USB: UHCI: improve scheduling of interrupt URBs

This patch (as1140) adds a little intelligence to the interrupt-URB
scheduler in uhci-hcd.  Right now the scheduler is stupid; every URB
having the same period is assigned to the same slot.  Thus a large
group of period-N URBs can fill their slot and cause -ENOSPC errors
even when all the lower-period slots are empty.

With the patch, if an URB doesn't fit in its assigned slot then the
scheduler will try using lower-period slots.  This will provide
greater flexibility.  As an example, the driver will be able to handle
more than just three or four mice, which the current driver cannot.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

diff --git a/drivers/usb/host/uhci-q.c b/drivers/usb/host/uhci-q.c
index 1f0c2cf..5631d89 100644
--- a/drivers/usb/host/uhci-q.c
+++ b/drivers/usb/host/uhci-q.c
@@ -1065,13 +1065,18 @@
 		}
 		if (exponent < 0)
 			return -EINVAL;
-		qh->period = 1 << exponent;
-		qh->skel = SKEL_INDEX(exponent);
 
-		/* For now, interrupt phase is fixed by the layout
-		 * of the QH lists. */
-		qh->phase = (qh->period / 2) & (MAX_PHASE - 1);
-		ret = uhci_check_bandwidth(uhci, qh);
+		/* If the slot is full, try a lower period */
+		do {
+			qh->period = 1 << exponent;
+			qh->skel = SKEL_INDEX(exponent);
+
+			/* For now, interrupt phase is fixed by the layout
+			 * of the QH lists.
+			 */
+			qh->phase = (qh->period / 2) & (MAX_PHASE - 1);
+			ret = uhci_check_bandwidth(uhci, qh);
+		} while (ret != 0 && --exponent >= 0);
 		if (ret)
 			return ret;
 	} else if (qh->period > urb->interval)