[PATCH 2/2] lockdep: annotate bcsp driver - v2

[Peter Zijlstra]

=============================================
[ INFO: possible recursive locking detected ]
2.6.18-1.2699.fc6 #1
---------------------------------------------
swapper/0 is trying to acquire lock:
 (&list->lock#3){+...}, at: [<c05ad307>] skb_dequeue+0x12/0x43

but task is already holding lock:
 (&list->lock#3){+...}, at: [<df98cd79>] bcsp_dequeue+0x6a/0x11e [hci_uart]

Moving a skb from the unack'ed to the rel(iable) list nests the two list locks.
Reliable packets are never moved the other way, hence no circular dependency
exists.

Signed-off-by: Peter Zijlstra <a.p.zijlstra@xxxxxxxxx>
---
 drivers/bluetooth/hci_bcsp.c |    4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

Index: linux-2.6/drivers/bluetooth/hci_bcsp.c
===================================================================
--- linux-2.6.orig/drivers/bluetooth/hci_bcsp.c
+++ linux-2.6/drivers/bluetooth/hci_bcsp.c
@@ -330,7 +330,7 @@ static struct sk_buff *bcsp_dequeue(stru
 	   reliable packet if the number of packets sent but not yet ack'ed
 	   is < than the winsize */
 
-	spin_lock_irqsave(&bcsp->unack.lock, flags);
+	spin_lock_irqsave_nested(&bcsp->unack.lock, flags, SINGLE_DEPTH_NESTING);
 
 	if (bcsp->unack.qlen < BCSP_TXWINSIZE && (skb = skb_dequeue(&bcsp->rel)) != NULL) {
 		struct sk_buff *nskb = bcsp_prepare_pkt(bcsp, skb->data, skb->len, bt_cb(skb)->pkt_type);
@@ -696,7 +696,7 @@ static void bcsp_timed_event(unsigned lo
 
 	BT_DBG("hu %p retransmitting %u pkts", hu, bcsp->unack.qlen);
 
-	spin_lock_irqsave(&bcsp->unack.lock, flags);
+	spin_lock_irqsave_nested(&bcsp->unack.lock, flags, SINGLE_DEPTH_NESTING);
 
 	while ((skb = __skb_dequeue_tail(&bcsp->unack)) != NULL) {
 		bcsp->msgq_txseq = (bcsp->msgq_txseq - 1) & 0x07;


-
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/