linux-kernel - Re: [PATCH v3 4/9] media: venus: vdec: add video decoder files

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <86442d1d-4a12-71c1-97fa-12bc73bb5045@linaro.org>
Date:   Fri, 18 Nov 2016 11:11:34 +0200
From:   Stanimir Varbanov <stanimir.varbanov@...aro.org>
To:     Stanimir Varbanov <stanimir.varbanov@...aro.org>,
        Hans Verkuil <hverkuil@...all.nl>,
        Mauro Carvalho Chehab <mchehab@...nel.org>
Cc:     Andy Gross <andy.gross@...aro.org>,
        Bjorn Andersson <bjorn.andersson@...aro.org>,
        Stephen Boyd <sboyd@...eaurora.org>,
        Srinivas Kandagatla <srinivas.kandagatla@...aro.org>,
        linux-media@...r.kernel.org, linux-kernel@...r.kernel.org,
        linux-arm-msm@...r.kernel.org
Subject: Re: [PATCH v3 4/9] media: venus: vdec: add video decoder files

Hi Hans,

>>> +
>>> +static int
>>> +vdec_reqbufs(struct file *file, void *fh, struct v4l2_requestbuffers *b)
>>> +{
>>> +	struct vb2_queue *queue = to_vb2q(file, b->type);
>>> +
>>> +	if (!queue)
>>> +		return -EINVAL;
>>> +
>>> +	return vb2_reqbufs(queue, b);
>>> +}
>>
>> Is there any reason why the v4l2_m2m_ioctl_reqbufs et al helper functions
>> can't be used? I strongly recommend that, unless there is a specific reason
>> why that won't work.
> 
> So that means I need to completely rewrite the v4l2 part and adopt it
> for mem2mem device APIs.
> 
> If that is what you meant I can invest some time to make a estimation
> what would be the changes and time needed. After that we can decide what
> to do - take the driver as is and port it to mem2mem device APIs later
> on or wait for the this transition to happen before merging.
> 

I made an attempt to adopt v4l2 part of the venus driver to m2m API's
and the result was ~300 less lines of code, but with the price of few
extensions in m2m APIs (and I still have issues with running
simultaneously multiple instances).

I have to add few functions/macros to iterate over a list (list_for_each
and friends). This is used to find the returned from decoder buffers by
address and associate them to vb2_buffer, because the decoder can change
the order of the output buffers.

The main problem I have is registering of the capture buffers before
session_start. This is requirement (disadvantage) of the firmware
implementation i.e. I need to announce capture buffers (address and size
of the buffer) to the firmware before start buffer interaction by
session_start.

So having that I think I will need one more week to stabilize the driver
to the state that it was before this m2m transition.

Thoughts?

-- 
regards,
Stan