09:32 PaulePanter: Hi. Is https://paste.debian.net/1082641/ a known issue?
18:04 AndrewR: huh, was looking at mozilla sources (windows hw decoder support) https://github.com/mozilla/gecko/blob/central/dom/media/platforms/wmf/DXVA2Manager.cpp - "List of NVidia Telsa GPU known to have broken NV12 rendering" - this includes my G92, according to comments .... I wonder why (hw limitation, incomplete wrt this bug _windows_ drivers?)
19:19 karolherbst: PaulePanter: huh
19:20 karolherbst: Lyude: any idea?
19:21 karolherbst: AndrewR: who knows
19:35 Lyude: karolherbst: huh? not sure what you're asking about
19:35 karolherbst: Lyude: PaulePanters paste
19:36 karolherbst: " DRM: DDC responded, but no EDID for DP-1" is the main error
19:36 karolherbst: wondering if you know anything about it
19:36 Lyude: yes, make sure this person isn't using a kvm switch
19:36 Lyude: also can you post the paste again? it's not in my chat log for some reason
20:04 karolherbst: Lyude: https://paste.debian.net/1082641/
20:05 karolherbst: ohh, maybe a broken cable
20:05 karolherbst: PaulePanter: do you have by any chance a replacement cable?
20:08 Lyude: yeesh, why are we WARNING on that
20:08 Lyude: also it looks like it is getting an edid, but the edid it's getting is corrupt
20:15 karolherbst: Lyude: could also be a broken edid on the hardware, but then the version is kind of weird
20:16 Lyude: yeah... my bet is still on a dumb kvm switch
20:17 karolherbst: I kind of get the feeling that a "smart" kvm switch would cause problems, not a dumb one :p
20:17 Lyude: well whether or not it's dumb or smart likely depends on who you're asking :p
20:18 karolherbst: true
20:18 karolherbst: PaulePanter: still there?
20:19 karolherbst: Lyude: but let's agree on something between the GPU and the display is causing issues :p
20:19 Lyude: hehe, of course
20:30 ajax: that's the transfer just not happening.
20:31 ajax: the edid looks like junk ascii
20:33 Lyude: ajax: from undefined memory contents I'm assuming
20:49 ajax: first 18 bytes are: lse,"61":false}mn
20:49 ajax: naughty kernel memory leak
20:50 ajax: "DDC responded but no EDID" is an enigmatic li'l message
20:51 karolherbst: ohh, interesting
20:53 ajax:sincerely hopes he's reading this code wrong or just has an old checkout
20:56 Lyude: ajax: which bit of code?
20:57 ajax: nouveau_connector_ddc_detect()'s DP path is just straight up wrong
20:57 Lyude: i feel like there is some WIP commit I have in one of my branches that is related to this
20:57 ajax: it's only probing that the aux channel is working, not that there's a DDC peer
20:57 Lyude: ahhhh, yeah I don't think I have a fix for that
20:58 Lyude: ajax: patches welcome, or I can fix it when I get a chance
21:00 ajax: does suggest a wonky kvm switch though. i could easily see a dp kvm implementing auxch to the bridge all the time but only routing ddc to the active host
21:00 ajax: so in a sense it's good this bug is there?