On Wed, Jun 01, 2011 at 09:14:05AM -0700, Dmitry Torokhov wrote: > On Tuesday, May 31, 2011 10:07:13 PM Matt Evans wrote: > > On 01/06/11 07:37, Dmitry Torokhov wrote: > > > No, what we were missingis that units of measurement for FS isoc > > > endpoints is frames, not microframes. I think the following patch is > > > what we need. > > > > Aha, great! I missed that unit-change subtlety in my patch. Though my > > test machine's unavailable ATM, the code will give the right result (3) > > for the bInterval (1) in my USB audio isoch endpoints. Happy for my > > patch to be kicked down the well in favour of this. :) > > > > > Thanks, > > > Dmitry > > > > > > > > > Subject: USB: xhci - fix interval calculation for FS isoc endpoints > > > > > > Full-speed isoc endpoints specify interval in exponent based form > > > in frames, not microframes, so we need to adjust accordingly. > > > > > > Reported-by: Matt Evans <matt@xxxxxxxxxx> > > > Signed-off-by: Dmitry Torokhov <dtor@xxxxxxxxxx> > > > > Acked-by: Matt Evans <matt@xxxxxxxxxx> > > Thanks Matt. > > > > > Should also CC stable to follow your original patch, no? > > > > I'll defer to Sarah on that - if she's happy with this version of the > patch then yes, I believe we should mark it for stable as well. Yep, the patch looks fine, and should fix Matt's problem with the host rejecting intervals less than 3. Go ahead and mark it for stable kernels back to 2.6.31. I'm still interested in whether the host would actually accept FS isochronous endpoints with an bInterval greater than 7, leading to a endpoint context Interval field > 10. I guess we'll cross that bridge when we come to it. I don't think many users will have full speed devices that need to be polled so infrequently. Sarah Sharp -- 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
