On Fri, Feb 1, 2013 at 1:57 PM, Jean Delvare <khali@xxxxxxxxxxxx> wrote:
On Fri, 1 Feb 2013 13:36:40 -0800, Guenter Roeck wrote:
> On Fri, Feb 01, 2013 at 01:18:22PM -0800, Vincent Palatin wrote:
> > When the kernel has device tree support, we avoid doing the auto-detection> (...)
> > as probing the busses might mess-up sensitive I2C devices or trigger long
> > timeouts on non-functional busses.
> As for the detect function, I am not entirely sure if we want or should getI can't think of any reason for doing that. Detection only happens on
> rid of it just because OF is configured. That may have unintended side effects.
> Either case, I would argue that if we do want to do that, it should be done
> for _all_ I2C devices and not just for the lm90 driver. In other words, it
> should be done in the I2C subsystem, maybe with a separate configuration option.
I2C buses which request it (by setting i2c_adapter.class.) I2C system
bus drivers for OF-based platforms would simply not set any class flag,
so detection won't be triggered on them.
Thanks for the advice.
I will try update the i2c adapter driver ( i2c-s3c2410.c in my case) to avoid hardcoding the class flags and maybe add some configuration depending on a device tree node property.
--
Vincent
_______________________________________________ lm-sensors mailing list lm-sensors@xxxxxxxxxxxxxx http://lists.lm-sensors.org/mailman/listinfo/lm-sensors
