diff mbox series

[v4,1/2] Bluetooth: hci_sync: Add a new quirk to skip HCI_FLT_CLEAR_ALL

Message ID 20220307200445.5554-1-swyterzone@gmail.com
State Accepted
Commit 94ea164ef6ec96c269fe50eea34c2449fe8df299
Headers show
Series [v4,1/2] Bluetooth: hci_sync: Add a new quirk to skip HCI_FLT_CLEAR_ALL | expand

Commit Message

Ismael Ferreras Morezuelas March 7, 2022, 8:04 p.m. UTC
Some controllers have problems with being sent a command to clear
all filtering. While the HCI code does not unconditionally
send a clear-all anymore at BR/EDR setup (after the state machine
refactor), there might be more ways of hitting these codepaths
in the future as the kernel develops.

Cc: stable@vger.kernel.org
Cc: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Ismael Ferreras Morezuelas <swyterzone@gmail.com>
---
 include/net/bluetooth/hci.h | 10 ++++++++++
 net/bluetooth/hci_sync.c    | 16 ++++++++++++++++
 2 files changed, 26 insertions(+)

Comments

Marcel Holtmann March 10, 2022, 9:56 a.m. UTC | #1
Hi Ismael,

> Some controllers have problems with being sent a command to clear
> all filtering. While the HCI code does not unconditionally
> send a clear-all anymore at BR/EDR setup (after the state machine
> refactor), there might be more ways of hitting these codepaths
> in the future as the kernel develops.
> 
> Cc: stable@vger.kernel.org
> Cc: Hans de Goede <hdegoede@redhat.com>
> Signed-off-by: Ismael Ferreras Morezuelas <swyterzone@gmail.com>
> ---
> include/net/bluetooth/hci.h | 10 ++++++++++
> net/bluetooth/hci_sync.c    | 16 ++++++++++++++++
> 2 files changed, 26 insertions(+)

both patches have been applied to bluetooth-next tree.

Regards

Marcel
diff mbox series

Patch

diff --git a/include/net/bluetooth/hci.h b/include/net/bluetooth/hci.h
index 35c073d44ec5..5cb095b09a94 100644
--- a/include/net/bluetooth/hci.h
+++ b/include/net/bluetooth/hci.h
@@ -255,6 +255,16 @@  enum {
 	 * during the hdev->setup vendor callback.
 	 */
 	HCI_QUIRK_BROKEN_READ_TRANSMIT_POWER,
+
+	/* When this quirk is set, HCI_OP_SET_EVENT_FLT requests with
+	 * HCI_FLT_CLEAR_ALL are ignored and event filtering is
+	 * completely avoided. A subset of the CSR controller
+	 * clones struggle with this and instantly lock up.
+	 *
+	 * Note that devices using this must (separately) disable
+	 * runtime suspend, because event filtering takes place there.
+	 */
+	HCI_QUIRK_BROKEN_FILTER_CLEAR_ALL,
 };
 
 /* HCI device flags */
diff --git a/net/bluetooth/hci_sync.c b/net/bluetooth/hci_sync.c
index e31d1150dc71..c3bdaf2de511 100644
--- a/net/bluetooth/hci_sync.c
+++ b/net/bluetooth/hci_sync.c
@@ -2812,6 +2812,9 @@  static int hci_set_event_filter_sync(struct hci_dev *hdev, u8 flt_type,
 	if (!hci_dev_test_flag(hdev, HCI_BREDR_ENABLED))
 		return 0;
 
+	if (test_bit(HCI_QUIRK_BROKEN_FILTER_CLEAR_ALL, &hdev->quirks))
+		return 0;
+
 	memset(&cp, 0, sizeof(cp));
 	cp.flt_type = flt_type;
 
@@ -2832,6 +2835,13 @@  static int hci_clear_event_filter_sync(struct hci_dev *hdev)
 	if (!hci_dev_test_flag(hdev, HCI_EVENT_FILTER_CONFIGURED))
 		return 0;
 
+	/* In theory the state machine should not reach here unless
+	 * a hci_set_event_filter_sync() call succeeds, but we do
+	 * the check both for parity and as a future reminder.
+	 */
+	if (test_bit(HCI_QUIRK_BROKEN_FILTER_CLEAR_ALL, &hdev->quirks))
+		return 0;
+
 	return hci_set_event_filter_sync(hdev, HCI_FLT_CLEAR_ALL, 0x00,
 					 BDADDR_ANY, 0x00);
 }
@@ -4831,6 +4841,12 @@  static int hci_update_event_filter_sync(struct hci_dev *hdev)
 	if (!hci_dev_test_flag(hdev, HCI_BREDR_ENABLED))
 		return 0;
 
+	/* Some fake CSR controllers lock up after setting this type of
+	 * filter, so avoid sending the request altogether.
+	 */
+	if (test_bit(HCI_QUIRK_BROKEN_FILTER_CLEAR_ALL, &hdev->quirks))
+		return 0;
+
 	/* Always clear event filter when starting */
 	hci_clear_event_filter_sync(hdev);