Re: [PATCH v2 1/7] drm/bridge: ps8640: Use atomic variants of drm_bridge_funcs

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Hi Sam,

On Fri, Oct 22, 2021 at 07:13:58PM +0200, Sam Ravnborg wrote:
> Hi Laurent,
> 
> > From a quick look only cadence/cdns-mhdp8546 subclass
> > drm_bridge_state and I wonder if the right thing to do would be to
> > implement fallback to the helpers if the bridge driver do not set
> > any of the .atomic_duplicate_state(), .atomic_destroy_state(), or .atomic_reset().
> > 
> > That would drop the following from a few bridges:
> >         .atomic_duplicate_state = drm_atomic_helper_bridge_duplicate_state,
> >         .atomic_destroy_state = drm_atomic_helper_bridge_destroy_state,
> >         .atomic_reset = drm_atomic_helper_bridge_reset,
> 
> To answer myself here. This would create a dependency from the core to
> the helpers which is not OK so idea dropped again.

I agree it would be nicer, but the dependency is likely a problem. That
being said, we have multiple types of helpers. The first set is the
modeset helpers, which were designed as one implementation of KMS
operations, with an opt-in API for drivers. The core should not depend
on those. There are however other helpers that are only default
implementations of some operations, without any dependency on other
components. The atomic state helpers fall in this category, they
implement .atomic_* operations of the drm_*_funcs structures, not
drm_*_helper_funcs. It could make sense to move them to the DRM core.

-- 
Regards,

Laurent Pinchart



[Index of Archives]     [Linux DRI Users]     [Linux Intel Graphics]     [Linux USB Devel]     [Video for Linux]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]     [XFree86]     [Linux USB Devel]     [Video for Linux]     [Linux Audio Users]     [Linux Kernel]     [Linux SCSI]     [XFree86]
  Powered by Linux