Re: [PATCH 1/2] i2c: add "reset" sysfs entry for adapters.

[Jean Delvare <khali-PUYAD+kWke1g9hUCZPvPmw@public.gmane.org>]

On Wed, 11 Feb 2009 14:00:29 +0100, Rodolfo Giometti wrote:
> On Tue, Feb 10, 2009 at 11:44:38AM +0100, Jean Delvare wrote:
> > On Mon, 9 Feb 2009 21:41:08 +0100, Rodolfo Giometti wrote:
> > > On Mon, Feb 09, 2009 at 05:17:20PM +0100, Jean Delvare wrote:
> > > > On Fri,  6 Feb 2009 16:23:17 +0100, Rodolfo Giometti wrote:
> > > > > It could happen that an i2c adapter may lock the bus due due
> > > > 
> > > > Typo: "due due" -> "due to".
> > > 
> > > Ok.
> > > 
> > > > > electrical problems, so the user may recover this stale state by using:
> > > > > 
> > > > > 	$ echo 1 > /sys/class/i2c-adapter/i2c-0/reset
> > > > 
> > > > What is the intended purpose, debugging? I would certainly hope that
> > > > drivers know by themselves when they need a reset. Requiring users to
> > > > reset themselves doesn't sound terribly friendly :(
> > > 
> > > Unluckely not all adapter does it by themselfes! At least not PXA
> > > one. I'm currently work on adding this new feature.
> > 
> > Well, if you are going to fix the i2c-pxa driver, and it's the only
> > driver using the manual reset feature, then do we really need to add
> > this feature?
> 
> I don't know about other controllers, I know about PXA one because I'm
> currently working on it. :)
> 
> > 
> > Don't get me wrong, if this feature is still useful then I am fine
> > adding it. But OTOH we still have to think twice when doing such core
> > changes, because they increase the memory footprint for all systems.
> 
> I agree. Maybe I can add such entry with a conditional #ifdef DEBUG
> and setting it off by default?

Yes, I would make it conditional upon CONFIG_I2C_DEBUG_BUS, and
possibly make it a private attribute of i2c-pxa if no other developers
are interested.

> > I am curious what other developers think. Me, I never felt the need for
> > a manual reset feature. Usually when I have managed to break an I2C
> > controller, cycling the driver would fix it, or if not, it was so
> > broken that I had to power-cycle the box anyway.
> > 
> > Are there other developers out there who consider the proposed manual
> > reset feature a good thing to have? Please speak up.
> 
> However, regarding PXA controller I found a way to automagically reset
> it when it locks. I'll send a patchset ASAP.

Well, then I'm not sure if we still care about this debugging patch...

-- 
Jean Delvare