linux-next: build failure after merge of the gpio tree

linux-next: build failure after merge of the gpio tree

[Stephen Rothwell]

Hi Linus,

After merging the gpio tree, today's linux-next build (x86_64
allmodconfig) failed like this:

drivers/bluetooth/hci_bcm.c: In function 'bcm_acpi_probe':
drivers/bluetooth/hci_bcm.c:429:9: error: too few arguments to function 'devm_gpiod_get'
  gpio = devm_gpiod_get(&pdev->dev, "device-wakeup");
         ^
In file included from drivers/bluetooth/hci_bcm.c:32:0:
include/linux/gpio/consumer.h:73:32: note: declared here
 struct gpio_desc *__must_check devm_gpiod_get(struct device *dev,
                                ^
drivers/bluetooth/hci_bcm.c:437:9: error: too few arguments to function 'devm_gpiod_get'
  gpio = devm_gpiod_get(&pdev->dev, "shutdown");
         ^
In file included from drivers/bluetooth/hci_bcm.c:32:0:
include/linux/gpio/consumer.h:73:32: note: declared here
 struct gpio_desc *__must_check devm_gpiod_get(struct device *dev,
                                ^

Caused by commit

  b17d1bf16cc7 ("gpio: make flags mandatory for gpiod_get functions")

Interacting with commit

  0395ffc1ee05 ("Bluetooth: hci_bcm: Add PM for BCM devices")

from the bluetooth tree.

I added the following merge fix patch:

From: Stephen Rothwell <sfr@canb.auug.org.au>
Date: Wed, 12 Aug 2015 16:55:58 +1000
Subject: [PATCH] Bluetooth: hci_bcm: fix for devm_gpiod_get API change

Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
---
 drivers/bluetooth/hci_bcm.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/bluetooth/hci_bcm.c b/drivers/bluetooth/hci_bcm.c
index cb7fe837f7d9..c3291a4ab2e8 100644
--- a/drivers/bluetooth/hci_bcm.c
+++ b/drivers/bluetooth/hci_bcm.c
@@ -426,7 +426,7 @@ static int bcm_acpi_probe(struct bcm_device *dev)
 
 	dev->clk = devm_clk_get(&pdev->dev, NULL);
 
-	gpio = devm_gpiod_get(&pdev->dev, "device-wakeup");
+	gpio = devm_gpiod_get(&pdev->dev, "device-wakeup", GPIOD_ASIS);
 	if (!IS_ERR(gpio)) {
 		ret = gpiod_direction_output(gpio, 0);
 		if (ret)
@@ -434,7 +434,7 @@ static int bcm_acpi_probe(struct bcm_device *dev)
 		dev->device_wakeup = gpio;
 	}
 
-	gpio = devm_gpiod_get(&pdev->dev, "shutdown");
+	gpio = devm_gpiod_get(&pdev->dev, "shutdown", GPIOD_ASIS);
 	if (!IS_ERR(gpio)) {
 		ret = gpiod_direction_output(gpio, 0);
 		if (ret)
-- 
2.5.0

-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au

[PATCH] Bluetooth: hci_bcm: improve use of gpios API

[Uwe Kleine-König]

devm_gpiod_get currently has an optional parameter to set initial
direction and value for the requested gpio. Make use of this to simplify
the driver and make it not fail to build when this parameter is made
mandatory (which is scheduled for 4.3-rc1).

Moreover use the _optional variant of devm_gpiod_get to simplify error
handling (which also gets more strict for free).

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
---
Hello,

this is needed on top of commit 0395ffc1ee05 ("Bluetooth: hci_bcm: Add
PM for BCM devices") incombination with b17d1bf16cc7 ("gpio: make flags
mandatory for gpiod_get functions") which is currently sitting in next.

Stephen fixed it up with a simpler patch, which works fine, but doesn't
benefit from the nice things devm_gpiod_get et al offer.

Best regards
Uwe

 drivers/bluetooth/hci_bcm.c | 24 +++++++++---------------
 1 file changed, 9 insertions(+), 15 deletions(-)

diff --git a/drivers/bluetooth/hci_bcm.c b/drivers/bluetooth/hci_bcm.c
index cb7fe837f7d9..6070303418d8 100644
--- a/drivers/bluetooth/hci_bcm.c
+++ b/drivers/bluetooth/hci_bcm.c
@@ -408,7 +408,6 @@ static int bcm_acpi_probe(struct bcm_device *dev)
 {
 	struct platform_device *pdev = dev->pdev;
 	const struct acpi_device_id *id;
-	struct gpio_desc *gpio;
 	struct acpi_device *adev;
 	LIST_HEAD(resources);
 	int ret;
@@ -426,21 +425,16 @@ static int bcm_acpi_probe(struct bcm_device *dev)
 
 	dev->clk = devm_clk_get(&pdev->dev, NULL);
 
-	gpio = devm_gpiod_get(&pdev->dev, "device-wakeup");
-	if (!IS_ERR(gpio)) {
-		ret = gpiod_direction_output(gpio, 0);
-		if (ret)
-			return ret;
-		dev->device_wakeup = gpio;
-	}
+	dev->device_wakeup = devm_gpiod_get_optional(&pdev->dev,
+						     "device-wakeup",
+						     GPIOD_OUT_LOW);
+	if (IS_ERR(dev->device_wakeup))
+		return PTR_ERR(dev->device_wakeup);
 
-	gpio = devm_gpiod_get(&pdev->dev, "shutdown");
-	if (!IS_ERR(gpio)) {
-		ret = gpiod_direction_output(gpio, 0);
-		if (ret)
-			return ret;
-		dev->shutdown = gpio;
-	}
+	dev->shutdown = devm_gpiod_get_optional(&pdev->dev, "shutdown",
+						GPIOD_OUT_LOW);
+	if (IS_ERR(dev->shutdown))
+		return PTR_ERR(dev->shutdown);
 
 	/* Make sure at-least one of the GPIO is defined and that
 	 * a name is specified for this instance
-- 
2.4.6

Re: [PATCH] Bluetooth: hci_bcm: improve use of gpios API

[Stephen Rothwell]

Hi Uwe,

On Wed, 12 Aug 2015 09:20:56 +0200 Uwe Kleine-König <u.kleine-koenig@pengutronix.de> wrote:
>
> devm_gpiod_get currently has an optional parameter to set initial
> direction and value for the requested gpio. Make use of this to simplify
> the driver and make it not fail to build when this parameter is made
> mandatory (which is scheduled for 4.3-rc1).
> 
> Moreover use the _optional variant of devm_gpiod_get to simplify error
> handling (which also gets more strict for free).
> 
> Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
> ---
> Hello,
> 
> this is needed on top of commit 0395ffc1ee05 ("Bluetooth: hci_bcm: Add
> PM for BCM devices") incombination with b17d1bf16cc7 ("gpio: make flags
> mandatory for gpiod_get functions") which is currently sitting in next.
> 
> Stephen fixed it up with a simpler patch, which works fine, but doesn't
> benefit from the nice things devm_gpiod_get et al offer.
> 
> Best regards
> Uwe
> 
>  drivers/bluetooth/hci_bcm.c | 24 +++++++++---------------
>  1 file changed, 9 insertions(+), 15 deletions(-)

I will use this instead of my simple patch tomorrow if the bluetooth
maintainer has not applied it.

-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au

Re: [PATCH] Bluetooth: hci_bcm: improve use of gpios API

[Marcel Holtmann]

Hi Uwe,

> devm_gpiod_get currently has an optional parameter to set initial
> direction and value for the requested gpio. Make use of this to simplify
> the driver and make it not fail to build when this parameter is made
> mandatory (which is scheduled for 4.3-rc1).
> 
> Moreover use the _optional variant of devm_gpiod_get to simplify error
> handling (which also gets more strict for free).
> 
> Signed-off-by: Uwe Kleine-König <u.kleine-koenig-bIcnvbaLZ9MEGnE8C9+IrQ@public.gmane.org>
> ---
> Hello,
> 
> this is needed on top of commit 0395ffc1ee05 ("Bluetooth: hci_bcm: Add
> PM for BCM devices") incombination with b17d1bf16cc7 ("gpio: make flags
> mandatory for gpiod_get functions") which is currently sitting in next.
> 
> Stephen fixed it up with a simpler patch, which works fine, but doesn't
> benefit from the nice things devm_gpiod_get et al offer.

patch has been applied to bluetooth-next tree.

Regards

Marcel