Skip to contentSkip to conversation

Bluetooth: ISO: Fix KCSAN data-races on iso_pi(sk)

PATCH Applied SeungJu Cheon 3 participants 5 messages Apr 21

Subsystems (4)
lore
Share & tools
lore
#0 lore raw

v2 cover Bluetooth: ISO: Fix KCSAN data-races on iso_pi(sk)

Found while auditing iso_pi(sk) field accesses after a KCSAN report. Patch 1/2 is the reported race on iso_pi(sk)->dst in iso_sock_connect(); patch 2/2 covers related races on other iso_pi(sk) fields accessed in iso_connect_{bis,cis}() and iso_connect_ind() that were found by inspection during the same audit.

Changes in v2:
- Patch 1/2: Use sa->iso_bdaddr directly instead of caching the bacmp() result in a local variable, as suggested by Luiz [1]. This avoids reading from iso_pi(sk) entirely for the broadcast check.

- Patch 2/2: No changes.

v1: https://lore.kernel.org/linux-bluetooth/[email protected]/

[1] https://lore.kernel.org/linux-bluetooth/CABBYNZLBoU3byfK_G+=sTkBx3wNwEh2X6_7dG4+4LFtrc3Skpw@mail.gmail.com/

SeungJu Cheon (2):
Bluetooth: ISO: Fix data-race on dst in iso_sock_connect()
Bluetooth: ISO: Fix data-race on iso_pi(sk) in socket and HCI event
paths

 net/bluetooth/iso.c | 56 +++++++++++++++++++++++++--------------------
 1 file changed, 31 insertions(+), 25 deletions(-)
Signature · 1 line
2.52.0
#1 lore raw

PATCH v2 1/2 Bluetooth: ISO: Fix data-race on dst in iso_sock_connect()

iso_sock_connect() copies the destination address into iso_pi(sk)->dst under lock_sock, then releases the lock and reads it back with bacmp() to decide between the CIS and BIS connect paths:

    lock_sock(sk);
    bacpy(&iso_pi(sk)->dst, &sa->iso_bdaddr);
    iso_pi(sk)->dst_type = sa->iso_bdaddr_type;
    release_sock(sk);

    if (bacmp(&iso_pi(sk)->dst, BDADDR_ANY))  // <- no lock held

This read after release_sock() races with any concurrent write to iso_pi(sk)->dst on the same socket.

Fix by reading the destination address directly from the local sockaddr argument (sa->iso_bdaddr) instead of iso_pi(sk)->dst. Since sa is a function-local argument, reading it requires no locking and avoids the race.

This patch addresses only the bacmp() race in iso_sock_connect(); other unprotected iso_pi(sk) accesses are fixed separately in the next patch.

KCSAN report:

BUG: KCSAN: data-race in memcmp+0x39/0xb0

race at unknown origin, with read to 0xffff8f96ea66dde3 of 1 bytes by task 549 on cpu 1:

 memcmp+0x39/0xb0
 iso_sock_connect+0x275/0xb40
 __sys_connect_file+0xbd/0xe0
 __sys_connect+0xe0/0x110
 __x64_sys_connect+0x40/0x50
 x64_sys_call+0xcad/0x1c60
 do_syscall_64+0x133/0x590
 entry_SYSCALL_64_after_hwframe+0x77/0x7f

value changed: 0x00 -> 0xee

Reported by Kernel Concurrency Sanitizer on:

CPU: 1 UID: 0 PID: 549 Comm: iso_race_combin Not tainted 7.0.0-08391-g1d51b370a0f8 #40 PREEMPT(lazy)
Fixes ccf74f2390d6 ("Bluetooth: Add BTPROTO_ISO socket type")
---
 net/bluetooth/iso.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
net/bluetooth/iso.c +1 -1
index be145e2736b7..290a1b9a9daa 100644
@@ -1193,7 +1193,7 @@ static int iso_sock_connect(struct socket *sock, struct sockaddr_unsized *addr,
 
 	release_sock(sk);
 
-	if (bacmp(&iso_pi(sk)->dst, BDADDR_ANY))
+	if (bacmp(&sa->iso_bdaddr, BDADDR_ANY))
 		err = iso_connect_cis(sk);
 	else
 		err = iso_connect_bis(sk);
Signature · 1 line
2.52.0
#2 lore raw

This is automated email and please do not reply to this email!

Dear submitter,

Thank you for submitting the patches to the linux bluetooth mailing list. This is a CI test results with your patch series:
PW Link:https://patchwork.kernel.org/project/bluetooth/list/?series=1083591

---Test result---

Test Summary:

CheckPatch                    FAIL      0.93 seconds
GitLint                       FAIL      0.37 seconds
SubjectPrefix                 PASS      0.38 seconds
BuildKernel                   PASS      20.60 seconds
CheckAllWarning               PASS      22.42 seconds
CheckSparse                   PASS      21.78 seconds
BuildKernel32                 PASS      20.58 seconds
TestRunnerSetup               PASS      419.00 seconds
TestRunner_iso-tester         PASS      26.91 seconds
IncrementalBuild              PASS      23.33 seconds

Details
##############################
Test: CheckPatch - FAIL
Desc: Run checkpatch.pl script
Output:
[v2,1/2] Bluetooth: ISO: Fix data-race on dst in iso_sock_connect()
WARNING: Prefer a maximum 75 chars per line (possible unwrapped commit description?) #134:
race at unknown origin, with read to 0xffff8f96ea66dde3 of 1 bytes by task 549 on cpu 1:

total: 0 errors, 1 warnings, 0 checks, 8 lines checked

NOTE: For some of the reported defects, checkpatch may be able to
mechanically convert to the typical style using --fix or --fix-inplace.

/github/workspace/src/patch/14531805.patch has style problems, please review.

NOTE: Ignored message types: UNKNOWN_COMMIT_ID

NOTE: If any of the errors are false positives, please report
them to the maintainer, see CHECKPATCH in MAINTAINERS.

[v2,2/2] Bluetooth: ISO: Fix data-race on iso_pi(sk) in socket and HCI event paths
WARNING: Prefer a maximum 75 chars per line (possible unwrapped commit description?) #156:

CPU: 1 UID: 0 PID: 334 Comm: iso_setup_race Not tainted 7.0.0-10949-g8541d8f725c6 #44 PREEMPT(lazy)

total: 0 errors, 1 warnings, 0 checks, 137 lines checked

NOTE: For some of the reported defects, checkpatch may be able to
mechanically convert to the typical style using --fix or --fix-inplace.

/github/workspace/src/patch/14531806.patch has style problems, please review.

NOTE: Ignored message types: UNKNOWN_COMMIT_ID

NOTE: If any of the errors are false positives, please report
them to the maintainer, see CHECKPATCH in MAINTAINERS.

##############################
Test: GitLint - FAIL
Desc: Run gitlint
Output:
[v2,1/2] Bluetooth: ISO: Fix data-race on dst in iso_sock_connect()

WARNING: I3 - ignore-body-lines: gitlint will be switching from using Python regex 'match' (match beginning) to 'search' (match anywhere) semantics. Please review your ignore-body-lines.regex option accordingly. To remove this warning, set general.regex-style-search=True. More details: https://jorisroovers.github.io/gitlint/configuration/#regex-style-search
31: B1 Line exceeds max length (88>80): "race at unknown origin, with read to 0xffff8f96ea66dde3 of 1 bytes by task 549 on cpu 1:"
44: B1 Line exceeds max length (100>80): "CPU: 1 UID: 0 PID: 549 Comm: iso_race_combin Not tainted 7.0.0-08391-g1d51b370a0f8 #40 PREEMPT(lazy)"
[v2,2/2] Bluetooth: ISO: Fix data-race on iso_pi(sk) in socket and HCI event paths

WARNING: I3 - ignore-body-lines: gitlint will be switching from using Python regex 'match' (match beginning) to 'search' (match anywhere) semantics. Please review your ignore-body-lines.regex option accordingly. To remove this warning, set general.regex-style-search=True. More details: https://jorisroovers.github.io/gitlint/configuration/#regex-style-search
1: T1 Title exceeds max length (82>80): "[v2,2/2] Bluetooth: ISO: Fix data-race on iso_pi(sk) in socket and HCI event paths"
53: B1 Line exceeds max length (99>80): "CPU: 1 UID: 0 PID: 334 Comm: iso_setup_race Not tainted 7.0.0-10949-g8541d8f725c6 #44 PREEMPT(lazy)"

https://github.com/bluez/bluetooth-next/pull/108

---
Regards,
Linux Bluetooth

#3 lore raw

PATCH v2 2/2 Bluetooth: ISO: Fix data-race on iso_pi(sk) in socket and HCI event paths

In reply to SeungJu Cheon

Several iso_pi(sk) fields (qos, qos_user_set, bc_sid, base, base_len, sync_handle, bc_num_bis) are written under lock_sock in iso_sock_setsockopt() and iso_sock_bind(), but read and written under hci_dev_lock only in two other paths:

- iso_connect_bis() / iso_connect_cis(), invoked from connect(2), read qos/base/bc_sid and reset qos to default_qos on the qos_user_set validation failure -- all without lock_sock.

- iso_connect_ind(), invoked from hci_rx_work, writes sync_handle, bc_sid, qos.bcast.encryption, bc_num_bis, base and base_len on PA_SYNC_ESTABLISHED / PAST_RECEIVED / BIG_INFO_ADV_REPORT / PER_ADV_REPORT events. The BIG_INFO handler additionally passes &iso_pi(sk)->qos together with sync_handle / bc_num_bis / bc_bis to hci_conn_big_create_sync() while setsockopt may be mutating them.

Acquire lock_sock around the affected accesses in both paths.

The locking order hci_dev_lock -> lock_sock matches the existing iso_conn_big_sync() precedent, whose comment documents the same requirement for hci_conn_big_create_sync(). The HCI connect/bind helpers do not wait for command completion -- they enqueue work via hci_cmd_sync_queue{,_once}() / hci_le_create_cis_pending() and return -- so the added hold time is comparable to iso_conn_big_sync().

KCSAN report:

BUG: KCSAN: data-race in iso_connect_cis / iso_sock_setsockopt

read to 0xffffa3ae8ce3cdc8 of 1 bytes by task 335 on cpu 0:

 iso_connect_cis+0x49f/0xa20
 iso_sock_connect+0x60e/0xb40
 __sys_connect_file+0xbd/0xe0
 __sys_connect+0xe0/0x110
 __x64_sys_connect+0x40/0x50
 x64_sys_call+0xcad/0x1c60
 do_syscall_64+0x133/0x590
 entry_SYSCALL_64_after_hwframe+0x77/0x7f

write to 0xffffa3ae8ce3cdc8 of 60 bytes by task 334 on cpu 1:

 iso_sock_setsockopt+0x69a/0x930
 do_sock_setsockopt+0xc3/0x170
 __sys_setsockopt+0xd1/0x130
 __x64_sys_setsockopt+0x64/0x80
 x64_sys_call+0x1547/0x1c60
 do_syscall_64+0x133/0x590
 entry_SYSCALL_64_after_hwframe+0x77/0x7f

Reported by Kernel Concurrency Sanitizer on:

CPU: 1 UID: 0 PID: 334 Comm: iso_setup_race Not tainted 7.0.0-10949-g8541d8f725c6 #44 PREEMPT(lazy)

The iso_connect_ind() races were found by inspection.

Fixes ccf74f2390d6 ("Bluetooth: Add BTPROTO_ISO socket type")
---
 net/bluetooth/iso.c | 54 +++++++++++++++++++++++++--------------------
 1 file changed, 30 insertions(+), 24 deletions(-)
net/bluetooth/iso.c +30 -24
index 290a1b9a9daa..7cb2864fe872 100644
@@ -347,6 +347,7 @@ static int iso_connect_bis(struct sock *sk)
 		return -EHOSTUNREACH;
 
 	hci_dev_lock(hdev);
+	lock_sock(sk);
 
 	if (!bis_capable(hdev)) {
 		err = -EOPNOTSUPP;
@@ -399,13 +400,9 @@ static int iso_connect_bis(struct sock *sk)
 		goto unlock;
 	}
 
-	lock_sock(sk);
-
 	err = iso_chan_add(conn, sk, NULL);
-	if (err) {
-		release_sock(sk);
+	if (err)
 		goto unlock;
-	}
 
 	/* Update source addr of the socket */
 	bacpy(&iso_pi(sk)->src, &hcon->src);
@@ -421,9 +418,8 @@ static int iso_connect_bis(struct sock *sk)
 		iso_sock_set_timer(sk, READ_ONCE(sk->sk_sndtimeo));
 	}
 
-	release_sock(sk);
-
 unlock:
+	release_sock(sk);
 	hci_dev_unlock(hdev);
 	hci_dev_put(hdev);
 	return err;
@@ -444,6 +440,7 @@ static int iso_connect_cis(struct sock *sk)
 		return -EHOSTUNREACH;
 
 	hci_dev_lock(hdev);
+	lock_sock(sk);
 
 	if (!cis_central_capable(hdev)) {
 		err = -EOPNOTSUPP;
@@ -498,13 +495,9 @@ static int iso_connect_cis(struct sock *sk)
 		goto unlock;
 	}
 
-	lock_sock(sk);
-
 	err = iso_chan_add(conn, sk, NULL);
-	if (err) {
-		release_sock(sk);
+	if (err)
 		goto unlock;
-	}
 
 	/* Update source addr of the socket */
 	bacpy(&iso_pi(sk)->src, &hcon->src);
@@ -520,9 +513,8 @@ static int iso_connect_cis(struct sock *sk)
 		iso_sock_set_timer(sk, READ_ONCE(sk->sk_sndtimeo));
 	}
 
-	release_sock(sk);
-
 unlock:
+	release_sock(sk);
 	hci_dev_unlock(hdev);
 	hci_dev_put(hdev);
 	return err;
@@ -2256,8 +2248,10 @@ int iso_connect_ind(struct hci_dev *hdev, bdaddr_t *bdaddr, __u8 *flags)
 		sk = iso_get_sock(hdev, &hdev->bdaddr, bdaddr, BT_LISTEN,
 				  iso_match_sid, ev1);
 		if (sk && !ev1->status) {
+			lock_sock(sk);
 			iso_pi(sk)->sync_handle = le16_to_cpu(ev1->handle);
 			iso_pi(sk)->bc_sid = ev1->sid;
+			release_sock(sk);
 		}
 
 		goto done;
@@ -2268,8 +2262,10 @@ int iso_connect_ind(struct hci_dev *hdev, bdaddr_t *bdaddr, __u8 *flags)
 		sk = iso_get_sock(hdev, &hdev->bdaddr, bdaddr, BT_LISTEN,
 				  iso_match_sid_past, ev1a);
 		if (sk && !ev1a->status) {
+			lock_sock(sk);
 			iso_pi(sk)->sync_handle = le16_to_cpu(ev1a->sync_handle);
 			iso_pi(sk)->bc_sid = ev1a->sid;
+			release_sock(sk);
 		}
 
 		goto done;
@@ -2296,27 +2292,35 @@ int iso_connect_ind(struct hci_dev *hdev, bdaddr_t *bdaddr, __u8 *flags)
 					  ev2);
 
 		if (sk) {
-			int err;
-			struct hci_conn	*hcon = iso_pi(sk)->conn->hcon;
+			int err = 0;
+			bool big_sync;
+			struct hci_conn *hcon;
 
+			lock_sock(sk);
+
+			hcon = iso_pi(sk)->conn->hcon;
 			iso_pi(sk)->qos.bcast.encryption = ev2->encryption;
 
 			if (ev2->num_bis < iso_pi(sk)->bc_num_bis)
 				iso_pi(sk)->bc_num_bis = ev2->num_bis;
 
-			if (!test_bit(BT_SK_DEFER_SETUP, &bt_sk(sk)->flags) &&
-			    !test_and_set_bit(BT_SK_BIG_SYNC, &iso_pi(sk)->flags)) {
+			big_sync = !test_bit(BT_SK_DEFER_SETUP, &bt_sk(sk)->flags) &&
+				   !test_and_set_bit(BT_SK_BIG_SYNC, &iso_pi(sk)->flags);
+
+			if (big_sync)
 				err = hci_conn_big_create_sync(hdev, hcon,
 							       &iso_pi(sk)->qos,
 							       iso_pi(sk)->sync_handle,
 							       iso_pi(sk)->bc_num_bis,
 							       iso_pi(sk)->bc_bis);
-				if (err) {
-					bt_dev_err(hdev, "hci_le_big_create_sync: %d",
-						   err);
-					sock_put(sk);
-					sk = NULL;
-				}
+
+			release_sock(sk);
+
+			if (big_sync && err) {
+				bt_dev_err(hdev, "hci_le_big_create_sync: %d",
+					   err);
+				sock_put(sk);
+				sk = NULL;
 			}
 		}
 
@@ -2370,8 +2374,10 @@ int iso_connect_ind(struct hci_dev *hdev, bdaddr_t *bdaddr, __u8 *flags)
 			if (!base || base_len > BASE_MAX_LENGTH)
 				goto done;
 
+			lock_sock(sk);
 			memcpy(iso_pi(sk)->base, base, base_len);
 			iso_pi(sk)->base_len = base_len;
+			release_sock(sk);
 		} else {
 			/* This is a PA data fragment. Keep pa_data_len set to 0
 			 * until all data has been reassembled.
Signature · 1 line
2.52.0
#4 lore raw
In reply to SeungJu Cheon
This series was applied to bluetooth/bluetooth-next.git (master) by Luiz Augusto von Dentz <[email protected]>:

Hello:

This series was applied to bluetooth/bluetooth-next.git (master) by Luiz Augusto von Dentz <[email protected]>:

Quoted 6 lines · On Tue, 21 Apr 2026 11:51:20 +0900 you wrote:

Found while auditing iso_pi(sk) field accesses after a KCSAN report. Patch 1/2 is the reported race on iso_pi(sk)->dst in iso_sock_connect(); patch 2/2 covers related races on other iso_pi(sk) fields accessed in iso_connect_{bis,cis}() and iso_connect_ind() that were found by inspection during the same audit.

Quoted 7 lines

Changes in v2:
- Patch 1/2: Use sa->iso_bdaddr directly instead of caching the bacmp() result in a local variable, as suggested by Luiz [1]. This avoids reading from iso_pi(sk) entirely for the broadcast check.

[...]

Here is the summary with links:
- [v2,1/2] Bluetooth: ISO: Fix data-race on dst in iso_sock_connect()

    https://git.kernel.org/bluetooth/bluetooth-next/c/20ca2749b31a
  - [v2,2/2] Bluetooth: ISO: Fix data-race on iso_pi(sk) in socket and HCI event paths
    https://git.kernel.org/bluetooth/bluetooth-next/c/66d4d518020b

You are awesome, thank you!

Signature · 2 lines
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html