00:08 Mangix: Venemo: I don't think that's right. For reference, I've gotten 1440p110 working on dual link DVI in Windows with patched drivers. 85 or 90 on non DC. Can't go over 75 or 80 with DC.
02:47 Mangix: Bah humbug. DC on 6.17.8 is broken as well.
02:50 Mangix: I'm guessing https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/drivers?id=7495962cbceb967e095233a5673ea71f3bcdee7e is missing.
02:51 Mangix: edp...yeah I'm using DP-MST
03:54 Mangix: hilarious thing I found out. My nightlight screen reddening thing works when DC is enabled. Does nothing with DC disabled.
05:54 Venemo: Mangix: 85 or 90 what?
07:17 Venemo: Mangix: also, what is broken on 6.17.8?
16:55 Mangix: Venemo: I mean there's a black screen. That is, when it boots, it goes black. BIOS and bootloader are fine.
16:55 Mangix: I have no idea if it even is operational.
17:10 johnny0: temporarily disable your graphical environment and blacklist amdgpu at boot with modprobe.blacklist=amdgpu -- you can then modprobe amdgpu after the machine boots to have a chance at getting the dmesg log
17:12 johnny0: either via sshing in or, with systemd, by blindly shutting gracefully then checking the previous log with journalctl -b -1
17:12 Venemo: it's likely the same crash that was already reported weeks ago
17:12 Venemo: I sent a patch series to the 6.17 maintainers with the fix. it seems they haven't included the fix
17:13 Venemo: they told me they would, but they haven't
17:14 Venemo: they included a few more fixes but just not the one that fixes the crash
17:14 Venemo: I don't know what more I can do
17:15 Venemo: Mangix: I am sorry for the trouble. the best I can recommend is try 6.18
17:26 johnny0: Venemo: oh, it does look like some are in the stable queue though, so maybe they'll get through for 6.17.9
18:23 Venemo: johnny0: where do you check this stable queue?
18:44 johnny0: https://git.kernel.org/pub/scm/linux/kernel/git/stable/stable-queue.git/log/
18:45 johnny0: you should have been emailed about them as they got added, but if you get an absurd number of emails per day they can easily get lost in the noise
19:00 Venemo: I got some emails, but there were too many to tell precisely
19:01 Venemo: johnny0: where do I look, on that site?
19:15 johnny0: if you click on the "fixes for all trees" commits after the 6.17.8 tag you can see which commits are in
19:16 johnny0: are in the queue for the next set of stable releases that is
19:20 Venemo: oh, I see they did backport it after all. that's nice
19:21 Venemo: I guess just need to wait for the next release then
19:51 Mangix: Venemo: these are your commits missing from 6.17 as of .8: https://gist.github.com/neheb/77958b0a46c90ec8c1893ec9b4ec3c2f
19:52 Mangix: the fastboot one seems most likely.
19:53 Mangix: anyway, I have no issue running a slightly older kernel. 6.19 seems to be the most interesting.
20:04 Venemo: Mangix: the fastboot one isn't relevant. the actual fix is "drm/amd/pm: Use pm_display_cfg in legacy DPM (v2)"
22:07 Caterpillar: do you know if https://paste.centos.org/view/66852728 already has a duplicate in https://gitlab.freedesktop.org/drm/amd/-/issues/ ? Unfortunately Freedesktop banned LLM crawlers so finding a duplicate is way more time demanding
22:22 Venemo: Caterpillar: that just looks like a generic gpu hang copied from a dmesg log. it isn't actionable
22:23 Caterpillar: Venemo: probably I should have collected /sys/class/drm/card1/device/devcoredump/data but I did not have the opportunity to do that