Linux Media Controller development
 help / color / mirror / Atom feed
From: Daniel Scheller <d.scheller.oss@gmail.com>
To: Ralph Metzler <rjkm@metzlerbros.de>
Cc: linux-media@vger.kernel.org, mchehab@kernel.org,
	mchehab@s-opensource.com, jasmin@anw.at
Subject: Re: [PATCH] [media] dvb-frontends/stv0910: prevent consecutive mutex_unlock()'s
Date: Sat, 21 Oct 2017 11:57:57 +0200	[thread overview]
Message-ID: <20171021115757.729e000f@audiostation.wuest.de> (raw)
In-Reply-To: <23019.4906.236885.50919@morden.metzler>

Am Sat, 21 Oct 2017 11:28:10 +0200
schrieb Ralph Metzler <rjkm@metzlerbros.de>:

> Daniel Scheller writes:
>  > From: Daniel Scheller <d.scheller@gmx.net>
>  > 
>  > When calling gate_ctrl() with enable=0 if previously the mutex
>  > wasn't locked (ie. on enable=1 failure and subdrivers not handling
>  > this properly, or by otherwise badly behaving drivers), the
>  > i2c_lock could be unlocked  
> 
> I think drivers and subdrivers should rather be fixed so that this
> cannot happen.

As long as stv6111 remains the only chip/driver interfacing with the
stv0910, that's an easy task. However, if other hardware has some other
stv0910+tunerchip combination, things get interesting. In a perfect
world with unicorns and such, every component interacts as intended,
but that's not the case, so I believe this should be handled at the
root.

> But to do this we will first need to define exactly how a failure in
> gate_ctrl() is supposed to be handled, both inside gate_ctrl() and
> by calling drivers.

Well, IMHO (and thats the intention) if gate_ctrl fails due to a
hardware/I2C problem, it isn't opened so there's no need to hold the
lock (since the gate isn't - exclusively - opened). For reasons stated
above this keeps things safe from deadlocking (and we want to avoid
that, even more than double unlocking).

>  > consecutively which isn't allowed. Prevent this by keeping track
>  > of the lock state, and actually call mutex_unlock() only when
>  > certain the lock is held.  
> 
> Why not use mutex_is_locked()?

Good catch (I should try harder finding out what the kernel API has to
offer...). If you prefer that, I'll respin with this and without the
var as v2.

> And there should be a debug message if it (tried double unlocking)
> happens.

Ok. Should IMHO go to dev_dbg then - if drivers don't catch that
situation, this may else lead do kernel log spam.

Best regards,
Daniel Scheller
-- 
https://github.com/herrnst

  reply	other threads:[~2017-10-21  9:58 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-10-21  8:36 [PATCH] [media] dvb-frontends/stv0910: prevent consecutive mutex_unlock()'s Daniel Scheller
2017-10-21  9:28 ` Ralph Metzler
2017-10-21  9:57   ` Daniel Scheller [this message]
2017-11-07  9:50     ` Mauro Carvalho Chehab

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20171021115757.729e000f@audiostation.wuest.de \
    --to=d.scheller.oss@gmail.com \
    --cc=jasmin@anw.at \
    --cc=linux-media@vger.kernel.org \
    --cc=mchehab@kernel.org \
    --cc=mchehab@s-opensource.com \
    --cc=rjkm@metzlerbros.de \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox