6.14-stable review patch. If anyone has any objections, please let me know.
------------------
From: Luiz Augusto von Dentz luiz.von.dentz@intel.com
[ Upstream commit 3a7fdfb7d876910cfe734488f553dbbc938f8f16 ]
Some controllers seems to generate HCI_EV_LE_DIRECT_ADV_REPORT even when scan_filter is not set to 0x02 or 0x03, which indicates that local privacy is enabled, causing them to be ignored thus breaking auto-connect logic:
< HCI Command: LE Set Scan Parameters (0x08|0x000b) plen 7 Type: Passive (0x00) Interval: 60.000 msec (0x0060) Window: 30.000 msec (0x0030) Own address type: Public (0x00) Filter policy: Ignore not in accept list (0x01) ...
HCI Event: LE Meta Event (0x3e) plen 18
LE Direct Advertising Report (0x0b) Num reports: 1 Event type: Connectable directed - ADV_DIRECT_IND (0x01) Address type: Random (0x01) Address: XX:XX:XX:XX:XX:XX (Static) Direct address type: Random (0x01) Direct address: XX:XX:XX:XX:XX:XX (Non-Resolvable) RSSI: -54 dBm (0xca)
So this attempts to mitigate the above problem by skipping checking of direct_addr if local privacy is not enabled.
Link: https://github.com/bluez/bluez/issues/1138 Fixes: e209e5ccc5ac ("Bluetooth: MGMT: Mark LL Privacy as stable") Signed-off-by: Luiz Augusto von Dentz luiz.von.dentz@intel.com Signed-off-by: Sasha Levin sashal@kernel.org --- net/bluetooth/hci_event.c | 17 ++++++++++------- 1 file changed, 10 insertions(+), 7 deletions(-)
diff --git a/net/bluetooth/hci_event.c b/net/bluetooth/hci_event.c index dd86b7a242b72..e2bfbcee06a80 100644 --- a/net/bluetooth/hci_event.c +++ b/net/bluetooth/hci_event.c @@ -6053,8 +6053,17 @@ static void process_adv_report(struct hci_dev *hdev, u8 type, bdaddr_t *bdaddr, * a LE Direct Advertising Report event. In that case it is * important to see if the address is matching the local * controller address. + * + * If local privacy is not enable the controller shall not be + * generating such event since according to its documentation it is only + * valid for filter_policy 0x02 and 0x03, but the fact that it did + * generate LE Direct Advertising Report means it is probably broken and + * won't generate any other event which can potentially break + * auto-connect logic so in case local privacy is not enable this + * ignores the direct_addr so it works as a regular report. */ - if (!hci_dev_test_flag(hdev, HCI_MESH) && direct_addr) { + if (!hci_dev_test_flag(hdev, HCI_MESH) && direct_addr && + hci_dev_test_flag(hdev, HCI_PRIVACY)) { direct_addr_type = ev_bdaddr_type(hdev, direct_addr_type, &bdaddr_resolved);
@@ -6064,12 +6073,6 @@ static void process_adv_report(struct hci_dev *hdev, u8 type, bdaddr_t *bdaddr, if (!hci_bdaddr_is_rpa(direct_addr, direct_addr_type)) return;
- /* If the controller is not using resolvable random - * addresses, then this report can be ignored. - */ - if (!hci_dev_test_flag(hdev, HCI_PRIVACY)) - return; - /* If the local IRK of the controller does not match * with the resolvable random address provided, then * this report can be ignored.