On Wed, 15 Apr 2009, David Brownell wrote: > On Wednesday 15 April 2009, Dan Streetman wrote: > > Yep, you are right on that point - so to adhere to spec, the TT > > scheduler really does need to rebalance the TT schedule to put all the > > interrupt transfers after isoc transfers. > > I'd not call that "rebalancing". Rebalancing is when the > arms of the schedule tree shift around. Section 4.12.2.5 of the EHCI spec is entitled "Rebalancing the Periodic Schedule", and the first thing it says is that system software must occasionally adjust a periodic queue head's S-mask and C-mask fields during operation. If you don't consider that to be rebalancing, what do you call it? > In this case, all that's required is to make sure insertions > follow the "ISO first" rule. Which they should already be > doing ... if they aren't, things will break. The problem is that sometimes there isn't room to put an isoc transfer first without moving an interrupt transfer out of the way. Otherwise the submission has to fail, even though there might be adequate bandwidth remaining. Nothing really wrong with this, but it is sub-optimal. Alan Stern -- 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
