linux-media.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: "Niklas Söderlund" <niklas.soderlund@ragnatech.se>
To: Dafna Hirschfeld <dafna.hirschfeld@collabora.com>
Cc: linux-media@vger.kernel.org, laurent.pinchart@ideasonboard.com,
	helen.koike@collabora.com, ezequiel@collabora.com,
	hverkuil@xs4all.nl, kernel@collabora.com, dafna3@gmail.com,
	sakari.ailus@linux.intel.com, mchehab@kernel.org,
	tfiga@chromium.org
Subject: Re: [PATCH v3 09/10] media: rkisp1: cap: simplify the link validation by compering the media bus code
Date: Wed, 30 Sep 2020 21:00:25 +0200	[thread overview]
Message-ID: <20200930190025.GH1516931@oden.dyn.berto.se> (raw)
In-Reply-To: <20200723132014.4597-10-dafna.hirschfeld@collabora.com>

Hi Dafna,

This commit is not just a simplification but a change of behavior.  The 
change is for the better but it broke capture of NV12 and NV21 formats 
in libcamera unexpectedly.

The issue at hand is that libcamera when configuring the pipeline 
retrieves the mbus code for the ISP (rkisp1_isp) source pad (2) and then 
propagates it to the resizer (rkisp_resizer_{main,self}path) sink pad 
(0) and then to the resizers source pad (1). Effectively propagating 
MEDIA_BUS_FMT_YUYV8_2X8 for all formats.

At this point if the video node (main or self) is configured with a 
YUV420 format (NV12, NV21, etc) and with this change applied the link 
validation will fail as MEDIA_BUS_FMT_YUYV8_1_5X8 !=  
MEDIA_BUS_FMT_YUYV8_2X8. Given the nature of how link validation is 
implemented it's VIDIOC_QBUF that returns a -EPIPE when it fails and 
libcamera lockup the capture session.

I will submit a fix for this to libcamera to bring it in sync with this 
change.

Would it be possible to ask that future changes to the rkisp1 driver be 
tested with libcamera so we can track and update both the kernel and 
user-space components of this driver at the same time and avoid nasty 
surprises? :-)

On 2020-07-23 15:20:13 +0200, Dafna Hirschfeld wrote:
> The capture has a mapping of the mbus code needed for each pixelformat.
> This can be used to simplify the link validation by comparing the mbus
> code in the capture with the code in the resizer.
> 
> Signed-off-by: Dafna Hirschfeld <dafna.hirschfeld@collabora.com>
> ---
>  drivers/staging/media/rkisp1/rkisp1-capture.c | 18 ++++--------------
>  1 file changed, 4 insertions(+), 14 deletions(-)
> 
> diff --git a/drivers/staging/media/rkisp1/rkisp1-capture.c b/drivers/staging/media/rkisp1/rkisp1-capture.c
> index 4dabd07a3da9..a5e2521577dd 100644
> --- a/drivers/staging/media/rkisp1/rkisp1-capture.c
> +++ b/drivers/staging/media/rkisp1/rkisp1-capture.c
> @@ -1255,22 +1255,11 @@ static int rkisp1_capture_link_validate(struct media_link *link)
>  	struct v4l2_subdev *sd =
>  		media_entity_to_v4l2_subdev(link->source->entity);
>  	struct rkisp1_capture *cap = video_get_drvdata(vdev);
> -	struct rkisp1_isp *isp = &cap->rkisp1->isp;
> -	u8 isp_pix_enc = isp->src_fmt->pixel_enc;
> -	u8 cap_pix_enc = cap->pix.info->pixel_enc;
> +	const struct rkisp1_capture_fmt_cfg *fmt =
> +		rkisp1_find_fmt_cfg(cap, cap->pix.fmt.pixelformat);
>  	struct v4l2_subdev_format sd_fmt;
>  	int ret;
>  
> -	if (cap_pix_enc != isp_pix_enc &&
> -	    !(isp_pix_enc == V4L2_PIXEL_ENC_YUV &&
> -	      cap_pix_enc == V4L2_PIXEL_ENC_RGB)) {
> -		dev_err(cap->rkisp1->dev,
> -			"format type mismatch in link '%s:%d->%s:%d'\n",
> -			link->source->entity->name, link->source->index,
> -			link->sink->entity->name, link->sink->index);
> -		return -EPIPE;
> -	}
> -
>  	sd_fmt.which = V4L2_SUBDEV_FORMAT_ACTIVE;
>  	sd_fmt.pad = link->source->index;
>  	ret = v4l2_subdev_call(sd, pad, get_fmt, NULL, &sd_fmt);
> @@ -1278,7 +1267,8 @@ static int rkisp1_capture_link_validate(struct media_link *link)
>  		return ret;
>  
>  	if (sd_fmt.format.height != cap->pix.fmt.height ||
> -	    sd_fmt.format.width != cap->pix.fmt.width)
> +	    sd_fmt.format.width != cap->pix.fmt.width ||
> +	    sd_fmt.format.code != fmt->mbus)
>  		return -EPIPE;
>  
>  	return 0;
> -- 
> 2.17.1
> 

-- 
Regards,
Niklas Söderlund

  reply	other threads:[~2020-09-30 19:00 UTC|newest]

Thread overview: 30+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-07-23 13:20 [PATCH v3 00/10] media: staging: rkisp1: add support to V4L2_CAP_IO_MC Dafna Hirschfeld
2020-07-23 13:20 ` [PATCH v3 01/10] media: staging: rkisp1: cap: change RGB24 format to XBGR32 Dafna Hirschfeld
2020-08-03 23:49   ` Helen Koike
2020-08-31 16:33     ` Dafna Hirschfeld
2020-07-23 13:20 ` [PATCH v3 02/10] media: staging: rkisp1: cap: remove unsupported formats Dafna Hirschfeld
2020-07-23 13:20 ` [PATCH v3 03/10] media: staging: rkisp1: cap: remove unsupported format YUV444 Dafna Hirschfeld
2020-08-03 23:49   ` Helen Koike
2020-07-23 13:20 ` [PATCH v3 04/10] media: staging: rkisp1: don't support bayer format on selfpath resizer Dafna Hirschfeld
2020-08-03 23:49   ` Helen Koike
2020-07-23 13:20 ` [PATCH v3 05/10] media: staging: rkisp1: add capability V4L2_CAP_IO_MC to capture devices Dafna Hirschfeld
2020-08-03 23:50   ` Helen Koike
2020-08-03 23:50   ` Helen Koike
2020-07-23 13:20 ` [PATCH v3 06/10] media: staging: rkisp1: add a helper function to enumerate supported mbus formats on capture Dafna Hirschfeld
2020-08-03 23:49   ` Helen Koike
2020-08-29 17:39     ` Dafna Hirschfeld
2020-08-30 14:43   ` Tomasz Figa
2020-08-31 13:02     ` Dafna Hirschfeld
2020-07-23 13:20 ` [PATCH v3 07/10] media: staging: rkisp1: rsz: enumerate the formats on the src pad according to the capture Dafna Hirschfeld
2020-08-03 23:50   ` Helen Koike
2020-08-29 17:53     ` Dafna Hirschfeld
2020-08-29 20:48       ` Tomasz Figa
2020-07-23 13:20 ` [PATCH v3 08/10] media: staging: rkisp1: rsz: Add support to more YUV encoded mbus codes on src pad Dafna Hirschfeld
2020-07-23 13:20 ` [PATCH v3 09/10] media: rkisp1: cap: simplify the link validation by compering the media bus code Dafna Hirschfeld
2020-09-30 19:00   ` Niklas Söderlund [this message]
2020-10-01  2:03     ` Laurent Pinchart
2020-10-01 19:36       ` Dafna Hirschfeld
2020-10-01 22:48         ` Laurent Pinchart
2020-10-02  7:04           ` Dafna Hirschfeld
2020-10-22 11:02           ` Helen Koike
2020-07-23 13:20 ` [PATCH v3 10/10] media: staging: rkisp1: fix configuration for GREY pixelformat Dafna Hirschfeld

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20200930190025.GH1516931@oden.dyn.berto.se \
    --to=niklas.soderlund@ragnatech.se \
    --cc=dafna.hirschfeld@collabora.com \
    --cc=dafna3@gmail.com \
    --cc=ezequiel@collabora.com \
    --cc=helen.koike@collabora.com \
    --cc=hverkuil@xs4all.nl \
    --cc=kernel@collabora.com \
    --cc=laurent.pinchart@ideasonboard.com \
    --cc=linux-media@vger.kernel.org \
    --cc=mchehab@kernel.org \
    --cc=sakari.ailus@linux.intel.com \
    --cc=tfiga@chromium.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).