Skip to contentSkip to conversation

prevent CVE-2012-2372 rds-ping causes kernel panic

PATCH In review Jay Fenlason 3 participants 3 messages Oct 22, 2012

Subsystems (3)
Networking InfiniBand RDS - reliable datagram sockets
lore
Share & tools
lore

Jay Fenlason

author
#0 lore raw

PATCH 1/1 prevent CVE-2012-2372 rds-ping causes kernel panic

If you use rds-ping of the local IP address of some Infiniband HCAs (QLogic, possibly others) the machine will panic with a BUG_ON due to an overly restrictive check. Loosen the restriction a bit

This should have gotten sent a while ago (it was first noticed in
https://bugzilla.redhat.com/show_bug.cgi?id=803936
and patched in kernel-2.6.32-275.el6) but I got confused about its embargo status and lost it.

Signed-off-by Jay Fenlason <[email protected]>
net/rds/ib_send.c +1 -1
index e590949..7920c85 100644
@@ -544,7 +544,7 @@ int rds_ib_xmit(struct rds_connection *conn, struct rds_message *rm,
 	int flow_controlled = 0;
 	int nr_sig = 0;
 
-	BUG_ON(off % RDS_FRAG_SIZE);
+	BUG_ON(!conn->c_loopback && off % RDS_FRAG_SIZE);
 	BUG_ON(hdr_off != 0 && hdr_off != sizeof(struct rds_header));
 
 	/* Do not send cong updates to IB loopback */

--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html

#1 lore raw

net/rds/ib_send.c

Quoted 12 lines · Jay Fenlason

If you use rds-ping of the local IP address of some Infiniband HCAs (QLogic, possibly others) the machine will panic with a BUG_ON due to an overly restrictive check. Loosen the restriction a bit

This should have gotten sent a while ago (it was first noticed in
https://bugzilla.redhat.com/show_bug.cgi?id=803936
and patched in kernel-2.6.32-275.el6) but I got confused about its embargo status and lost it.

Signed-off-by Jay Fenlason <[email protected]>
net/rds/ib_send.c +0 -0

<formletter>

This is not the correct way to submit patches for inclusion in the stable kernel tree. Please read Documentation/stable_kernel_rules.txt for how to do this properly.

</formletter>
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html

#2 lore raw
In reply to Jay Fenlason
Quoted 16 lines · Jay Fenlason

If you use rds-ping of the local IP address of some Infiniband HCAs (QLogic, possibly others) the machine will panic with a BUG_ON due to an overly restrictive check. Loosen the restriction a bit

This should have gotten sent a while ago (it was first noticed in
https://bugzilla.redhat.com/show_bug.cgi?id=803936
and patched in kernel-2.6.32-275.el6) but I got confused about its embargo status and lost it.

Signed-off-by Jay Fenlason <[email protected]>
net/rds/ib_send.c +0 -0
index e590949..7920c85 100644

net/rds/ib_send.c int rds_ib_xmit(struct rds_connection *conn, struct rds_message *rm, · line 550

Quoted 8 lines
net/rds/ib_send.c +1 -1
@@ -544,7 +544,7 @@ int rds_ib_xmit(struct rds_connection *conn, struct rds_message *rm,
  	int flow_controlled = 0;
  	int nr_sig = 0;
 
-	BUG_ON(off % RDS_FRAG_SIZE);
+	BUG_ON(!conn->c_loopback && off % RDS_FRAG_SIZE);
  	BUG_ON(hdr_off != 0 && hdr_off != sizeof(struct rds_header));
 
  	/* Do not send cong updates to IB loopback */

Hello Jay,

The recommended approach for checking preconditions is to use WARN_ON_ONCE(), not BUG_ON(). Linus explained this last month in a message posted on the LKML (http://lkml.org/lkml/2012/9/27/461).

Bart.

--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html