linux-kernel - Re: [PATCH v8 08/10] gpu: host1x: mipi: Keep MIPI clock enabled till calibration is done

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]

Date:   Thu, 6 Aug 2020 11:44:04 -0700
From:   Sowjanya Komatineni <skomatineni@...dia.com>
To:     Dmitry Osipenko <digetx@...il.com>,
        Thierry Reding <thierry.reding@...il.com>
CC:     <jonathanh@...dia.com>, <frankc@...dia.com>, <hverkuil@...all.nl>,
        <sakari.ailus@....fi>, <robh+dt@...nel.org>,
        <helen.koike@...labora.com>, <gregkh@...uxfoundation.org>,
        <linux-media@...r.kernel.org>, <devicetree@...r.kernel.org>,
        <linux-tegra@...r.kernel.org>, <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH v8 08/10] gpu: host1x: mipi: Keep MIPI clock enabled till
 calibration is done


On 8/6/20 11:18 AM, Dmitry Osipenko wrote:
> 06.08.2020 21:07, Sowjanya Komatineni пишет:
>> On 8/6/20 11:01 AM, Dmitry Osipenko wrote:
>>> 06.08.2020 20:52, Sowjanya Komatineni пишет:
>>> ...
>>>> Right mutex_unlock should happen at end of finish_calibration.
>>>>
>>>> With keeping mutex locked in start, we dont have to check for active to
>>>> be 0 to issue start as mutex will keep it locked and other pads
>>>> calibration can only go thru when current one is done.
>>>>
>>>> So instead of below sequence, its simpler to do this way?
>>>>
>>>> start_calibration()
>>>>
>>>> - mutex_lock
>>>>
>>>> - wait for 72uS after start
>>>>
>>>> finish_calibration()
>>>>
>>>> - keep check for ACTIVE = 0 and DONE = 1
>>> I think only the DONE bits which correspond to the mipi_device->pads
>>> bitmask should be awaited.
>> As next START can't be triggered when auto cal is ACTIVE, we should keep
>> this in finish.
>>
>> As we do mutex_unlock only at end of finish, other pads calibrations
>> dont go thru till the one in process is finished.
>>
>> So in this case ACTIVE applies to current selected pads that are under
>> calibration.
> Should be better to check only the relevant bits in order to catch bugs,
> otherwise you may get a DONE status from the irrelevant pads.
tegra_mipi_device is separate for DSI and CSI channels. mutex lock used 
during calibrate is device specific lock.
So, it will not prevent other devices to hold till on going calibration 
is done unless we add wait for active bit before triggering start.

Currently we wait for active bit at end during calibration done check 
after start trigger. But when other devices go thru calibration in 
parallel as lock is device specific and not common lock for all devices 
it will trigger start but MIPI calibration logic ignore if previous 
calibration is still in progress.

Need to serialize calibration start requests from different devices 
based on ACTIVE bit.

>>> Perhaps the start_calibration() also needs to be changed to not touch
>>> the MIPI_CAL_CONFIG bits of the unrelated pads?
>> Driver already takes care of programming corresponding pads config only.
> It writes 0 to the config of the unrelated pads, which probably isn't
> nice if some pads use periodic auto-calibration.
>
> https://elixir.bootlin.com/linux/v5.8/source/drivers/gpu/host1x/mipi.c#L350
>
> Although looks like auto-calibration isn't supported by the current driver.

Yes we don't use auto-calibration.

Only common bit here is MIPI_CAL_CTRL start. All others are pad specific 
currently.