Re: [patch] Real-Time Preemption, -RT-2.6.12-rc6-V0.7.48-00

[Ingo Molnar]

* William Weston <weston@xxxxxxxxx> wrote:

> Fortunately, there were some dumps (all ns83820 tx related) left in 
> the logs before the machine locked:

i'm not sure it's related to the lockup - but there is a generic bug in 
the driver, which in ns83820_tx_timeout() does:

	local_irq_save(flags);
	...
	CALL do_tx_done()

		spin_lock_irq(&dev->tx_lock);
		...
		spin_unlock_irq(&dev->tx_lock); // [1]
	...
	local_irq_restore(flags); // [2]

this leads to interrupts being enabled at [1], while the intention was 
to enable them at [2]. To solve this bug we can do the tx-locking in 
ns83820_tx_timeout(). (local_irq_disable() use in ns83820_tx_timeout() 
was probably SMP-unsafe too, but needs a rare race.)

find the patch below - it's also included in the -50-05 -RT tree i just 
uploaded. Can you confirm that you dont get the warnings in the -50-05 
(and later) -RT kernels?

	Ingo

Index: drivers/net/ns83820.c
===================================================================
--- drivers/net/ns83820.c.orig
+++ drivers/net/ns83820.c
@@ -1013,8 +1013,6 @@ static void do_tx_done(struct net_device
 	struct ns83820 *dev = PRIV(ndev);
 	u32 cmdsts, tx_done_idx, *desc;
 
-	spin_lock_irq(&dev->tx_lock);
-
 	dprintk("do_tx_done(%p)\n", ndev);
 	tx_done_idx = dev->tx_done_idx;
 	desc = dev->tx_descs + (tx_done_idx * DESC_SIZE);
@@ -1070,7 +1068,6 @@ static void do_tx_done(struct net_device
 		netif_start_queue(ndev);
 		netif_wake_queue(ndev);
 	}
-	spin_unlock_irq(&dev->tx_lock);
 }
 
 static void ns83820_cleanup_tx(struct ns83820 *dev)
@@ -1371,7 +1368,9 @@ static void ns83820_do_isr(struct net_de
 	 * work has accumulated
 	 */
 	if ((ISR_TXDESC | ISR_TXIDLE | ISR_TXOK | ISR_TXERR) & isr) {
+		spin_lock_irq(&dev->tx_lock);
 		do_tx_done(ndev);
+		spin_unlock_irq(&dev->tx_lock);
 
 		/* Disable TxOk if there are no outstanding tx packets.
 		 */
@@ -1456,7 +1455,7 @@ static void ns83820_tx_timeout(struct ne
         u32 tx_done_idx, *desc;
 	unsigned long flags;
 
-	local_irq_save(flags);
+	spin_lock_irqsave(&dev->tx_lock, flags);
 
 	tx_done_idx = dev->tx_done_idx;
 	desc = dev->tx_descs + (tx_done_idx * DESC_SIZE);
@@ -1483,7 +1482,7 @@ static void ns83820_tx_timeout(struct ne
 		ndev->name,
 		tx_done_idx, dev->tx_free_idx, le32_to_cpu(desc[DESC_CMDSTS]));
 
-	local_irq_restore(flags);
+	spin_unlock_irqrestore(&dev->tx_lock, flags);
 }
 
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/