summaryrefslogtreecommitdiffstats
path: root/drivers/media/usb/usbtv/usbtv-video.c
diff options
context:
space:
mode:
authorLubomir Rintel <lkundrak@v3.sk>2016-06-01 10:04:07 -0300
committerMauro Carvalho Chehab <mchehab@s-opensource.com>2016-07-08 16:16:58 -0300
commit8ed50eb42041070b6d7a3d4a7736374d12e463bb (patch)
treee1128970b9446d5d9c581b1a44c077436423a89f /drivers/media/usb/usbtv/usbtv-video.c
parent5ae1e2b24f8f90e17a0e15e267f32a1eaa1281eb (diff)
downloadlinux-8ed50eb42041070b6d7a3d4a7736374d12e463bb.tar.bz2
[media] usbtv: improve a comment
Patrick Keshishian improved the explanation of the protocol when porting the driver to OpenBSD. Given it's a reverse engineering one and there's no documetnation it might be helpful to whoever hacks on the driver. Signed-off-by: Patrick Keshishian <sidster@boxsoft.com> Signed-off-by: Lubomir Rintel <lkundrak@v3.sk> Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Diffstat (limited to 'drivers/media/usb/usbtv/usbtv-video.c')
-rw-r--r--drivers/media/usb/usbtv/usbtv-video.c19
1 files changed, 17 insertions, 2 deletions
diff --git a/drivers/media/usb/usbtv/usbtv-video.c b/drivers/media/usb/usbtv/usbtv-video.c
index 256956b5499c..2a089756c988 100644
--- a/drivers/media/usb/usbtv/usbtv-video.c
+++ b/drivers/media/usb/usbtv/usbtv-video.c
@@ -265,8 +265,23 @@ static int usbtv_setup_capture(struct usbtv *usbtv)
/* Copy data from chunk into a frame buffer, deinterlacing the data
* into every second line. Unfortunately, they don't align nicely into
* 720 pixel lines, as the chunk is 240 words long, which is 480 pixels.
- * Therefore, we break down the chunk into two halves before copyting,
- * so that we can interleave a line if needed. */
+ * Therefore, we break down the chunk into two halves before copying,
+ * so that we can interleave a line if needed.
+ *
+ * Each "chunk" is 240 words; a word in this context equals 4 bytes.
+ * Image format is YUYV/YUV 4:2:2, consisting of Y Cr Y Cb, defining two
+ * pixels, the Cr and Cb shared between the two pixels, but each having
+ * separate Y values. Thus, the 240 words equal 480 pixels. It therefore,
+ * takes 1.5 chunks to make a 720 pixel-wide line for the frame.
+ * The image is interlaced, so there is a "scan" of odd lines, followed
+ * by "scan" of even numbered lines.
+ *
+ * Following code is writing the chunks in correct sequence, skipping
+ * the rows based on "odd" value.
+ * line 1: chunk[0][ 0..479] chunk[0][480..959] chunk[1][ 0..479]
+ * line 3: chunk[1][480..959] chunk[2][ 0..479] chunk[2][480..959]
+ * ...etc.
+ */
static void usbtv_chunk_to_vbuf(u32 *frame, __be32 *src, int chunk_no, int odd)
{
int half;