Re: Some 7.1-post fixes that might be worth picking up rather sooner than later

Thorsten Leemhuis <[email protected]> Tue, 7 Jul 2026 12:19:20 +0200
Newsgroups dev.linux.lists.regressions,org.kernel.vger.stable
Message-ID <[email protected]>
[CCing a few people]

On 6/22/26 07:32, Greg KH wrote:
> On Fri, Jun 19, 2026 at 11:43:41AM +0200, Greg KH wrote:
>> On Fri, Jun 19, 2026 at 08:04:35AM +0200, Thorsten Leemhuis wrote:
>>> Hi Stable Team! From the regressions point I think it might be nice to
>>> pick up the following changes for the next round of stable updates (e.g.
>>> 7.1.2), as they seem to fix regressions I've seen multiple people report
>>> with 7.1:
>>> [...]
>>> * 12f58a6caad3be ("drm/amd/display: Fix Color Manager (3DLUT, Shaper,
>>> Blend)") [v7.1-post]
> 
> This doesn't apply to 7.1.y, and would need a working backport.
Just a quick status update twimc:

I pointed that out in
https://gitlab.freedesktop.org/drm/amd/-/work_items/5396 , but nothing
happened from the AMD side afaics. They have much on their plate, I
fully understand that, I guess it fell through the cracks (maybe this
mail helps). Thing is: the backport of the revert is quite big, so
nobody else (including me) did yet dare to submit it themselves.

So two weeks later the regression caused by e56e3cff2a1bb2
("drm/amd/display: Sync dcn42 with DC 3.2.373") [v7.1-rc1] is still
unfixed in 7.1.y as far as I can see it -- a regression that is known
since more than two months now, as the revert to fix it (12f58a6caad3be,
mentioned in the quote above) was submitted already on 2026-04-29, but
only made it to mainline during the merge window for 7.2 (this is
another thing that afaics fell through the cracks; sadly I only became
aware of the regression after 7.1 was out, otherwise I would have made
noise earlier to get it included in 7.1).

This all seems rather unfortunate.

Luckily is seems the KDE Plasma developer turned off support for the
color pipeline stuff in Kwin 6.7.1
(https://invent.kde.org/plasma/kwin/-/commit/9079a417b821f80c0d9e3bc5014a388e0e340f82
), which apparently avoids the problem for many users (see the
gitlab.freedesktop ticket linked earlier).

Ciao, Thorsten