On Mon, 11 Mar 2013, Adrian Bassett wrote:
> > You know, the 6e0c3339a6 commit consists of two hunks, and they are
> > pretty much independent of one another. You could try reverting each
> > of the hunks separately to narrow things down. Maybe only one of them
> > is responsible for your problem.
>
> Have now been able to do this.
>
> Using a vanilla 3.8.2 as base but with the following commit reverted
> as it's scheduled for removal in 3.8.3 anyway:
>
> 55bcdce8a8228223ec4d17d8ded8134ed265d2c5 - 'USB: EHCI: remove ASS/PSS polling timeout'
>
> Then 6e0c3339a6f19d748f16091d0a05adeb1e1f822b - 'USB: EHCI: unlink
> one async QH at a time' - was also reverted and two kernels were
> built. The first, 'A', incorporated the first part of 6e0c3339
> (changes to drivers/usb/host/ehci-q.c starting at line 1197) while
> the second, 'B', had that code reverted in favour of the remainder of
> the original composite patch.
>
> What I found was that 'A' was entirely un-problematic in that I could
> not trigger the problem behaviour at all. On the other hand, 'B'
> could be made to trigger at first suspend/resume attempt. Moreover,
> whilst previously the system with the composite patch in would always
> return after a few minutes, with 'B' I eventually gave up waiting and
> forced a re-boot. Perhaps the first half of the original patch in
> some way mitigates the effects of the second.
Ah, that explains a lot. Indeed, there does appear to be a bug.
I haven't tried out this patch myself, but I think it will fix your
problem. It is based on 3.8 and should apply to 3.8.2. Let me know
what happens.
Alan Stern
Index: 3.8/drivers/usb/host/ehci-hcd.c
===================================================================
--- 3.8.orig/drivers/usb/host/ehci-hcd.c
+++ 3.8/drivers/usb/host/ehci-hcd.c
@@ -302,6 +302,7 @@ static void ehci_quiesce (struct ehci_hc
static void end_unlink_async(struct ehci_hcd *ehci);
static void unlink_empty_async(struct ehci_hcd *ehci);
+static void unlink_empty_async_suspended(struct ehci_hcd *ehci);
static void ehci_work(struct ehci_hcd *ehci);
static void start_unlink_intr(struct ehci_hcd *ehci, struct ehci_qh *qh);
static void end_unlink_intr(struct ehci_hcd *ehci, struct ehci_qh *qh);
Index: 3.8/drivers/usb/host/ehci-hub.c
===================================================================
--- 3.8.orig/drivers/usb/host/ehci-hub.c
+++ 3.8/drivers/usb/host/ehci-hub.c
@@ -328,7 +328,7 @@ static int ehci_bus_suspend (struct usb_
ehci->rh_state = EHCI_RH_SUSPENDED;
end_unlink_async(ehci);
- unlink_empty_async(ehci);
+ unlink_empty_async_suspended(ehci);
ehci_handle_intr_unlinks(ehci);
end_free_itds(ehci);
Index: 3.8/drivers/usb/host/ehci-q.c
===================================================================
--- 3.8.orig/drivers/usb/host/ehci-q.c
+++ 3.8/drivers/usb/host/ehci-q.c
@@ -1316,6 +1316,20 @@ static void unlink_empty_async(struct eh
}
}
+/* The root hub is suspended; unlink all the async QHs */
+static void unlink_empty_async_suspended(struct ehci_hcd *ehci)
+{
+ struct ehci_qh *qh, *qh_to_unlink;
+
+ qh = ehci->async->qh_next.qh;
+ while (qh) {
+ qh_to_unlink = qh;
+ qh = qh->qh_next.qh;
+ WARN_ON(!list_empty(&qh_to_unlink->qtd_list));
+ start_unlink_async(ehci, qh_to_unlink);
+ }
+}
+
/* makes sure the async qh will become idle */
/* caller must own ehci->lock */
--
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
