drm/doc/rfc: Mark long running workload as complete.
authorMatthew Brost <matthew.brost@intel.com>
Fri, 1 Dec 2023 04:21:51 +0000 (23:21 -0500)
committerRodrigo Vivi <rodrigo.vivi@intel.com>
Thu, 7 Dec 2023 16:00:30 +0000 (11:00 -0500)
commita85607e3cfc5032daf64efc5f6b5387f2271b073
tree925e9c3f1886ed36561e322d3ac4b9931cddf6a1
parent76385d493c2137460ee7735a5d3a494099c35188
drm/doc/rfc: Mark long running workload as complete.

No DRM scheduler changes required, drivers just return NULL in run_job
vfunc.

The rough consensus is that no helper or extra scaffolding is needed
around long-running jobs and no further changes to drm-scheduler.

At least for now. Other drivers that currently do long-running workloads
have no plat to use drm-scheduler. Besides, the current consensus is
that this solution of simply returning NULL to the run_job function should
work without extra code duplication or complication.

On top of that, this item was already a non-blocking one for upstreaming Xe,
so let's move that to the 'Completed' section and revisit the long-running
solution as a community after Xe is integrated in DRM.

Signed-off-by: Matthew Brost <matthew.brost@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20231201042158.80009-2-rodrigo.vivi@intel.com
Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Documentation/gpu/rfc/xe.rst