> -----Original Message-----
> From: Nikula, Jani <[email protected]>
> Sent: Wednesday, February 25, 2026 11:42 PM
> To: Roper, Matthew D <[email protected]>; Shankar, Uma
> <[email protected]>
> Cc: [email protected]; [email protected];
> [email protected]; Vivi, Rodrigo <[email protected]>; Thomas
> Hellström <[email protected]>; Brost, Matthew
> <[email protected]>
> Subject: Re: [PATCH] drm/xe/compat: Remove unused i915_reg.h from compat
> header
> 
> On Wed, 25 Feb 2026, Matt Roper <[email protected]> wrote:
> > On Wed, Feb 25, 2026 at 04:40:23PM +0000, Shankar, Uma wrote:
> >> Pushed to drm-xe-next. Thanks for the review.
> >
> > Even though this is an Xe patch it might have been better to send this
> > one through drm-intel-next rather than drm-xe-next since drm-xe-next
> > doesn't have the prerequisite display patches yet to make this header
> > removal possible.  So at the moment the build of drm-xe-next is broken.
> > Not a huge issue since drm-tip is fine, and drm-xe-next will likely be
> > fixed soon whenever a backmerge happens, but I figured I should point
> > it out.

I did compile check before merge but somehow had a config with xe build 
disabled (face palm ☹).
Thanks Matt for pointing this out.

> On the contrary, I think it is an issue that warrants a revert.
> 
> It's a hard requirement that committers *always* build the branch they apply
> patches to before pushing, no exceptions. CI doesn't guarantee that at all.
> 
> With the build already broken, you can't do that, and more breakage can slip 
> in
> until it's fixed.
> 
> It'll take a drm-intel-next pull request to drm-next, and then a drm-next 
> backmerge
> to drm-xe-next to fix this, and we don't generally want to rush that to 
> hotfix build
> issues like this. The build failure
> *before* pushing is an indication a backmerge might be needed, and that's the
> normal order of things.

Thanks Jani, shall I send a revert on drm-xe-next and push through 
drm-intel-next instead ?
Apologies for the noise on this.

Regards,
Uma Shankar

> 
> BR,
> Jani.
> 
> 
> --
> Jani Nikula, Intel

Reply via email to