The g_fmt logic is currently broken, as it is not returning the same imagesize as previoulsy calculated by s_fmt. Fix it by just re-using whatever it was defined at s_fmt, if this was called before. Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@xxxxxxxxxx> --- To mailbombing on a large number of people, only mailing lists were C/C on the cover. See [PATCH 0/7] at: https://lore.kernel.org/all/cover.1636026959.git.mchehab+huawei@xxxxxxxxxx/ .../staging/media/atomisp/pci/atomisp_ioctl.c | 64 ++----------------- 1 file changed, 7 insertions(+), 57 deletions(-) diff --git a/drivers/staging/media/atomisp/pci/atomisp_ioctl.c b/drivers/staging/media/atomisp/pci/atomisp_ioctl.c index 1e6da6116a06..2fb64d5cbead 100644 --- a/drivers/staging/media/atomisp/pci/atomisp_ioctl.c +++ b/drivers/staging/media/atomisp/pci/atomisp_ioctl.c @@ -934,75 +934,25 @@ static int atomisp_try_fmt_cap(struct file *file, void *fh, static int atomisp_g_fmt_cap(struct file *file, void *fh, struct v4l2_format *f) { - struct v4l2_subdev_format fmt = { - .which = V4L2_SUBDEV_FORMAT_ACTIVE - }; struct video_device *vdev = video_devdata(file); struct atomisp_device *isp = video_get_drvdata(vdev); - struct v4l2_fmtdesc fmtdesc = { 0 }; struct atomisp_video_pipe *pipe; - struct atomisp_sub_device *asd; - struct v4l2_subdev *camera; - u32 depth; - int ret; rt_mutex_lock(&isp->mutex); pipe = atomisp_to_video_pipe(vdev); rt_mutex_unlock(&isp->mutex); f->fmt.pix = pipe->pix; - if (!f->fmt.pix.width) { - asd = atomisp_to_video_pipe(vdev)->asd; - if (!asd) - return -EINVAL; - camera = isp->inputs[asd->input_curr].camera; - if(!camera) - return -EINVAL; + /* If s_fmt was issued, just return whatever is was previouly set */ + if (f->fmt.pix.sizeimage) + return 0; - ret = atomisp_enum_fmt_cap(file, fh, &fmtdesc); - if (ret) - return ret; + f->fmt.pix.pixelformat = V4L2_PIX_FMT_YUYV; + f->fmt.pix.width = 10000; + f->fmt.pix.height = 10000; - rt_mutex_lock(&isp->mutex); - ret = v4l2_subdev_call(isp->inputs[asd->input_curr].camera, - pad, get_fmt, NULL, &fmt); - rt_mutex_unlock(&isp->mutex); - if (ret) - return ret; - - 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); - f->fmt.pix.bytesperline = DIV_ROUND_UP(f->fmt.pix.width * depth, 8); - f->fmt.pix.sizeimage = PAGE_ALIGN(f->fmt.pix.height * f->fmt.pix.bytesperline); - - /* - * FIXME: do we need to setup this differently, depending on the - * sensor or the pipeline? - */ - f->fmt.pix.colorspace = V4L2_COLORSPACE_REC709; - f->fmt.pix.ycbcr_enc = V4L2_YCBCR_ENC_709; - f->fmt.pix.xfer_func = V4L2_XFER_FUNC_709; - - return 0; + return atomisp_try_fmt_cap(file, fh, f); } static int atomisp_s_fmt_cap(struct file *file, void *fh, -- 2.31.1