08:36MrCooper: hwentlan_: if the HW cursor is visible on a rotated plane, do the cursor contents not get rotated?
08:49emersion: it gets rotated iirc
08:58MrCooper: then we have a problem, since the cursor planes have no rotation property
08:59emersion: i have a patch that fixes this
08:59emersion: sent a long time ago
09:07MrCooper: actually, you removed the rotation property in 1347385fe187 ("drm/amd/display: don't expose rotation prop for cursor plane")
09:08emersion: i removed it because it was exposed and a no-op
09:09emersion: my ML patch re-adds it and checks it matches the underlying plane's
09:09emersion: can't remember if there is a check for the underlying plane rotation rn
09:11MrCooper: removing the property kind of made things worse; with the property, user space doing the right thing would have worked out accidentally at least, but now user space will likely double-rotate the cursor contents
09:13emersion: i don't really agree
09:15MrCooper: anyway, let's land your patch fixing it ASAP
09:18emersion: hm and now i can't find it back
09:21emersion: ah this https://patchwork.freedesktop.org/patch/423452/
09:22emersion: i think other patches from this series are missing too
09:22emersion: "amd/display: add cursor check for YUV underlying pipe"
09:22emersion: "amd/display: add cursor alpha and blend mode checks"
09:23emersion: "amd/display: re-introduce cursor plane rotation prop"
09:23emersion: these 3 we have in steamos but have not been applied upstream
09:25emersion: maybe it has to do with the chromeos mess
09:25emersion: i don't remember tbh
09:26emersion: (at some point it was unclear whether chromeos using fancy KMS props on primary plane + legacy setcursor ioctl was fine or not)
10:12emersion: rebasing that series now
10:42MrCooper: cool
10:53emersion: there https://patchwork.freedesktop.org/series/87840/
15:28emersion: hm, apparently i got confused in the past because current behavior seems correct
15:31emersion: ah, we just tested the same thing at the same time it seems MrCooper
15:31MrCooper: heh, great minds think alike
15:32emersion: so, no idea how i got confused
15:32MrCooper: I'm actually glad, would have been quite messy to fix otherwise
15:33emersion: indeed!