01:49Tashtari: So I seem to be getting a kernel panic to do with nouveau that is triggered by dragging or resizing windows in X.
01:50Tashtari: Console output here: http://paste.debian.net/1129972/
01:50Tashtari: Does this look in any way familiar to anyone? =)
02:01Tashtari: Card is a GT218.
02:04imirkin: that's a new one...
02:04imirkin: skeggsb: --^
02:05imirkin: some kind of fail in g84_fifo_uevent_fini .. looks like use-after-free: RAX: 6b6b6b6b6b6b6b6b
02:07HdkR: rax = `kkkkkkkk` good stuff
02:13imirkin: 6b is the linux poison value
02:13imirkin: set on slab free or whatever
04:55Tashtari: Anything else I can do by way of troubleshooting in the meantime?
05:20Tashtari: I tried going down to one monitor (from two) just to see if that kept the problem from appearing... it did not.
05:20Tashtari: Console output the second here: http://paste.debian.net/1129978/
05:21Tashtari: It is slightly different but the trace is still the same as far down as nvkm_fifo_uevent, and includes the g84_fifo_uevent_fini that imirkin mentioned.
05:23Tashtari: As before, the thing that trips it off seems to be dragging windows in X11.
07:31sigod: apart from nouveau.noaccel=1 kernel parameter is there a way to allow acceleration in x and disable opengl or does that just do the same thing in terms of stopping the freeze in a gtx 660
07:31imirkin: sigod: you're aware that you can use blob firmware, right?
07:31sigod: stopping the context switching error i mean
07:32sigod: not on parabola linux
07:32imirkin: not sure what that means
07:32imirkin: but you can set up pretty much anything on any distro
07:32imirkin: nothing about linux is distro-specific
07:33sigod: the blob firmware isnt supported on parabola because it's libre
07:33imirkin: but you can fix that by installing a linux kernel ...
07:33imirkin: either way, i think that any accel will kill the board
07:33sigod: but then it wouldnt be the libre kernel
07:33imirkin: but you can always try disabling GL only
07:34imirkin: remove nouveau_dri.so
07:34imirkin: or stick LIBGL_ALWAYS_SOFTWARE=1 into your /etc/environment
07:34sigod: yea that's what i was reading on a post you wrote
07:34imirkin: noaccel still does accel of fbdev btw
07:34sigod: won't nouveau dri.so get reinstalled in updates
07:35sigod: i was just thinking if it didnt crash without opengl then i could use compton to stop tearing in videos
07:37sigod: nouveau.noaccel=1 made the system rock solid though
08:17sigod: imirkin, does xrender require hardware acceleration?
08:19sigod: for compton
09:13sigod: https://github.com/yshui/picom/wiki/Vsync-Situation - this made me understand the situation
15:56Tashtari: Now that we're in a slightly different time-zone-awake-time... :) I'm getting a kernel panic that appears to come from nouveau triggered by dragging or resizing windows in X11. My card is a GT218.
15:56Tashtari: Console output: http://paste.debian.net/1129972/
15:57Tashtari: And another of the same: http://paste.debian.net/1129978/
15:58Tashtari: Seems to happen whether there's one monitor plugged in or two.
15:58Tashtari: I definitely need a nudge in the right direction here. :) Even if it's more debugging steps that I can do to narrow it down. Can anyone help?
16:05imirkin_: sadly the list of people who can help don't change.
16:06imirkin_: really you want skeggsb, who's based out of eastern side of australia.
16:06imirkin_: i.e. asleep right now
16:47Tashtari: imirkin_: Ok, I'll wait on them.
16:48imirkin_: cosurgi: how goes the testing?
19:13cosurgi: imirkin_: no problems so far.
19:14cosurgi: imirkin_: no EAGAIN messages in the logs either.
19:14cosurgi: usually ot was over a month. We have to be patient ;)
19:19imirkin_: ahh ok
19:19imirkin_: yeah, i thought your repro was faster.
20:21cosurgi: imirkin_: first my desktop reopening repro is too slow. Now the bug reproduction is too slow. And I'm working like crazy ;)
20:23imirkin_: MORE! FASTER!
20:24cosurgi:sheds sweat, tears and blood.
20:49imirkin_: should see a doctor about that last one
20:50Pie_Mage: and a therapist about the middle one