09:42airlied[d]: gfxstrand[d]: just fyi my turing laptop is 3 hrs into a 4hr cts run, seems to be doing fine, booted on i915 gpu and testing on nvk as secondary,
10:00asdqueerfromeu[d]: https://cdn.discordapp.com/attachments/1034184951790305330/1277568453615816704/systemd-journal-nouveau.txt?ex=66cda3d7&is=66cc5257&hm=79f6a6d940f3ca540eab1aadd0418a4c126b03dedf117a61b33193dca73a63ee&
10:00asdqueerfromeu[d]: These nouveau instabilities really require a NVK port to OGK 🍩
20:30airlied[d]: gfxstrand[d]: no idea if this is helpful https://paste.centos.org/view/raw/b8597901
20:30airlied[d]: but the laptop run did finish without incident
20:30airlied[d]: asdqueerfromeu[d]: what caused that?
20:35asdqueerfromeu[d]: airlied[d]: I launched Mutter Wayland after KWin Wayland is already running (after I ran a fullscreen game (which used radeonsi) on the Mutter session it defaulted to the nouveau monitor and the whole laptop kind of froze but the game's sound continued to play)
20:38airlied[d]: some fence got seriously mangled in all that, not sure ogk would save you
20:38gfxstrand[d]: airlied[d]: Those dEQP-VK.pipeline.fast_linked_library.multisample.min_sample_shading.* fails are odd
20:41asdqueerfromeu[d]: asdqueerfromeu[d]: After that fail I force powered off the laptop (after trying REISUB) but I got a completely different issue (a DRM oops which caused the system to partially hang after launching an OpenGL application when GDM is enabled)
20:42gfxstrand[d]: The descriptor buffer capture/replay fails are a CTS bug that's been fixed
20:42airlied[d]: I do have an old cts, I'll update it
20:44asdqueerfromeu[d]: https://cdn.discordapp.com/attachments/1034184951790305330/1277730317020692583/systemd-journal-ttm.txt?ex=66ce3a96&is=66cce916&hm=0072b015495baf02345a66c77c4355a7017f1153469f4f478f3430ed4d9bc06d&
20:44asdqueerfromeu[d]: asdqueerfromeu[d]: Here's a journalctl snippet for reference (I'm not sure if this is related to that nouveau crash though)
20:48gfxstrand[d]: airlied[d]: The fix has been merged into an ancient branch but that merge hasn't propagated to main yet
20:48gfxstrand[d]: I just pulled main today and had to cherry-pick,