linux-kernel - Re: [PATCH 0/2] i2c-dev: Don't let userspace block adapter

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: <20160706153449.GL2671@ubuntu>
Date:	Wed, 6 Jul 2016 08:34:49 -0700
From:	Viresh Kumar <viresh.kumar@...aro.org>
To:	Lars-Peter Clausen <lars@...afoo.de>
Cc:	Wolfram Sang <wsa@...-dreams.de>, Jean Delvare <jdelvare@...e.com>,
	linux-i2c@...r.kernel.org, linux-kernel@...r.kernel.org,
	gregkh@...uxfoundation.org, Johan Hovold <johan@...nel.org>,
	Alex Elder <elder@...aro.org>
Subject: Re: [PATCH 0/2] i2c-dev: Don't let userspace block adapter

On 06-07-16, 16:41, Lars-Peter Clausen wrote:
> On 07/06/2016 04:57 AM, Viresh Kumar wrote:
> > Hi Wolfram/Jean,
> > 
> > I am part of the kernel team for Google's projectara [1], where we are
> > building a module smart phone.
> > 
> > This series tries to fix one of the problems we hit on our system as we
> > are required to hotplug pretty much every thing on the phone and so this
> > fixes hotplug issues with i2c-dev.
> > 
> > As described in the second patch, the current implementation of i2c-dev
> > file operations doesn't let the modules (hardware attached to the phone)
> > eject from the phone as the cleanup path for the module hasn't finished
> > yet (i2c adapter not removed).
> > 
> > We can't let the userspace block the kernel devices forever in such
> > cases.
> > 
> > I was able to test them on the ARA phone with kernel 3.10 only and not
> > mainline.
> 
> This sounds like you want hot-unplug. This is currently not support by the
> I2C framework for adapters. A better approach compared to this series might
> be to implement full hot-unplug support for I2C adapters. This will probably
> also be useful for additional usecases.

Yeah, we need hot-unplug.

Hmm, doing that would require more knowledge of the framework and I am
afraid I don't have it right now, not that it can't be done :)

-- 
viresh