Skip to contentSkip to conversation

cifs: Revert bogus fix for CVE-2024-54680 and its followup commit.

PATCH New Kuniyuki Iwashima 1 participant 3 messages Apr 2, 2025

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

v2 cover cifs: Revert bogus fix for CVE-2024-54680 and its followup commit.

Commit e9f2517a3e18 ("smb: client: fix TCP timers deadlock after rmmod") was not only a bogus fix for LOCKDEP issue but also introduced a real TCP socket leak.

I'm working on the LOCKDEP fix on the networking side, so let's revert the commit and its followup fix.

For details, please see each commit.

Changes:
v2: Add Cc: stable

Kuniyuki Iwashima (2):
Revert "smb: client: Fix netns refcount imbalance causing leaks and
use-after-free"
Revert "smb: client: fix TCP timers deadlock after rmmod"

 fs/smb/client/connect.c | 34 +++++++++-------------------------
 1 file changed, 9 insertions(+), 25 deletions(-)
Signature · 1 line
2.48.1
#1 lore raw

PATCH v2 1/2 Revert "smb: client: Fix netns refcount imbalance causing leaks and use-after-free"

This reverts commit 4e7f1644f2ac6d01dc584f6301c3b1d5aac4eaef.

The commit e9f2517a3e18 ("smb: client: fix TCP timers deadlock after rmmod") is not only a bogus fix for LOCKDEP null-ptr-deref but also introduces a real issue, TCP sockets leak, which will be explained in detail in the next revert.

Also, CNA assigned CVE-2024-54680 to it but is rejecting it. [0]

Thus, we are reverting the commit and its follow-up commit 4e7f1644f2ac ("smb: client: Fix netns refcount imbalance causing leaks and use-after-free").

Fixes 4e7f1644f2ac ("smb: client: Fix netns refcount imbalance causing leaks and use-after-free")
---
 fs/smb/client/connect.c | 16 ++++++++--------
 1 file changed, 8 insertions(+), 8 deletions(-)
fs/smb/client/connect.c +8 -8
index 10a7c28d2d44..137a611c5ab0 100644
@@ -300,7 +300,6 @@ cifs_abort_connection(struct TCP_Server_Info *server)
 			 server->ssocket->flags);
 		sock_release(server->ssocket);
 		server->ssocket = NULL;
-		put_net(cifs_net_ns(server));
 	}
 	server->sequence_number = 0;
 	server->session_estab = false;
@@ -3367,12 +3366,8 @@ generic_ip_connect(struct TCP_Server_Info *server)
 		/*
 		 * Grab netns reference for the socket.
 		 *
-		 * This reference will be released in several situations:
-		 * - In the failure path before the cifsd thread is started.
-		 * - In the all place where server->socket is released, it is
-		 *   also set to NULL.
-		 * - Ultimately in clean_demultiplex_info(), during the final
-		 *   teardown.
+		 * It'll be released here, on error, or in clean_demultiplex_info() upon server
+		 * teardown.
 		 */
 		get_net(net);
 
@@ -3388,8 +3383,10 @@ generic_ip_connect(struct TCP_Server_Info *server)
 	}
 
 	rc = bind_socket(server);
-	if (rc < 0)
+	if (rc < 0) {
+		put_net(cifs_net_ns(server));
 		return rc;
+	}
 
 	/*
 	 * Eventually check for other socket options to change from
@@ -3444,6 +3441,9 @@ generic_ip_connect(struct TCP_Server_Info *server)
 	    (server->rfc1001_sessinit == -1 && sport == htons(RFC1001_PORT)))
 		rc = ip_rfc1001_connect(server);
 
+	if (rc < 0)
+		put_net(cifs_net_ns(server));
+
 	return rc;
 }
 
Signature · 1 line
2.48.1
#2 lore raw

PATCH v2 2/2 Revert "smb: client: fix TCP timers deadlock after rmmod"

In reply to Kuniyuki Iwashima

This reverts commit e9f2517a3e18a54a3943c098d2226b245d488801.

Commit e9f2517a3e18 ("smb: client: fix TCP timers deadlock after rmmod") is intended to fix a null-ptr-deref in LOCKDEP, which is mentioned as CVE-2024-54680, but is actually did not fix anything; The issue can be reproduced on top of it. [0]

Also, it reverted the change by commit ef7134c7fc48 ("smb: client:
Fix use-after-free of network namespace.") and introduced a real issue by reviving the kernel TCP socket.

When a reconnect happens for a CIFS connection, the socket state transitions to FIN_WAIT_1. Then, inet_csk_clear_xmit_timers_sync() in tcp_close() stops all timers for the socket.

If an incoming FIN packet is lost, the socket will stay at FIN_WAIT_1 forever, and such sockets could be leaked up to net.ipv4.tcp_max_orphans.

Usually, FIN can be retransmitted by the peer, but if the peer aborts the connection, the issue comes into reality.

I warned about this privately by pointing out the exact report [1], but the bogus fix was finally merged.

So, we should not stop the timers to finally kill the connection on our side in that case, meaning we must not use a kernel socket for TCP whose sk->sk_net_refcnt is 0.

The kernel socket does not have a reference to its netns to make it possible to tear down netns without cleaning up every resource in it.

For example, tunnel devices use a UDP socket internally, but we can destroy netns without removing such devices and let it complete during exit. Otherwise, netns would be leaked when the last application died.

However, this is problematic for TCP sockets because TCP has timers to close the connection gracefully even after the socket is close()d. The lifetime of the socket and its netns is different from the lifetime of the underlying connection.

If the socket user does not maintain the netns lifetime, the timer could be fired after the socket is close()d and its netns is freed up, resulting in use-after-free.

Actually, we have seen so many similar issues and converted such sockets to have a reference to netns.

That's why I converted the CIFS client socket to have a reference to netns (sk->sk_net_refcnt == 1), which is somehow mentioned as out-of-scope of CIFS and technically wrong in e9f2517a3e18, but **is in-scope and right fix**.

Regarding the LOCKDEP issue, we can prevent the module unload by bumping the module refcount when switching the LOCKDDEP key in sock_lock_init_class_and_name(). [2]

For a while, let's revert the bogus fix.

Note that now we can use sk_net_refcnt_upgrade() for the socket conversion, but I'll do so later separately to make backport easy.

---
 fs/smb/client/connect.c | 36 ++++++++++--------------------------
 1 file changed, 10 insertions(+), 26 deletions(-)
fs/smb/client/connect.c +10 -26
index 137a611c5ab0..989d8808260b 100644
@@ -1073,13 +1073,9 @@ clean_demultiplex_info(struct TCP_Server_Info *server)
 	msleep(125);
 	if (cifs_rdma_enabled(server))
 		smbd_destroy(server);
-
 	if (server->ssocket) {
 		sock_release(server->ssocket);
 		server->ssocket = NULL;
-
-		/* Release netns reference for the socket. */
-		put_net(cifs_net_ns(server));
 	}
 
 	if (!list_empty(&server->pending_mid_q)) {
@@ -1127,7 +1123,6 @@ clean_demultiplex_info(struct TCP_Server_Info *server)
 		 */
 	}
 
-	/* Release netns reference for this server. */
 	put_net(cifs_net_ns(server));
 	kfree(server->leaf_fullpath);
 	kfree(server->hostname);
@@ -1773,8 +1768,6 @@ cifs_get_tcp_session(struct smb3_fs_context *ctx,
 
 	tcp_ses->ops = ctx->ops;
 	tcp_ses->vals = ctx->vals;
-
-	/* Grab netns reference for this server. */
 	cifs_set_net_ns(tcp_ses, get_net(current->nsproxy->net_ns));
 
 	tcp_ses->sign = ctx->sign;
@@ -1902,7 +1895,6 @@ cifs_get_tcp_session(struct smb3_fs_context *ctx,
 out_err_crypto_release:
 	cifs_crypto_secmech_release(tcp_ses);
 
-	/* Release netns reference for this server. */
 	put_net(cifs_net_ns(tcp_ses));
 
 out_err:
@@ -1911,10 +1903,8 @@ cifs_get_tcp_session(struct smb3_fs_context *ctx,
 			cifs_put_tcp_session(tcp_ses->primary_server, false);
 		kfree(tcp_ses->hostname);
 		kfree(tcp_ses->leaf_fullpath);
-		if (tcp_ses->ssocket) {
+		if (tcp_ses->ssocket)
 			sock_release(tcp_ses->ssocket);
-			put_net(cifs_net_ns(tcp_ses));
-		}
 		kfree(tcp_ses);
 	}
 	return ERR_PTR(rc);
@@ -3356,20 +3346,20 @@ generic_ip_connect(struct TCP_Server_Info *server)
 		socket = server->ssocket;
 	} else {
 		struct net *net = cifs_net_ns(server);
+		struct sock *sk;
 
-		rc = sock_create_kern(net, sfamily, SOCK_STREAM, IPPROTO_TCP, &server->ssocket);
+		rc = __sock_create(net, sfamily, SOCK_STREAM,
+				   IPPROTO_TCP, &server->ssocket, 1);
 		if (rc < 0) {
 			cifs_server_dbg(VFS, "Error %d creating socket\n", rc);
 			return rc;
 		}
 
-		/*
-		 * Grab netns reference for the socket.
-		 *
-		 * It'll be released here, on error, or in clean_demultiplex_info() upon server
-		 * teardown.
-		 */
-		get_net(net);
+		sk = server->ssocket->sk;
+		__netns_tracker_free(net, &sk->ns_tracker, false);
+		sk->sk_net_refcnt = 1;
+		get_net_track(net, &sk->ns_tracker, GFP_KERNEL);
+		sock_inuse_add(net, 1);
 
 		/* BB other socket options to set KEEPALIVE, NODELAY? */
 		cifs_dbg(FYI, "Socket created\n");
@@ -3383,10 +3373,8 @@ generic_ip_connect(struct TCP_Server_Info *server)
 	}
 
 	rc = bind_socket(server);
-	if (rc < 0) {
-		put_net(cifs_net_ns(server));
+	if (rc < 0)
 		return rc;
-	}
 
 	/*
 	 * Eventually check for other socket options to change from
@@ -3423,7 +3411,6 @@ generic_ip_connect(struct TCP_Server_Info *server)
 	if (rc < 0) {
 		cifs_dbg(FYI, "Error %d connecting to server\n", rc);
 		trace_smb3_connect_err(server->hostname, server->conn_id, &server->dstaddr, rc);
-		put_net(cifs_net_ns(server));
 		sock_release(socket);
 		server->ssocket = NULL;
 		return rc;
@@ -3441,9 +3428,6 @@ generic_ip_connect(struct TCP_Server_Info *server)
 	    (server->rfc1001_sessinit == -1 && sport == htons(RFC1001_PORT)))
 		rc = ip_rfc1001_connect(server);
 
-	if (rc < 0)
-		put_net(cifs_net_ns(server));
-
 	return rc;
 }
 
Signature · 1 line
2.48.1