On Mon, May 21, 2001 at 04:52:58PM -0400, Johannes Erdfelt wrote:
> On Mon, May 21, 2001, David Brownell <[EMAIL PROTECTED]> wrote:
> > > That is, I want to move several URBs for one
> > > endpoint, and _when_they_are_done_, I want to move an URB for another
> > > endpoint. And, it doesn't matter how the beginning URBs end, I still want
> > > to move the last one.
> >
> > So far as I know, the only way to make that happen is to have the
> > completion function(s) for that first endpoint trigger the behavior
> > you want (by submitting urbs for the other endpoint).
>
> This could be done with ->next. Or you can do it in the completion
> handler.
Wait... different answer here.... can this be done with ->next? People
seem to have different opinions here. Some say ->next is for ISO only.
What I'd _hope_ for is this: I construct a chain of arbitrary URBs in a
linked list via the ->next pointer. I send the chain to the API, and now I
go to sleep while I wait for a completion message from the last URB.
Meanwhile, the URBs are processed by executing the head of the queue, and
(when that's finished), executing the next item on the queue.
Is that possible? Yes, I can do this via submitting URBs from the
completion function. But, then I have to do a lot of management in the
completion handler, which I don't want to do. I'd like to set everything
up ahead of time and let it run, so that the time between URBs is
minimized.
Yes, I realize that I could try to prep the next URB while one is running,
but then there are synchronization problems. So, if I'm preping them all
in advance anyway....
Matt
--
Matthew Dharm Home: [EMAIL PROTECTED]
Maintainer, Linux USB Mass Storage Driver
Oh BAY-bee.
-- Dust Puppy to Greg
User Friendly, 12/13/1997
PGP signature
