linux-kernel - Re: [PATCH] Add virtio gpu driver.

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <20160531100041.22293fb1@eldfell>
Date:	Tue, 31 May 2016 10:00:41 +0300
From:	Pekka Paalanen <ppaalanen@...il.com>
To:	Gerd Hoffmann <kraxel@...hat.com>
Cc:	Daniel Vetter <daniel@...ll.ch>, virtio-dev@...ts.oasis-open.org,
	"Michael S. Tsirkin" <mst@...hat.com>,
	"open list:ABI/API" <linux-api@...r.kernel.org>,
	Rusty Russell <rusty@...tcorp.com.au>,
	open list <linux-kernel@...r.kernel.org>,
	"open list:DRM DRIVERS" <dri-devel@...ts.freedesktop.org>,
	"open list:VIRTIO CORE, NET..." 
	<virtualization@...ts.linux-foundation.org>,
	Dave Airlie <airlied@...hat.com>
Subject: Re: [PATCH] Add virtio gpu driver.

On Tue, 31 May 2016 08:29:20 +0200
Gerd Hoffmann <kraxel@...hat.com> wrote:

>   Hi,
> 
> > > Why attach the hotspot to the plane?  Wouldn't it make more sense to
> > > make it a framebuffer property?  
> > 
> > We don't have properties on the framebuffer. I guess you /could/ just add
> > it internally to struct drm_framebuffer, and not bother exposing to
> > userspace. I guess that would be a lot simpler,  
> 
> Yes.  I can simply stick the hotspot into drm_framebuffer in
> drm_mode_cursor_universal() and pick up the values in the driver's plane
> update function.
> 
> >  but it also means that
> > atomic userspace can't use hotspots before we add properties to fbs. And
> > doing that is a bit tricky since drm_framebuffer objects are meant to be
> > invariant - this assumption is deeply in-grained into the code all over
> > the place, everything just compares pointers when semantically it means to
> > compare the entire fb (including backing storage pointer/offsets and
> > everything).  
> 
> Hmm, the hotspot location for a given cursor image is invariant too, so
> I don't think that would be a big issue.
> 
> I'd expect userspace define a bunch of cursors, then switch between them
> (instead of defining a single cursor, then constantly updating it).

Except updating a single cursor (well, two alternating buffers) is
exactly what Weston does, since there is no "set of cursors". On
Wayland, a cursor is just a regular surface like any other with
arbitrary content from a client, except it happens to be associated
with a pointer device.

Furthermore, in Weston a cursor plane is not special in any way. *Any*
client surface can go on the cursor plane if it fits. Universal planes,
and all that.

That's one existing userspace. I suppose that is sub-optimal for
virtual drivers, isn't it? But what else could Weston do without having
separate paths for "normal DRM" vs. "virtual DRM"?


Thanks,
pq

Content of type "application/pgp-signature" skipped