Re: [PATCH v1] Bluetooth: hci_qca: Make bt_en and susclk not mandatory for QCA Rome

[Rocky Liao <rjliao@codeaurora.org>]

在 2020-03-04 16:53，Johan Hovold 写道：
> On Wed, Mar 04, 2020 at 09:54:29AM +0800, Rocky Liao wrote:
>> On some platforms the bt_en pin and susclk are default on and there
>> is no exposed resource to control them. This patch makes the bt_en
>> and susclk not mandatory to have BT work. It also will not set the
>> HCI_QUIRK_NON_PERSISTENT_SETUP and shutdown() callback if bt_en is
>> not available.
>> 
>> Signed-off-by: Rocky Liao <rjliao@codeaurora.org>
>> ---
>>  drivers/bluetooth/hci_qca.c | 47 
>> ++++++++++++++++++++-----------------
>>  1 file changed, 26 insertions(+), 21 deletions(-)
>> 
>> diff --git a/drivers/bluetooth/hci_qca.c b/drivers/bluetooth/hci_qca.c
>> index bf436d6e638e..325baa046c3a 100644
>> --- a/drivers/bluetooth/hci_qca.c
>> +++ b/drivers/bluetooth/hci_qca.c
>> @@ -1562,9 +1562,11 @@ static int qca_power_on(struct hci_dev *hdev)
>>  		ret = qca_wcn3990_init(hu);
>>  	} else {
>>  		qcadev = serdev_device_get_drvdata(hu->serdev);
>> -		gpiod_set_value_cansleep(qcadev->bt_en, 1);
>> -		/* Controller needs time to bootup. */
>> -		msleep(150);
>> +		if (!IS_ERR(qcadev->bt_en)) {
>> +			gpiod_set_value_cansleep(qcadev->bt_en, 1);
>> +			/* Controller needs time to bootup. */
>> +			msleep(150);
>> +		}
>>  	}
>> 
>>  	return ret;
>> @@ -1750,7 +1752,7 @@ static void qca_power_shutdown(struct hci_uart 
>> *hu)
>>  		host_set_baudrate(hu, 2400);
>>  		qca_send_power_pulse(hu, false);
>>  		qca_regulator_disable(qcadev);
>> -	} else {
>> +	} else if (!IS_ERR(qcadev->bt_en)) {
>>  		gpiod_set_value_cansleep(qcadev->bt_en, 0);
>>  	}
>>  }
>> @@ -1852,6 +1854,7 @@ static int qca_serdev_probe(struct serdev_device 
>> *serdev)
>>  	struct hci_dev *hdev;
>>  	const struct qca_vreg_data *data;
>>  	int err;
>> +	bool power_ctrl_enabled = true;
>> 
>>  	qcadev = devm_kzalloc(&serdev->dev, sizeof(*qcadev), GFP_KERNEL);
>>  	if (!qcadev)
>> @@ -1901,35 +1904,37 @@ static int qca_serdev_probe(struct 
>> serdev_device *serdev)
>>  		qcadev->bt_en = devm_gpiod_get(&serdev->dev, "enable",
>>  					       GPIOD_OUT_LOW);
> 
> Shouldn't you use devm_gpiod_get_optional()?
> 
>>  		if (IS_ERR(qcadev->bt_en)) {
>> -			dev_err(&serdev->dev, "failed to acquire enable gpio\n");
>> -			return PTR_ERR(qcadev->bt_en);
>> +			dev_warn(&serdev->dev, "failed to acquire enable gpio\n");
>> +			power_ctrl_enabled = false;
> 
> And bail out on errors, but treat NULL as !port_ctrl_enabled?
> 
>>  		}
>> 
>>  		qcadev->susclk = devm_clk_get(&serdev->dev, NULL);
> 
> And devm_clk_get_optional() here.
> 
> Etc.
> 
> Also does the devicetree binding need to be updated to reflect this
> change?
> 
OK, I will update to use the suggested API.
The devicetress binding doesn't need to be updated as they are already 
optional there.

> Johan