Roland> You know, looking at the ipoib code, I can't even recreate
Roland> why xmit_lock is taken in the set_multicast_list method
Roland> anyway, or how it works at all -- it seems
Roland> set_multicast_list will always be called with xmit_lock
Roland> already held. What the heck is going on?
Never mind -- I see that the set_multicast_list work needs to be
deferred to process context, so ipoib_mcast_restart_task() doesn't run
directly from the call to set_multicast_list.
I guess the fix in the current kernel is just something like the
below. And in the netif_tx_lock() patch, the local_irq_save() /
local_irq_restore() calls can just be removed.
Am I on the right track?
Anyway I won't push the patch below since the bug is harmless right
now and it can be fixed up as part of the netif_tx_lock() patch.
Thanks,
Roland
diff --git a/drivers/infiniband/ulp/ipoib/ipoib_multicast.c
b/drivers/infiniband/ulp/ipoib/ipoib_multicast.c
index ec41c8f..5f3eaf1 100644
--- a/drivers/infiniband/ulp/ipoib/ipoib_multicast.c
+++ b/drivers/infiniband/ulp/ipoib/ipoib_multicast.c
@@ -814,13 +814,12 @@ void ipoib_mcast_restart_task(void *dev_
struct dev_mc_list *mclist;
struct ipoib_mcast *mcast, *tmcast;
LIST_HEAD(remove_list);
- unsigned long flags;
ipoib_dbg_mcast(priv, "restarting multicast task\n");
ipoib_mcast_stop_thread(dev, 0);
- spin_lock_irqsave(&dev->xmit_lock, flags);
+ spin_lock_bh(&dev->xmit_lock);
spin_lock(&priv->lock);
/*
@@ -895,7 +894,7 @@ void ipoib_mcast_restart_task(void *dev_
}
spin_unlock(&priv->lock);
- spin_unlock_irqrestore(&dev->xmit_lock, flags);
+ spin_unlock_bh(&dev->xmit_lock);
/* We have to cancel outside of the spinlock */
list_for_each_entry_safe(mcast, tmcast, &remove_list, list) {
-
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to [EMAIL PROTECTED]
More majordomo info at http://vger.kernel.org/majordomo-info.html
