Hi Prashant,
Thanks for the patch, but I have one small comment:
On 03/20/2015 07:33 AM, Prashant Laddha wrote:
> This patch adds support for calculating v4l2_bt_timings based on
> CVT and GTF standards. The timings are calculated for a given
> standard, CVT or GTF using a set of parameters- width, height,
> refresh rate and flags like whether it is an interlaced format,
> and whether to use reduced blanking.
>
> CVT Modeline calculation -
> Implements Coordinated Video Timings (CVT) Standard Ver 1.2 Feb 08,
> 2013. The timing calculations are based on VESA CVT Generator Rev 1.2
> by Graham Loveridge May 28, 2013.
>
> GTF modeline calculation -
> Implements Generalized Timing Formula (GTF) Standard Ver 1.1 Sept 02,
> 1999. The timing calculations are based on GTF timing spreadsheet by
> Andy Morrish. The default GTF timings are used if flag for reduced
> blanking is false, otherwise secondary GTF timings are used.
>
> Suggested by: Hans Verkuil <hans.verkuil@xxxxxxxxx>
>
> Cc: Hans Verkuil <hans.verkuil@xxxxxxxxx>
> Signed-off-by: Prashant Laddha <prladdha@xxxxxxxxx>
> ---
> utils/v4l2-ctl/Makefile.am | 3 +-
> utils/v4l2-ctl/v4l2-ctl-modes.cpp | 512 ++++++++++++++++++++++++++++++++++++++
> utils/v4l2-ctl/v4l2-ctl.h | 8 +
> 3 files changed, 522 insertions(+), 1 deletion(-)
> create mode 100644 utils/v4l2-ctl/v4l2-ctl-modes.cpp
>
> diff --git a/utils/v4l2-ctl/v4l2-ctl-modes.cpp b/utils/v4l2-ctl/v4l2-ctl-modes.cpp
> new file mode 100644
> index 0000000..5157fea
> --- /dev/null
> +++ b/utils/v4l2-ctl/v4l2-ctl-modes.cpp
<snip>
> + if (cvt->interlaced == V4L2_DV_INTERLACED) {
> + cvt->il_vfrontporch = v_fp;
> + cvt->il_vsync = v_sync;
> + cvt->il_vbackporch = v_bp;
> + /* For interlaced format, add half lines to front and back
> + * porches of odd and even fields respectively */
> + cvt->flags |= V4L2_DV_FL_HALF_LINE;
> + cvt->vfrontporch += 1;
> + cvt->il_vbackporch += 1;
This isn't right, you should do the +1 only for the il_vbackporch. Otherwise
V4L2_DV_BT_FRAME_HEIGHT(bt) would be one too big.
The HALF_LINE flag means that, if drivers support it, they can add a half-line
to the vfrontporch of the odd field and subtract a half-line for the vbackporch
of the even field.
BTW, the V4L2_DV_FL_HALF_LINE documentation in the spec should be improved to
state the above. Right now it doesn't specify to which porch the half-lines go.
This same bug is in the detect_gtf function.
Regards,
Hans
--
To unsubscribe from this list: send the line "unsubscribe linux-media" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
