[PATCH] drm/i915/ring_submission: Fix timeline left held on VMA alloc error

Janusz Krzysztofik janusz.krzysztofik at linux.intel.com
Thu Jun 12 09:45:46 UTC 2025


On Thursday, 12 June 2025 11:35:31 CEST Jani Nikula wrote:
> On Thu, 12 Jun 2025, Janusz Krzysztofik <janusz.krzysztofik at linux.intel.com> wrote:
> > On Wednesday, 11 June 2025 22:54:40 CEST Andi Shyti wrote:
> >> Hi Nitin,
> >> 
> >> On Wed, Jun 11, 2025 at 03:45:30PM +0000, Gote, Nitin R wrote:
> >> > [...]
> >> > > Subject: [PATCH] drm/i915/ring_submission: Fix timeline left held on VMA alloc
> >> > > error
> >> > >
> >> > 
> >> > Generally, it's preferred to use "drm/i915/gt:" file path over "drm/i915/ring_submission:"  file name in the commit title.
> >> 
> >> good observation, I missed it. I agree with Nitin on this, it can
> >> be fixed before merging.
> >
> > I'm not sure.  I found no single word on the *subsystem* component of the 
> > canonical patch format subject line (or commit message) expected to reflect 
> > any directory structure in case of DRM.
> 
> It's not about the directory structure, though, but rather about
> (admittedly unwritten) conventions. Usually about driver components,
> features or platforms.
> 
> See:
> 
> $ git log --since={5years} --no-merges --pretty=%s -- "<PATH>" | sed 's/:.*//' | sort | uniq -c | sort -rn
> 
> Where "<PATH>" is drivers/gpu/drm/i915/gt/intel_ring_submission.c or
> drivers/gpu/drm/i915/gt.
> 
> "ring" or "submission" is just not there in the prefix, at all.

I see.  Is there a convention for designating old, pre-execlists *platforms* 
as affected subsystem / area?  Or is describing it in the summary phrase of 
the commit message the only way?

Thanks,
Janusz

> 
> 
> BR,
> Jani.
> 
> > However, if you think it should for 
> > some reason, or you just don't recognize i915 ring submission as a good 
> > candidate for the subsystem component of the commit message, then I'm OK with 
> > drm/i915/gt, but then, the summary phrase of the commit message seems too 
> > general for the whole GT subsystem, not pointing to ring submission as the 
> > only submission method out of the three that's affected, and needs to be 
> > rephrased, I believe, while still kept short enough.  Maybe "Fix *legacy* 
> > timeline held on VMA alloc error" (with the 'left' word dropped)?
> >
> > Thanks,
> > Janusz
> >
> >> 
> >> Andi
> >> 
> >
> >
> >
> >
> 
> 






More information about the Intel-gfx mailing list