On Mon, May 30, 2016 at 10:02:23AM +0200, Marcus Glocker wrote:
> I was in the middle of testing a uvideo(4) mmap queue diff when I
> noticed that our video(1) tool doesn't support the mmap method to grab
> frames. This diff adds it and also makes mmap the default method.
>
> Some test reports with different uvideo(4) devices would be welcome.
> It would be good if you could switch between different resolutions and
> also regression test the read method. E.g.:
>
> $ video
> $ video -s 800x600
> $ video -s 1600x1200
> $ video -g
> $ video -g -s 1024x768
> ...
>
> Thanks,
> Marcus
>
A quick test of the built-in camera of my T420, shows no regression when
directly compared to output of video(1) without your patch. I tested
each of the supported frame sizes and rates for about 10 seconds both
with and without -g.
$ dmesg | grep video
uvideo0 at uhub2 port 6 configuration 1 interface 0 "Chicony Electronics Co.,
Ltd. Integrated Camera" rev 2.00/7.52 addr 3
video0 at uvideo0
$ video -v
video device /dev/video:
encodings: yuy2
frame sizes (width x height, in pixels) and rates (in frames per second):
320x240: 30, 15
352x288: 30, 15
424x240: 30, 15
640x360: 30, 15
640x480: 30, 15
800x448: 15
960x544: 10
1280x720: 10
controls: brightness, contrast, saturation, hue, gamma, sharpness
Xv adaptor 0, Intel(R) Textured Video:
encodings: yuy2, uyvy
max size: 1600x900
