[PATCH] iso pipe support for usbatm
rkagan at mail.ru
Fri Mar 25 06:35:02 EST 2005
On Fri, Mar 25, 2005 at 12:23:06PM +0100, castet.matthieu at free.fr wrote:
> Selon Roman Kagan <rkagan at mail.ru>:
> > What's the point to delay it then? I would think that the test for the
> > urb status could be used to discard the data in this urb (by setting
> > ->filled_cells to zero). Delaying it until the next urb arrives doesn't
> > seem to make any sense.
> This is what I was trying to said...
> So we finally agree that we should always call tasklet_schedule and use the
Sure, that's what my patch did, don't remember myself arguing against it
> urb->status to force the filled_cells to 0 in case of errors ?
Ups, forgot about this guy. Will resend the patch in a minute.
More information about the Usbatm