On Fri, Oct 11, 2002, David Brownell <[EMAIL PROTECTED]> wrote: > > >>All the other HCs apply a certain degree of round-robin logic to > >>promote fairness in silicon. UHCI has to do that in software. > >>I'd leave this part in. > > > > OHCI only does one TD per frame per transfer? > > No, but it only tries one transfer before going to the > next endpoint. Normally that won't mean an entire TD, > since (for example) a bulk TD might describe 4KBytes > (more than one transfer).
Sorry, I should have said packet. I have UHCI on the brain :) > OHCI has a control/bulk service ratio, and the HC will try > N control endpoints (from 1 to 4) per bulk endpoint. > > EHCI is similar, but without such a ratio: one transfer, > then move to the next QH. But EHCI 1.0 controllers can > support a "park" mode where up to 3 TDs are processed before > they look at the next QH. Presumably that's useful. > > In both cases the drivers try to keep the queues as > short as possible, so it's quite possible for the HC > to run through one endpoint's queue without looking at > another endpoint. So it does a maximum of one packet per endpoint per frame for low speed devices? Or does it only do one packet and then move on, but if there's still time left in the frame, there's the chance of doing another one? Does it differentiate between high speed and low speed control? JE ------------------------------------------------------- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf _______________________________________________ [EMAIL PROTECTED] To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
