Re: rmmod e1000 hangs (Was Re: 2.6.22-rc2-mm1)

[Herbert Xu <herbert@gondor.apana.org.au>]

On Fri, May 25, 2007 at 07:48:14PM -0400, Jeff Garzik wrote:
> 
> Is this for -stable or upstream?  I got confused with all the patches 
> flying about.
> 
> Send it to me, if it's for upstream.

It's for upstream.  Once you take then I'll rediff it for stable.

[E1000]: Restore netif_poll_enable call but make sure IRQs are off

This restores the previously removed netif_poll_enable call in
e1000_open.  It's needed on all but the first call to e1000_open
for a NIC as e1000_close always calls netif_poll_disable.

netif_poll_enable can only be called safely if no polls have been
scheduled.  This should be the case as long as we don't enter our
IRQ handler.

In order to guarantee this we explicitly disable IRQs as early
as possible when we're probing the NIC.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

Cheers,
-- 
Visit Openswan at http://www.openswan.org/
Email: Herbert Xu ~{PmV>HI~} <herbert@gondor.apana.org.au>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt
--
diff --git a/drivers/net/e1000/e1000_main.c b/drivers/net/e1000/e1000_main.c
index cbc7feb..9ec35b7 100644
--- a/drivers/net/e1000/e1000_main.c
+++ b/drivers/net/e1000/e1000_main.c
@@ -1325,7 +1325,10 @@ e1000_sw_init(struct e1000_adapter *adapter)
 	spin_lock_init(&adapter->tx_queue_lock);
 #endif
 
-	atomic_set(&adapter->irq_sem, 1);
+	/* Explicitly disable IRQ since the NIC can be in any state. */
+	atomic_set(&adapter->irq_sem, 0);
+	e1000_irq_disable(adapter);
+
 	spin_lock_init(&adapter->stats_lock);
 
 	set_bit(__E1000_DOWN, &adapter->flags);
@@ -1431,6 +1434,10 @@ e1000_open(struct net_device *netdev)
 	/* From here on the code is the same as e1000_up() */
 	clear_bit(__E1000_DOWN, &adapter->flags);
 
+#ifdef CONFIG_E1000_NAPI
+	netif_poll_enable(netdev);
+#endif
+
 	e1000_irq_enable(adapter);
 
 	/* fire a link status change interrupt to start the watchdog */