From e8b80056d07bc849777d032bc699a59cc28f9ddb Mon Sep 17 00:00:00 2001 From: Alan Stern Date: Tue, 5 Jul 2011 12:34:05 -0400 Subject: [PATCH 2/2] USB: EHCI: go back to using the system clock for QH unlinks This patch (as1477) fixes a problem affecting a few types of EHCI controller. Contrary to what one might expect, these controllers automatically stop their internal frame counter when no ports are enabled. Since ehci-hcd currently relies on the frame counter for determining when it should unlink QHs from the async schedule, those controllers run into trouble: The frame counter stops and the QHs never get unlinked. Some systems have also experienced other problems traced back to commit b963801164618e25fbdc0cd452ce49c3628b46c8 (USB: ehci-hcd unlink speedups), which made the original switch from using the system clock to using the frame counter. It never became clear what the reason was for these problems, but evidently it is related to use of the frame counter. To fix all these problems, this patch more or less reverts that commit and goes back to using the system clock. But this can't be done cleanly because other changes have since been made to the scan_async() subroutine. One of these changes involved the tricky logic that tries to avoid rescanning QHs that have already been seen when the scanning loop is restarted, which happens whenever an URB is given back. Switching back to clock-based unlinks would make this logic even more complicated. Therefore the new code doesn't rescan the entire async list whenever a giveback occurs. Instead it rescans only the current QH and continues on from there. This requires the use of a separate pointer to keep track of the next QH to scan, since the current QH may be unlinked while the scanning is in progress. That new pointer must be global, so that it can be adjusted forward whenever the _next_ QH gets unlinked. (uhci-hcd uses this same trick.) Simplification of the scanning loop removes a level of indentation, which accounts for the size of the patch. The amount of code changed is relatively small, and it isn't exactly a reversion of the b963801164 commit. This fixes Bugzilla #32432. Signed-off-by: Alan Stern CC: Tested-by: Matej Kenda Signed-off-by: Greg Kroah-Hartman --- drivers/usb/host/ehci-hcd.c | 8 ++--- drivers/usb/host/ehci-q.c | 82 +++++++++++++++++++++---------------------- drivers/usb/host/ehci.h | 3 +- 3 files changed, 45 insertions(+), 48 deletions(-) diff --git a/drivers/usb/host/ehci-hcd.c b/drivers/usb/host/ehci-hcd.c index 761bbb3..8e5b995 100644 --- a/drivers/usb/host/ehci-hcd.c +++ b/drivers/usb/host/ehci-hcd.c @@ -83,7 +83,8 @@ static const char hcd_name [] = "ehci_hcd"; #define EHCI_IAA_MSECS 10 /* arbitrary */ #define EHCI_IO_JIFFIES (HZ/10) /* io watchdog > irq_thresh */ #define EHCI_ASYNC_JIFFIES (HZ/20) /* async idle timeout */ -#define EHCI_SHRINK_FRAMES 5 /* async qh unlink delay */ +#define EHCI_SHRINK_JIFFIES (DIV_ROUND_UP(HZ, 200) + 1) + /* 200-ms async qh unlink delay */ /* Initial IRQ latency: faster than hw default */ static int log2_irq_thresh = 0; // 0 to 6 @@ -138,10 +139,7 @@ timer_action(struct ehci_hcd *ehci, enum ehci_timer_action action) break; /* case TIMER_ASYNC_SHRINK: */ default: - /* add a jiffie since we synch against the - * 8 KHz uframe counter. - */ - t = DIV_ROUND_UP(EHCI_SHRINK_FRAMES * HZ, 1000) + 1; + t = EHCI_SHRINK_JIFFIES; break; } mod_timer(&ehci->watchdog, t + jiffies); diff --git a/drivers/usb/host/ehci-q.c b/drivers/usb/host/ehci-q.c index 0079610..a664203 100644 --- a/drivers/usb/host/ehci-q.c +++ b/drivers/usb/host/ehci-q.c @@ -1226,6 +1226,8 @@ static void start_unlink_async (struct ehci_hcd *ehci, struct ehci_qh *qh) prev->hw->hw_next = qh->hw->hw_next; prev->qh_next = qh->qh_next; + if (ehci->qh_scan_next == qh) + ehci->qh_scan_next = qh->qh_next.qh; wmb (); /* If the controller isn't running, we don't have to wait for it */ @@ -1251,53 +1253,49 @@ static void scan_async (struct ehci_hcd *ehci) struct ehci_qh *qh; enum ehci_timer_action action = TIMER_IO_WATCHDOG; - ehci->stamp = ehci_readl(ehci, &ehci->regs->frame_index); timer_action_done (ehci, TIMER_ASYNC_SHRINK); -rescan: stopped = !HC_IS_RUNNING(ehci_to_hcd(ehci)->state); - qh = ehci->async->qh_next.qh; - if (likely (qh != NULL)) { - do { - /* clean any finished work for this qh */ - if (!list_empty(&qh->qtd_list) && (stopped || - qh->stamp != ehci->stamp)) { - int temp; - - /* unlinks could happen here; completion - * reporting drops the lock. rescan using - * the latest schedule, but don't rescan - * qhs we already finished (no looping) - * unless the controller is stopped. - */ - qh = qh_get (qh); - qh->stamp = ehci->stamp; - temp = qh_completions (ehci, qh); - if (qh->needs_rescan) - unlink_async(ehci, qh); - qh_put (qh); - if (temp != 0) { - goto rescan; - } - } - /* unlink idle entries, reducing DMA usage as well - * as HCD schedule-scanning costs. delay for any qh - * we just scanned, there's a not-unusual case that it - * doesn't stay idle for long. - * (plus, avoids some kind of re-activation race.) + ehci->qh_scan_next = ehci->async->qh_next.qh; + while (ehci->qh_scan_next) { + qh = ehci->qh_scan_next; + ehci->qh_scan_next = qh->qh_next.qh; + rescan: + /* clean any finished work for this qh */ + if (!list_empty(&qh->qtd_list)) { + int temp; + + /* + * Unlinks could happen here; completion reporting + * drops the lock. That's why ehci->qh_scan_next + * always holds the next qh to scan; if the next qh + * gets unlinked then ehci->qh_scan_next is adjusted + * in start_unlink_async(). */ - if (list_empty(&qh->qtd_list) - && qh->qh_state == QH_STATE_LINKED) { - if (!ehci->reclaim && (stopped || - ((ehci->stamp - qh->stamp) & 0x1fff) - >= EHCI_SHRINK_FRAMES * 8)) - start_unlink_async(ehci, qh); - else - action = TIMER_ASYNC_SHRINK; - } + qh = qh_get(qh); + temp = qh_completions(ehci, qh); + if (qh->needs_rescan) + unlink_async(ehci, qh); + qh->unlink_time = jiffies + EHCI_SHRINK_JIFFIES; + qh_put(qh); + if (temp != 0) + goto rescan; + } - qh = qh->qh_next.qh; - } while (qh); + /* unlink idle entries, reducing DMA usage as well + * as HCD schedule-scanning costs. delay for any qh + * we just scanned, there's a not-unusual case that it + * doesn't stay idle for long. + * (plus, avoids some kind of re-activation race.) + */ + if (list_empty(&qh->qtd_list) + && qh->qh_state == QH_STATE_LINKED) { + if (!ehci->reclaim && (stopped || + time_after_eq(jiffies, qh->unlink_time))) + start_unlink_async(ehci, qh); + else + action = TIMER_ASYNC_SHRINK; + } } if (action == TIMER_ASYNC_SHRINK) timer_action (ehci, TIMER_ASYNC_SHRINK); diff --git a/drivers/usb/host/ehci.h b/drivers/usb/host/ehci.h index c2c2f9d..23d9a5e 100644 --- a/drivers/usb/host/ehci.h +++ b/drivers/usb/host/ehci.h @@ -74,6 +74,7 @@ struct ehci_hcd { /* one per controller */ /* async schedule support */ struct ehci_qh *async; struct ehci_qh *reclaim; + struct ehci_qh *qh_scan_next; unsigned scanning : 1; /* periodic schedule support */ @@ -116,7 +117,6 @@ struct ehci_hcd { /* one per controller */ struct timer_list iaa_watchdog; struct timer_list watchdog; unsigned long actions; - unsigned stamp; unsigned periodic_stamp; unsigned random_frame; unsigned long next_statechange; @@ -337,6 +337,7 @@ struct ehci_qh { struct ehci_qh *reclaim; /* next to reclaim */ struct ehci_hcd *ehci; + unsigned long unlink_time; /* * Do NOT use atomic operations for QH refcounting. On some CPUs -- 1.7.6