12:42lumag: mlankhorst, it seems drm-misc-next isn't getting pulled into linux-next. narmstrong pointed out that this is because of the pending patches in drm-misc-next-fixes.
12:57zmike: eric_engestrom: does the zink+nvk job in ci have a new enough kernel for rebar? cc gfxstrand karolherbst
14:30eric_engestrom: zmike: https://gitlab.freedesktop.org/mesa/mesa/-/blob/main/src/nouveau/ci/gitlab-ci-inc.yml#L57
14:30eric_engestrom: > This is https://gitlab.freedesktop.org/gfxstrand/linux/-/commits/nvk at d6820f36f375ece63226b26568be3a893985dc60 nouveau: set placement to original placement on uvmm validate.
14:30eric_engestrom: no clue about rebar though
14:30zmike: gfxstrand: ^ ?
14:31eric_engestrom: (just did a git fetch, no new commits to that branch since I last built from it)
14:37zmike: ok
15:09gfxstrand: eric_engestrom: That patch fixes power management and OOM issues. Not sure what you're asking, though.
15:10zmike: asking if rebar is enabled by that point
15:10gfxstrand: rebar has nothing to do with that
15:10eric_engestrom: it's not about that commit, but about that branch
15:10gfxstrand: rebar has been on for forever, assuming you've got it enabled in BIOS
15:10zmike: oh
15:10zmike: huh
15:10zmike: maybe I misunderstood then
15:11gfxstrand: There is a kernel patch that gives us a query for ReBAR but that's been in for a while now
15:11zmike: maybe that's it
15:11zmike: either way thx
15:13eric_engestrom: btw we get a _lot_ of hangs, is that something that's actively being worked on?
15:15gfxstrand: What GPU?
15:15gfxstrand: What do the hangs look like?
15:17daniels: https://gitlab.freedesktop.org/mesa/mesa/-/pipelines?page=1&scope=all&source=schedule
15:17daniels: https://gitlab.freedesktop.org/mesa/mesa/-/pipelines/1195586
15:18daniels: https://gitlab.freedesktop.org/mesa/mesa/-/jobs/59601342
15:21eric_engestrom: thanks daniels :)
15:22eric_engestrom: yeah `nouveau 0000:2d:00.0: deqp-vk[28345]: job timeout, channel 336 killed!` is what we usually see
15:22eric_engestrom: and GA106 is the gpu
15:22eric_engestrom: (3060 IIRC)
15:23eric_engestrom:has to go away, hopefully that's enough info, sorry for dropping that and leaving :(
15:23gfxstrand: My runs are pretty clean. Not 100% clean but I don't see a lot of those
15:24eric_engestrom: most runs don't get any job timeout, but maybe 30% do?
15:24eric_engestrom: (pulling that number out of thin air)
15:24gfxstrand: That's believable
15:25gfxstrand: I think we still have some arbitrary timeout kernel bugs
15:27eric_engestrom: is there a way to increase the timeout?
15:27eric_engestrom: something like FREEDRENO_HANGCHECK_MS
15:28eric_engestrom: (/me really needs to go now)
15:28gfxstrand: Not that I'm aware of
15:28gfxstrand: And I don't think it's a real timeout. I think nouveau.ko is missing an interrupt
20:03airlied: lumag: not getting pulled due to vc4 warnings
20:05airlied: eric_engestrom: csn you test eith a kernel patch easily?
20:55DemiMarie: @_oftc_alyssa:matrix.org: congratulations!!!!!