Received-SPF: pass (google.com: domain of linux-bluetooth-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
From:   Sonny Sasaka <sonnysasaka@chromium.org>
To:     linux-bluetooth@vger.kernel.org
Cc:     Sonny Sasaka <sonnysasaka@chromium.org>
Subject: [PATCH] Bluetooth: Handle Inquiry Cancel error after Inquiry Complete
Date:   Wed,  6 May 2020 12:55:03 -0700
Message-Id: <20200506195503.5837-1-sonnysasaka@chromium.org>
In-Reply-To: <89566A98-F0CE-4E70-A344-915E81FB0C14@holtmann.org>
References: <89566A98-F0CE-4E70-A344-915E81FB0C14@holtmann.org>
Sender: linux-bluetooth-owner@vger.kernel.org
Precedence: bulk

After sending Inquiry Cancel command to the controller, it is possible
that Inquiry Complete event comes before Inquiry Cancel command complete
event. In this case the Inquiry Cancel command will have status of
Command Disallowed since there is no Inquiry session to be cancelled.
This case should not be treated as error, otherwise we can reach an
inconsistent state.

Example of a btmon trace when this happened:

< HCI Command: Inquiry Cancel (0x01|0x0002) plen 0
> HCI Event: Inquiry Complete (0x01) plen 1
        Status: Success (0x00)
> HCI Event: Command Complete (0x0e) plen 4
      Inquiry Cancel (0x01|0x0002) ncmd 1
        Status: Command Disallowed (0x0c)
---
 net/bluetooth/hci_event.c | 20 ++++++++++++++++++--
 1 file changed, 18 insertions(+), 2 deletions(-)

diff --git a/net/bluetooth/hci_event.c b/net/bluetooth/hci_event.c
index 3e7badb3ac2d..e8047175de10 100644
--- a/net/bluetooth/hci_event.c
+++ b/net/bluetooth/hci_event.c
@@ -41,12 +41,28 @@
 
 /* Handle HCI Event packets */
 
-static void hci_cc_inquiry_cancel(struct hci_dev *hdev, struct sk_buff *skb)
+static void hci_cc_inquiry_cancel(struct hci_dev *hdev, struct sk_buff *skb,
+				  u8 *new_status)
 {
 	__u8 status = *((__u8 *) skb->data);
 
 	BT_DBG("%s status 0x%2.2x", hdev->name, status);
 
+	/* It is possible that we receive Inquiry Complete event right
+	 * before we receive Inquiry Cancel Command Complete event, in
+	 * which case the latter event should have status of Command
+	 * Disallowed (0x0c). This should not be treated as error, since
+	 * we actually achieve what Inquiry Cancel wants to achieve,
+	 * which is to end the last Inquiry session.
+	 */
+	if (status == 0x0c && !test_bit(HCI_INQUIRY, &hdev->flags)) {
+		bt_dev_err(hdev, "Ignoring error of "
+				 "HCI Inquiry Cancel command");
+		status = 0x00;
+	}
+
+	*new_status = status;
+
 	if (status)
 		return;
 
@@ -3036,7 +3052,7 @@ static void hci_cmd_complete_evt(struct hci_dev *hdev, struct sk_buff *skb,
 
 	switch (*opcode) {
 	case HCI_OP_INQUIRY_CANCEL:
-		hci_cc_inquiry_cancel(hdev, skb);
+		hci_cc_inquiry_cancel(hdev, skb, status);
 		break;
 
 	case HCI_OP_PERIODIC_INQ:
-- 
2.17.1