On Wed, 30 Dec 2009, Huang, FrankR wrote:
> Alan,
>
> The mouse is still hang with this patch.
> The done_head is also 0 after the udelay(10).
> I suspect when the done_head is 0 , this interrupt is not
> belonging the ohci controller.
I guess you are right. The controller appears to set the WDH bit
_before_ setting ohci->hcca->done_head, which is a violation of section
6.5.2 in the OHCI specification.
> Based on my modification, I omit this
> interrupt when done_head is 0. Is that ok?
The idea is okay, but your modification can be improved. Does this
patch solve the problem?
Alan Stern
Index: usb-2.6/drivers/usb/host/ohci-hcd.c
===================================================================
--- usb-2.6.orig/drivers/usb/host/ohci-hcd.c
+++ usb-2.6/drivers/usb/host/ohci-hcd.c
@@ -831,10 +831,19 @@ static irqreturn_t ohci_irq (struct usb_
usb_hcd_resume_root_hub(hcd);
}
+ /* Some controllers violate the OHCI spec by setting OHCI_INTR_WDH
+ * before writing ohci->hcca->done_head. They don't generate an
+ * early interrupt, but if the IRQ line is shared then we might see
+ * the status bit prematurely.
+ */
if (ints & OHCI_INTR_WDH) {
- spin_lock (&ohci->lock);
- dl_done_list (ohci);
- spin_unlock (&ohci->lock);
+ if (unlikely((ohci->hcca->done_head == 0)) {
+ ints &= ~OHCI_INTR_WDH;
+ } else {
+ spin_lock(&ohci->lock);
+ dl_done_list(ohci);
+ spin_unlock(&ohci->lock);
+ }
}
if (quirk_zfmicro(ohci) && (ints & OHCI_INTR_SF)) {
--
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
