lkml.org 
[lkml]   [2018]   [Sep]   [6]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
SubjectRe: [PATCH v8 4/8] media: platform: Add Cedrus VPU decoder driver
From
Date
On 09/06/2018 09:01 AM, Hans Verkuil wrote:
> On 09/05/2018 06:29 PM, Paul Kocialkowski wrote:
>> Hi and thanks for the review!
>>
>> Le lundi 03 septembre 2018 à 11:11 +0200, Hans Verkuil a écrit :
>>> On 08/28/2018 09:34 AM, Paul Kocialkowski wrote:
>>>> +static int cedrus_queue_setup(struct vb2_queue *vq, unsigned int *nbufs,
>>>> + unsigned int *nplanes, unsigned int sizes[],
>>>> + struct device *alloc_devs[])
>>>> +{
>>>> + struct cedrus_ctx *ctx = vb2_get_drv_priv(vq);
>>>> + struct cedrus_dev *dev = ctx->dev;
>>>> + struct v4l2_pix_format_mplane *mplane_fmt;
>>>> + struct cedrus_format *fmt;
>>>> + unsigned int i;
>>>> +
>>>> + switch (vq->type) {
>>>> + case V4L2_BUF_TYPE_VIDEO_OUTPUT_MPLANE:
>>>> + mplane_fmt = &ctx->src_fmt;
>>>> + fmt = cedrus_find_format(mplane_fmt->pixelformat,
>>>> + CEDRUS_DECODE_SRC,
>>>> + dev->capabilities);
>>>> + break;
>>>> +
>>>> + case V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE:
>>>> + mplane_fmt = &ctx->dst_fmt;
>>>> + fmt = cedrus_find_format(mplane_fmt->pixelformat,
>>>> + CEDRUS_DECODE_DST,
>>>> + dev->capabilities);
>>>> + break;
>>>> +
>>>> + default:
>>>> + return -EINVAL;
>>>> + }
>>>> +
>>>> + if (!fmt)
>>>> + return -EINVAL;
>>>> +
>>>> + if (fmt->num_buffers == 1) {
>>>> + sizes[0] = 0;
>>>> +
>>>> + for (i = 0; i < fmt->num_planes; i++)
>>>> + sizes[0] += mplane_fmt->plane_fmt[i].sizeimage;
>>>> + } else if (fmt->num_buffers == fmt->num_planes) {
>>>> + for (i = 0; i < fmt->num_planes; i++)
>>>> + sizes[i] = mplane_fmt->plane_fmt[i].sizeimage;
>>>> + } else {
>>>> + return -EINVAL;
>>>> + }
>>>> +
>>>> + *nplanes = fmt->num_buffers;
>>>
>>> This code does not take VIDIOC_CREATE_BUFFERS into account.
>>>
>>> If it is called from that ioctl, then *nplanes is non-zero and you need
>>> to check if *nplanes equals fmt->num_buffers and that sizes[n] is >=
>>> the required size of the format. If so, then return 0, otherwise return
>>> -EINVAL.
>>
>> Thanks for spotting this, I'll fix it as you suggested in the next
>> revision.
>>
>>> Doesn't v4l2-compliance fail on that? Or is that test skipped because this
>>> is a decoder for which streaming is not supported (yet)?
>>
>> Apparently, v4l2-compliance doesn't fail since I'm getting:
>> test VIDIOC_REQBUFS/CREATE_BUFS/QUERYBUF: OK
>
> It is tested, but only with the -s option. I'll see if I can improve the
> tests.

I've improved the tests. v4l2-compliance should now fail when run (without the
-s option) against this driver. Can you check that that is indeed the case?

Thanks!

Hans

\
 
 \ /
  Last update: 2018-09-06 09:23    [W:0.060 / U:0.184 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site