05:11imirkin: ok. that stupid GTF-GL45.gtf33.GL3Tests.explicit_attrib_location.explicit_attrib_location_pipeline is figured out. now one more GTF test left...
05:11imirkin: GTF-GL45.gtf30.GL3Tests.transform_feedback.transform_feedback_overflow
05:36imirkin: probably a task for next weekend.
08:00airlied: skeggsb: hey, running a G98 machine under F30 (5.4.7) this thing is a crashfest assuming due to OOM or other stuff
08:06airlied: skeggsb: https://people.freedesktop.org/~airlied/scratch/nouveau-logs.gz has some traces from crashes
21:13imirkin_: hm ... looks like you run out of vram. but there's an additional unusual issue
21:13imirkin_: anyways, running out of vram = major fail ... you really just want to stick to the nouveau ddx and !gnome/kde if you want things to work
21:13imirkin_: airlied: --^
21:14imirkin_: unfortunately redhat includes a patch to explicitly avoid the nouveau ddx
21:14imirkin_: which debian has also seemingly copied
21:14imirkin_: nouveau 0000:05:00.0: fb: trapped read at 0000e0a000 on channel -1 [0fee0000 unknown] engine 06 [BAR] client 08 [PFIFO_READ] subclient 00 [FB] reason
21:14imirkin_: 00000002 [PAGE_NOT_PRESENT]
21:14imirkin_: that's a weird error. as si TTM buffer eviction failed
22:46pedahzur: Howdy! https://bugs.freedesktop.org/ says it's not in use, and to go to https://gitlab.freedesktop.org/ for bugs. But Nouveau isn't listed there. https://nouveau.freedesktop.org/wiki/Bugs/ still inks to bugs.freedesktop.org. Where is one supposed to open bugs for Nouveau?
22:47imirkin_: here's good. or mailing list.
22:50pedahzur: imirkin_: Well, I can't paste my kernel log here. :) Is there no bug tracker any more?
22:50imirkin_: pastebin
22:50pedahzur: I'll try the mailing list.
22:50imirkin_: bugzilla got shut down, apparently
22:50imirkin_: i'm variously annoyed by gitlab
22:50pedahzur: Nouveau didn't migrate to gitlab.freedesktop.org?
22:50pedahzur: ah
22:50imirkin_: so i don't monitor anything that goes on in gitlab
22:51pedahzur: Well, there isn't a nouveau project in gitlab, soooo. :)
22:51pedahzur: subscribed to the list...waiting for confirmation e-mail
22:52imirkin_: before you waste your time...
22:52imirkin_: is this like "random hangs, sometimes"
22:52imirkin_: coz if so, not much i can do to help
22:55tanriol: imirkin, is misrendering worth reporting and, if so, what information is needed about it? I saw some problems a few months ago, IIRC, but that was a closed-source Steam game, so not really sure what information I can collect about it.
22:55imirkin_: tanriol: well, you could mention the game. might have access to it.
22:55imirkin_: also which GPU family
22:56imirkin_: and you can make an apitrace, although it's a bit of a pain with steam, i think there are guides
23:02pedahzur: imirkin_: Posted to mailing list. It's a "screen doesn't come back" after suspend issue. But on an old (supported) card.
23:02imirkin_: ah, that's legit
23:02imirkin_: did it ever work?
23:03pedahzur: If there's something I can put in the post-sleep script to re-init the display, I'd be happy.
23:03imirkin_: like in an old kernel or whatever?
23:03pedahzur: Not that I know of. Ubuntu 19.10 install.
23:03pedahzur: I might have an old ISO around I could try.
23:03imirkin_: i still haven't gotten the mail
23:04pedahzur: Neither have I. :)
23:05pedahzur: Not in the archives yet either.
23:08imirkin_: probably has to be approved by the approrpriate authorities
23:08tanriol: imirkin_: It was observed with Duskers. Unable to reproduce at the moment as they crash for some reason before entering gameplay. The GPU is GM107.
23:09pedahzur: imirkin_: Very likely, but usually Mailman will let the sender know it's in moderation. So, not sure.
23:09imirkin_: tanriol: well, i just pushed a few fixes, some of which are GM10x+-specific. they're minor though, unlikely to be hit in practical scenarios.
23:12tanriol: So, IIUC, apitrace is a good idea... ok, will try the next time I reproduce something.
23:12imirkin_: tanriol: unfortunately i don't think that i have Duskers in my accessible collection -- i'm on a few of the "free for mesa developers" plans, but i don't think this is included in those
23:13imirkin_: i get all the valve games, and all the games ported by ... one of the porting guys, i forget which
23:43imirkin_: pedahzur: so yeah, dunno when that mail will get approved. if you want me to look in the meanwhile, pastebin something and link here, for example.
23:51pedahzur: imirkin_: Yeah, just got back a "awaits moderator approval" e-mail. :)
23:51pedahzur: Ha! "Message body is too big: 169728 bytes with a limit of 100 KB" I included the kernel log and video.rom. :)
23:51pedahzur: Hold on....
23:52imirkin_: yeah, this situation kinda stinks
23:52imirkin_: sigh
23:52pedahzur: imirkin_: https://gist.github.com/jkugler/8c6f4cac6de9978f7ffac57676abe833
23:53imirkin_: how is the screen connected?
23:53pedahzur: imirkin_: Is there a command I can issue to re-init the driver/card without kill my X session? I could invoke that from a /etc/pm/sleep.d/ script.
23:53imirkin_:guesses DP
23:53pedahzur: imirkin_: Laptop. Direct connection. Latitude E6510.
23:54imirkin_: oh, internal screen? what is the name of the connector? eDP-1 or LVDS-1?
23:54pedahzur: eDP-1 is the one that works with xrandr. :)
23:54imirkin_: yeah
23:54imirkin_: so you have 2 separate problems
23:54imirkin_: congratulations
23:55imirkin_: first off, it seems like you get this on resume: [ 1292.693447] nouveau 0000:01:00.0: devinit: 0x00005ea9[0]: script needs OR link
23:55imirkin_: secondly you get
23:55imirkin_: [ 1293.952413] nouveau 0000:01:00.0: disp: outp 00:0006:0344: link rate unsupported by sink
23:55imirkin_: [ 1293.955810] nouveau 0000:01:00.0: disp: outp 00:0006:0344: training failed
23:55imirkin_: which is just very unfortunate.
23:56imirkin_: i suspect doing a "chvt 1; chvt 7" or whatever could force a fresh modeset
23:56imirkin_: (un)fortunately, i'm not very well versed in all this vbios script stuff, esp around displays
23:57imirkin_: that's skeggsb's thing ... he'll definitely need the vbios, but also if you could get some more debug info -- just boot with nouveau.debug=debug drm.debug=0x1e should be enough
23:57imirkin_: and this is another one of these !@#$% 1366x768 panels, which i'm sure isn't helping