Both hci_find_adv_instance and hci_remove_adv_instance have a comment
above their function definition saying that these two functions require
the caller to hold the hdev->lock lock. However, hci_le_ext_adv_term_evt
does not acquire that lock and neither does its caller hci_le_meta_evt
(hci_le_meta_evt calls hci_le_ext_adv_term_evt via an indirect function
call because of the lookup in hci_le_ev_table).
The other event handlers all acquire and release the hdev->lock and they
follow the rule that hci_find_adv_instance and hci_remove_adv_instance
must be called while holding the hdev->lock lock.
The solution is to make sure hci_le_ext_adv_term_evt also acquires and
releases the hdev->lock lock. The check on ev->status which logs a
warning and does an early return is not covered by the lock because
other functions also access ev->status without holding the lock.
Signed-off-by: Niels Dossche <[email protected]>
---
net/bluetooth/hci_event.c | 19 ++++++++++++-------
1 file changed, 12 insertions(+), 7 deletions(-)
diff --git a/net/bluetooth/hci_event.c b/net/bluetooth/hci_event.c
index fc30f4c03d29..3bf048d0df37 100644
--- a/net/bluetooth/hci_event.c
+++ b/net/bluetooth/hci_event.c
@@ -5670,8 +5670,6 @@ static void hci_le_ext_adv_term_evt(struct hci_dev *hdev, void *data,
bt_dev_dbg(hdev, "status 0x%2.2x", ev->status);
- adv = hci_find_adv_instance(hdev, ev->handle);
-
/* The Bluetooth Core 5.3 specification clearly states that this event
* shall not be sent when the Host disables the advertising set. So in
* case of HCI_ERROR_CANCELLED_BY_HOST, just ignore the event.
@@ -5684,9 +5682,13 @@ static void hci_le_ext_adv_term_evt(struct hci_dev *hdev, void *data,
return;
}
+ hci_dev_lock(hdev);
+
+ adv = hci_find_adv_instance(hdev, ev->handle);
+
if (ev->status) {
if (!adv)
- return;
+ goto unlock;
/* Remove advertising as it has been terminated */
hci_remove_adv_instance(hdev, ev->handle);
@@ -5694,12 +5696,12 @@ static void hci_le_ext_adv_term_evt(struct hci_dev *hdev, void *data,
list_for_each_entry_safe(adv, n, &hdev->adv_instances, list) {
if (adv->enabled)
- return;
+ goto unlock;
}
/* We are no longer advertising, clear HCI_LE_ADV */
hci_dev_clear_flag(hdev, HCI_LE_ADV);
- return;
+ goto unlock;
}
if (adv)
@@ -5714,16 +5716,19 @@ static void hci_le_ext_adv_term_evt(struct hci_dev *hdev, void *data,
if (hdev->adv_addr_type != ADDR_LE_DEV_RANDOM ||
bacmp(&conn->resp_addr, BDADDR_ANY))
- return;
+ goto unlock;
if (!ev->handle) {
bacpy(&conn->resp_addr, &hdev->random_addr);
- return;
+ goto unlock;
}
if (adv)
bacpy(&conn->resp_addr, &adv->random_addr);
}
+
+unlock:
+ hci_dev_unlock(hdev);
}
static void hci_le_conn_update_complete_evt(struct hci_dev *hdev, void *data,
--
2.35.1
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=612647
---Test result---
Test Summary:
CheckPatch PASS 1.60 seconds
GitLint PASS 1.04 seconds
SubjectPrefix PASS 0.85 seconds
BuildKernel PASS 29.24 seconds
BuildKernel32 PASS 25.92 seconds
Incremental Build with patchesPASS 36.10 seconds
TestRunner: Setup PASS 457.23 seconds
TestRunner: l2cap-tester PASS 12.96 seconds
TestRunner: bnep-tester PASS 6.09 seconds
TestRunner: mgmt-tester PASS 98.82 seconds
TestRunner: rfcomm-tester PASS 6.91 seconds
TestRunner: sco-tester PASS 7.28 seconds
TestRunner: smp-tester PASS 6.88 seconds
TestRunner: userchan-tester PASS 6.17 seconds
---
Regards,
Linux Bluetooth
Hi Niels,
> Both hci_find_adv_instance and hci_remove_adv_instance have a comment
> above their function definition saying that these two functions require
> the caller to hold the hdev->lock lock. However, hci_le_ext_adv_term_evt
> does not acquire that lock and neither does its caller hci_le_meta_evt
> (hci_le_meta_evt calls hci_le_ext_adv_term_evt via an indirect function
> call because of the lookup in hci_le_ev_table).
>
> The other event handlers all acquire and release the hdev->lock and they
> follow the rule that hci_find_adv_instance and hci_remove_adv_instance
> must be called while holding the hdev->lock lock.
>
> The solution is to make sure hci_le_ext_adv_term_evt also acquires and
> releases the hdev->lock lock. The check on ev->status which logs a
> warning and does an early return is not covered by the lock because
> other functions also access ev->status without holding the lock.
>
> Signed-off-by: Niels Dossche <[email protected]>
> ---
> net/bluetooth/hci_event.c | 19 ++++++++++++-------
> 1 file changed, 12 insertions(+), 7 deletions(-)
patch has been applied to bluetooth-next tree.
Regards
Marcel