Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]


Groups > linux.kernel > #1447335 > unrolled thread

[PATCH v2] [media] vb2: include lengths in dmabuf qbuf debug message

Started byJavier Martinez Canillas <javier@osg.samsung.com>
First post2016-07-20 18:10 +0200
Last post2016-07-21 01:30 +0200
Articles 5 — 2 participants

Back to article view | Back to linux.kernel


Contents

  [PATCH v2] [media] vb2: include lengths in dmabuf qbuf debug message Javier Martinez Canillas <javier@osg.samsung.com> - 2016-07-20 18:10 +0200
    Re: [PATCH v2] [media] vb2: include lengths in dmabuf qbuf debug  message Sakari Ailus <sakari.ailus@iki.fi> - 2016-07-20 22:00 +0200
      Re: [PATCH v2] [media] vb2: include lengths in dmabuf qbuf debug  message Javier Martinez Canillas <javier@osg.samsung.com> - 2016-07-21 01:00 +0200
        Re: [PATCH v2] [media] vb2: include lengths in dmabuf qbuf debug  message Sakari Ailus <sakari.ailus@iki.fi> - 2016-07-21 01:10 +0200
          Re: [PATCH v2] [media] vb2: include lengths in dmabuf qbuf debug  message Javier Martinez Canillas <javier@osg.samsung.com> - 2016-07-21 01:30 +0200

#1447335 — [PATCH v2] [media] vb2: include lengths in dmabuf qbuf debug message

FromJavier Martinez Canillas <javier@osg.samsung.com>
Date2016-07-20 18:10 +0200
Subject[PATCH v2] [media] vb2: include lengths in dmabuf qbuf debug message
Message-ID<rX2en-3Bs-37@gated-at.bofh.it>
If the VIDIOC_QBUF ioctl fails due a wrong dmabuf length, it's
useful to get the invalid and minimum lengths as a debug info.

Before this patch:

vb2-core: __qbuf_dmabuf: invalid dmabuf length for plane 1

After this patch:

vb2-core: __qbuf_dmabuf: invalid dmabuf length 221248 for plane 1, minimum length 410880

Signed-off-by: Javier Martinez Canillas <javier@osg.samsung.com>

---

Changes in v2:
- Use %u instead of %d (Sakari Ailus)
- Include min_length (Sakari Ailus)

 drivers/media/v4l2-core/videobuf2-core.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/drivers/media/v4l2-core/videobuf2-core.c b/drivers/media/v4l2-core/videobuf2-core.c
index b6fbc04f9699..bbba50d6e1ad 100644
--- a/drivers/media/v4l2-core/videobuf2-core.c
+++ b/drivers/media/v4l2-core/videobuf2-core.c
@@ -1227,8 +1227,10 @@ static int __qbuf_dmabuf(struct vb2_buffer *vb, const void *pb)
 			planes[plane].length = dbuf->size;
 
 		if (planes[plane].length < vb->planes[plane].min_length) {
-			dprintk(1, "invalid dmabuf length for plane %d\n",
-				plane);
+			dprintk(1, "invalid dmabuf length %u for plane %d, "
+				"minimum length %u\n",
+				planes[plane].length, plane,
+				vb->planes[plane].min_length);
 			dma_buf_put(dbuf);
 			ret = -EINVAL;
 			goto err;
-- 
2.5.5

[toc] | [next] | [standalone]


#1447425 — Re: [PATCH v2] [media] vb2: include lengths in dmabuf qbuf debug message

FromSakari Ailus <sakari.ailus@iki.fi>
Date2016-07-20 22:00 +0200
SubjectRe: [PATCH v2] [media] vb2: include lengths in dmabuf qbuf debug message
Message-ID<rX5OV-5J8-5@gated-at.bofh.it>
In reply to#1447335
On Wed, Jul 20, 2016 at 12:07:55PM -0400, Javier Martinez Canillas wrote:
> If the VIDIOC_QBUF ioctl fails due a wrong dmabuf length, it's
> useful to get the invalid and minimum lengths as a debug info.
> 
> Before this patch:
> 
> vb2-core: __qbuf_dmabuf: invalid dmabuf length for plane 1
> 
> After this patch:
> 
> vb2-core: __qbuf_dmabuf: invalid dmabuf length 221248 for plane 1, minimum length 410880
> 
> Signed-off-by: Javier Martinez Canillas <javier@osg.samsung.com>
> 
> ---
> 
> Changes in v2:
> - Use %u instead of %d (Sakari Ailus)
> - Include min_length (Sakari Ailus)
> 
>  drivers/media/v4l2-core/videobuf2-core.c | 6 ++++--
>  1 file changed, 4 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/media/v4l2-core/videobuf2-core.c b/drivers/media/v4l2-core/videobuf2-core.c
> index b6fbc04f9699..bbba50d6e1ad 100644
> --- a/drivers/media/v4l2-core/videobuf2-core.c
> +++ b/drivers/media/v4l2-core/videobuf2-core.c
> @@ -1227,8 +1227,10 @@ static int __qbuf_dmabuf(struct vb2_buffer *vb, const void *pb)
>  			planes[plane].length = dbuf->size;
>  
>  		if (planes[plane].length < vb->planes[plane].min_length) {
> -			dprintk(1, "invalid dmabuf length for plane %d\n",
> -				plane);
> +			dprintk(1, "invalid dmabuf length %u for plane %d, "
> +				"minimum length %u\n",

You shouldn't split strings. It breaks grep.

With that changed,

Acked-by: Sakari Ailus <sakari.ailus@linux.intel.com>

> +				planes[plane].length, plane,
> +				vb->planes[plane].min_length);
>  			dma_buf_put(dbuf);
>  			ret = -EINVAL;
>  			goto err;

-- 
Sakari Ailus
e-mail: sakari.ailus@iki.fi	XMPP: sailus@retiisi.org.uk

[toc] | [prev] | [next] | [standalone]


#1447526 — Re: [PATCH v2] [media] vb2: include lengths in dmabuf qbuf debug message

FromJavier Martinez Canillas <javier@osg.samsung.com>
Date2016-07-21 01:00 +0200
SubjectRe: [PATCH v2] [media] vb2: include lengths in dmabuf qbuf debug message
Message-ID<rX8D8-7wq-13@gated-at.bofh.it>
In reply to#1447425
Hello Sakari,

On 07/20/2016 03:52 PM, Sakari Ailus wrote:
> On Wed, Jul 20, 2016 at 12:07:55PM -0400, Javier Martinez Canillas wrote:
>> If the VIDIOC_QBUF ioctl fails due a wrong dmabuf length, it's
>> useful to get the invalid and minimum lengths as a debug info.
>>
>> Before this patch:
>>
>> vb2-core: __qbuf_dmabuf: invalid dmabuf length for plane 1
>>
>> After this patch:
>>
>> vb2-core: __qbuf_dmabuf: invalid dmabuf length 221248 for plane 1, minimum length 410880
>>
>> Signed-off-by: Javier Martinez Canillas <javier@osg.samsung.com>
>>
>> ---
>>
>> Changes in v2:
>> - Use %u instead of %d (Sakari Ailus)
>> - Include min_length (Sakari Ailus)
>>
>>  drivers/media/v4l2-core/videobuf2-core.c | 6 ++++--
>>  1 file changed, 4 insertions(+), 2 deletions(-)
>>
>> diff --git a/drivers/media/v4l2-core/videobuf2-core.c b/drivers/media/v4l2-core/videobuf2-core.c
>> index b6fbc04f9699..bbba50d6e1ad 100644
>> --- a/drivers/media/v4l2-core/videobuf2-core.c
>> +++ b/drivers/media/v4l2-core/videobuf2-core.c
>> @@ -1227,8 +1227,10 @@ static int __qbuf_dmabuf(struct vb2_buffer *vb, const void *pb)
>>  			planes[plane].length = dbuf->size;
>>  
>>  		if (planes[plane].length < vb->planes[plane].min_length) {
>> -			dprintk(1, "invalid dmabuf length for plane %d\n",
>> -				plane);
>> +			dprintk(1, "invalid dmabuf length %u for plane %d, "
>> +				"minimum length %u\n",
> 
> You shouldn't split strings. It breaks grep.
> 

Yes I know but if I didn't split the line, it would had been longer than
the max 80 character per line convention. On those situations I follow
what's already done in the file for consistency and most strings in the
videobuf2-core file, whose lines are over 80 characters, are being split.

But if having a longer line is preferred, I'll happily re-spin the patch.

> With that changed,
> 
> Acked-by: Sakari Ailus <sakari.ailus@linux.intel.com>
> 
>> +				planes[plane].length, plane,
>> +				vb->planes[plane].min_length);
>>  			dma_buf_put(dbuf);
>>  			ret = -EINVAL;
>>  			goto err;
> 

Best regards,
-- 
Javier Martinez Canillas
Open Source Group
Samsung Research America

[toc] | [prev] | [next] | [standalone]


#1447529 — Re: [PATCH v2] [media] vb2: include lengths in dmabuf qbuf debug message

FromSakari Ailus <sakari.ailus@iki.fi>
Date2016-07-21 01:10 +0200
SubjectRe: [PATCH v2] [media] vb2: include lengths in dmabuf qbuf debug message
Message-ID<rX8MO-7OR-27@gated-at.bofh.it>
In reply to#1447526
Hi Javier,

On Wed, Jul 20, 2016 at 06:56:52PM -0400, Javier Martinez Canillas wrote:
> Hello Sakari,
> 
> On 07/20/2016 03:52 PM, Sakari Ailus wrote:
> > On Wed, Jul 20, 2016 at 12:07:55PM -0400, Javier Martinez Canillas wrote:
> >> If the VIDIOC_QBUF ioctl fails due a wrong dmabuf length, it's
> >> useful to get the invalid and minimum lengths as a debug info.
> >>
> >> Before this patch:
> >>
> >> vb2-core: __qbuf_dmabuf: invalid dmabuf length for plane 1
> >>
> >> After this patch:
> >>
> >> vb2-core: __qbuf_dmabuf: invalid dmabuf length 221248 for plane 1, minimum length 410880
> >>
> >> Signed-off-by: Javier Martinez Canillas <javier@osg.samsung.com>
> >>
> >> ---
> >>
> >> Changes in v2:
> >> - Use %u instead of %d (Sakari Ailus)
> >> - Include min_length (Sakari Ailus)
> >>
> >>  drivers/media/v4l2-core/videobuf2-core.c | 6 ++++--
> >>  1 file changed, 4 insertions(+), 2 deletions(-)
> >>
> >> diff --git a/drivers/media/v4l2-core/videobuf2-core.c b/drivers/media/v4l2-core/videobuf2-core.c
> >> index b6fbc04f9699..bbba50d6e1ad 100644
> >> --- a/drivers/media/v4l2-core/videobuf2-core.c
> >> +++ b/drivers/media/v4l2-core/videobuf2-core.c
> >> @@ -1227,8 +1227,10 @@ static int __qbuf_dmabuf(struct vb2_buffer *vb, const void *pb)
> >>  			planes[plane].length = dbuf->size;
> >>  
> >>  		if (planes[plane].length < vb->planes[plane].min_length) {
> >> -			dprintk(1, "invalid dmabuf length for plane %d\n",
> >> -				plane);
> >> +			dprintk(1, "invalid dmabuf length %u for plane %d, "
> >> +				"minimum length %u\n",
> > 
> > You shouldn't split strings. It breaks grep.
> > 
> 
> Yes I know but if I didn't split the line, it would had been longer than
> the max 80 character per line convention. On those situations I follow
> what's already done in the file for consistency and most strings in the
> videobuf2-core file, whose lines are over 80 characters, are being split.
> 
> But if having a longer line is preferred, I'll happily re-spin the patch.

I guess in videobuf2's case it's that the strings contain lots of format
specifiers --- it's not that useful to be able to grep those. I think this
case is a borderline one.

Up to you. You've got my ack.

> > With that changed,
> > 
> > Acked-by: Sakari Ailus <sakari.ailus@linux.intel.com>
> > 
> >> +				planes[plane].length, plane,
> >> +				vb->planes[plane].min_length);
> >>  			dma_buf_put(dbuf);
> >>  			ret = -EINVAL;
> >>  			goto err;
> > 
> 
> Best regards,
> -- 
> Javier Martinez Canillas
> Open Source Group
> Samsung Research America

-- 
Kind regards,

Sakari Ailus
e-mail: sakari.ailus@iki.fi	XMPP: sailus@retiisi.org.uk

[toc] | [prev] | [next] | [standalone]


#1447537 — Re: [PATCH v2] [media] vb2: include lengths in dmabuf qbuf debug message

FromJavier Martinez Canillas <javier@osg.samsung.com>
Date2016-07-21 01:30 +0200
SubjectRe: [PATCH v2] [media] vb2: include lengths in dmabuf qbuf debug message
Message-ID<rX969-7Vf-1@gated-at.bofh.it>
In reply to#1447529
Hello Sakari,

On 07/20/2016 07:08 PM, Sakari Ailus wrote:
> Hi Javier,
> 
> On Wed, Jul 20, 2016 at 06:56:52PM -0400, Javier Martinez Canillas wrote:
>> Hello Sakari,
>>
>> On 07/20/2016 03:52 PM, Sakari Ailus wrote:
>>> On Wed, Jul 20, 2016 at 12:07:55PM -0400, Javier Martinez Canillas wrote:
>>>> If the VIDIOC_QBUF ioctl fails due a wrong dmabuf length, it's
>>>> useful to get the invalid and minimum lengths as a debug info.
>>>>
>>>> Before this patch:
>>>>
>>>> vb2-core: __qbuf_dmabuf: invalid dmabuf length for plane 1
>>>>
>>>> After this patch:
>>>>
>>>> vb2-core: __qbuf_dmabuf: invalid dmabuf length 221248 for plane 1, minimum length 410880
>>>>
>>>> Signed-off-by: Javier Martinez Canillas <javier@osg.samsung.com>
>>>>
>>>> ---
>>>>
>>>> Changes in v2:
>>>> - Use %u instead of %d (Sakari Ailus)
>>>> - Include min_length (Sakari Ailus)
>>>>
>>>>  drivers/media/v4l2-core/videobuf2-core.c | 6 ++++--
>>>>  1 file changed, 4 insertions(+), 2 deletions(-)
>>>>
>>>> diff --git a/drivers/media/v4l2-core/videobuf2-core.c b/drivers/media/v4l2-core/videobuf2-core.c
>>>> index b6fbc04f9699..bbba50d6e1ad 100644
>>>> --- a/drivers/media/v4l2-core/videobuf2-core.c
>>>> +++ b/drivers/media/v4l2-core/videobuf2-core.c
>>>> @@ -1227,8 +1227,10 @@ static int __qbuf_dmabuf(struct vb2_buffer *vb, const void *pb)
>>>>  			planes[plane].length = dbuf->size;
>>>>  
>>>>  		if (planes[plane].length < vb->planes[plane].min_length) {
>>>> -			dprintk(1, "invalid dmabuf length for plane %d\n",
>>>> -				plane);
>>>> +			dprintk(1, "invalid dmabuf length %u for plane %d, "
>>>> +				"minimum length %u\n",
>>>
>>> You shouldn't split strings. It breaks grep.
>>>
>>
>> Yes I know but if I didn't split the line, it would had been longer than
>> the max 80 character per line convention. On those situations I follow
>> what's already done in the file for consistency and most strings in the
>> videobuf2-core file, whose lines are over 80 characters, are being split.
>>
>> But if having a longer line is preferred, I'll happily re-spin the patch.
> 
> I guess in videobuf2's case it's that the strings contain lots of format
> specifiers --- it's not that useful to be able to grep those. I think this
> case is a borderline one.
>

Yes, lots of format specifiers as you said. Which also makes harder to grep
the messages verbatim and someone would had to grep a sub-string anyways or
figure the specifier for each variable to be able to grep the complete line.

So I would prefer to kept the line split as is if you don't mind.
 
> Up to you. You've got my ack.
>

Thanks a lot.
 
>>> With that changed,
>>>
>>> Acked-by: Sakari Ailus <sakari.ailus@linux.intel.com>
>>>
>>>> +				planes[plane].length, plane,
>>>> +				vb->planes[plane].min_length);
>>>>  			dma_buf_put(dbuf);
>>>>  			ret = -EINVAL;
>>>>  			goto err;
>>>
>>
>> Best regards,
>> -- 
>> Javier Martinez Canillas
>> Open Source Group
>> Samsung Research America
> 

Best regards,
-- 
Javier Martinez Canillas
Open Source Group
Samsung Research America

[toc] | [prev] | [standalone]


Back to top | Article view | linux.kernel


csiph-web