linux-kernel - Re: [4/5] i2c: davinci: use bus recovery infrastructure

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:	Mon, 24 Nov 2014 21:07:54 +0100
From:	Uwe Kleine-König 
	<u.kleine-koenig@...gutronix.de>
To:	Grygorii Strashko <grygorii.strashko@...com>
Cc:	Wolfram Sang <wsa@...-dreams.de>, linux-i2c@...r.kernel.org,
	linux-kernel@...r.kernel.org, Sekhar Nori <nsekhar@...com>,
	Kevin Hilman <khilman@...prootsystems.com>,
	Santosh Shilimkar <ssantosh@...nel.org>,
	Murali Karicheri <m-karicheri2@...com>
Subject: Re: [4/5] i2c: davinci: use bus recovery infrastructure

Hi Grygorii,

On Mon, Nov 24, 2014 at 03:26:10PM +0200, Grygorii Strashko wrote:
> On 11/23/2014 10:36 PM, Uwe Kleine-König wrote:
> > On Fri, Nov 21, 2014 at 09:33:22PM +0200, Grygorii Strashko wrote:
> >> On 11/21/2014 09:07 PM, Uwe Kleine-König wrote:
> >>> On Thu, Nov 20, 2014 at 12:03:07PM +0200, Grygorii Strashko wrote:
> >>> Just another general comment about the driver that doesn't influence the
> >>> correctness of this patch: The i2c-davinci driver is quite quick to
> >>> reset the bus. I wonder how often this reset triggers. Is the bus in
> >>> question less "stable" than others?
> >>
> >> In comparison to ..? :)
> > In comparison to other bus drivers in other SoCs. I know this might be
> > hard to answer. I just wonder where the reason for this has to be
> > located. Strange hardware? Software bug? Or is this SoC just operating
> > with strange slaves more often than others?
> 
> Davinci driver does reset in two cases:
> - when I2C transaction isn't completed due to timeout (no irq received)
> - when BB is detected
> both cases are reasonable, because in 1st case HW state is undefined
> in 2d case - Davinci I2C supports only master mode and if BB detected
> we need perform some recovery procedure.
note that for multi-master scenarios bus recovery might be bad. So that
has not necessarily something to do with "master mode only".

> Also, this patch doesn't introduce functional changes - it's just code
> reworking intended to reuse I2C bus recovery infrastructure
Yeah this is fine, and my concern shouldn't stop your patch from getting
in as it is an improvement for sure. Still I only had to handle BB once
in my linux lifetime, and that was for a different reason. (Machine
doing massive i2c transfers and so a client was often holding the bus
busy when the watchdog resetted the cpu.) So I just wondered how often
the need for recovery arises.

Best regards
Uwe

-- 
Pengutronix e.K.                           | Uwe Kleine-König            |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/