Hi Thomas,
that's just an optimization. DRI2 can't send buffers allocated by the
client to the X server for displaying, e.g. all allocations must be done
by X.
So in the simple case of just rendering directly into the output buffer
without any advanced filters I just delayed the rendering until we got
the real displayable buffer from the X server and rendered directly into
that one.
This saves us an extra copy of the video frames and so quite a bit of
memory bandwith.
It's not necessary any more with DRI3, so we are on the verge of
removing that.
Regards,
Christian.
Am 24.10.2016 um 07:03 schrieb Thomas Hellstrom:
Hi, Christian,
I remember some time ago you briefly explained what the delayed
rendering in st/vdpau is all about, but I can't seem to find that email.
Could you point me to a description or briefly detail what it is all
about, and also why it doesn't seem to be needed when vdpau uses filters?
Thanks,
Thomas
_______________________________________________
mesa-dev mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/mesa-dev
