I sent a patch for this fix to Greg on Saturday. It hasn't made it into
the master branch yet because Greg hasn't queued it yet. The patch is
on the xhci-2.6.35-bugfixes branch if you need it.
Sarah Sharp
On Wed, Jul 14, 2010 at 07:54:52PM +0530, Ramya Desai wrote:
> Dear Sarah,
>
> We found a small bug in the xhci-mem.c file which is related to the
> setting of the MULT field in the Endpoint Context Data Structure. The
> existing code sets the MULT field for all the SUPER SPEED endpoints.
> However, according to the 6.2.3 section of the xHCI 0.96
> specification, the MULT field shall be “0” for all endpoint types
> except for SS Isochronous. When we are testing the BULK endpoints with
> MaxStreams value set to 5, we encountered this error. However, after
> applying the patch, the device is working fine without any issues.
>
> Here is patch which fixes the error in xhci-mem.c file.
>
> diff -urNB drivers/usb/host/xhci-mem.c drivers/usb/host-modified/xhci-mem.c
> --- drivers/usb/host/xhci-mem.c 2010-07-08 12:22:48.000000000 +0530
> +++ drivers/usb/host-modified/xhci-mem.c 2010-07-14
> 18:57:51.252160556 +0530
> @@ -1006,11 +1006,13 @@
> * High speed endpoint descriptors can define "the number of additional
> * transaction opportunities per microframe", but that goes in the Max Burst
> * endpoint context field.
> + * According to 6.2.3 section of the xHCI_0.96 spec, the mult field (9:8 bits)
> + * shall be "0" for all endpoint types except for SS Isochronous.
> */
> static inline u32 xhci_get_endpoint_mult(struct usb_device *udev,
> struct usb_host_endpoint *ep)
> {
> - if (udev->speed != USB_SPEED_SUPER)
> + if ((udev->speed != USB_SPEED_SUPER) ||
> (!usb_endpoint_xfer_isoc(&ep->desc)))
> return 0;
> return ep->ss_ep_comp.bmAttributes;
> }
>
> Please let me know, if there are any concerns.
>
> Thanks and Regards,
> Ramya
--
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
