| Bug ID | 107889 |
|---|---|
| Summary | colors are wrong on one screen with Radeon RX Vega M GL and K= ernel 4.18 |
| Product | DRI |
| Version | XOrg git |
| Hardware | x86-64 (AMD64) |
| OS | All |
| Status | NEW |
| Severity | normal |
| Priority | medium |
| Component | DRM/AMDgpu |
| Assignee | dri-devel@lists.freedesktop.org |
| Reporter | mchehab@infradead.org |
Created attachment 141508 [details]
Picture with the 3 monitors. The colors at the center one are wrong
I have a 3 monitors setup on a NUC8i7HNK (with latest BIOS: 0049).
I'm currently using Kernel 4.17 + DRM patches for 4.18-rc1. Everything works
OK. However, upgrading the Kernel to 4.18.5 (kernel-4.18.5-200.fc28.x86_64)
causes that the center (and default) monitor to mangle colors. The same iss=
ue
also happens with upstream Kernel 4.18 vanilla.
I'm enclosing a picture of what's happening there.
Created attachment 141509 <=
a href=3D"attachment.cgi?id=3D141509&action=3Dedit" title=3D"output of =
xrandr">[details]
output of xrandr
Created attachment =
141510 [details]
Xorg logs with Kernel 4.18.5
Created attachment 141511 [details]
Xorg logs with Kernel 4.17.3 + drm-next patches submitted for 4.18-rc1
| What | Removed | Added |
|---|---|---|
| Attachment #141511 mime type | text/x-log | text/plain |
Can you bisect? Also please attach your dmesg output.
Created attachm=
ent 141526 [details]
Dmesg with Kernel 4.18.0 vanilla
(In reply to Alex Deucher from comment #4) > Can you bisect? Also please attach your dmesg o= utput. Bisect didn't work. There is a regression between 4.17 and 4.18-rc4 on ext4= fs that causes it fail to mount a file system: set 11 08:16:49 coco.lan kernel: EXT4-fs (dm-13): ext4_check_descriptors: B= lock bitmap for group 0 overlaps block group descriptors set 11 08:16:49 coco.lan kernel: EXT4-fs (dm-13): group descriptors corrupt= ed!
(In reply to Mauro Carvalho Chehab from comment #5) > Created attachment 141526 [details] > Dmesg with Kernel 4.18.0 vanilla Forgot to mention, but the bug also happens on 4.18.0 vanilla. So, the bug is somewhere between drm-next-2018-06-06-1, e. g. between this changeset: commit 568cf2e6aa0c762f14d2d0d481a006f93c63ab7a Merge: 74860cbfdd41 ebe1d22b57b8 Author: Dave Airlie <airlied&#= 64;redhat.com> Date: Wed May 30 11:06:17 2018 +1000 Merge tag 'drm-amdkfd-next-2018-05-28' of git://people.freedesktop.org/~gabbayo/linux into drm-next - Build amdkfd's related files inside amdgpu only if amdkfd is built - Fix compile warning - Print info message in case ASIC is not supported by amdkfd Signed-off-by: Dave Airlie <airlied@redhat.com> Link: https://patchwork.freedesktop.org/patch/msgid/20180528= 110435.GA17960@odedg-x270 and Kernel 4.18.
(In reply to Mauro Carvalho Chehab from comment #6) > Bisect didn't work. There is a regression betwee= n 4.17 and 4.18-rc4 on > ext4fs that causes it fail to mount a file system: There are ways to deal with that, e.g.: * Manually apply the fix for the ext4 regression when necessary * Run "git bisect skip" for commits which aren't testable * Manually find another commit which can be tested
(In reply to Michel D=C3=A4nzer from comment #8) > (In reply to Mauro Carvalho Chehab from comment #6) > > Bisect didn't work. There is a regression between 4.17 and 4.18-r= c4 on > > ext4fs that causes it fail to mount a file system: >=20 > There are ways to deal with that, e.g.: >=20 > * Manually apply the fix for the ext4 regression when necessary > * Run "git bisect skip" for commits which aren't testable > * Manually find another commit which can be tested I tried to bisect, but I was unable to find the patch with broke it. I susp= ect it could be some sort of race condition, perhaps when retrieving EDID from = the 3 monitors. Which buggy Kernels, the collor has a chance to be mangled duri= ng X11 start and after DPMS screen on. Sometimes, it goes fine, sometimes, it = gets messed. Anyway, since yesterday I'm using Kernel 4.20, with the latest linux-firmwa= re. It seems that the issue vanished, but I need to keep it running for a while= to be certain.
| What | Removed | Added |
|---|---|---|
| Resolution | --- | MOVED |
| Status | NEW | RESOLVED |
-- GitLab Migration Automatic Message -- This bug has been migrated to freedesktop.org's GitLab instance and has been closed from further activity. You can subscribe and participate further through the new bug through this = link to our GitLab instance: https://gitlab.freedesktop.org/drm/amd/issues/521.