On Tue, Jul 12, 2011 at 10:35:53PM +0300, Péter Ujfalusi wrote:
Just to avoid confusion (in my part): if I split this patch to one big + several small incremental patches, which eventually lead to the current situation is what you were asking for?
The rest of the patch looked fine from a code point of view so probably just leave that as-is and improve the changelog - the delayed_work was the only real issue I saw from a code point of view and the bit I was suggesting to split out into a separate patch.
With the patches I can document the reasoning behind the non standard workarounds. Probably if we see them spread out, we might be able to find better ways for them...
Yeah, and like I say if it were controlled by the machine driver it'd be better.