2015-06-29 21:19:21

by Benoit Parrot

Subject: [Patch v3 1/1] media: am437x-vpfe: Requested frame size and fmt overwritten by current sensor setting

Upon a S_FMT the input/requested frame size and pixel format is
overwritten by the current sub-device settings.
Fix this so application can actually set the frame size and format.

Fixes: 417d2e507edc ("[media] media: platform: add VPFE capture driver support for AM437X")
Cc: <[email protected]> # v4.0+
Signed-off-by: Benoit Parrot <[email protected]>
---
Changes since v2:
- fix the stable commit reference syntax

drivers/media/platform/am437x/am437x-vpfe.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/media/platform/am437x/am437x-vpfe.c b/drivers/media/platform/am437x/am437x-vpfe.c
index eb25c43da126..0fa62c50f62d 100644
--- a/drivers/media/platform/am437x/am437x-vpfe.c
+++ b/drivers/media/platform/am437x/am437x-vpfe.c
@@ -1584,7 +1584,7 @@ static int vpfe_s_fmt(struct file *file, void *priv,
return -EBUSY;
}

- ret = vpfe_try_fmt(file, priv, fmt);
+ ret = vpfe_try_fmt(file, priv, &format);
if (ret)
return ret;

--
1.8.5.1

2015-07-07 11:13:40

by Lad, Prabhakar

Subject: Re: [Patch v3 1/1] media: am437x-vpfe: Requested frame size and fmt overwritten by current sensor setting

On Mon, Jun 29, 2015 at 10:19 PM, Benoit Parrot <[email protected]> wrote:
> Upon a S_FMT the input/requested frame size and pixel format is
> overwritten by the current sub-device settings.
> Fix this so application can actually set the frame size and format.
>
> Fixes: 417d2e507edc ("[media] media: platform: add VPFE capture driver support for AM437X")
> Cc: <[email protected]> # v4.0+
> Signed-off-by: Benoit Parrot <[email protected]>

Acked-by: Lad, Prabhakar <[email protected]>

Cheers,
--Prabhakar Lad

> ---
> Changes since v2:
> - fix the stable commit reference syntax
>
> drivers/media/platform/am437x/am437x-vpfe.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/media/platform/am437x/am437x-vpfe.c b/drivers/media/platform/am437x/am437x-vpfe.c
> index eb25c43da126..0fa62c50f62d 100644
> --- a/drivers/media/platform/am437x/am437x-vpfe.c
> +++ b/drivers/media/platform/am437x/am437x-vpfe.c
> @@ -1584,7 +1584,7 @@ static int vpfe_s_fmt(struct file *file, void *priv,
> return -EBUSY;
> }
>
> - ret = vpfe_try_fmt(file, priv, fmt);
> + ret = vpfe_try_fmt(file, priv, &format);
> if (ret)
> return ret;
>
> --
> 1.8.5.1
>

2015-07-13 09:20:58

by Hans Verkuil

Subject: Re: [Patch v3 1/1] media: am437x-vpfe: Requested frame size and fmt overwritten by current sensor setting

On 06/29/2015 11:19 PM, Benoit Parrot wrote:
> Upon a S_FMT the input/requested frame size and pixel format is
> overwritten by the current sub-device settings.
> Fix this so application can actually set the frame size and format.
>
> Fixes: 417d2e507edc ("[media] media: platform: add VPFE capture driver support for AM437X")
> Cc: <[email protected]> # v4.0+
> Signed-off-by: Benoit Parrot <[email protected]>
> ---
> Changes since v2:
> - fix the stable commit reference syntax
>
> drivers/media/platform/am437x/am437x-vpfe.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/media/platform/am437x/am437x-vpfe.c b/drivers/media/platform/am437x/am437x-vpfe.c
> index eb25c43da126..0fa62c50f62d 100644
> --- a/drivers/media/platform/am437x/am437x-vpfe.c
> +++ b/drivers/media/platform/am437x/am437x-vpfe.c
> @@ -1584,7 +1584,7 @@ static int vpfe_s_fmt(struct file *file, void *priv,
> return -EBUSY;
> }
>
> - ret = vpfe_try_fmt(file, priv, fmt);
> + ret = vpfe_try_fmt(file, priv, &format);
> if (ret)
> return ret;
>
>

I'm sorry, but this is wrong. The actual bug is not in s_fmt but in try_fmt.

try_fmt is not actually attempting to 'try' the format, but it just gets the
current subdev settings.

Instead it should do what s_fmt is doing, except for actually setting the format.

Frankly the am437x code is a bit of a mess in how it implements try and s_fmt.

Now, I am going to merge this patch anyway since it is clear from the code
that &format was intended. But this code certainly needs some more TLC since
this patch only addresses a small part of a much bigger problem.

Regards,

Hans