[PATCH] drm: Reject dumb buffers when driver/device doesn't support modesetting

Ville Syrjälä ville.syrjala at linux.intel.com
Wed Mar 18 20:44:33 UTC 2020


On Wed, Mar 18, 2020 at 01:31:07PM -0700, Matt Roper wrote:
> On Wed, Mar 18, 2020 at 05:49:59PM +0200, Ville Syrjala wrote:
> > From: Ville Syrjälä <ville.syrjala at linux.intel.com>
> > 
> > Currently a driver must not provide a .dumb_create() hook in the
> > drm_driver structure if it wants to declare dumb buffers as not
> > supported. So if the same driver wants to support both modeset
> > and non-modeset devices it would require two distinct drm_driver
> > structures in order to reject the dumb buffer operations on the
> > non-modeset devices. That's rather tedious, so let's make life
> > easier for such drivers by also checking for the DRIVER_MODESET
> > flag before we declare dumb buffers as supported. Now all the
> > driver has to do is clear the flag for any device that can't
> > do modesetting.
> 
> Will this be a problem for vgem?  I thought it exposed dumb buffers
> without modesetting support?

Well that's disappointing.

-- 
Ville Syrjälä
Intel


More information about the dri-devel mailing list