On Thu, Apr 2, 2026 at 4:59 AM Thorsten Leemhuis
<[email protected]> wrote:
>
> Lo! Me again, sorry for pestering you folks about workflow issues again
> after doing so just a few weeks ago. :-/
>
> On 4/1/26 20:44, Alex Deucher wrote:
> >
> > More stuff for 7.1.
> >
> > The following changes since commit a51973c5dff8a0f01cc7d1b2007306ea0004fa16:
> >
> >   Merge tag 'drm-xe-next-2026-03-26-1' of 
> > https://gitlab.freedesktop.org/drm/xe/kernel into drm-next (2026-03-30 
> > 06:04:59 +1000)
> >
> > are available in the Git repository at:
> >
> >   https://gitlab.freedesktop.org/agd5f/linux.git 
> > tags/amd-drm-next-7.1-2026-04-01
> > [...]
>
> I noticed that this lacks "drm/amd/display: Wire up
> dcn10_dio_construct() for all pre-DCN401 generations", which is a fix
> that was posted 10 days ago[1] for a 7.0-rc1 regression that at least
> three people hit[2]; a fix that now seems to be in "this week's display
> driver promotion cycle before merging"[3] at AMD.
>
> Is this delay really needed / worth it for a regression fix at this
> point of the our devel cycle?
>
> And yes, I understand that this might be shared code that AMD wants to
> test internally first. Still asking, because at the same time it would
> be nice to (a) fix the regressions rather sooner than later and (b) not
> fix in the last minute.
>
> Are "pre-DCN401" chips even tested in this "driver promotion cycle"? And
> how often does it find problems anyway? Can't we just apply the fix and
> revert it quickly later in case AMD find problems (yes, I understand
> that this complicates things, but I wonder if that might be worth it if
> this is rare)?
>

There are always new fixes.  Worse case it ends up in 7.0.1.  If it
causes other regressions, then we end up introducing a new regression
in rc7. Pick your poison I guess.

Alex

> Ciao, Thorsten
>
> [1]
> https://lore.kernel.org/all/[email protected]/
>
> [2]
> https://lore.kernel.org/all/caokstbs1vnbdjrze3m2v-nc%[email protected]/
> https://gitlab.freedesktop.org/drm/amd/-/work_items/5052
>
> [3]
> https://lore.kernel.org/all/[email protected]/

Reply via email to