From: Sean Wang sean.wang@mediatek.com
[ Upstream commit db3f1f9b5d88d8d7f9eaa486f71784dd319285ff ]
Currently, there is a loop in btmtksdio_txrx_work() which iteratively executes until the variable int_status is zero.
But the variable int_status should be masked out with the actual interrupt sources (MTK_REG_CHISR bit 0-15) before we check the loop condition. Otherwise, RX_PKT_LEN (MTK_REG_CHISR bit 16-31) which is read-only and unclearable would cause the loop to get stuck on some chipsets like MT7663s.
Fixes: 26270bc189ea ("Bluetooth: btmtksdio: move interrupt service to work") Signed-off-by: Sean Wang sean.wang@mediatek.com Signed-off-by: Marcel Holtmann marcel@holtmann.org Signed-off-by: Sasha Levin sashal@kernel.org --- drivers/bluetooth/btmtksdio.c | 2 ++ 1 file changed, 2 insertions(+)
diff --git a/drivers/bluetooth/btmtksdio.c b/drivers/bluetooth/btmtksdio.c index 72e00264d9f1..86a52eb77e01 100644 --- a/drivers/bluetooth/btmtksdio.c +++ b/drivers/bluetooth/btmtksdio.c @@ -92,6 +92,7 @@ MODULE_DEVICE_TABLE(sdio, btmtksdio_table); #define TX_EMPTY BIT(2) #define TX_FIFO_OVERFLOW BIT(8) #define FW_MAILBOX_INT BIT(15) +#define INT_MASK GENMASK(15, 0) #define RX_PKT_LEN GENMASK(31, 16)
#define MTK_REG_CSICR 0xc0 @@ -565,6 +566,7 @@ static void btmtksdio_txrx_work(struct work_struct *work) * FIFO. */ sdio_writel(bdev->func, int_status, MTK_REG_CHISR, NULL); + int_status &= INT_MASK;
if ((int_status & FW_MAILBOX_INT) && bdev->data->chipid == 0x7921) {