[PATCH] Bluetooth: btusb: Disable runtime suspend on Realtek devices

[PATCH] Bluetooth: btusb: Disable runtime suspend on Realtek devices

[Kai-Heng Feng]

After commit 9e45524a0111 ("Bluetooth: btusb: Fix suspend issue for
Realtek devices") both WiFi and Bluetooth stop working after reboot:
[   34.322617] usb 1-8: reset full-speed USB device number 3 using xhci_hcd
[   34.450401] usb 1-8: device descriptor read/64, error -71
[   34.694375] usb 1-8: device descriptor read/64, error -71
...
[   44.599111] rtw_pci 0000:02:00.0: failed to poll offset=0x5 mask=0x3 value=0x0
[   44.599113] rtw_pci 0000:02:00.0: mac power on failed
[   44.599114] rtw_pci 0000:02:00.0: failed to power on mac
[   44.599114] rtw_pci 0000:02:00.0: leave idle state failed
[   44.599492] rtw_pci 0000:02:00.0: failed to leave ips state
[   44.599493] rtw_pci 0000:02:00.0: failed to leave idle state

That commit removed USB_QUIRK_RESET_RESUME, which not only resets the USB
device after resume, it also prevents the device from being runtime
suspended by USB core. My experiment shows if the Realtek btusb device
ever runtime suspends once, the entire wireless module becomes useless
after reboot.

So let's explicitly disable runtime suspend on Realtek btusb device for
now.

Fixes: 9e45524a0111 ("Bluetooth: btusb: Fix suspend issue for Realtek devices")
Signed-off-by: Kai-Heng Feng <kai.heng.feng@canonical.com>
---
 drivers/bluetooth/btusb.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/drivers/bluetooth/btusb.c b/drivers/bluetooth/btusb.c
index d9cd0677d41c..0eaeca0a64fb 100644
--- a/drivers/bluetooth/btusb.c
+++ b/drivers/bluetooth/btusb.c
@@ -3832,6 +3832,10 @@ static int btusb_probe(struct usb_interface *intf,
 		 * (DEVICE_REMOTE_WAKEUP)
 		 */
 		set_bit(BTUSB_WAKEUP_DISABLE, &data->flags);
+
+		err = usb_autopm_get_interface(intf);
+		if (err < 0)
+			goto out_free_dev;
 	}
 
 	if (id->driver_info & BTUSB_AMP) {
-- 
2.17.1

Re: [PATCH] Bluetooth: btusb: Disable runtime suspend on Realtek devices

[Marcel Holtmann]

Hi Kai-Heng,

> After commit 9e45524a0111 ("Bluetooth: btusb: Fix suspend issue for
> Realtek devices") both WiFi and Bluetooth stop working after reboot:
> [   34.322617] usb 1-8: reset full-speed USB device number 3 using xhci_hcd
> [   34.450401] usb 1-8: device descriptor read/64, error -71
> [   34.694375] usb 1-8: device descriptor read/64, error -71
> ...
> [   44.599111] rtw_pci 0000:02:00.0: failed to poll offset=0x5 mask=0x3 value=0x0
> [   44.599113] rtw_pci 0000:02:00.0: mac power on failed
> [   44.599114] rtw_pci 0000:02:00.0: failed to power on mac
> [   44.599114] rtw_pci 0000:02:00.0: leave idle state failed
> [   44.599492] rtw_pci 0000:02:00.0: failed to leave ips state
> [   44.599493] rtw_pci 0000:02:00.0: failed to leave idle state
> 
> That commit removed USB_QUIRK_RESET_RESUME, which not only resets the USB
> device after resume, it also prevents the device from being runtime
> suspended by USB core. My experiment shows if the Realtek btusb device
> ever runtime suspends once, the entire wireless module becomes useless
> after reboot.
> 
> So let's explicitly disable runtime suspend on Realtek btusb device for
> now.
> 
> Fixes: 9e45524a0111 ("Bluetooth: btusb: Fix suspend issue for Realtek devices")
> Signed-off-by: Kai-Heng Feng <kai.heng.feng@canonical.com>
> ---
> drivers/bluetooth/btusb.c | 4 ++++
> 1 file changed, 4 insertions(+)

patch has been applied to bluetooth-next tree.

Regards

Marcel