On Tue, Dec 28, 2010 at 07:14:19PM +0200, Felipe Balbi wrote:

quoted text
> +static void twl4030_sih_bus_sync_unlock(unsigned int irq)
> +{
> +       struct sih_agent        *agent = get_irq_chip_data(irq);
> +
> +       mutex_unlock(&agent->irq_lock);
> +}

I suspect you need to do some sort of sync with the hardware here - the
_sync bit of the name comes from the fact that the mask and unmask stuff
is still called with IRQs disabled and so can't touch and I2C chip, this
is called after reenabling them give a chance for the updates done to
be reflected in the hardware.  The implementation everyone else has done
is to update a register cache in the other functions then write that
out here before dropping the mutex.

quoted text
>  static struct irq_chip twl4030_sih_irq_chip = {
>  	.name		= "twl4030",
>  	.mask		= twl4030_sih_mask,
>  	.unmask		= twl4030_sih_unmask,
>  	.set_type	= twl4030_sih_set_type,
> +	.bus_lock	= twl4030_sih_bus_lock,
> +	.bus_sync_unlock = twl4030_sih_bus_sync_unlock,
>  };

I just realised that this collides with the conversion to the irq_
versions that has been done on the driver in -next by either myself or
Lennart (we both submitted essentially the same patches and a couple of
his went in) - that was a purely mechanical conversion that didn't
address any of the issues this patch addresses but they're touching the
same code.
--
unsubscribe notice
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/