Unexpected files in the drm-intel tree

Unexpected files in the drm-intel tree

[Mark Brown]

[-- Attachment #1: Type: text/plain, Size: 154 bytes --]

Commit

  ac3baea883dab ("drm/panel: find_panel_by_fwnode() return a counted reference")

added this unexpected file:

 drivers/gpu/drm/drm_panel.c.orig


[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

Re: Unexpected files in the drm-intel tree

[Jani Nikula]

On Tue, 18 Aug 2026, Mark Brown <broonie@kernel.org> wrote:
> Commit
>
>   ac3baea883dab ("drm/panel: find_panel_by_fwnode() return a counted reference")
>
> added this unexpected file:
>
>  drivers/gpu/drm/drm_panel.c.orig

So this has been reported before for drm-misc [1] and drm-xe [2] trees,
and has been fixed in the drm-misc tree by commit 651b0bd0922b
("drm/panel: remove accidentally committed .orig file").

Similar to what Joonas said in [3], this should really only be reported
once for the tree *and* branch where it originated. There's not a lot of
value in reporting again whenever something gets merged somewhere else.


BR,
Jani.


[1] https://lore.kernel.org/r/al9cE1PAYd9jyCfZ@sirena.org.uk
[2] https://lore.kernel.org/r/annCRqegoQ9PBClJ@sirena.org.uk
[3] https://lore.kernel.org/r/178712337932.24680.17671628229026350089@jlahtine-mobl

-- 
Jani Nikula, Intel

Re: Unexpected files in the drm-intel tree

[Mark Brown]

[-- Attachment #1: Type: text/plain, Size: 1929 bytes --]

On Wed, Aug 19, 2026 at 12:31:07PM +0300, Jani Nikula wrote:
> On Tue, 18 Aug 2026, Mark Brown <broonie@kernel.org> wrote:

> > Commit

> >   ac3baea883dab ("drm/panel: find_panel_by_fwnode() return a counted reference")

> > added this unexpected file:

> >  drivers/gpu/drm/drm_panel.c.orig

> So this has been reported before for drm-misc [1] and drm-xe [2] trees,
> and has been fixed in the drm-misc tree by commit 651b0bd0922b
> ("drm/panel: remove accidentally committed .orig file").

> Similar to what Joonas said in [3], this should really only be reported
> once for the tree *and* branch where it originated. There's not a lot of
> value in reporting again whenever something gets merged somewhere else.

There's no database of issues here, the check that's being done is based
purely on the new commits and diffstat that just appeared when we fetched
the tree.  Deduplicating things is in general a depresssingly hard
problem even for basic stuff (eg, is this a new issue or a recurrance of
an old issue?), and the much more common case is that there's downstream
trees sending problematic commits to their upstreams which the upstreams
would like to know about.  DRM is quite unusual in the amount of
sideways merging that goes on between parallel trees, and can also be on
the higher side for the length of time issues get left for once they've
been merged into other trees.

What I would have hoped would have happened here is that whichever tree
originally had the issue would have fixed it before they sent it to be
merged elsewhere, then it wouldn't pop up in all the other trees, or if
the original issue already got merged then a fix would be sent promptly
before other trees start picking things up.  In this case it looks like
-next first saw this on July 21st and the fix wasn't posted until 10th
August (then applied on the 11th) so there was unfortunately plenty of
time for things to propagate around.

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 488 bytes --]