3c59x: Fix deadlock in vortex_error()

This fixes a bug introduced in commit de847272149365363a6043a963a6f42fb91566e2
"3c59x: Use fine-grained locks for MII and windowed register access".

vortex_interrupt() holds vp->window_lock over multiple register accesses to reduce locking overhead. However it also needs to call vortex_error() sometimes, and that uses the regular functions for access to windowed registers, which will try to acquire window_lock again.

Therefore, drop window_lock around the call to vortex_error() and set the window afterward reacquiring the lock. Since vortex_error() may call vortex_rx(), which *does* require its caller to hold window_lock, lift that call up into vortex_interrupt(). This also removes the potential for calling vortex_rx() on a later-generation NIC.

---
Dave,

This is the regression I mentioned before. 3c59x should be good after this.

Ben.

 drivers/net/3c59x.c |   10 +++++++---
 1 files changed, 7 insertions(+), 3 deletions(-)

index a045559..85671ad 100644
@@ -1994,10 +1994,9 @@ vortex_error(struct net_device *dev, int status)
 		}
 	}
 
-	if (status & RxEarly) {				/* Rx early is unused. */
-		vortex_rx(dev);
+	if (status & RxEarly)				/* Rx early is unused. */
 		iowrite16(AckIntr | RxEarly, ioaddr + EL3_CMD);
-	}
+
 	if (status & StatsFull) {			/* Empty statistics. */
 		static int DoneDidThat;
 		if (vortex_debug > 4)

@@ -2298,7 +2297,12 @@ vortex_interrupt(int irq, void *dev_id)
 		if (status & (HostError | RxEarly | StatsFull | TxComplete | IntReq)) {
 			if (status == 0xffff)
 				break;
+			if (status & RxEarly)
+				vortex_rx(dev);
+			spin_unlock(&vp->window_lock);
 			vortex_error(dev, status);
+			spin_lock(&vp->window_lock);
+			window_set(vp, 7);
 		}
 
 		if (--work_done < 0) {

1.7.1

From: Ben Hutchings <[email protected]>
Date: Tue, 07 Sep 2010 02:27:10 +0100

Excellent, applied, thanks!