linux-kernel - [PATCH 5.10 023/663] Bluetooth: hci_uart: Fix a race for write

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Mon,  1 Mar 2021 17:04:31 +0100
From:   Greg Kroah-Hartman <gregkh@...uxfoundation.org>
To:     linux-kernel@...r.kernel.org
Cc:     Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        stable@...r.kernel.org, Claire Chang <tientzu@...omium.org>,
        Marcel Holtmann <marcel@...tmann.org>,
        Sasha Levin <sashal@...nel.org>
Subject: [PATCH 5.10 023/663] Bluetooth: hci_uart: Fix a race for write_work scheduling

From: Claire Chang <tientzu@...omium.org>

[ Upstream commit afe0b1c86458f121b085271e4f3034017a90d4a3 ]

In hci_uart_write_work, there is a loop/goto checking the value of
HCI_UART_TX_WAKEUP. If HCI_UART_TX_WAKEUP is set again, it keeps trying
hci_uart_dequeue; otherwise, it clears HCI_UART_SENDING and returns.

In hci_uart_tx_wakeup, if HCI_UART_SENDING is already set, it sets
HCI_UART_TX_WAKEUP, skips schedule_work and assumes the running/pending
hci_uart_write_work worker will do hci_uart_dequeue properly.

However, if the HCI_UART_SENDING check in hci_uart_tx_wakeup is done after
the loop breaks, but before HCI_UART_SENDING is cleared in
hci_uart_write_work, the schedule_work is skipped incorrectly.

Fix this race by changing the order of HCI_UART_SENDING and
HCI_UART_TX_WAKEUP modification.

Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2")
Fixes: 82f5169bf3d3 ("Bluetooth: hci_uart: add serdev driver support library")
Signed-off-by: Claire Chang <tientzu@...omium.org>
Signed-off-by: Marcel Holtmann <marcel@...tmann.org>
Signed-off-by: Sasha Levin <sashal@...nel.org>
---
 drivers/bluetooth/hci_ldisc.c  | 7 +++----
 drivers/bluetooth/hci_serdev.c | 4 ++--
 2 files changed, 5 insertions(+), 6 deletions(-)

diff --git a/drivers/bluetooth/hci_ldisc.c b/drivers/bluetooth/hci_ldisc.c
index f83d67eafc9f0..8be4d807d1370 100644
--- a/drivers/bluetooth/hci_ldisc.c
+++ b/drivers/bluetooth/hci_ldisc.c
@@ -127,10 +127,9 @@ int hci_uart_tx_wakeup(struct hci_uart *hu)
 	if (!test_bit(HCI_UART_PROTO_READY, &hu->flags))
 		goto no_schedule;
 
-	if (test_and_set_bit(HCI_UART_SENDING, &hu->tx_state)) {
-		set_bit(HCI_UART_TX_WAKEUP, &hu->tx_state);
+	set_bit(HCI_UART_TX_WAKEUP, &hu->tx_state);
+	if (test_and_set_bit(HCI_UART_SENDING, &hu->tx_state))
 		goto no_schedule;
-	}
 
 	BT_DBG("");
 
@@ -174,10 +173,10 @@ restart:
 		kfree_skb(skb);
 	}
 
+	clear_bit(HCI_UART_SENDING, &hu->tx_state);
 	if (test_bit(HCI_UART_TX_WAKEUP, &hu->tx_state))
 		goto restart;
 
-	clear_bit(HCI_UART_SENDING, &hu->tx_state);
 	wake_up_bit(&hu->tx_state, HCI_UART_SENDING);
 }
 
diff --git a/drivers/bluetooth/hci_serdev.c b/drivers/bluetooth/hci_serdev.c
index ef96ad06fa54e..9e03402ef1b37 100644
--- a/drivers/bluetooth/hci_serdev.c
+++ b/drivers/bluetooth/hci_serdev.c
@@ -83,9 +83,9 @@ static void hci_uart_write_work(struct work_struct *work)
 			hci_uart_tx_complete(hu, hci_skb_pkt_type(skb));
 			kfree_skb(skb);
 		}
-	} while (test_bit(HCI_UART_TX_WAKEUP, &hu->tx_state));
 
-	clear_bit(HCI_UART_SENDING, &hu->tx_state);
+		clear_bit(HCI_UART_SENDING, &hu->tx_state);
+	} while (test_bit(HCI_UART_TX_WAKEUP, &hu->tx_state));
 }
 
 /* ------- Interface to HCI layer ------ */
-- 
2.27.0