On Wed, Oct 10, 2001, Jean Tourrilhes <[EMAIL PROTECTED]> wrote:
> Now that uhci support ZERO_PACKET (thanks JE), I'm going to
> try to make irda-usb working on it.
> One of the problem I have is that uhci does the timeout
> differently than all the other drivers (usb-uhci and usb-ohci), and
> the two way are not compatible (usb-uhci and usb-ohci use relative
> timeout, uhci use absolute timeout). Therefore, I need from the driver
> to detect which HCI is running. Depending on which HCI is detected, I
> would format the timeout appropriately before submitting the URB.
> I can't really use use a #define, because I have both a UHCI
> and OHCI card in my PC and a dongle behing each card, so I need to
> have one driver instance in each mode.
>
> Of course, there is another solution for the long term, but
> it's going to degenerate in long discussions...
>
> P.S.: By the way, when I run irda-usb simultaneously on top of
> usb-uhci and usb-ohci, I've got nice oops. I'll try 2.4.11 later...
The real fix is not to workaround specific bugs, but to make all of the
HCI's work the same way.
Does this patch fix the problem for you?
JE
--- linux-2.4.11/drivers/usb/uhci.c.orig Wed Oct 10 18:23:27 2001
+++ linux-2.4.11/drivers/usb/uhci.c Wed Oct 10 18:24:23 2001
@@ -1955,7 +1955,7 @@
uhci_fsbr_timeout(uhci, u);
/* Check if the URB timed out */
- if (u->timeout && time_after_eq(jiffies, u->timeout)) {
+ if (u->timeout && time_after_eq(jiffies, urbp->inserttime +
+u->timeout)) {
list_del(&u->urb_list);
list_add_tail(&u->urb_list, &list);
}
_______________________________________________
[EMAIL PROTECTED]
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
