00:16imirkin: pmoreau: have you made any progress on the nv50 compute stuff beyond what you had a while back? did you get a chance to integrate my change?
01:29dviola: is pmu power management related?
01:32imirkin: but really it's just a cpu
01:32imirkin: with a rtos
01:33imirkin: that cpu happens to have access to some things which are important for power management
01:33dviola: I tried the oldest archlinux isos I could find and the issue still happens with those
01:33imirkin: no reason to think it wouldn't
01:34imirkin: the vbios must be leaving the board in a weird state on init
02:17dviola: imirkin: your comment about "leaving the board in a weird state" made me try something else:
02:19dviola: moving this reset code: https://github.com/torvalds/linux/blob/master/drivers/gpu/drm/nouveau/nvkm/subdev/pmu/base.c#L112-L114
02:19dviola: before this comment: https://github.com/torvalds/linux/blob/master/drivers/gpu/drm/nouveau/nvkm/subdev/pmu/base.c#L105
02:20dviola: that makes the timeout disappear as well
02:20imirkin: my guess is that's not entirely safe.
02:21imirkin: but skeggsb would know :)
02:22dviola: that kinda reminded me of this problem: https://github.com/torvalds/linux/commit/ee50c130c82175eaa0820c96b6d3763928af2241#diff-a7904e73523acf0a46a1f8b3c815120a9b4058273a12c0af83879beb2c3086ee
02:23dviola: my suspend/resume would be broken until I changed the order of that line
03:22dviola: imirkin: thanks for the suggestion btw
03:47az: Hi, I've had a crash while using Nouveau drivers on Ubuntu 20.04 when I run an Android emulator. It works with nvidia propriety drivers.
03:47az: The emulator itself is the Android Studio emulator; @beephone is a "pixel" device profile, running the android-28 image.
03:50HdkR: That's a known issue with multithreaded GL in nouveau
03:51HdkR: You can't really work around it in Nouveau atm
04:12dviola: imirkin, skeggsb: this looks interesting as well: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/drivers/gpu/drm/nouveau/nvkm/engine/gr/gf100.c#n1988
04:13dviola: though I'm not sure if it's the same issue...
04:15dviola: I need to go, later
04:27az: HdkR, thanks
13:31dviola: I thought the code inside that if (reset) was not being triggered for my card
13:31dviola: but I added a printk and it appears in dmesg, uhm
23:02pmoreau: @imirkin: No progress and no changes made, sadly. :-/
23:13imirkin: pmoreau: ok thanks