patch "USB: EHCI: reorganize isochronous scheduler routine" added to gregkh-2.6 tree

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



This is a note to let you know that I've just added the patch titled

    USB: EHCI: reorganize isochronous scheduler routine

to my gregkh-2.6 tree which can be found in directory form at:
    http://www.kernel.org/pub/linux/kernel/people/gregkh/gregkh-2.6/patches/
 and in git form at:
    git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/patches.git

The filename of this patch is:
    usb-ehci-reorganize-isochronous-scheduler-routine.patch

The patch will show up in the next release of the linux-next tree
(usually sometime within the next 24 hours during the week.)

If this patch meets the merge guidelines for a bugfix, it should be
merged into Linus's tree before the next major kernel release.
If not, it will be merged into Linus's tree during the next merge window.

Either way, you will probably be copied on the patch when it gets sent
to Linus for merging so that others can see what is happening in kernel
development.

If you have any questions about this process, please let me know.


>From stern+4c290172@xxxxxxxxxxxxxxxxxxx  Thu Jul 22 15:52:14 2010
Date: Wed, 14 Jul 2010 11:03:53 -0400 (EDT)
From: Alan Stern <stern@xxxxxxxxxxxxxxxxxxx>
To: Greg KH <greg@xxxxxxxxx>
cc: David Brownell <david-b@xxxxxxxxxxx>, USB list <linux-usb@xxxxxxxxxxxxxxx>
Subject: USB: EHCI: reorganize isochronous scheduler routine
Message-ID: <Pine.LNX.4.44L0.1007141058150.1634-100000@xxxxxxxxxxxxxxxxxxxx>

This patch (as1408) rearranges the scheduling code in ehci-hcd, partly
to improve its structure, but mainly to change the way it works.
Whether or not a transfer exceeds the hardware schedule length will
now be determined by looking at the last frame the transfer would use,
instead of the first available frame following the end of the transfer.

The benefit of this change is that it allows the driver to accept
valid URBs which would otherwise be rejected.  For example, suppose
the schedule length is 1024 frames, the endpoint period is 256 frames,
and a four-packet URB is submitted.  The four transfers would occupy
slots that are 0, 256, 512, and 768 frames past the current frame
(plus an extra slop factor).  These don't exceed the 1024-frame limit,
so the URB should be accepted.  But the current code notices that the
next available slot would be 1024 frames (plus slop) in the future,
which is beyond the limit, and so the URB is rejected unnecessarily.

Signed-off-by: Alan Stern <stern@xxxxxxxxxxxxxxxxxxx>
CC: David Brownell <david-b@xxxxxxxxxxx>
Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxx>


---
 drivers/usb/host/ehci-sched.c |  110 +++++++++++++++++++++---------------------
 1 file changed, 57 insertions(+), 53 deletions(-)

--- a/drivers/usb/host/ehci-sched.c
+++ b/drivers/usb/host/ehci-sched.c
@@ -1413,13 +1413,6 @@ iso_stream_schedule (
 		goto fail;
 	}
 
-	if (stream->depth + span > mod) {
-		ehci_dbg (ehci, "request %p would overflow (%d+%d>%d)\n",
-			urb, stream->depth, span, mod);
-		status = -EFBIG;
-		goto fail;
-	}
-
 	now = ehci_readl(ehci, &ehci->regs->frame_index) & (mod - 1);
 
 	/* Typical case: reuse current schedule, stream is still active.
@@ -1428,7 +1421,7 @@ iso_stream_schedule (
 	 * slot in the schedule, implicitly assuming URB_ISO_ASAP.
 	 */
 	if (likely (!list_empty (&stream->td_list))) {
-		start = stream->next_uframe;
+		u32	excess;
 
 		/* For high speed devices, allow scheduling within the
 		 * isochronous scheduling threshold.  For full speed devices
@@ -1440,21 +1433,23 @@ iso_stream_schedule (
 		else
 			next = now;
 
-		/* Fell behind (by up to twice the slop amount)? */
-		if (((start - next) & (mod - 1)) >=
-				mod - 2 * SCHEDULE_SLOP)
-			start += period * DIV_ROUND_UP(
-					(next - start) & (mod - 1),
-					period);
-
-		/* Tried to schedule too far into the future? */
-		if (unlikely(((start - now) & (mod - 1)) + span
-					>= mod - 2 * SCHEDULE_SLOP)) {
+		/* Fell behind (by up to twice the slop amount)?
+		 * We decide based on the time of the last currently-scheduled
+		 * slot, not the time of the next available slot.
+		 */
+		excess = (stream->next_uframe - period - next) & (mod - 1);
+		if (excess >= mod - 2 * SCHEDULE_SLOP)
+			start = next + excess - mod + period *
+					DIV_ROUND_UP(mod - excess, period);
+		else
+			start = next + excess + period;
+		if (start - now >= mod) {
+			ehci_dbg(ehci, "request %p would overflow (%d+%d >= %d)\n",
+					urb, start - now - period, period,
+					mod);
 			status = -EFBIG;
 			goto fail;
 		}
-		stream->next_uframe = start;
-		goto ready;
 	}
 
 	/* need to schedule; when's the next (u)frame we could start?
@@ -1463,51 +1458,60 @@ iso_stream_schedule (
 	 * can also help high bandwidth if the dma and irq loads don't
 	 * jump until after the queue is primed.
 	 */
-	start = SCHEDULE_SLOP + (now & ~0x07);
-	start &= mod - 1;
-	stream->next_uframe = start;
-
-	/* NOTE:  assumes URB_ISO_ASAP, to limit complexity/bugs */
-
-	/* find a uframe slot with enough bandwidth */
-	for (; start < (stream->next_uframe + period); start++) {
-		int		enough_space;
-
-		/* check schedule: enough space? */
-		if (stream->highspeed)
-			enough_space = itd_slot_ok (ehci, mod, start,
-					stream->usecs, period);
-		else {
-			if ((start % 8) >= 6)
-				continue;
-			enough_space = sitd_slot_ok (ehci, mod, stream,
-					start, sched, period);
+	else {
+		start = SCHEDULE_SLOP + (now & ~0x07);
+
+		/* NOTE:  assumes URB_ISO_ASAP, to limit complexity/bugs */
+
+		/* find a uframe slot with enough bandwidth */
+		next = start + period;
+		for (; start < next; start++) {
+
+			/* check schedule: enough space? */
+			if (stream->highspeed) {
+				if (itd_slot_ok(ehci, mod, start,
+						stream->usecs, period))
+					break;
+			} else {
+				if ((start % 8) >= 6)
+					continue;
+				if (sitd_slot_ok(ehci, mod, stream,
+						start, sched, period))
+					break;
+			}
 		}
 
-		/* schedule it here if there's enough bandwidth */
-		if (enough_space) {
-			stream->next_uframe = start & (mod - 1);
-			goto ready;
+		/* no room in the schedule */
+		if (start == next) {
+			ehci_dbg(ehci, "iso resched full %p (now %d max %d)\n",
+				urb, now, now + mod);
+			status = -ENOSPC;
+			goto fail;
 		}
 	}
 
-	/* no room in the schedule */
-	ehci_dbg (ehci, "iso %ssched full %p (now %d max %d)\n",
-		list_empty (&stream->td_list) ? "" : "re",
-		urb, now, now + mod);
-	status = -ENOSPC;
+	/* Tried to schedule too far into the future? */
+	if (unlikely(start - now + span - period
+				>= mod - 2 * SCHEDULE_SLOP)) {
+		ehci_dbg(ehci, "request %p would overflow (%d+%d >= %d)\n",
+				urb, start - now, span - period,
+				mod - 2 * SCHEDULE_SLOP);
+		status = -EFBIG;
+		goto fail;
+	}
 
-fail:
-	iso_sched_free (stream, sched);
-	urb->hcpriv = NULL;
-	return status;
+	stream->next_uframe = start & (mod - 1);
 
-ready:
 	/* report high speed start in uframes; full speed, in frames */
 	urb->start_frame = stream->next_uframe;
 	if (!stream->highspeed)
 		urb->start_frame >>= 3;
 	return 0;
+
+ fail:
+	iso_sched_free(stream, sched);
+	urb->hcpriv = NULL;
+	return status;
 }
 
 /*-------------------------------------------------------------------------*/

--
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


[Index of Archives]     [Linux Media]     [Linux Input]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]     [Old Linux USB Devel Archive]

  Powered by Linux