From: Sean Wang <sean.wang@mediatek.com>

Date: Wed, 12 Jan 2022 15:39:17 +0800

Subject: Bluetooth: btmtksdio: mask out interrupt status

Patch-mainline: v5.18-rc1

Git-commit: db3f1f9b5d88d8d7f9eaa486f71784dd319285ff

References: jsc#PED-1407

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>

Acked-by: Lee, Chun-Yi <jlee@suse.com>

---

 drivers/bluetooth/btmtksdio.c |    2 ++

 1 file changed, 2 insertions(+)

--- a/drivers/bluetooth/btmtksdio.c

+++ b/drivers/bluetooth/btmtksdio.c

@@ -92,6 +92,7 @@ MODULE_DEVICE_TABLE(sdio, btmtksdio_tabl

 #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 w

 		 * 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) {