SUNRPC: Fix race corrupting rpc upcall

If rpc_queue_upcall() adds a new upcall to the rpci->pipe list just
after rpc_pipe_release calls rpc_purge_list(), but before it calls
gss_pipe_release (as rpci->ops->release_pipe(inode)), then the latter
will free a message without deleting it from the rpci->pipe list.

We will be left with a freed object on the rpc->pipe list.  Most
frequent symptoms are kernel crashes in rpc.gssd system calls on the
pipe in question.

Reported-by: J. Bruce Fields <bfields@redhat.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
Cc: stable@kernel.org

1 files changed, 5 insertions, 4 deletions

diff --git a/net/sunrpc/auth_gss/auth_gss.c b/net/sunrpc/auth_gss/auth_gss.c
index dcfc66bab2bb..12c485982814 100644
--- a/net/sunrpc/auth_gss/auth_gss.c
+++ b/net/sunrpc/auth_gss/auth_gss.c
@@ -745,17 +745,18 @@ gss_pipe_release(struct inode *inode)
         struct rpc_inode *rpci = RPC_I(inode);
         struct gss_upcall_msg *gss_msg;
 
+restart:
         spin_lock(&inode->i_lock);
-        while (!list_empty(&rpci->in_downcall)) {
+        list_for_each_entry(gss_msg, &rpci->in_downcall, list) {
 
-                gss_msg = list_entry(rpci->in_downcall.next,
-                                struct gss_upcall_msg, list);
+                if (!list_empty(&gss_msg->msg.list))
+                        continue;
                 gss_msg->msg.errno = -EPIPE;
                 atomic_inc(&gss_msg->count);
                 __gss_unhash_msg(gss_msg);
                 spin_unlock(&inode->i_lock);
                 gss_release_msg(gss_msg);
-                spin_lock(&inode->i_lock);
+                goto restart;
         }
         spin_unlock(&inode->i_lock);