Re: [PATCH 2/2] drm: add DRIVER_RENDER_ONLY

[Christian König <ckoenig.leichtzumerken@xxxxxxxxx>]

Am 11.10.22 um 13:27 schrieb Thomas Zimmermann:
> Hi
>
> Am 11.10.22 um 13:04 schrieb Christian König:
> > This allows to suppress creating the primary node. Useful for drivers
> > which don't expose any display functionality, but are render only by
> > design.
> >
> > Signed-off-by: Christian König <christian.koenig@xxxxxxx>
> > ---
> >   drivers/gpu/drm/drm_drv.c   | 10 ++++++----
> >   drivers/gpu/drm/drm_prime.c |  2 +-
> >   include/drm/drm_drv.h       |  7 +++++++
> >   3 files changed, 14 insertions(+), 5 deletions(-)
> >
> > diff --git a/drivers/gpu/drm/drm_drv.c b/drivers/gpu/drm/drm_drv.c
> > index d81783f43452..4a525f78a932 100644
> > --- a/drivers/gpu/drm/drm_drv.c
> > +++ b/drivers/gpu/drm/drm_drv.c
> > @@ -634,9 +634,11 @@ static int drm_dev_init(struct drm_device *dev,
> >               goto err;
> >       }
> >   -    ret = drm_minor_alloc(dev, DRM_MINOR_PRIMARY);
> > -    if (ret)
> > -        goto err;
> > +    if (!drm_core_check_feature(dev, DRIVER_RENDER_ONLY)) {
> > +        ret = drm_minor_alloc(dev, DRM_MINOR_PRIMARY);
> > +        if (ret)
> > +            goto err;
> > +    }
> >         ret = drm_legacy_create_map_hash(dev);
> >       if (ret)
> > @@ -902,7 +904,7 @@ int drm_dev_register(struct drm_device *dev, 
> > unsigned long flags)
> >            driver->name, driver->major, driver->minor,
> >            driver->patchlevel, driver->date,
> >            dev->dev ? dev_name(dev->dev) : "virtual device",
> > -         dev->primary->index);
> > +         (dev->primary ?: dev->render)->index);
> >         goto out_unlock;
> >   diff --git a/drivers/gpu/drm/drm_prime.c b/drivers/gpu/drm/drm_prime.c
> > index a3f180653b8b..4afd3f15b135 100644
> > --- a/drivers/gpu/drm/drm_prime.c
> > +++ b/drivers/gpu/drm/drm_prime.c
> > @@ -740,7 +740,7 @@ int drm_gem_prime_mmap(struct drm_gem_object 
> > *obj, struct vm_area_struct *vma)
> >       }
> >         /* Used by drm_gem_mmap() to lookup the GEM object */
> > -    priv->minor = obj->dev->primary;
> > +    priv->minor = obj->dev->primary ?: obj->dev->render;
> >       fil->private_data = priv;
> >         ret = drm_vma_node_allow(&obj->vma_node, priv);
> > diff --git a/include/drm/drm_drv.h b/include/drm/drm_drv.h
> > index f6159acb8856..0a1450d47ca2 100644
> > --- a/include/drm/drm_drv.h
> > +++ b/include/drm/drm_drv.h
> > @@ -94,6 +94,13 @@ enum drm_driver_feature {
> >        * synchronization of command submission.
> >        */
> >       DRIVER_SYNCOBJ_TIMELINE         = BIT(6),
> > +    /**
> > +     * @DRIVER_RENDER_ONLY
> > +     *
> > +     * Driver is a render node only driver and don't want to support 
> > the
> > +     * primary interface.
> > +     */
> > +    DRIVER_RENDER_ONLY        = BIT(7),
>
> Is it really necessary to add a new flag?  If a driver sets DRIVER_GEM 
> and omits DRIVER_MODESET wouldn't this have the same meaning?

Mhm, good point. I haven't thought about that.

Let me double check if anybody is actually using this.

Thanks,
Christian.

> Best regards
> Thomas
>
> >         /* IMPORTANT: Below are all the legacy flags, add new ones 
> > above. */