Commit
cde1a8a99287 ("Bluetooth: btusb: Fix and detect most of the
Chinese Bluetooth controllers") made the detection of fake controllers
more generic fixing it for much of the newer fakes / clones.
But this does not work for a fake CSR controller with a bcdDevice
value of 0x0134, which was correctly identified as fake before
this change.
Add an extra check for this special case, checking for a combination
of a bcdDevice value of 0x0134, together with a lmp_subver of 0x0c5c
and a hci_ver of BLUETOOTH_VER_2_0.
The chip inside this fake dongle is marked as with "clockwise cw6629d".
Fixes:
cde1a8a99287 ("Bluetooth: btusb: Fix and detect most of the Chinese Bluetooth controllers")
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
static int btusb_setup_csr(struct hci_dev *hdev)
{
+ struct btusb_data *data = hci_get_drvdata(hdev);
+ u16 bcdDevice = le16_to_cpu(data->udev->descriptor.bcdDevice);
struct hci_rp_read_local_version *rp;
struct sk_buff *skb;
bool is_fake = false;
le16_to_cpu(rp->hci_ver) > BLUETOOTH_VER_4_0)
is_fake = true;
+ /* Other clones which beat all the above checks */
+ else if (bcdDevice == 0x0134 &&
+ le16_to_cpu(rp->lmp_subver) == 0x0c5c &&
+ le16_to_cpu(rp->hci_ver) == BLUETOOTH_VER_2_0)
+ is_fake = true;
+
if (is_fake) {
bt_dev_warn(hdev, "CSR: Unbranded CSR clone detected; adding workarounds...");