On Thu, Jan 27, 2011 at 09:03:47AM +0530, Sebastian Andrzej Siewior wrote: > * Russell King - ARM Linux | 2011-01-26 08:44:37 [+0000]: > > >So what's happening with this patch? Are you going to submit it for -rc? > > I probably am but I will hide this workaround behind a module option. > The problem with this workaround is that it drastically increases the > number of interrupts genrated by the controller and this does not make > me happpy. > I come up with a way which will probably shorten the race window but it > does not close it. I found a nasty way how to find out which bits were > overwritten but it does not work for all patterns. > So I think that I add a printk to let the user know about this > workaround and if he needs, he can enable it via a module option. The choice seems to be between a USB interface which locks up and doesn't behave properly vs one which consumes lots of interrupts. Isn't it possible to only enable SOF interrupts only when we have URBs waiting, and turn them off while idle? If there's nothing queued which might appear on the 'done' list then there's no point having a stream of SOF interrupts. -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
