02:06fudgespinner: so, about DRI. what handles which version of DRI to be used?
02:07imirkin: the client
02:07fudgespinner: as in, Xorg?
02:07imirkin: the client will query the available protocols, etc
02:07imirkin: Xorg is the server.
02:07imirkin: it makes things available.
02:07imirkin: the client picks something.
02:07imirkin: (client = an application, generally)
02:08fudgespinner: ah, so Xorg options to force DRI version are just telling the client the maximum version of DRI that it can be used?
02:08imirkin: well, DRI1 and DRI2 are fundamentally different
02:09imirkin: DRI3 actually uses the DRI2 protocol
02:09imirkin: (but it requires some version of the protocol)
02:10fudgespinner: I see...
02:10imirkin: xdpyinfo will list the available protocols, but probably not the sub-versions
10:31marex: Hi, I have a problem and it just got worse recently
10:32marex: so, before recent update on debian unstable/experimental, I had a machine running linux 5.10.y and mesa likely 21.1.y , with nouveau obviously , and when I started sway and suspended the box, it took a while
10:33marex: likely because sway held on to some dmabuf or whatever
10:33marex: if I stopped sway and suspended, it suspended and resumed fine
10:33marex: well now, after updating to linux 5.14.y and mesa 21.2.2 , suspend/resume even on console is buggy with image corruption of the console after resume
10:34marex: and I get a lot of these
10:34marex: [ 101.279313] nouveau 0000:01:00.0: gr: 00000030 [ILLEGAL_MTHD ILLEGAL_CLASS] ch 1 [003fbce000 DRM] subc 3 class 0000 mthd 085c data 00000420
10:34marex: # lspci -s 01:00.0
10:34marex: 01:00.0 VGA compatible controller: NVIDIA Corporation GT216GLM [Quadro FX 880M] (rev a2)
10:34marex: # lspci -ns 01:00.0
10:34marex: 01:00.0 0300: 10de:0a3c (rev a2)
14:39karolherbst: Sooo.. trying to push patches from MLs through gitlab and then drm-misc
14:39karolherbst: first MR here: https://gitlab.freedesktop.org/drm/nouveau/-/merge_requests/5
14:40karolherbst: imirkin, Lyude, skeggsb_ (and others): please take a look and leave any comments or questions, even if it's process related
14:40karolherbst: just picked some patches from the ML which are uncontroversial, next batch might be more substantial (and maybe even doing -next instead of just -fixes)
16:19karolherbst: nice.. got the power now to mark patches as accepted on patchwork for nouveau :)
21:29karolherbst: imirkin: btw.. I hacked up the patchwork tools in such a way, that I find all duplicates. Trying to scan all drm related patches so that the patch list for nouveau is clean :)
21:29karolherbst:still fears there are like 2000+ patches ignored
21:49Ermine: How many active nouveau devs are here?
21:55karolherbst: Ermine: depends on the question
21:56karolherbst: but usually the answer is between 1-4