08:24 MrCooper: anholt robclark: mesa-cheza failure: https://gitlab.freedesktop.org/mesa/mesa/-/jobs/1646935
08:46 hakzsam: MrCooper: right, should I retry?
08:46 MrCooper: I already did
08:48 hakzsam: it consistently fails then?
08:51 MrCooper: it passed
08:51 MrCooper: are you looking at a different pipeline maybe?
08:53 hakzsam: yeah, I was looking at the one you pasted.
08:53 MrCooper: (note that retrying results in a different job ID)
08:53 hakzsam: sure
09:47 mripard_: sravn: if it's a fixing a weird build situation, it should probably be merged as a fix
09:47 mripard_: can you cherry-pick it on drm-misc-fixes, or does it need some stuff that are only in drm-misc-next ?
10:19 danvet: mripard_, just pushed drm-fixes to -rc2 <- airlied
12:29 mripard_: danvet: I forwarded drm-misc-next this morning too
18:49 sravn: pcercuei: No matter if this is fb or mode then you are right, there is no such callback today. And we should add it so the display driver have the possibility to request the panel to adjust to the selected fb pixel format, bus format and mode
18:53 sravn: mripard_: If someone babysits me I can cherry-pick it. I think there are no dependencies
18:55 sravn: mripard_: git checkout drm-misc-fixes; git cherry-pick <sha-1> works for me.
18:55 sravn: mripard_: Is it as simple as pushing drm-misc-fixes now?
18:56 sravn: mripard_: After build tests ofc
19:03 pcercuei: sravn: I'll take a look
19:39 sravn: pcercuei: Thanks. I just now saw I fat-fingered this as I had written this the other day and scrolling with the mouse made me resend the same text
21:16 pcercuei: I'm annoyed by the fact that CONFIG_I2C is a hard dependency of DRM
21:18 imirkin: edid and dp rely on it
21:18 pcercuei: my system has neither of these
21:20 imirkin: yeah, probably could be made optional, but just pointing out why it's not necessarily going to be trivial
21:50 agd5f: pcercuei, a lot of GPUs have components connected by i2c as well (e.g., voltage regulators, various display encoders (e.g., DVO to TMDS bridges), etc.)
21:51 pcercuei: agd5f: a lot of *GPUs*, yes
21:52 airlied: pcercuei: feel free to scratch the itch, it might be possible to move stuff into files which deal with i2c
21:52 pcercuei: Here I'm dealing with an embedded system with just a LCDC + panel
21:52 airlied: the edid stuff being the main one
21:52 airlied: and displayport
21:52 pcercuei: I'll push it on my very large TODO list
21:53 airlied: like it's not something you'll find anyone else going to jump on, since for most people plugging a monitor in is a necessity
21:55 glennk: now i'm curious about an embedded board that doesn't have i2c somewhere on it...
21:57 imirkin: all spi
22:01 idr: jekstrand, dj-death: Is there any use in having multiple "intrinsic memory_barrier_shared () ()" in a row?
22:04 idr: Hm... it looks we trim that to a single WAIT instruction in the backend.
22:26 jekstrand: idr: cmarcelo has a pass to remove redundant barriers
22:26 idr: Heh... I should have guessed that somebody would. :)