[PATCH 0/2] 3.1-git ipoib: fix NULL pointer dereference

[PATCH 0/2] 3.1-git ipoib: fix NULL pointer dereference

[Bernd Schubert]

The following series  
- fixes a NULL pointer dereferences in ipoib_start_xmit()
- renames variable 'n' to 'dst_neigh' to make the code better readable

Question: I am unsure about the usage of likely() here, as I run into the else branch,
maybe it is not that unlikely and likely() should be removed?

---

Bernd Schubert (2):
      Fix possible Null pointer dereference in ipoib_start_xmit()
      Rename 'n' into a longer variable name.


 drivers/infiniband/ulp/ipoib/ipoib_main.c |   17 ++++++++++-------
 1 files changed, 10 insertions(+), 7 deletions(-)

--
Bernd Schubert
Fraunhofer ITWM / FhGFS

--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH 1/2] Fix possible Null pointer dereference in ipoib_start_xmit()

[Bernd Schubert]

This will fix https://bugzilla.kernel.org/show_bug.cgi?id=41212

fslab2 login: [  114.392408] EXT4-fs (sdc): barriers disabled
[  114.449737] EXT4-fs (sdc): mounted filesystem with writeback data mode.
Opts: journal_async_commit,barrier=0,data=writeback
[  240.944030] BUG: unable to handle kernel NULL pointer dereference at
0000000000000040
[  240.948007] IP: [<ffffffffa0366ce9>] ipoib_start_xmit+0x39/0x280 [ib_ipoib]
[...]
[  240.948007] Call Trace:
[  240.948007]  <IRQ>
[  240.948007]  [<ffffffff812cd5e0>] dev_hard_start_xmit+0x2a0/0x590
[  240.948007]  [<ffffffff8131f680>] ? arp_create+0x70/0x200
[  240.948007]  [<ffffffff812e8e1f>] sch_direct_xmit+0xef/0x1c0

Signed-off-by: Bernd Schubert <bernd.schubert-mPn0NPGs4xGatNDF+KUbs4QuADTiUCJX@public.gmane.org>
---
 drivers/infiniband/ulp/ipoib/ipoib_main.c |    8 +++++---
 1 files changed, 5 insertions(+), 3 deletions(-)

diff --git a/drivers/infiniband/ulp/ipoib/ipoib_main.c b/drivers/infiniband/ulp/ipoib/ipoib_main.c
index 43f89ba..fe89c46 100644
--- a/drivers/infiniband/ulp/ipoib/ipoib_main.c
+++ b/drivers/infiniband/ulp/ipoib/ipoib_main.c
@@ -717,11 +717,13 @@ static int ipoib_start_xmit(struct sk_buff *skb, struct net_device *dev)
 {
 	struct ipoib_dev_priv *priv = netdev_priv(dev);
 	struct ipoib_neigh *neigh;
-	struct neighbour *n;
+	struct neighbour *n = NULL;
 	unsigned long flags;
 
-	n = dst_get_neighbour(skb_dst(skb));
-	if (likely(skb_dst(skb) && n)) {
+	if (likely(skb_dst(skb)))
+		n = dst_get_neighbour(skb_dst(skb));
+
+	if (likely(n)) {
 		if (unlikely(!*to_ipoib_neigh(n))) {
 			ipoib_path_lookup(skb, dev);
 			return NETDEV_TX_OK;

--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH 2/2] Rename 'n' into a longer variable name.

[Bernd Schubert]

When it comes to me variable names consisting of a single letter
should be forbidden by coding style guide lines, as it is rather
difficult to search for single letter, such as 'n'.

Rename struct neighbour *n to dst_neigh


Signed-off-by: Bernd Schubert <bernd.schubert-mPn0NPGs4xGatNDF+KUbs4QuADTiUCJX@public.gmane.org>
---
 drivers/infiniband/ulp/ipoib/ipoib_main.c |   15 ++++++++-------
 1 files changed, 8 insertions(+), 7 deletions(-)

diff --git a/drivers/infiniband/ulp/ipoib/ipoib_main.c b/drivers/infiniband/ulp/ipoib/ipoib_main.c
index fe89c46..189d4cb 100644
--- a/drivers/infiniband/ulp/ipoib/ipoib_main.c
+++ b/drivers/infiniband/ulp/ipoib/ipoib_main.c
@@ -717,22 +717,22 @@ static int ipoib_start_xmit(struct sk_buff *skb, struct net_device *dev)
 {
 	struct ipoib_dev_priv *priv = netdev_priv(dev);
 	struct ipoib_neigh *neigh;
-	struct neighbour *n = NULL;
+	struct neighbour *dst_neigh = NULL;
 	unsigned long flags;
 
 	if (likely(skb_dst(skb)))
-		n = dst_get_neighbour(skb_dst(skb));
+		dst_neigh = dst_get_neighbour(skb_dst(skb));
 
-	if (likely(n)) {
-		if (unlikely(!*to_ipoib_neigh(n))) {
+	if (likely(dst_neigh)) {
+		if (unlikely(!*to_ipoib_neigh(dst_neigh))) {
 			ipoib_path_lookup(skb, dev);
 			return NETDEV_TX_OK;
 		}
 
-		neigh = *to_ipoib_neigh(n);
+		neigh = *to_ipoib_neigh(dst_neigh);
 
 		if (unlikely((memcmp(&neigh->dgid.raw,
-				     n->ha + 4,
+				     dst_neigh->ha + 4,
 				     sizeof(union ib_gid))) ||
 			     (neigh->dev != dev))) {
 			spin_lock_irqsave(&priv->lock, flags);
@@ -758,7 +758,8 @@ static int ipoib_start_xmit(struct sk_buff *skb, struct net_device *dev)
 				return NETDEV_TX_OK;
 			}
 		} else if (neigh->ah) {
-			ipoib_send(dev, skb, neigh->ah, IPOIB_QPN(n->ha));
+			ipoib_send(dev, skb, neigh->ah,
+				   IPOIB_QPN(dst_neigh->ha));
 			return NETDEV_TX_OK;
 		}
 

--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH 1/2] Fix possible Null pointer dereference in ipoib_start_xmit()

[Roland Dreier]

Thanks, applied.
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH 2/2] Rename 'n' into a longer variable name.

[David Miller]

From: Bernd Schubert <bernd.schubert-mPn0NPGs4xGatNDF+KUbs4QuADTiUCJX@public.gmane.org>
Date: Tue, 16 Aug 2011 12:56:59 +0200

> When it comes to me variable names consisting of a single letter
> should be forbidden by coding style guide lines, as it is rather
> difficult to search for single letter, such as 'n'.

I totally disagree and use such things aggressively myself.

I am not applying this patch.
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html