19:13 pac85: is there some way of running the pre merge tests on an MR without assigning marge (and ideally filtering out further so I can run just the zink tests)? Someone gave me an incantation involving some script in the mesa repo ages ago but I can't find it anymore
19:16 glehmann: pac85: ./bin/ci/ci_run_n_monitor.sh --target ".*zink.*" --include-stage ".*"
19:17 pac85: glehmann: thank you!
19:17 glehmann: you have to setup the gitlab token to use it though
19:17 pac85: oh right
19:17 glehmann: or provide it with --token to the script
20:42 silurian_invader: mlankhorst: I think what I have to do is store the event in atomic_flush and then send it in the vblank IRQ
20:46 mlankhorst: yeah
20:50 silurian_invader: but I find it confusing that this functionality isn't already in the drm core
20:51 silurian_invader: or how it differs from just arming the event
20:52 silurian_invader: like, the event is armed under event_lock and sent in drm_crtc_handle_vblank/drm_handle_vblank_events
20:53 silurian_invader: and the only difference is that arming calls drm_vblank_passed on the sequence, and doing it manually doesn't
20:55 silurian_invader: I don't really understand the scenario outlined in 632c6e4edef1 ("drm/vblank: Fix flip event vblank count")
20:55 silurian_invader: like, why does it even matter what the counter is?
20:57 silurian_invader: and wouldn't the dangerous case be if the interrupt fired too early?
22:02 cwabbott: gfxstrand: since you didn't comment I assume you're ok with https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/37550 ?
23:22 ngcortes: mareko, there are a few thousand outstanding deqp failures on intel bdw from a patch you merged on mesa main