On Fri, 12 Jun 2009, Julie Zhu wrote:
> Hello, all,
>
> We found this on a PowerPC embedded platform using an FPGA-based USB
> host controller. The Linux kernel is 2.6.29.
>
> When disconnect the device during an interrupt transfer, we found that
> periodic schedule is disabled, but quickly enabled again. It is later
> disabled again. The bogus enable happens in ehci_urb_dequeue():
>
> /* reschedule QH iff another request is queued */
> if (!list_empty (&qh->qtd_list)
> && HC_IS_RUNNING (hcd->state)) {
> printk(KERN_INFO "qh_schedule in urb_dequeue %x\n",
> hcd->state);
>
> rc = qh_schedule(ehci, qh);
>
> The qh_schedule() then calls qh_link_periodic(), which calls
> enable_periodic().
>
> It is unclear how qh->qtd_list should be updated properly for the
> disconnect, although ep->urb_list is cleared.
How can ep->urb_list be cleared without qh->qtd_list also getting
cleared?
> Is this a bug?
Probably not. The periodic schedule should be enabled as long as any
periodic URBs are queued. The fact that they are addressed to a device
which has been unplugged doesn't matter -- usb_disconnect() will unlink
the remaining URBs for the unplugged device quickly enough.
Alan Stern
--
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
