linux-kernel - Re: [PATCH 2/4] i2c: stm32f7: recover the bus on access timeout

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 for Android: free password hash cracker in your pocket

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

Message-ID: <20211130085618.GA668426@gnbcxd0016.gnb.st.com>
Date:   Tue, 30 Nov 2021 09:56:18 +0100
From:   Alain Volmat <alain.volmat@...s.st.com>
To:     Wolfram Sang <wsa@...nel.org>, <pierre-yves.mordret@...s.st.com>,
        <alexandre.torgue@...s.st.com>, <linux-i2c@...r.kernel.org>,
        <linux-stm32@...md-mailman.stormreply.com>,
        <linux-arm-kernel@...ts.infradead.org>,
        <linux-kernel@...r.kernel.org>, <fabrice.gasnier@...s.st.com>,
        <amelie.delaunay@...s.st.com>
Subject: Re: [PATCH 2/4] i2c: stm32f7: recover the bus on access timeout

Hi Alain,

On Mon, Nov 29, 2021 at 01:52:30PM +0100, Wolfram Sang wrote:
> Hi Alain,
> 
> > > > +		stm32f7_i2c_wait_free_bus(i2c_dev);
> > > 
> > > This does only a controller reset, not a bus recovery with 9 toggling
> > > pulses, or?
> > 
> > indeed. I might better rework this and at the same time introduce the
> > bus recovery mechanism via the bus recovery callback in this driver.
> > Please don't merge this patch and I will rework that.
> 
> Wait a sec. Resetting a controller at the end of a failed transfer might
> make sense if the controller is otherwise in an confused state.
> 
> Full bus recovery (9 pulses) should be done at the beginning of a
> transfer when SDA is low, though.
> 
> So, I'd actually suggest to apply this patch and add full bus recovery
> based on SDA low at the beginning of a transfer seperately.
> 
> What doo you think?

I just checked again.  Indeed, this patch is here to handle cases when
communication went bad with a device leading to controller being left in
a confused state.  This is done to put it back in a working state.

I agree with you on the fact to decouple this with the 9 pulses bus
recovery and first apply this one first.

Thanks.
Alain

> 
> All the best,
> 
>    Wolfram
>