[PATCH v3 1/2] drm/gem: drm_gem_dumb_map_offset(): reject dma-buf

Laurent Pinchart laurent.pinchart at ideasonboard.com
Thu Aug 17 19:40:52 UTC 2017


Hi Chris,

On Thursday 17 Aug 2017 20:09:20 Chris Wilson wrote:
> Quoting Laurent Pinchart (2017-08-17 20:01:40)
> > On Thursday 17 Aug 2017 19:30:44 Chris Wilson wrote:
> > > Quoting Laurent Pinchart (2017-08-17 18:56:09)
> > > 
> > > > On Thursday 17 Aug 2017 18:21:30 Noralf Trønnes wrote:
> > > > > Reject mapping an imported dma-buf since is's an invalid use-case.
> > > > > 
> > > > > Cc: Philipp Zabel <p.zabel at pengutronix.de>
> > > > > Cc: Laurent Pinchart <laurent.pinchart at ideasonboard.com>
> > > > > Cc: Sean Paul <seanpaul at chromium.org>
> > > > > Cc: Daniel Vetter <daniel.vetter at ffwll.ch>
> > > > > Signed-off-by: Noralf Trønnes <noralf at tronnes.org>
> > > > > ---
> > > > > 
> > > > >  drivers/gpu/drm/drm_gem.c | 6 ++++++
> > > > >  1 file changed, 6 insertions(+)
> > > > > 
> > > > > diff --git a/drivers/gpu/drm/drm_gem.c b/drivers/gpu/drm/drm_gem.c
> > > > > index ad4e9cf..8da5801 100644
> > > > > --- a/drivers/gpu/drm/drm_gem.c
> > > > > +++ b/drivers/gpu/drm/drm_gem.c
> > > > > @@ -333,6 +333,12 @@ int drm_gem_dumb_map_offset(struct drm_file
> > > > > *file,
> > > > > struct drm_device *dev, if (!obj)
> > > > > 
> > > > >               return -ENOENT;
> > > > > 
> > > > > +     /* Don't allow imported objects to be mapped */
> > > > > +     if (obj->import_attach) {
> > > > > +             ret = -EINVAL;
> > > > > +             goto out;
> > > > > +     }
> > > > > +
> > > > 
> > > > Wouldn't it be better to move this check to
> > > > drm_gem_create_mmap_offset_size() to reject mapping of all imported
> > > > dmabuf, not just the ones associated with dumb buffers ?
> > > 
> > > No, we use that and we don't ban mapping an imported object.
> > 
> > Do you use that with driver-specific buffers only or with dumb buffers too
> > ?
> > 
> For dma-buf we import? The current presumption is that the sg_table is
> backed by struct page. How would we know otherwise? I am actually not
> sure what would happen if we tried to use another mmio bar from the GPU.
> 
> For dumb buffers we create, they are just ordinary buffers.

My question was whether you need to mmap the dma-buf you import through the 
dumb buffers API (which this patch prevents for drivers using the GEM dumb 
helpers), or only through your driver-specific buffer API.

-- 
Regards,

Laurent Pinchart



More information about the dri-devel mailing list