linux-kernel - Re: [RFC PATCH v9 6/9] media: tegra: Add Tegra210 Video input driver

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: <8f01724e-95d4-e6d8-9e2d-bce4b8dd0177@nvidia.com>
Date:   Thu, 23 Apr 2020 18:08:27 -0700
From:   Sowjanya Komatineni <skomatineni@...dia.com>
To:     Dmitry Osipenko <digetx@...il.com>, <thierry.reding@...il.com>,
        <jonathanh@...dia.com>, <frankc@...dia.com>, <hverkuil@...all.nl>,
        <sakari.ailus@....fi>, <helen.koike@...labora.com>
CC:     <sboyd@...nel.org>, <linux-media@...r.kernel.org>,
        <devicetree@...r.kernel.org>, <linux-clk@...r.kernel.org>,
        <linux-tegra@...r.kernel.org>, <linux-kernel@...r.kernel.org>
Subject: Re: [RFC PATCH v9 6/9] media: tegra: Add Tegra210 Video input driver


On 4/23/20 5:51 PM, Sowjanya Komatineni wrote:
>
> On 4/23/20 5:42 PM, Dmitry Osipenko wrote:
>> External email: Use caution opening links or attachments
>>
>>
>> 24.04.2020 02:50, Sowjanya Komatineni пишет:
>>> On 4/23/20 4:25 PM, Dmitry Osipenko wrote:
>>>> External email: Use caution opening links or attachments
>>>>
>>>>
>>>> 24.04.2020 02:20, Sowjanya Komatineni пишет:
>>>>> On 4/23/20 4:19 PM, Sowjanya Komatineni wrote:
>>>>>> On 4/23/20 4:16 PM, Dmitry Osipenko wrote:
>>>>>>> External email: Use caution opening links or attachments
>>>>>>>
>>>>>>>
>>>>>>> 22.04.2020 09:18, Sowjanya Komatineni пишет:
>>>>>>>> +static int chan_capture_kthread_start(void *data)
>>>>>>>> +{
>>>>>>>> +     struct tegra_vi_channel *chan = data;
>>>>>>>> +     struct tegra_channel_buffer *buf;
>>>>>>>> +     int err = 0;
>>>>>>>> +
>>>>>>>> +     set_freezable();
>>>>>>>> +
>>>>>>>> +     while (1) {
>>>>>>>> +             try_to_freeze();
>>>>>>>> +
>>>>>>>> + wait_event_interruptible(chan->start_wait,
>>>>>>>> + !list_empty(&chan->capture) ||
>>>>>>>> + kthread_should_stop());
>>>>>>>> +
>>>>>>>> +             if (kthread_should_stop())
>>>>>>>> +                     break;
>>>>>>>> +
>>>>>>>> +             /*
>>>>>>>> +              * Source is not streaming if error is non-zero.
>>>>>>>> +              * So, do not dequeue buffers on capture error.
>>>>>>>> +              */
>>>>>>>> +             if (err)
>>>>>>>> +                     continue;
>>>>>>> This will result in an endless loop, I suppose it wasn't the
>>>>>>> intention.
>>>>>> no it will not. on error we report vb2_queue_error which will do
>>>>>> streaming stop request.
>>>>>>
>>>>>> So thread will be stopped on streaming stop request thru kthread 
>>>>>> stop
>>>>>> signal
>>>>> To be clear on error it reports vb2 queue error and waits for stop
>>>>> streaming to happen
>>>> If thread should exit on error, then it should do it on the actual
>>>> error. Otherwise it looks very error-prone.
>>> When v4l2 drivers indicate fatal error through vb2_queue_error, queue
>>> error flag  is set and wakes up all processes waiting on queue along
>>> with polling reporting  EPOLLERR and also reporting error for queuing
>>> and dequeuing buffers. Stream stop will surely happen which stops the
>>> thread.
>> This doesn't explain what is the point of continuing to loop instead of
>> exiting immediately on error.
>
> We are using 2 threads and when capture start error happens, we can 
> stop capture_start thread immediately but capture_finish thread will 
> still run for any outstanding buffers.
>
> So, as it makes no diff stopping both threads during stream stop which 
> will definitely happen on error and when we don't dequeue buffers
>
Also there will be an issue if we break on error immediately during 
stop_streaming -> kthread_stop()

As stop streaming can happen any time, we do kthread_stop and in case of 
error if we stop thread and on stop streaming kthread_stop might crash 
as kthread_stop can only be called on running thread