Am Samstag, 28. August 2004 04:48 schrieb Alan Stern:
> On Fri, 27 Aug 2004, Oliver Neukum wrote:
>
> > > > > @@ -668,7 +668,7 @@
> > > > > INTBUFFERSIZE,
> > > > > int_callback,
> > > > > kaweth,
> > > > > - HZ/4);
> > > > > + 8);
> > > > > kaweth->irq_urb->transfer_dma = kaweth->intbufferhandle;
> > > > > kaweth->irq_urb->transfer_flags |= URB_NO_TRANSFER_DMA_MAP;
>
> > > That field can't handle such a big number, so I just guesed as to what
> > > you really ment it to be. Have a better suggestion?
> >
> > Yes. Revert it. Sparse here has overreached the limit of its usefullness.
> > The device is always full speed, never high speed. The shift won't be used.
> > The value is what the vendor recommends.
>
> I see this is part of a call to usb_fill_int_urb(). Since the device
> always runs at full speed, the value should be a time interval given in
> frames, i.e., milliseconds. If you intend the URB to fire every 250ms,
> then the value should be 250 -- not HZ/4, which can vary considerably
> among platforms. Alternatively, you could copy the value from the
> endpoint's descriptor.
You are right. It should be 250. I just found HZ/4 so much more expressive
that I didn't consider portability :-( The problem here is that the value
from the descriptors is incorrect and needs to be overridden. I'll just
make a patch.
Regards
Oliver
-------------------------------------------------------
This SF.Net email is sponsored by BEA Weblogic Workshop
FREE Java Enterprise J2EE developer tools!
Get your free copy of BEA WebLogic Workshop 8.1 today.
http://ads.osdn.com/?ad_id=5047&alloc_id=10808&op=click
_______________________________________________
[EMAIL PROTECTED]
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
