[PATCH] Bluetooth: avoid silent hci_bcm ACPI PM regression

[PATCH] Bluetooth: avoid silent hci_bcm ACPI PM regression

[Johan Hovold]

The hci_bcm platform-device hack which was used to implement
power management for ACPI devices is being replaced by a
serial-device-bus implementation.

Unfortunately, when the corresponding change to the ACPI code lands (a
change that will stop enumerating and registering the serial-device-node
child as a platform device) PM will break silently unless serdev
TTY-port controller support has been enabled. Specifically, hciattach
(btattach) would still succeed, but power management would no longer
work.

Although this is strictly a runtime dependency, let's make the driver
depend on SERIAL_DEV_CTRL_TTYPORT, which is the particular serdev
controller implementation used by the ACPI devices currently managed by
this driver, to avoid breaking PM without anyone noticing.

Note that the driver already has a (build-time) dependency on the serdev
bus code.

Signed-off-by: Johan Hovold <johan@kernel.org>
---
 drivers/bluetooth/Kconfig | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/bluetooth/Kconfig b/drivers/bluetooth/Kconfig
index fae5a74dc737..082e1c7329de 100644
--- a/drivers/bluetooth/Kconfig
+++ b/drivers/bluetooth/Kconfig
@@ -169,6 +169,7 @@ config BT_HCIUART_BCM
 	bool "Broadcom protocol support"
 	depends on BT_HCIUART
 	depends on BT_HCIUART_SERDEV
+	depends on (!ACPI || SERIAL_DEV_CTRL_TTYPORT)
 	select BT_HCIUART_H4
 	select BT_BCM
 	help
-- 
2.14.2

Re: [PATCH] Bluetooth: avoid silent hci_bcm ACPI PM regression

[Marcel Holtmann]

Hi Johan,

> The hci_bcm platform-device hack which was used to implement
> power management for ACPI devices is being replaced by a
> serial-device-bus implementation.
> 
> Unfortunately, when the corresponding change to the ACPI code lands (a
> change that will stop enumerating and registering the serial-device-node
> child as a platform device) PM will break silently unless serdev
> TTY-port controller support has been enabled. Specifically, hciattach
> (btattach) would still succeed, but power management would no longer
> work.
> 
> Although this is strictly a runtime dependency, let's make the driver
> depend on SERIAL_DEV_CTRL_TTYPORT, which is the particular serdev
> controller implementation used by the ACPI devices currently managed by
> this driver, to avoid breaking PM without anyone noticing.
> 
> Note that the driver already has a (build-time) dependency on the serdev
> bus code.
> 
> Signed-off-by: Johan Hovold <johan@kernel.org>
> ---
> drivers/bluetooth/Kconfig | 1 +
> 1 file changed, 1 insertion(+)

patch has been applied to bluetooth-next tree.

Regards

Marcel