linux-kernel - Re: [PATCH 8/9] bq27x00: Add miscdevice for each battery with ioctl for reading registers

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:	Fri, 25 Nov 2011 23:48:39 +0100
From:	Pali Rohár <pali.rohar@...il.com>
To:	Anton Vorontsov <cbouatmailru@...il.com>
Cc:	linux-kernel@...r.kernel.org, Lars-Peter Clausen <lars@...afoo.de>,
	syed rafiuddin <rafiuddin.sayed@...il.com>,
	Rodolfo Giometti <giometti@...ux.it>,
	David Woodhouse <dwmw2@...radead.org>
Subject: Re: [PATCH 8/9] bq27x00: Add miscdevice for each battery with ioctl for reading registers

On Saturday 26 November 2011 02:04:48 Anton Vorontsov wrote:
> On Fri, Nov 25, 2011 at 09:54:04PM +0100, Pali Rohár wrote:
> > On Saturday 26 November 2011 00:46:26 Anton Vorontsov wrote:
> > > On Fri, Nov 25, 2011 at 09:30:28PM +0100, Pali Rohár wrote:
> > > [...]
> > > 
> > > > This interface is not only for BME. Also some popular bq27200.sh
> > > > script which print bq registers in human readable form needs
> > > > this interface (with LD_PRELOAD library).
> > > > 
> > > > Link for that shell script http://enivax.net/jk/n900/bq.tar
> > > 
> > > That might be a good excuse to have the raw interface. Although,
> > > as this is for debugging purposes only, the same effect can be
> > > accomplished by unloading bq module and using i2c userspace
> > > interface directly... I guess.
> > 
> > Yes, unloading bq module and then starting script working. But I think
> > that we could have some interface how to access directly to i2c when
> > some i2c module for chip is loaded.
> 
> This would be not safe as this might (in case of RW registers)
> break kernel's driver behaviour (well, in bq case you only allow
> reading, so not problem in this particular case).
> 
> What would be more practical, is to allow I2C core to provide
> userspace interface even for already bound I2C devices.
> 
> That could be some kind of CONFIG_I2C_UNSAFE_DEBUG: when
> selected I2C core would allow access to all I2C devices. But still,
> the niche for such a feature is tiny, so I doubt that it is worth
> doing at all.
> 
> In any case, I just think that being able to access already bound
> I2C devices from userspace might be a good thing for debugging,
> but having such an interface per-driver is impractical.
> 
> Thanks,

Yes, you are right that per-driver access is bad. So I will use this ioctl 
interface patch only in maemo specified kernels (eg. maemo kernel-power).

-- 
Pali Rohár
pali.rohar@...il.com
Download attachment "signature.asc" of type "application/pgp-signature" (199 bytes)