FAILED: patch "[PATCH] ksmbd: reset rcount per connection in" failed to apply to 6.12-stable tree

PATCH New [email protected] 2 participants 4 messages Apr 24 – Apr 25

Subsystems (2)

Stable branch Kernel SMB3 server (ksmbd)

lore

Share & tools

lore

[email protected]

author maintainer

2026-04-24 09:44 UTC #0 lore raw

The patch below does not apply to the 6.12-stable tree.
If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <[email protected]>.

To reproduce the conflict and resubmit, you may use the following commands:

git fetch https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/ linux-6.12.y
git checkout FETCH_HEAD
git cherry-pick -x def036ef87f8641c1c525d5ae17438d7a1006491
# <resolve conflicts, build, test, etc.>
git commit -s
git send-email --to '<[email protected]>' --in-reply-to '2026042410-cinch-frostbite-b760@gregkh' --subject-prefix 'PATCH 6.12.y' HEAD^..

Possible dependencies:

thanks,

greg k-h

------------------ original commit in Linus's tree ------------------

From def036ef87f8641c1c525d5ae17438d7a1006491 Mon Sep 17 00:00:00 2001
From: DaeMyung Kang <[email protected]>
Date: Sun, 19 Apr 2026 02:28:44 +0900
Subject: [PATCH] ksmbd: reset rcount per connection in
ksmbd_conn_wait_idle_sess_id()

rcount is intended to be connection-specific: 2 for curr_conn, 1 for every other connection sharing the same session. However, it is initialised only once before the hash iteration and is never reset. After the loop visits curr_conn, later sibling connections are also checked against rcount == 2, so a sibling with req_running == 1 is incorrectly treated as idle. This makes the outcome depend on the hash iteration order: whether a given sibling is checked against the loose (< 2) or the strict (< 1) threshold is decided by whether it happens to be visited before or after curr_conn.

The function's contract is "wait until every connection sharing this session is idle" so that destroy_previous_session() can safely tear the session down. The latched rcount violates that contract and reopens the teardown race window the wait logic was meant to close:
destroy_previous_session() may proceed before sibling channels have actually quiesced, overlapping session teardown with in-flight work on those connections.

Recompute rcount inside the loop so each connection is compared against its own threshold regardless of iteration order.

This is a code-inspection fix for an iteration-order-dependent logic error; a targeted reproducer would require SMB3 multichannel with in-flight work on a sibling channel landing after curr_conn in hash order, which is not something that can be triggered reliably.

Fixes 76e98a158b20 ("ksmbd: fix race condition between destroy_previous_session() and smb2 operations()")

Cc [email protected]

Signed-off-by DaeMyung Kang <[email protected]>

Acked-by Namjae Jeon <[email protected]>

Signed-off-by Steve French <[email protected]>

fs/smb/server/connection.c +2 -3

index a26899d12df1..b5e077f272cf 100644
@@ -237,7 +237,7 @@ int ksmbd_conn_wait_idle_sess_id(struct ksmbd_conn *curr_conn, u64 sess_id)
 {
 	struct ksmbd_conn *conn;
 	int rc, retry_count = 0, max_timeout = 120;
-	int rcount = 1, bkt;
+	int rcount, bkt;
 
 retry_idle:
 	if (retry_count >= max_timeout)

@@ -246,8 +246,7 @@ int ksmbd_conn_wait_idle_sess_id(struct ksmbd_conn *curr_conn, u64 sess_id)
 	down_read(&conn_list_lock);
 	hash_for_each(conn_list, bkt, conn, hlist) {
 		if (conn->binding || xa_load(&conn->sessions, sess_id)) {
-			if (conn == curr_conn)
-				rcount = 2;
+			rcount = (conn == curr_conn) ? 2 : 1;
 			if (atomic_read(&conn->req_running) >= rcount) {
 				rc = wait_event_timeout(conn->req_running_q,
 					atomic_read(&conn->req_running) < rcount,

Sasha Levin

maintainer

2026-04-25 09:07 UTC #1 lore raw

PATCH 1/3 ksmbd: use msleep instaed of schedule_timeout_interruptible()

From: Namjae Jeon <[email protected]>

[ Upstream commit f75f8bdd4ff4830abe31a1b94892eb12b85b9535 ]

use msleep instaed of schedule_timeout_interruptible() to guarantee the task delays as expected.

Signed-off-by Namjae Jeon <[email protected]>

Signed-off-by Steve French <[email protected]>

Stable-dep-of: def036ef87f8 ("ksmbd: reset rcount per connection in ksmbd_conn_wait_idle_sess_id()")

Signed-off-by Sasha Levin <[email protected]>

---
 fs/smb/server/connection.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

fs/smb/server/connection.c +1 -1

index 2537b52eea3e6..1b96d58183cdf 100644
@@ -495,7 +495,7 @@ static void stop_sessions(void)
 	up_read(&conn_list_lock);
 
 	if (!list_empty(&conn_list)) {
-		schedule_timeout_interruptible(HZ / 10); /* 100ms */
+		msleep(100);
 		goto again;
 	}
 }

Signature · 1 line

2.53.0

Sasha Levin

maintainer

2026-04-25 09:07 UTC #2 lore raw

PATCH 2/3 ksmbd: replace connection list with hash table

From: Namjae Jeon <[email protected]>

[ Upstream commit 0bcc831be535269556f59cb70396f7e34f03a276 ]

Replace connection list with hash table to improve lookup performance.

Signed-off-by Namjae Jeon <[email protected]>

Signed-off-by Steve French <[email protected]>

Stable-dep-of: def036ef87f8 ("ksmbd: reset rcount per connection in ksmbd_conn_wait_idle_sess_id()")

Signed-off-by Sasha Levin <[email protected]>

---
 fs/smb/server/connection.c     | 23 +++++++++++------------
 fs/smb/server/connection.h     |  6 ++++--
 fs/smb/server/smb2pdu.c        |  4 ++--
 fs/smb/server/transport_rdma.c |  5 +++++
 fs/smb/server/transport_tcp.c  | 25 +++++++++++++++++++++----
 5 files changed, 43 insertions(+), 20 deletions(-)

fs/smb/server/connection.c +11 ~~-12~~

index 1b96d58183cdf..076f704b37a7b 100644
@@ -19,7 +19,7 @@ static DEFINE_MUTEX(init_lock);
 
 static struct ksmbd_conn_ops default_conn_ops;
 
-LIST_HEAD(conn_list);
+DEFINE_HASHTABLE(conn_list, CONN_HASH_BITS);
 DECLARE_RWSEM(conn_list_lock);
 
 /**

@@ -33,7 +33,7 @@ DECLARE_RWSEM(conn_list_lock);
 void ksmbd_conn_free(struct ksmbd_conn *conn)
 {
 	down_write(&conn_list_lock);
-	list_del(&conn->conns_list);
+	hash_del(&conn->hlist);
 	up_write(&conn_list_lock);
 
 	xa_destroy(&conn->sessions);

@@ -78,7 +78,6 @@ struct ksmbd_conn *ksmbd_conn_alloc(void)
 
 	init_waitqueue_head(&conn->req_running_q);
 	init_waitqueue_head(&conn->r_count_q);
-	INIT_LIST_HEAD(&conn->conns_list);
 	INIT_LIST_HEAD(&conn->requests);
 	INIT_LIST_HEAD(&conn->async_requests);
 	spin_lock_init(&conn->request_lock);

@@ -91,19 +90,17 @@ struct ksmbd_conn *ksmbd_conn_alloc(void)
 
 	init_rwsem(&conn->session_lock);
 
-	down_write(&conn_list_lock);
-	list_add(&conn->conns_list, &conn_list);
-	up_write(&conn_list_lock);
 	return conn;
 }
 
 bool ksmbd_conn_lookup_dialect(struct ksmbd_conn *c)
 {
 	struct ksmbd_conn *t;
+	int bkt;
 	bool ret = false;
 
 	down_read(&conn_list_lock);
-	list_for_each_entry(t, &conn_list, conns_list) {
+	hash_for_each(conn_list, bkt, t, hlist) {
 		if (memcmp(t->ClientGUID, c->ClientGUID, SMB2_CLIENT_GUID_SIZE))
 			continue;

@@ -164,9 +161,10 @@ void ksmbd_conn_unlock(struct ksmbd_conn *conn)
 void ksmbd_all_conn_set_status(u64 sess_id, u32 status)
 {
 	struct ksmbd_conn *conn;
+	int bkt;
 
 	down_read(&conn_list_lock);
-	list_for_each_entry(conn, &conn_list, conns_list) {
+	hash_for_each(conn_list, bkt, conn, hlist) {
 		if (conn->binding || xa_load(&conn->sessions, sess_id))
 			WRITE_ONCE(conn->status, status);
 	}

@@ -182,14 +180,14 @@ int ksmbd_conn_wait_idle_sess_id(struct ksmbd_conn *curr_conn, u64 sess_id)
 {
 	struct ksmbd_conn *conn;
 	int rc, retry_count = 0, max_timeout = 120;
-	int rcount = 1;
+	int rcount = 1, bkt;
 
 retry_idle:
 	if (retry_count >= max_timeout)
 		return -EIO;
 
 	down_read(&conn_list_lock);
-	list_for_each_entry(conn, &conn_list, conns_list) {
+	hash_for_each(conn_list, bkt, conn, hlist) {
 		if (conn->binding || xa_load(&conn->sessions, sess_id)) {
 			if (conn == curr_conn)
 				rcount = 2;

@@ -480,10 +478,11 @@ static void stop_sessions(void)
 {
 	struct ksmbd_conn *conn;
 	struct ksmbd_transport *t;
+	int bkt;
 
 again:
 	down_read(&conn_list_lock);
-	list_for_each_entry(conn, &conn_list, conns_list) {
+	hash_for_each(conn_list, bkt, conn, hlist) {
 		t = conn->transport;
 		ksmbd_conn_set_exiting(conn);
 		if (t->ops->shutdown) {

@@ -494,7 +493,7 @@ static void stop_sessions(void)
 	}
 	up_read(&conn_list_lock);
 
-	if (!list_empty(&conn_list)) {
+	if (!hash_empty(conn_list)) {
 		msleep(100);
 		goto again;
 	}

fs/smb/server/connection.h +4 -2

index 2aa8084bb5930..78633ad1175c8 100644
@@ -52,11 +52,12 @@ struct ksmbd_conn {
 		u8			inet6_addr[16];
 #endif
 	};
+	unsigned int			inet_hash;
 	char				*request_buf;
 	struct ksmbd_transport		*transport;
 	struct nls_table		*local_nls;
 	struct unicode_map		*um;
-	struct list_head		conns_list;
+	struct hlist_node		hlist;
 	struct rw_semaphore		session_lock;
 	/* smb session 1 per user */
 	struct xarray			sessions;

@@ -151,7 +152,8 @@ struct ksmbd_transport {
 #define KSMBD_TCP_SEND_TIMEOUT	(5 * HZ)
 #define KSMBD_TCP_PEER_SOCKADDR(c)	((struct sockaddr *)&((c)->peer_addr))
 
-extern struct list_head conn_list;
+#define CONN_HASH_BITS	12
+extern DECLARE_HASHTABLE(conn_list, CONN_HASH_BITS);
 extern struct rw_semaphore conn_list_lock;
 
 bool ksmbd_conn_alive(struct ksmbd_conn *conn);

fs/smb/server/smb2pdu.c +2 -2

index 733c0bace6182..83b83b6d950cd 100644
@@ -7425,7 +7425,7 @@ int smb2_lock(struct ksmbd_work *work)
 	int nolock = 0;
 	LIST_HEAD(lock_list);
 	LIST_HEAD(rollback_list);
-	int prior_lock = 0;
+	int prior_lock = 0, bkt;
 
 	WORK_BUFFERS(work, req, rsp);

@@ -7535,7 +7535,7 @@ int smb2_lock(struct ksmbd_work *work)
 		nolock = 1;
 		/* check locks in connection list */
 		down_read(&conn_list_lock);
-		list_for_each_entry(conn, &conn_list, conns_list) {
+		hash_for_each(conn_list, bkt, conn, hlist) {
 			spin_lock(&conn->llist_lock);
 			list_for_each_entry_safe(cmp_lock, tmp2, &conn->lock_list, clist) {
 				if (file_inode(cmp_lock->fl->c.flc_file) !=

fs/smb/server/transport_rdma.c +5 -0

index d15c33480e0a9..3976009596d99 100644
@@ -381,6 +381,11 @@ static struct smb_direct_transport *alloc_transport(struct rdma_cm_id *cm_id)
 	conn = ksmbd_conn_alloc();
 	if (!conn)
 		goto err;
+
+	down_write(&conn_list_lock);
+	hash_add(conn_list, &conn->hlist, 0);
+	up_write(&conn_list_lock);
+
 	conn->transport = KSMBD_TRANS(t);
 	KSMBD_TRANS(t)->conn = conn;
 	KSMBD_TRANS(t)->ops = &ksmbd_smb_direct_transport_ops;

fs/smb/server/transport_tcp.c +21 -4

index 9ad9b50bf93e8..7da25c16051ce 100644
@@ -89,13 +89,21 @@ static struct tcp_transport *alloc_transport(struct socket *client_sk)
 	}
 
 #if IS_ENABLED(CONFIG_IPV6)
-	if (client_sk->sk->sk_family == AF_INET6)
+	if (client_sk->sk->sk_family == AF_INET6) {
 		memcpy(&conn->inet6_addr, &client_sk->sk->sk_v6_daddr, 16);
-	else
+		conn->inet_hash = ipv6_addr_hash(&client_sk->sk->sk_v6_daddr);
+	} else {
 		conn->inet_addr = inet_sk(client_sk->sk)->inet_daddr;
+		conn->inet_hash = ipv4_addr_hash(inet_sk(client_sk->sk)->inet_daddr);
+	}
 #else
 	conn->inet_addr = inet_sk(client_sk->sk)->inet_daddr;
+	conn->inet_hash = ipv4_addr_hash(inet_sk(client_sk->sk)->inet_daddr);
 #endif
+	down_write(&conn_list_lock);
+	hash_add(conn_list, &conn->hlist, conn->inet_hash);
+	up_write(&conn_list_lock);
+
 	conn->transport = KSMBD_TRANS(t);
 	KSMBD_TRANS(t)->conn = conn;
 	KSMBD_TRANS(t)->ops = &ksmbd_tcp_transport_ops;

@@ -240,7 +248,7 @@ static int ksmbd_kthread_fn(void *p)
 	struct socket *client_sk = NULL;
 	struct interface *iface = (struct interface *)p;
 	struct ksmbd_conn *conn;
-	int ret;
+	int ret, inet_hash;
 	unsigned int max_ip_conns;
 
 	while (!kthread_should_stop()) {

@@ -265,9 +273,18 @@ static int ksmbd_kthread_fn(void *p)
 		/*
 		 * Limits repeated connections from clients with the same IP.
 		 */
+#if IS_ENABLED(CONFIG_IPV6)
+		if (client_sk->sk->sk_family == AF_INET6)
+			inet_hash = ipv6_addr_hash(&client_sk->sk->sk_v6_daddr);
+		else
+			inet_hash = ipv4_addr_hash(inet_sk(client_sk->sk)->inet_daddr);
+#else
+		inet_hash = ipv4_addr_hash(inet_sk(client_sk->sk)->inet_daddr);
+#endif
+
 		max_ip_conns = 0;
 		down_read(&conn_list_lock);
-		list_for_each_entry(conn, &conn_list, conns_list) {
+		hash_for_each_possible(conn_list, conn, hlist, inet_hash) {
 #if IS_ENABLED(CONFIG_IPV6)
 			if (client_sk->sk->sk_family == AF_INET6) {
 				if (memcmp(&client_sk->sk->sk_v6_daddr,

Signature · 1 line

2.53.0

Sasha Levin

maintainer

2026-04-25 09:07 UTC #3 lore raw

PATCH 3/3 ksmbd: reset rcount per connection in ksmbd_conn_wait_idle_sess_id()

In reply to Sasha Levin

From: DaeMyung Kang <[email protected]>

[ Upstream commit def036ef87f8641c1c525d5ae17438d7a1006491 ]

Recompute rcount inside the loop so each connection is compared against its own threshold regardless of iteration order.

Fixes 76e98a158b20 ("ksmbd: fix race condition between destroy_previous_session() and smb2 operations()")

Cc [email protected]

Signed-off-by DaeMyung Kang <[email protected]>

Acked-by Namjae Jeon <[email protected]>

Signed-off-by Steve French <[email protected]>

Signed-off-by Sasha Levin <[email protected]>

---
 fs/smb/server/connection.c | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)

fs/smb/server/connection.c +2 -3

index 076f704b37a7b..d38d9c0272c29 100644
@@ -180,7 +180,7 @@ int ksmbd_conn_wait_idle_sess_id(struct ksmbd_conn *curr_conn, u64 sess_id)
 {
 	struct ksmbd_conn *conn;
 	int rc, retry_count = 0, max_timeout = 120;
-	int rcount = 1, bkt;
+	int rcount, bkt;
 
 retry_idle:
 	if (retry_count >= max_timeout)

@@ -189,8 +189,7 @@ int ksmbd_conn_wait_idle_sess_id(struct ksmbd_conn *curr_conn, u64 sess_id)
 	down_read(&conn_list_lock);
 	hash_for_each(conn_list, bkt, conn, hlist) {
 		if (conn->binding || xa_load(&conn->sessions, sess_id)) {
-			if (conn == curr_conn)
-				rcount = 2;
+			rcount = (conn == curr_conn) ? 2 : 1;
 			if (atomic_read(&conn->req_running) >= rcount) {
 				rc = wait_event_timeout(conn->req_running_q,
 					atomic_read(&conn->req_running) < rcount,

Signature · 1 line

2.53.0