Ver Fonte

3c59x: Fix shared IRQ handling

As its first order of business, boomerang_interrupt() checks whether
the device really has any pending interrupts. If it does not,
it does nothing and returns, but it still returns IRQ_HANDLED.

This is wrong: interrupt was not handled, IRQ handlers of other
devices sharing this IRQ line need to be called.

vortex_interrupt() has it right: it returns IRQ_NONE in this case
via IRQ_RETVAL(0).

Do the same in boomerang_interrupt().

Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
CC: David S. Miller <davem@davemloft.net>
CC: linux-kernel@vger.kernel.org
CC: netdev@vger.kernel.org
Signed-off-by: David S. Miller <davem@davemloft.net>
Denys Vlasenko há 10 anos atrás
pai
commit
4eed4d8ff9
1 ficheiros alterados com 3 adições e 1 exclusões
  1. 3 1
      drivers/net/ethernet/3com/3c59x.c

+ 3 - 1
drivers/net/ethernet/3com/3c59x.c

@@ -2382,6 +2382,7 @@ boomerang_interrupt(int irq, void *dev_id)
 	void __iomem *ioaddr;
 	int status;
 	int work_done = max_interrupt_work;
+	int handled = 0;
 
 	ioaddr = vp->ioaddr;
 
@@ -2400,6 +2401,7 @@ boomerang_interrupt(int irq, void *dev_id)
 
 	if ((status & IntLatch) == 0)
 		goto handler_exit;		/* No interrupt: shared IRQs can cause this */
+	handled = 1;
 
 	if (status == 0xffff) {		/* h/w no longer present (hotplug)? */
 		if (vortex_debug > 1)
@@ -2501,7 +2503,7 @@ boomerang_interrupt(int irq, void *dev_id)
 handler_exit:
 	vp->handling_irq = 0;
 	spin_unlock(&vp->lock);
-	return IRQ_HANDLED;
+	return IRQ_RETVAL(handled);
 }
 
 static int vortex_rx(struct net_device *dev)