Re: [PATCH v2 0/3] Improve error handling in the rcar-vin driver

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Hi Niklas, Hans,

On Tue, Jul 05, 2022 at 11:46:22AM +0200, Niklas Söderlund wrote:
> Hi Michael,
> 
> Thanks for your persistent work with this series.

Thank you for the feedback!

> On 2022-06-28 20:00:19 +0200, Michael Rodin wrote:
> > Hello,
> > 
> > this series is a followup to the other series [1] started by Niklas Söderlund
> > where only the first patch has been merged. The overall idea is to be more
> > compliant with the Renesas hardware manual which requires a reset or stop
> > of capture in the VIN module before reset of CSI2. Another goal is to be
> > more resilient with respect to non-critical CSI2 errors so the driver does
> > not end in an endless restart loop. Compared to the previous version [2] of
> > this series the patch 3 is replaced based on the conclusion in [3] so now
> > userspace has to take care of figuring out if a transfer error was harmless
> > or unrecoverable. Other patches are adapted accordingly so no assumptions
> > about criticality of transfer errors are made in the kernel and the
> > decision is left up to userspace.
> 
> I like this solution as it truly pushes the decision to user-space. What 
> bugs me a little bit is that we don't have a way to communicate errors 
> that we know are unrecoverable (it was for this case the work in this 
> area started) and ones that could be recoverable (the use-case added on 
> top).

Yes, it's not nice that V4L2_EVENT_XFER_ERROR does not tell userspace
whether an error is recoverable (i.e. the event can be ignored) or not
(i.e. a restart of streaming is required) but the other possible option
would be (as concluded in [3]) to implement a frame timeout monitoring
thread in v4l2 core. I am not sure if it is possible to implement this
second option cleanly...

> I would also like to hear what Hans thinks as he had good suggestions 
> for how to handle the cases we know can't be recovers in [4].

A a new function vb2_queue_error_with_event() suggested by Hans seems to be
redundant now, since it would not be used by rcar-vin (unless we implement
frame timeout monitoring in the v4l2 core). Or do you have an idea, which
drivers could be the first users of it, e.g. staging/media/imx I mentioned
before?

> > 
> > [1] https://urldefense.proofpoint.com/v2/url?u=https-3A__lore.kernel.org_linux-2Drenesas-2Dsoc_20211108160220.767586-2D1-2Dniklas.soderlund-2Brenesas-40ragnatech.se_&d=DwIDAw&c=euGZstcaTDllvimEN8b7jXrwqOf-v5A_CdpgnVfiiMM&r=sWsgk3pKkv5GeIDM2RZlPY8TjNFU2D0oBeOj6QNBadE&m=ecX7IwfatUO7SNPiyQ6x_8K9t2eWJf3y8GNuNHJ_0W0&s=Cli6jADEgMmCOLVoFekRRXzmty9WBXtoSF9utZJNMXY&e=
> > [2] https://urldefense.proofpoint.com/v2/url?u=https-3A__lore.kernel.org_all_1652983210-2D1194-2D1-2Dgit-2Dsend-2Demail-2Dmrodin-40de.adit-2Djv.com_&d=DwIDAw&c=euGZstcaTDllvimEN8b7jXrwqOf-v5A_CdpgnVfiiMM&r=sWsgk3pKkv5GeIDM2RZlPY8TjNFU2D0oBeOj6QNBadE&m=ecX7IwfatUO7SNPiyQ6x_8K9t2eWJf3y8GNuNHJ_0W0&s=6CysfSY0OoAenEwCzigeyPOb8vyaa4GgzkJSR-ny83U&e=
> > [3] https://urldefense.proofpoint.com/v2/url?u=https-3A__lore.kernel.org_all_YqEO3-252FKekkZhVjW-2B-40oden.dyn.berto.se_&d=DwIDAw&c=euGZstcaTDllvimEN8b7jXrwqOf-v5A_CdpgnVfiiMM&r=sWsgk3pKkv5GeIDM2RZlPY8TjNFU2D0oBeOj6QNBadE&m=ecX7IwfatUO7SNPiyQ6x_8K9t2eWJf3y8GNuNHJ_0W0&s=67JE_QR4x7omrtC7wzbpn2OgW75TAR80-R8WQyE-bVo&e=
> 
> 4. https://urldefense.proofpoint.com/v2/url?u=https-3A__lore.kernel.org_all_1fddc966-2D5a23-2D63b4-2D185e-2Dc17aa6d65b54-40xs4all.nl_&d=DwIDAw&c=euGZstcaTDllvimEN8b7jXrwqOf-v5A_CdpgnVfiiMM&r=sWsgk3pKkv5GeIDM2RZlPY8TjNFU2D0oBeOj6QNBadE&m=ecX7IwfatUO7SNPiyQ6x_8K9t2eWJf3y8GNuNHJ_0W0&s=I18yWgde2UKZY4AiwB5s-Lf12eebHOcHFZFOlTcO2oQ&e=
> 
> -- 
> Kind Regards,
> Niklas Söderlund

-- 
Best Regards,
Michael



[Index of Archives]     [Linux Samsung SOC]     [Linux Wireless]     [Linux Kernel]     [ATH6KL]     [Linux Bluetooth]     [Linux Netdev]     [Kernel Newbies]     [IDE]     [Security]     [Git]     [Netfilter]     [Bugtraq]     [Yosemite News]     [MIPS Linux]     [ARM Linux]     [Linux Security]     [Linux RAID]     [Linux ATA RAID]     [Samba]     [Device Mapper]

  Powered by Linux