10:47 karolherbst: okay... nice
10:48 karolherbst: no regressions inside the CTS.. and glcts -n "*multithread*" also works :D yay
10:48 karolherbst: I still suspect random multithreading state issues, but at least it shouldn't crash anymore
10:49 karolherbst: let me verify I also don't regress those android tests.. those are quite a lot
10:52 karolherbst: imirkin: soo.. if you got some time, 5 patches I think are ready to get merged: https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10752
10:52 gitlab-bot: Mesa issue (Merge request) 10752 in mesa "nouveau: multithreading fixes part 1" [Nouveau, Opened]
10:53 karolherbst: not a lot of changes but still some :)
16:59 imirkin: karolherbst: will try to look on sunday
17:45 karolherbst: cool, thanks!
19:00 karolherbst: mhh.. all aosp + GLES CTS + GL CTS tests: [110751/110751] Pass: 101871 Fail: 735 Skip: 8003 ExpectedFail: 0 UnexpectedPass: 0 Crash: 87 Timeout: 24 Missing: 0 Flake: 31 Duration: 5:41:14 Remaining: 0
19:00 karolherbst: curious about those timeouts
19:01 imirkin: wow, 5h? that seems like a lot.
19:01 karolherbst: check the number of tests
19:01 imirkin: that's also a lot of tests... yea
19:01 karolherbst: I have a script which merges like all mustpass files together :D
19:01 karolherbst: and removes dups
19:01 karolherbst: and pushes it through the parallel deqp runner
19:01 imirkin: which gen bw?
19:01 karolherbst: turing
19:01 imirkin: k
19:01 karolherbst: but that's with my MT fixes
19:02 karolherbst: but those timeouts also happen on main
19:02 karolherbst: I set the timeout to one minute
19:03 karolherbst: in 3 hours I have the result from the run on main :D
20:35 karolherbst: cleaning up my other patches will be a little painful though :/
20:35 karolherbst: imirkin: any opinions on reworking the state tracking we do in nvc0? I suspect that with my changes it will be bogus, due to the nature how submissions are made, but I don't think there is a pressing need to fix it atm as long as we don't necessarily hit bugs
21:22 imirkin: "reworking"?
21:22 karolherbst: well.. okay.. so here is the deal
21:22 karolherbst: with my rework each context has it's own push buffer they are building up
21:22 imirkin: there are a number of things it tries to optimize for
21:22 imirkin: so it's important to maintain those
21:23 karolherbst: so one context can build up its buffer base on one state, and the next on the previous one
21:23 imirkin: per-context pushbufs are fine, but very tricky to actually implement.
21:23 karolherbst: but
21:23 karolherbst: the latter context could submit the pushbuffer before the previous one
21:23 imirkin: that's why i have the screen-level graph state
21:23 karolherbst: sure
21:23 karolherbst: but we update the state when building the buffer
21:23 karolherbst: and the order of building the buffer doesn't have to match the order of submission
21:23 imirkin: yeah, but there's one pushbuf, so it all works out.
21:24 karolherbst: sure.. atm... but it's also crashing :D
21:24 karolherbst: so it might be fine on the GPU once you manage to submit it without damage from races
21:24 karolherbst: which.. normally doesn't happen
21:24 imirkin: anyways
21:24 imirkin: i gtg now
21:24 karolherbst: yeah
21:24 karolherbst: have fun
21:24 imirkin: but this is why per-context pushbufs are very tricky to do
21:24 imirkin: and i was suggesting not going in that direction.
21:24 karolherbst: I know
21:25 karolherbst: I am thinking about the problem but don't have a good solution yet
21:25 imirkin: (well, it's not THE reason why, but one of)
21:25 karolherbst: and anyway.. doesn't matter
21:25 karolherbst: today we crash
21:25 karolherbst: with my changes there might be glitches
21:25 karolherbst: which.. we had before anyway
21:25 karolherbst: so....
21:25 karolherbst: fixing this isn't exactly the highest priority atm
21:27 karolherbst: I simply want to come up with a plan :D
22:26 karolherbst: mhh.. on main: [110751/110751] Pass: 101855 Fail: 764 Skip: 8003 ExpectedFail: 0 UnexpectedPass: 0 Crash: 67 Timeout: 24 Missing: 0 Flake: 38 Duration: 5:46:02 Remaining: 0
22:26 karolherbst: 20 crashes less..
22:26 karolherbst: interesting
22:27 karolherbst: probably something dump.. will look into it tomorrow
22:27 RSpliet: it already is tomorrow!
22:27 karolherbst: ehhh
22:27 karolherbst: I shouldn't work on the weekend anyway
22:27 RSpliet: (a paradox in four words :-D)
22:28 karolherbst: and yesterday was today
22:29 karolherbst: or the other way around? mhh