v4l2videodec: use visible size, not coded size, for downstream negotiation filter
authorPhilipp Zabel <p.zabel@pengutronix.de>
Fri, 13 May 2016 12:58:41 +0000 (14:58 +0200)
committerNicolas Dufresne <nicolas.dufresne@collabora.com>
Mon, 6 Jun 2016 21:35:47 +0000 (17:35 -0400)
commit1f31715c98616791293d0f9539d43845d7e0343e
treedf54a6e29adf69c44c387915ca707d3de3268638
parent74f020fd2f1dc645efe35a7ba1f951f9c5ee7c4c
v4l2videodec: use visible size, not coded size, for downstream negotiation filter

gst_v4l2_probe_caps() returns the coded size, not the visible size. Subtract
the known padding from probed caps with the coded size before using them as
filter for caps negotiation with downstream elements.

https://bugzilla.gnome.org/show_bug.cgi?id=766382
sys/v4l2/gstv4l2videodec.c