[PATCH net v2 0/2] net/smc: fix use-after-free in smc_rx_pipe_buf_release()

[PATCH net v2 0/2] net/smc: fix use-after-free in smc_rx_pipe_buf_release()

[Hidayath Khan]

smc_rx_pipe_buf_release() tests sk_state before taking the socket lock
and then dereferences conn->rmb_desc and conn->lgr. A concurrent close
runs smc_conn_free() in between, which releases those structures. On the
is_reg_err path smcr_buf_unuse() frees the descriptor outright, so this
is a use-after-free.

Patch 2/2 fixes this by taking the socket lock first and testing
conn->freed instead. smc_conn_free() sets that flag before releasing
anything, under the same lock, so the two paths exclude each other.

Patch 1/2 is a prerequisite. conn->freed shares a byte with killed and
out_of_sync as single-bit bitfields. out_of_sync is written from the
receive tasklet without the socket lock, so a concurrent store to freed
from process context can be lost in the read-modify-write. Patch 1/2
gives each flag its own byte so stores do not interfere.

Hidayath Khan (2):
  net/smc: stop killed, freed and out_of_sync sharing a byte
  net/smc: fix use-after-free in smc_rx_pipe_buf_release()

 net/smc/smc.h    |  6 +++---
 net/smc/smc_rx.c | 11 +++++------
 2 files changed, 8 insertions(+), 9 deletions(-)

---
v2:
- Patch 1/2 is new. It is a prerequisite for 2/2.
- Patch 2/2 replaces the conn->rmb_desc = NULL approach from v1 with a
  conn->freed check. The NULL store exposed smc_cdc_msg_recv_action()
  and the sock_diag dump to NULL dereferences on paths that do not take
  the socket lock.

-- 
2.52.0

[PATCH net v2 1/2] net/smc: stop killed, freed and out_of_sync sharing a byte

[Hidayath Khan]

The three connection state flags are single-bit bitfields, so they occupy
one byte of struct smc_connection and every store to one is a
read-modify-write of the other two:

    u8  killed : 1;
    u8  freed : 1;
    u8  out_of_sync : 1;

They are not written under a common lock. smc_cdc_msg_validate() sets
out_of_sync from the receive tasklet, while smc_conn_kill() sets killed
from process context under lock_sock(), and the receive path does not defer
to the backlog when the socket is owned -- smc_cdc_msg_recv() takes only
bh_lock_sock().

Give each flag its own byte so a store no longer touches its neighbours.
All readers test them as booleans and are unchanged. struct smc_connection
grows by two bytes.

Fixes: b286a0651e44 ("net/smc: handle incoming CDC validation message")
Cc: stable@vger.kernel.org
Reviewed-by: Mahanta Jambigi <mjambigi@linux.ibm.com>
Signed-off-by: Hidayath Khan <hidayath@linux.ibm.com>
---
New in v2.

 net/smc/smc.h | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/net/smc/smc.h b/net/smc/smc.h
index 52145df83f6e..427b6d63b993 100644
--- a/net/smc/smc.h
+++ b/net/smc/smc.h
@@ -277,9 +277,9 @@ struct smc_connection {
 						 * 0 for SMC-R, 32 for SMC-D
 						 */
 	u64			peer_token;	/* SMC-D token of peer */
-	u8			killed : 1;	/* abnormal termination */
-	u8			freed : 1;	/* normal termination */
-	u8			out_of_sync : 1; /* out of sync with peer */
+	u8			killed;		/* abnormal termination */
+	u8			freed;		/* normal termination */
+	u8			out_of_sync;	/* out of sync with peer */
 };
 
 struct smc_sock {				/* smc sock container */
-- 
2.52.0

[PATCH net v2 2/2] net/smc: fix use-after-free in smc_rx_pipe_buf_release()

[Hidayath Khan]

smc_rx_splice() hands RMB pages to a pipe and takes a socket reference
per entry so the smc_sock stays alive until the reader finishes. The
connection does not: a concurrent close runs smc_conn_free(), which
releases the receive buffer back to the link group pool.

smc_rx_pipe_buf_release() tests sk_state before taking the socket lock.
The state can change between the test and the lock, and
smc_rx_update_cons() then dereferences conn->rmb_desc and walks
conn->lgr, which smc_conn_free() has already released. On the
is_reg_err path smcr_buf_unuse() frees the descriptor outright, so
this is a use-after-free.

Take the socket lock first and test conn->freed instead.
smc_conn_free() sets that flag before releasing anything, and every
caller holds the socket lock. The two paths exclude each other: either
the pipe release runs first with everything valid, or it sees the flag
and skips the update.

Fixes: 9014db202cb7 ("smc: add support for splice()")
Cc: stable@vger.kernel.org
Reviewed-by: Mahanta Jambigi <mjambigi@linux.ibm.com>
Signed-off-by: Hidayath Khan <hidayath@linux.ibm.com>
---
v2: Replace conn->rmb_desc = NULL with a conn->freed check under the
    socket lock. The NULL store exposed two paths without a NULL check:
    smc_cdc_msg_recv_action() in softirq and the sock_diag dump without
    the socket lock.

 net/smc/smc_rx.c | 11 +++++------
 1 file changed, 5 insertions(+), 6 deletions(-)

diff --git a/net/smc/smc_rx.c b/net/smc/smc_rx.c
index 5c9e4d8b57de..197fddc6271f 100644
--- a/net/smc/smc_rx.c
+++ b/net/smc/smc_rx.c
@@ -115,16 +115,15 @@ static void smc_rx_pipe_buf_release(struct pipe_inode_info *pipe,
 				    struct pipe_buffer *buf)
 {
 	struct smc_spd_priv *priv = (struct smc_spd_priv *)buf->private;
+	struct smc_connection *conn = &priv->smc->conn;
 	struct smc_sock *smc = priv->smc;
-	struct smc_connection *conn;
 	struct sock *sk = &smc->sk;
 
-	if (sk->sk_state == SMC_CLOSED ||
-	    sk->sk_state == SMC_PEERFINCLOSEWAIT ||
-	    sk->sk_state == SMC_APPFINCLOSEWAIT)
-		goto out;
-	conn = &smc->conn;
 	lock_sock(sk);
+	if (conn->freed) {
+		release_sock(sk);
+		goto out;
+	}
 	smc_rx_update_cons(smc, priv->len);
 	release_sock(sk);
 	if (atomic_sub_and_test(priv->len, &conn->splice_pending))
-- 
2.52.0