leandrof/linux
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

sctp: fix the issue sctp_diag uses lock_sock in rcu_read_lock

Browse Source 
When sctp dumps all the ep->assocs, it needs to lock_sock first,
but now it locks sock in rcu_read_lock, and lock_sock may sleep,
which would break rcu_read_lock.

This patch is to get and hold one sock when traversing the list.
After that and get out of rcu_read_lock, lock and dump it. Then
it will traverse the list again to get the next one until all
sctp socks are dumped.

For sctp_diag_dump_one, it fixes this issue by holding asoc and
moving cb() out of rcu_read_lock in sctp_transport_lookup_process.

Fixes: 8f840e47f1 ("sctp: add the sctp_diag.c file")
Signed-off-by: Xin Long <lucien.xin@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This commit is contained in:
Xin Long
2016-09-29 02:55:44 +08:00
committed by David S. Miller
parent 75b005b949
commit 1cceda7849
2 changed files with 47 additions and 21 deletions
Download Patch File Download Diff File Expand all files Collapse all files

10
net/sctp/socket.c
View File

@@ -4469,17 +4469,21 @@ int sctp_transport_lookup_process(int (*cb)(struct sctp_transport *, void *),
const union sctp_addr *paddr, void *p)
{
struct sctp_transport *transport;
int err = 0;
int err = -ENOENT;
rcu_read_lock();
transport = sctp_addrs_lookup_transport(net, laddr, paddr);
if (!transport || !sctp_transport_hold(transport))
goto out;
err = cb(transport, p);
sctp_association_hold(transport->asoc);
sctp_transport_put(transport);
out:
rcu_read_unlock();
err = cb(transport, p);
sctp_association_put(transport->asoc);
out:
return err;
}
EXPORT_SYMBOL_GPL(sctp_transport_lookup_process);