[PATCH 1/2] media: atomisp: better describe get_frame_info issues

[PATCH 1/2] media: atomisp: better describe get_frame_info issues

[Mauro Carvalho Chehab]

When atomisp is used by a normal client, it fails to get
frame info. However, the information is confusing and misleading,
as there are several wrappers for such function, and the error
could be on different places.

So, improve the error log in order to allow narrowing down
where the error is actually occuring.

Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
---
 .../staging/media/atomisp/pci/atomisp_cmd.c   |  4 +-
 .../media/atomisp/pci/atomisp_compat_css20.c  | 67 ++++++++++---------
 2 files changed, 39 insertions(+), 32 deletions(-)

diff --git a/drivers/staging/media/atomisp/pci/atomisp_cmd.c b/drivers/staging/media/atomisp/pci/atomisp_cmd.c
index a30dfcce54dd..32cae6908465 100644
--- a/drivers/staging/media/atomisp/pci/atomisp_cmd.c
+++ b/drivers/staging/media/atomisp/pci/atomisp_cmd.c
@@ -5444,9 +5444,9 @@ static int atomisp_set_fmt_to_isp(struct video_device *vdev,
 	else
 		ret = get_frame_info(asd, output_info);
 	if (ret) {
-		dev_err(isp->dev, "get_frame_info %ux%u (padded to %u) returned %d\n",
+		dev_err(isp->dev, "__get_frame_info %ux%u (padded to %u) returned %d\n",
 			pix->width, pix->height, pix->bytesperline, ret);
-		return -EINVAL;
+		return ret;
 	}
 
 	atomisp_update_grid_info(asd, pipe_id, source_pad);
diff --git a/drivers/staging/media/atomisp/pci/atomisp_compat_css20.c b/drivers/staging/media/atomisp/pci/atomisp_compat_css20.c
index 99a632f33d2d..1309855bb6c8 100644
--- a/drivers/staging/media/atomisp/pci/atomisp_compat_css20.c
+++ b/drivers/staging/media/atomisp/pci/atomisp_compat_css20.c
@@ -2657,41 +2657,48 @@ static int __get_frame_info(struct atomisp_sub_device *asd,
 	if (__destroy_pipes(asd, true))
 		dev_warn(isp->dev, "destroy pipe failed.\n");
 
-	if (__create_pipes(asd))
+	if (__create_pipes(asd)) {
+		dev_err(isp->dev, "can't create pipes\n");
 		return -EINVAL;
+	}
+
+	if (__create_streams(asd)) {
+		dev_err(isp->dev, "can't create streams\n");
+		goto stream_err;
+	}
 
-	if (__create_streams(asd))
+	ret = ia_css_pipe_get_info(asd->stream_env[stream_index].pipes[pipe_id],
+				   &p_info);
+	if (ret) {
+		dev_err(isp->dev, "can't get info from pipe\n");
 		goto stream_err;
+	}
 
-	ret = ia_css_pipe_get_info(
-		  asd->stream_env[stream_index]
-		  .pipes[pipe_id], &p_info);
-	if (!ret) {
-		switch (type) {
-		case ATOMISP_CSS_VF_FRAME:
-			*info = p_info.vf_output_info[0];
-			dev_dbg(isp->dev, "getting vf frame info.\n");
-			break;
-		case ATOMISP_CSS_SECOND_VF_FRAME:
-			*info = p_info.vf_output_info[1];
-			dev_dbg(isp->dev, "getting second vf frame info.\n");
-			break;
-		case ATOMISP_CSS_OUTPUT_FRAME:
-			*info = p_info.output_info[0];
-			dev_dbg(isp->dev, "getting main frame info.\n");
-			break;
-		case ATOMISP_CSS_SECOND_OUTPUT_FRAME:
-			*info = p_info.output_info[1];
-			dev_dbg(isp->dev, "getting second main frame info.\n");
-			break;
-		case ATOMISP_CSS_RAW_FRAME:
-			*info = p_info.raw_output_info;
-			dev_dbg(isp->dev, "getting raw frame info.\n");
-		}
-		dev_dbg(isp->dev, "get frame info: w=%d, h=%d, num_invalid_frames %d.\n",
-			info->res.width, info->res.height, p_info.num_invalid_frames);
-		return 0;
+	switch (type) {
+	case ATOMISP_CSS_VF_FRAME:
+		*info = p_info.vf_output_info[0];
+		dev_dbg(isp->dev, "getting vf frame info.\n");
+		break;
+	case ATOMISP_CSS_SECOND_VF_FRAME:
+		*info = p_info.vf_output_info[1];
+		dev_dbg(isp->dev, "getting second vf frame info.\n");
+		break;
+	case ATOMISP_CSS_OUTPUT_FRAME:
+		*info = p_info.output_info[0];
+		dev_dbg(isp->dev, "getting main frame info.\n");
+		break;
+	case ATOMISP_CSS_SECOND_OUTPUT_FRAME:
+		*info = p_info.output_info[1];
+		dev_dbg(isp->dev, "getting second main frame info.\n");
+		break;
+	case ATOMISP_CSS_RAW_FRAME:
+		*info = p_info.raw_output_info;
+		dev_dbg(isp->dev, "getting raw frame info.\n");
 	}
+	dev_dbg(isp->dev, "get frame info: w=%d, h=%d, num_invalid_frames %d.\n",
+		info->res.width, info->res.height, p_info.num_invalid_frames);
+
+	return 0;
 
 stream_err:
 	__destroy_pipes(asd, true);
-- 
2.31.1

[PATCH 2/2] media: atomisp: set default mode

[Mauro Carvalho Chehab]

Without setting a default mode at open(), applications that
don't call VIDIOC_SET_PARM with a custom atomisp parameters
won't work, as the pipeline won't be set:

	atomisp-isp2 0000:00:03.0: can't create streams
	atomisp-isp2 0000:00:03.0: __get_frame_info 1600x1200 (padded to 0) returned -22

So, as an step to allow generic apps to use this driver, put
the device's run_mode in preview after open.

After this patch, using v4l2grab starts to work:

	$ v4l2grab -D -f 'NV12' -x 1600 -y 1200 -d /dev/video2 -u
	$ nvt/raw2pnm -x1600 -y1200 -f NV12 out017.raw out017.pnm
	$ feh out017.pnm

Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
---
 drivers/staging/media/atomisp/pci/atomisp_fops.c | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/drivers/staging/media/atomisp/pci/atomisp_fops.c b/drivers/staging/media/atomisp/pci/atomisp_fops.c
index 72cbdce2142a..7df982c80b1a 100644
--- a/drivers/staging/media/atomisp/pci/atomisp_fops.c
+++ b/drivers/staging/media/atomisp/pci/atomisp_fops.c
@@ -893,6 +893,11 @@ static int atomisp_open(struct file *file)
 	else
 		pipe->users++;
 	rt_mutex_unlock(&isp->mutex);
+
+	/* Ensure that a mode is set */
+	if (asd)
+		v4l2_ctrl_s_ctrl(asd->run_mode, ATOMISP_RUN_MODE_PREVIEW);
+
 	return 0;
 
 css_error:
-- 
2.31.1

Re: [PATCH 1/2] media: atomisp: better describe get_frame_info issues

[Andy Shevchenko]

On Fri, Oct 29, 2021 at 11:50 AM Mauro Carvalho Chehab
<mchehab+huawei@kernel.org> wrote:
>
> When atomisp is used by a normal client, it fails to get
> frame info. However, the information is confusing and misleading,
> as there are several wrappers for such function, and the error
> could be on different places.
>
> So, improve the error log in order to allow narrowing down
> where the error is actually occuring.

...

> +       switch (type) {
> +       case ATOMISP_CSS_VF_FRAME:
> +               *info = p_info.vf_output_info[0];
> +               dev_dbg(isp->dev, "getting vf frame info.\n");
> +               break;
> +       case ATOMISP_CSS_SECOND_VF_FRAME:
> +               *info = p_info.vf_output_info[1];
> +               dev_dbg(isp->dev, "getting second vf frame info.\n");
> +               break;
> +       case ATOMISP_CSS_OUTPUT_FRAME:
> +               *info = p_info.output_info[0];
> +               dev_dbg(isp->dev, "getting main frame info.\n");
> +               break;
> +       case ATOMISP_CSS_SECOND_OUTPUT_FRAME:
> +               *info = p_info.output_info[1];
> +               dev_dbg(isp->dev, "getting second main frame info.\n");
> +               break;
> +       case ATOMISP_CSS_RAW_FRAME:
> +               *info = p_info.raw_output_info;
> +               dev_dbg(isp->dev, "getting raw frame info.\n");

Can we get break; here followed by default case?

>         }
> +       dev_dbg(isp->dev, "get frame info: w=%d, h=%d, num_invalid_frames %d.\n",
> +               info->res.width, info->res.height, p_info.num_invalid_frames);
> +
> +       return 0;


-- 
With Best Regards,
Andy Shevchenko

Re: [PATCH 2/2] media: atomisp: set default mode

[Mauro Carvalho Chehab]

Em Fri, 29 Oct 2021 09:49:36 +0100
Mauro Carvalho Chehab <mchehab+huawei@kernel.org> escreveu:

> Without setting a default mode at open(), applications that
> don't call VIDIOC_SET_PARM with a custom atomisp parameters
> won't work, as the pipeline won't be set:
> 
> 	atomisp-isp2 0000:00:03.0: can't create streams
> 	atomisp-isp2 0000:00:03.0: __get_frame_info 1600x1200 (padded to 0) returned -22
> 
> So, as an step to allow generic apps to use this driver, put
> the device's run_mode in preview after open.
> 
> After this patch, using v4l2grab starts to work:
> 
> 	$ v4l2grab -D -f 'NV12' -x 1600 -y 1200 -d /dev/video2 -u
> 	$ nvt/raw2pnm -x1600 -y1200 -f NV12 out017.raw out017.pnm
> 	$ feh out017.pnm

Added support for YUYV at v4l2grab (at contrib/test, under v4l-utils). 
The above can now be just:

	$ v4l2grab -f YUYV  -x 1600 -y 1200 -d /dev/video2 -u
	$ feh out017.pnm

(on a side note, the colorspace conversion from YUYV seems ackward,
but this can be adjusted later on at the toolset and/or at atomisp)

> 
> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
> ---
>  drivers/staging/media/atomisp/pci/atomisp_fops.c | 5 +++++
>  1 file changed, 5 insertions(+)
> 
> diff --git a/drivers/staging/media/atomisp/pci/atomisp_fops.c b/drivers/staging/media/atomisp/pci/atomisp_fops.c
> index 72cbdce2142a..7df982c80b1a 100644
> --- a/drivers/staging/media/atomisp/pci/atomisp_fops.c
> +++ b/drivers/staging/media/atomisp/pci/atomisp_fops.c
> @@ -893,6 +893,11 @@ static int atomisp_open(struct file *file)
>  	else
>  		pipe->users++;
>  	rt_mutex_unlock(&isp->mutex);
> +
> +	/* Ensure that a mode is set */
> +	if (asd)
> +		v4l2_ctrl_s_ctrl(asd->run_mode, ATOMISP_RUN_MODE_PREVIEW);
> +
>  	return 0;
>  
>  css_error:

Re: [PATCH 1/2] media: atomisp: better describe get_frame_info issues

[Mauro Carvalho Chehab]

Em Fri, 29 Oct 2021 12:01:57 +0300
Andy Shevchenko <andy.shevchenko@gmail.com> escreveu:

> > +       switch (type) {
> > +       case ATOMISP_CSS_VF_FRAME:
> > +               *info = p_info.vf_output_info[0];
> > +               dev_dbg(isp->dev, "getting vf frame info.\n");
> > +               break;
> > +       case ATOMISP_CSS_SECOND_VF_FRAME:
> > +               *info = p_info.vf_output_info[1];
> > +               dev_dbg(isp->dev, "getting second vf frame info.\n");
> > +               break;
> > +       case ATOMISP_CSS_OUTPUT_FRAME:
> > +               *info = p_info.output_info[0];
> > +               dev_dbg(isp->dev, "getting main frame info.\n");
> > +               break;
> > +       case ATOMISP_CSS_SECOND_OUTPUT_FRAME:
> > +               *info = p_info.output_info[1];
> > +               dev_dbg(isp->dev, "getting second main frame info.\n");
> > +               break;
> > +       case ATOMISP_CSS_RAW_FRAME:
> > +               *info = p_info.raw_output_info;
> > +               dev_dbg(isp->dev, "getting raw frame info.\n");  
> 
> Can we get break; here followed by default case?

Surely. I'll do such change on a separate patch.

Regards,
Mauro