The atomisp sensors and logic adds an extra pad lines/columns, called "dvs envelope". It also uses an extra 12 lines/columns at the sensor for BYT. As those are not visible to userspace, the V4L2 API should decrement such values when reporting the current resolution. Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@xxxxxxxxxx> --- drivers/staging/media/atomisp/pci/atomisp_ioctl.c | 14 ++++++++++++++ 1 file changed, 14 insertions(+) diff --git a/drivers/staging/media/atomisp/pci/atomisp_ioctl.c b/drivers/staging/media/atomisp/pci/atomisp_ioctl.c index 8f0a9a69f075..d92d41dd0ade 100644 --- a/drivers/staging/media/atomisp/pci/atomisp_ioctl.c +++ b/drivers/staging/media/atomisp/pci/atomisp_ioctl.c @@ -876,6 +876,20 @@ static int atomisp_g_fmt_cap(struct file *file, void *fh, v4l2_fill_pix_format(&f->fmt.pix, &fmt.format); f->fmt.pix.pixelformat = fmtdesc.pixelformat; + + /* + * HACK: The atomisp does something different here, as it + * seems to set the sensor to a slightly higher resolution than + * the visible ones. That seems to be needed by atomisp's ISP + * in order to properly handle the frames. So, s_fmt adds 16 + * extra columns/lines. See atomisp_subdev_set_selection(). + * + * Yet, the V4L2 userspace API doesn't expect those, so it + * needs to be decremented when reporting the visible + * resolution to userspace. + */ + f->fmt.pix.width -= pad_w; + f->fmt.pix.height -= pad_h; } depth = atomisp_get_pixel_depth(f->fmt.pix.pixelformat); -- 2.31.1
