hi! We had an interesting accessibility bug pop-up last night and I was unaware if there were any upstream hits that might be related: LP: #2039722
[16:04] -ubottu:#ubuntu-desktop- Launchpad bug 2039722 in gnome-settings-daemon (Ubuntu) "'Large Text' accessibility option cannot be turned off. Text is too large" [High, Confirmed] https://launchpad.net/bugs/2039722
[16:10] <lissyx> FTR I'm getting some amdgpu crashes on 23.10
[16:11] <Eickmeyer> lissyx: Sounds like a kernel regression. 😞
[16:11] <lissyx> Eickmeyer, I'm unsure, it's when using firefox
[16:13] <Eickmeyer> lissyx: Yeah, it might have something to do with GPU acceleration being attempted while using the kernel module, at least that's my suspicion. 
[16:14] <Eickmeyer> Sadly, I can't reproduce, I only have Nvidia.
[16:15] <lissyx> Eickmeyer, interesting, I was investigating that we now have h264 + vp9 hw decoding on firefox
[16:16] <Eickmeyer> lissyx: Yeah, that only serves to confirm my suspicion. Since the amdgpu modules are baked-in to the kernel, and you're not the first person I've seen having trouble with amdgpu so far, I highly suspect a kernel regression.
[16:22] <lissyx> I have a few logs
[16:22] <lissyx> first crash was also triggered by firefox :(
[16:23] <Eickmeyer> I suggest `ubuntu-bug linux` and attach those logs.
[16:27] <lissyx> https://bugs.launchpad.net/ubuntu/+source/linux/+bug/2039868
[16:27] -ubottu:#ubuntu-desktop- Launchpad bug 2039868 in linux (Ubuntu) "amdgpu reset during usage of firefox" [Undecided, New]
[16:29] <Eickmeyer> I'd attach the logs you got also unless they were automatically attached by apport.
[16:31] <lissyx> Eickmeyer, dmesg they are
[16:32] <Eickmeyer> Aha! Yep, they certainly are!
[16:36] <Eickmeyer> lissyx: with my very rudimentary knowledge of it, I can see that it was triggered by Firefox, but don't believe Firefox did anything out of the ordinary. It looks to me like the GPU timed-out, which tells me it failed to respond to the kernel in time, which leads me to blame the kernel.
[16:37] <lissyx> Eickmeyer, or it would not be the first time our code exercises mesa in a way that crashes
[16:37] <Eickmeyer> I guess that's possible too, but my clue is this:   116.875717] [drm:amdgpu_job_timedout [amdgpu]] *ERROR* ring gfx_low timeout, signaled seq=11496, emitted seq=11498
[16:38] <Eickmeyer> Then it begins resetting the GPU.
[16:38] <lissyx> 6.2 −> 6.5
[16:39] <Eickmeyer> That's a good point. Does this happen on 23.04 (6.2)?
[16:41] <lissyx> https://bugzilla.kernel.org/show_bug.cgi?id=201957
[16:41] -ubottu:#ubuntu-desktop- bugzilla.kernel.org bug 201957 in Drivers "amdgpu: ring gfx timeout" [Blocking, Resolved: Answered]
[16:41] <lissyx> Eickmeyer, I never had the issue on 23.04
[16:43] <Eickmeyer> Ah. Looks like from that bug, people were having no problem on 6.2, but had issues on 6.5. So, it's a regression in 6.5, which is in 23.10.
[16:44] <Eickmeyer> amdgpu.mcbp=0
[16:44]  * Eickmeyer cringes
[16:47] <Eickmeyer> lissyx: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/2039868/comments/4
[16:47] -ubottu:#ubuntu-desktop- Launchpad bug 2039868 in linux (Ubuntu) "amdgpu reset during usage of firefox" [Undecided, Confirmed]
[16:48] <lissyx> on https://gitlab.freedesktop.org/drm/amd/-/issues/2848 there are mentions of fixes on 6.6 kernel
[16:48] -ubottu:#ubuntu-desktop- Issue 2848 in drm/amd "amdgpu: [gfxhub0] no-retry page fault results in permanent system freeze (kernel 6.5.2-arch1-1)" [Opened]
[16:48] <lissyx> and some patch proposal for mesa?
[16:48] <lissyx> here mesa patch: https://gitlab.freedesktop.org/drm/amd/-/issues/2848#note_2095536
[16:48] <lissyx> hm mesa or kernel?
[16:48] <Eickmeyer> Oh nice. Good info to add.
[16:48] <lissyx> have to leave
[16:49] <Eickmeyer> Alright, I'll comment the bug accordingly. Have a nice evening!