Skip to contentSkip to conversation

bcm43xx: Fix low-traffic netdev watchdog TX timeouts

PATCH New Larry Finger 1 participant 1 message Oct 28, 2006

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 bcm43xx: Fix low-traffic netdev watchdog TX timeouts

From: Michael Buesch <[email protected]>

This fixes a netdev watchdog timeout problem.
The software needs to call netif_tx_disable before running the hardware calibration code. The problem condition can be shown by the following timegraph.

|---5secs - ~10 jiffies time---|---|OOPS
^                              ^
last real TX                   periodic work stops netif

At OOPS, the following happens:
The watchdog timer triggers, because the timeout of 5secs is over. The watchdog first checks for stopped TX. _Usually_ TX is only stopped from the TX handler to indicate a full TX queue. But this is different. We need to stop TX here, regardless of the TX queue state. So the watchdog recognizes the stopped device and assumes it is stopped due to full TX queues (Which is a _wrong_ assumption in this case). It then tests how far the last TX has been in the past. If it's more than 5secs (which is the case for low or no traffic), it will fire a TX timeout.

---

John,

Please apply this patch to wireless-2.6, and push it upstream to 2.6.19. It seems to have cured _all_ remaining cases of watchdog timeouts. This patch replaces the one sent earlier by Michael under the subject "Re: Fwd: [PATCH] bcm43xx: (hopefully) fix watchdog timeouts." BTW, that simplified version has been sent to -stable. It isn't quite as good as this one, but fixes most timeout cases and is more appropriate for stable.

Larry

drivers/net/wireless/bcm43xx/bcm43xx_main.c +15 -1
Index: wireless-2.6/drivers/net/wireless/bcm43xx/bcm43xx_main.c
===================================================================
@@ -3232,9 +3232,11 @@ static int estimate_periodic_work_badnes
 static void bcm43xx_periodic_work_handler(void *d)
 {
 	struct bcm43xx_private *bcm = d;
+	struct net_device *net_dev = bcm->net_dev;
 	unsigned long flags;
 	u32 savedirqs = 0;
 	int badness;
+	unsigned long orig_trans_start = 0;
 
 	mutex_lock(&bcm->mutex);
 	badness = estimate_periodic_work_badness(bcm->periodic_state);
@@ -3242,7 +3244,18 @@ static void bcm43xx_periodic_work_handle
 		/* Periodic work will take a long time, so we want it to
 		 * be preemtible.
 		 */
-		netif_tx_disable(bcm->net_dev);
+
+		netif_tx_lock_bh(net_dev);
+		/* We must fake a started transmission here, as we are going to
+		 * disable TX. If we wouldn't fake a TX, it would be possible to
+		 * trigger the netdev watchdog, if the last real TX is already
+		 * some time on the past (slightly less than 5secs)
+		 */
+		orig_trans_start = net_dev->trans_start;
+		net_dev->trans_start = jiffies;
+		netif_stop_queue(net_dev);
+		netif_tx_unlock_bh(net_dev);
+
 		spin_lock_irqsave(&bcm->irq_lock, flags);
 		bcm43xx_mac_suspend(bcm);
 		if (bcm43xx_using_pio(bcm))
@@ -3267,6 +3280,7 @@ static void bcm43xx_periodic_work_handle
 			bcm43xx_pio_thaw_txqueues(bcm);
 		bcm43xx_mac_enable(bcm);
 		netif_wake_queue(bcm->net_dev);
+		net_dev->trans_start = orig_trans_start;
 	}
 	mmiowb();
 	spin_unlock_irqrestore(&bcm->irq_lock, flags);

---