[Spice-devel] [PATCH xf86-video-qxl] Provide an implementation for put_image in dfps mode.

[Jeremy White]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi Christophe,

Thanks for the review.

On 12/11/2015 03:47 AM, Christophe Fergeau wrote:
> On Wed, Dec 02, 2015 at 02:14:32PM -0600, Jeremy White wrote:
>> This prevents the fallback from calling prepare_access against
>> the whole screen, which in turn keeps us from transmitting the
>> whole screen more than necessary.
>> 
>> Signed-off-by: Jeremy White <jwhite at codeweavers.com> --- 
>> src/dfps.c | 13 ++++++++++--- 1 file changed, 10 insertions(+), 3
>> deletions(-)
>> 
>> diff --git a/src/dfps.c b/src/dfps.c index 568ba5e..f113f27
>> 100644 --- a/src/dfps.c +++ b/src/dfps.c @@ -270,15 +270,22 @@
>> static Bool dfps_put_image (PixmapPtr dest, int x, int y, int w,
>> int h, char *src, int src_pitch) { dfps_info_t *info; +    FbBits
>> *dst; +    FbStride dst_stride; +    int dst_bpp;
>> 
>> if (!(info = dfps_get_info (dest))) return FALSE;
>> 
>> dfps_update_box(&info->updated_region, x, x + w, y, y + h);
>> 
>> -    /* We can avoid doing the put image ourselves, as the uxa
>> driver -       will fall back and do it for us if we return false
>> */ -    return FALSE; +    fbPrepareAccess(dest);
> 
> Wondering if this should be uxa_prepare_access() rather than 
> fbPrepareAccess? Not familiar at all with all of this, so I don't
> really know...

No, it shouldn't :-/.  The dfps mode is effectively a frame buffer
driver that periodically sends draws along.

Long term, it probably should be a completely new Xorg driver; it
would be much simpler and the code would be a lot more clear.

Cheers,

Jeremy

> 
> Acked-by: Christophe Fergeau <cfergeau at redhat.com>
> 
> Christophe
> 

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1

iEYEARECAAYFAlZrFn8ACgkQq6U76EE4uydJnACcDLzmthlkitxqWjEzCbG2f7NO
NfUAnj/A/aPzeeiP6NE4djy8T/fwag4i
=3qzu
-----END PGP SIGNATURE-----