22:44 <tobhe> heyo, there seems to be a bug with recent versions of mesa in panfrost on mediatek chromebooks. postmarket os seems to have a workaround here https://gitlab.postmarketos.org/postmarketOS/pmaports/-/merge_requests/6470. I couldn't find an upstream bug for this so I thought it might be worth bringing up here

22:45 <tobhe> I suspect https://github.com/hexdump0815/imagebuilder/issues/314 is the same issue, and so is https://bugs.launchpad.net/ubuntu/+source/mesa/+bug/2103601

22:46 <tobhe> some more info from pmos https://postmarketos.org/edge/2025/05/05/Graphical-issue-with-Mesa-25-on-Mediatek-SoCs/

23:08 <diederik> tobhe: Similar to https://bugs.debian.org/1099032 (reported upstream @ https://gitlab.freedesktop.org/mesa/mesa/-/issues/12718) ?

23:27 <tobhe> could be except on the chromebook in the screen doesn't just flicker and just just totally weird

23:28 <tobhe> as if it shows the same image multiple times in smaller

23:30 <tobhe> see screenshot in https://github.com/hexdump0815/imagebuilder/issues/314

23:31 <tobhe> reading the debian bug it sounds like it might be the same issue though. It also is a G52 so that checks out

23:37 <diederik> yeah, the 'afbc' sounded familiar and Sjoerd mentioned similar issues with MediaTek devices

23:37 <tobhe> commented on the bug and added all the links I shared here

23:41 <tobhe> diederik: I talked to tjaalton earlier and he'd be happy to fix it in Debian if we can figure out an easy fix

23:43 <diederik> nice! tjaalton is a very good maintainer :)

23:44 <tobhe> not sure if it is as easy reverting the commit mentioned in the debian tracker

23:45 <tobhe> e0f48568c7f2a638213dee317fde03c23775556f

23:46 <diederik> Oh great, you reported in the upstream bug tracker. That seems to *me* more useful then (just) in the Debian bug

23:50 <tobhe> well yes, I think we can safely assume that it isn't debian only at this point ;)

23:54 <diederik> true. I am/was mostly worried that upstream assumed the bug had been 'abandoned' (assuming they didn't periodically check the Debian bug themselves for updates)

23:55 <diederik> And I think it's reasonable to assume they know best what the best short and long term fix for it is or could be

23:56 <diederik> (I don't know if there are 'side effects' to reverting that commit)

23:57 <tobhe> the commit looks pretty small and isolated, so I'll just test and see if it helps I guess