Skip to contentSkip to conversation

target: fix a race condition between login_work and the login thread

PATCH Applied Maurizio Lombardi 3 participants 4 messages Nov 15, 2022 – Dec 1, 2022 SCSI target

lore
Share & tools
lore
#0 lore raw

PATCH v3 1/1 target: fix a race condition between login_work and the login thread

In case a malicious initiator sends some random data immediately after a login PDU; the iscsi_target_sk_data_ready() callback will schedule the login_work and, at the same time,
the negotiation may end without clearing the LOGIN_FLAGS_INITIAL_PDU flag (because no additional PDU exchanges are required to complete the login).

The login has been completed but the login_work function will find the LOGIN_FLAGS_INITIAL_PDU flag set and will never stop from rescheduling itself;
at this point, if the initiator drops the connection, the iscsit_conn structure will be freed, login_work will dereference a released socket structure and the kernel crashes.

BUG: kernel NULL pointer dereference, address: 0000000000000230
PF: supervisor write access in kernel mode
PF: error_code(0x0002) - not-present page
Workqueue: events iscsi_target_do_login_rx [iscsi_target_mod]
RIP: 0010:_raw_read_lock_bh+0x15/0x30
Call trace:
 iscsi_target_do_login_rx+0x75/0x3f0 [iscsi_target_mod]
 process_one_work+0x1e8/0x3c0

Fix this bug by forcing login_work to stop after the login has been completed and the socket callbacks have been restored.

Add a comment to clearify the return values of iscsi_target_do_login()

v3: cancel_delayed_work_sync() should be called by
iscsi_target_start_negotiation(), because the latter is only executed in login_thread context

V2: remove an unnecessary call to cancel_delayed_work();
fix a potential race condition in iscsi_start_negotiation() and in iscsi_target_do_login_rx()'s error paths

---
 drivers/target/iscsi/iscsi_target_nego.c | 12 ++++++++++--
 1 file changed, 10 insertions(+), 2 deletions(-)
drivers/target/iscsi/iscsi_target_nego.c +10 -2
index f2919319ad38..ff49c8f3fe24 100644
@@ -1018,6 +1018,13 @@ static int iscsi_target_handle_csg_one(struct iscsit_conn *conn, struct iscsi_lo
 	return 0;
 }
 
+/*
+ * RETURN VALUE:
+ *
+ *  1 = Login successful
+ * -1 = Login failed
+ *  0 = More PDU exchanges required
+ */
 static int iscsi_target_do_login(struct iscsit_conn *conn, struct iscsi_login *login)
 {
 	int pdu_count = 0;
@@ -1363,12 +1370,13 @@ int iscsi_target_start_negotiation(
 		ret = -1;
 
 	if (ret < 0) {
-		cancel_delayed_work_sync(&conn->login_work);
 		iscsi_target_restore_sock_callbacks(conn);
 		iscsi_remove_failed_auth_entry(conn);
 	}
-	if (ret != 0)
+	if (ret != 0) {
+		cancel_delayed_work_sync(&conn->login_work);
 		iscsi_target_nego_release(conn);
+	}
 
 	return ret;
 }
Signature · 1 line
2.31.1
Reviewed-by on v3 1/1
#1 lore raw
Quoted 31 lines · Maurizio Lombardi

In case a malicious initiator sends some random data immediately after a login PDU; the iscsi_target_sk_data_ready() callback will schedule the login_work and, at the same time,
the negotiation may end without clearing the LOGIN_FLAGS_INITIAL_PDU flag (because no additional PDU exchanges are required to complete the login).

The login has been completed but the login_work function will find the LOGIN_FLAGS_INITIAL_PDU flag set and will never stop from rescheduling itself;
at this point, if the initiator drops the connection, the iscsit_conn structure will be freed, login_work will dereference a released socket structure and the kernel crashes.

BUG: kernel NULL pointer dereference, address: 0000000000000230
PF: supervisor write access in kernel mode
PF: error_code(0x0002) - not-present page
Workqueue: events iscsi_target_do_login_rx [iscsi_target_mod]
RIP: 0010:_raw_read_lock_bh+0x15/0x30
Call trace:
 iscsi_target_do_login_rx+0x75/0x3f0 [iscsi_target_mod]
 process_one_work+0x1e8/0x3c0

Fix this bug by forcing login_work to stop after the login has been completed and the socket callbacks have been restored.

Add a comment to clearify the return values of iscsi_target_do_login()

v3: cancel_delayed_work_sync() should be called by
iscsi_target_start_negotiation(), because the latter is only executed in login_thread context

Quoted 5 lines

V2: remove an unnecessary call to cancel_delayed_work();
fix a potential race condition in iscsi_start_negotiation() and in iscsi_target_do_login_rx()'s error paths

#2 lore raw
In reply to Maurizio Lombardi

Maurizio,

Quoted 5 lines

In case a malicious initiator sends some random data immediately after a login PDU; the iscsi_target_sk_data_ready() callback will schedule the login_work and, at the same time, the negotiation may end without clearing the LOGIN_FLAGS_INITIAL_PDU flag (because no additional PDU exchanges are required to complete the login).

Applied to 6.2/scsi-staging, thanks!

Signature · 1 line
Martin K. Petersen	Oracle Linux Engineering
#3 lore raw
In reply to Maurizio Lombardi
Quoted 6 lines · Maurizio Lombardi

In case a malicious initiator sends some random data immediately after a login PDU; the iscsi_target_sk_data_ready() callback will schedule the login_work and, at the same time,
the negotiation may end without clearing the LOGIN_FLAGS_INITIAL_PDU flag (because no additional PDU exchanges are required to complete the login).

Quoted 8 lines

The login has been completed but the login_work function will find the LOGIN_FLAGS_INITIAL_PDU flag set and will never stop from rescheduling itself;
at this point, if the initiator drops the connection, the iscsit_conn structure will be freed, login_work will dereference a released socket structure and the kernel crashes.

[...]

Applied to 6.2/scsi-queue, thanks!

[1/1] target: fix a race condition between login_work and the login thread

      https://git.kernel.org/mkp/scsi/c/fec1b2fa62c1
Signature · 1 line
Martin K. Petersen	Oracle Linux Engineering