[PATCH v5 33/48] staging: nvec: Register with kernel power-off handler

[PATCH v5 33/48] staging: nvec: Register with kernel power-off handler

[Guenter Roeck]

Register with kernel power-off handler instead of setting pm_power_off
directly. Register with default priority since we don't know any better.

Cc: Julian Andres Klode <jak@jak-linux.org>
Cc: Marc Dietrich <marvin24@gmx.de>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
---
v5:
- Rebase to v3.18-rc3
v4:
- Do not use notifiers but internal functions and data structures to manage
  the list of power-off handlers. Drop unused parameters from callbacks, and
  make the power-off function type void
v3:
- Replace poweroff in all newly introduced variables and in text
  with power_off or power-off as appropriate
- Replace POWEROFF_PRIORITY_xxx with POWER_OFF_PRIORITY_xxx
v2:
- Use define to specify poweroff handler priority
- Use dev_warn instead of dev_err

 drivers/staging/nvec/nvec.c | 21 ++++++++++++---------
 drivers/staging/nvec/nvec.h |  2 ++
 2 files changed, 14 insertions(+), 9 deletions(-)

diff --git a/drivers/staging/nvec/nvec.c b/drivers/staging/nvec/nvec.c
index a93208a..acf2ea6 100644
--- a/drivers/staging/nvec/nvec.c
+++ b/drivers/staging/nvec/nvec.c
@@ -33,6 +33,7 @@
 #include <linux/mfd/core.h>
 #include <linux/mutex.h>
 #include <linux/notifier.h>
+#include <linux/pm.h>
 #include <linux/slab.h>
 #include <linux/spinlock.h>
 #include <linux/workqueue.h>
@@ -80,8 +81,6 @@ enum nvec_sleep_subcmds {
 #define LID_SWITCH BIT(1)
 #define PWR_BUTTON BIT(15)
 
-static struct nvec_chip *nvec_power_handle;
-
 static const struct mfd_cell nvec_devices[] = {
 	{
 		.name = "nvec-kbd",
@@ -759,12 +758,14 @@ static void nvec_disable_i2c_slave(struct nvec_chip *nvec)
 }
 #endif
 
-static void nvec_power_off(void)
+static void nvec_power_off(struct power_off_handler_block *this)
 {
+	struct nvec_chip *nvec = container_of(this, struct nvec_chip,
+					      power_off_hb);
 	char ap_pwr_down[] = { NVEC_SLEEP, AP_PWR_DOWN };
 
-	nvec_toggle_global_events(nvec_power_handle, false);
-	nvec_write_async(nvec_power_handle, ap_pwr_down, 2);
+	nvec_toggle_global_events(nvec, false);
+	nvec_write_async(nvec, ap_pwr_down, 2);
 }
 
 /*
@@ -878,8 +879,11 @@ static int tegra_nvec_probe(struct platform_device *pdev)
 	nvec->nvec_status_notifier.notifier_call = nvec_status_notifier;
 	nvec_register_notifier(nvec, &nvec->nvec_status_notifier, 0);
 
-	nvec_power_handle = nvec;
-	pm_power_off = nvec_power_off;
+	nvec->power_off_hb.handler = nvec_power_off;
+	nvec->power_off_hb.priority = POWER_OFF_PRIORITY_DEFAULT;
+	ret = register_power_off_handler(&nvec->power_off_hb);
+	if (ret)
+		dev_warn(nvec->dev, "Failed to register power-off handler\n");
 
 	/* Get Firmware Version */
 	msg = nvec_write_sync(nvec, get_firmware_version, 2);
@@ -914,13 +918,12 @@ static int tegra_nvec_remove(struct platform_device *pdev)
 {
 	struct nvec_chip *nvec = platform_get_drvdata(pdev);
 
+	unregister_power_off_handler(&nvec->power_off_hb);
 	nvec_toggle_global_events(nvec, false);
 	mfd_remove_devices(nvec->dev);
 	nvec_unregister_notifier(nvec, &nvec->nvec_status_notifier);
 	cancel_work_sync(&nvec->rx_work);
 	cancel_work_sync(&nvec->tx_work);
-	/* FIXME: needs check wether nvec is responsible for power off */
-	pm_power_off = NULL;
 
 	return 0;
 }
diff --git a/drivers/staging/nvec/nvec.h b/drivers/staging/nvec/nvec.h
index e271375..d4145da 100644
--- a/drivers/staging/nvec/nvec.h
+++ b/drivers/staging/nvec/nvec.h
@@ -163,6 +163,8 @@ struct nvec_chip {
 	struct nvec_msg *last_sync_msg;
 
 	int state;
+
+	struct power_off_handler_block power_off_hb;
 };
 
 extern int nvec_write_async(struct nvec_chip *nvec, const unsigned char *data,
-- 
1.9.1

Re: [PATCH v5 33/48] staging: nvec: Register with kernel power-off handler

[Marc Dietrich]

Am Donnerstag 06 November 2014, 08:43:17 schrieb Guenter Roeck:
> Register with kernel power-off handler instead of setting pm_power_off
> directly. Register with default priority since we don't know any better.
> 
> Cc: Julian Andres Klode <jak-4HMq4SXA452hPH1hqNUYSQ@public.gmane.org>
> Cc: Marc Dietrich <marvin24-Mmb7MZpHnFY@public.gmane.org>
> Cc: Greg Kroah-Hartman <gregkh-hQyY1W1yCW8ekmWlsbkhG0B+6BGkLq7r@public.gmane.org>
> Acked-by: Greg Kroah-Hartman <gregkh-hQyY1W1yCW8ekmWlsbkhG0B+6BGkLq7r@public.gmane.org>
> Signed-off-by: Guenter Roeck <linux-0h96xk9xTtrk1uMJSBkQmQ@public.gmane.org

ok, this version (V5) actually switches the device off in contrast to the 
first implementation. Somehow I have to find out why it works at all without 
irqs.

Acked-by: Marc Dietrich <mavin24-Mmb7MZpHnFY@public.gmane.org>

> ---
> v5:
> - Rebase to v3.18-rc3
> v4:
> - Do not use notifiers but internal functions and data structures to manage
>   the list of power-off handlers. Drop unused parameters from callbacks, and
> make the power-off function type void
> v3:
> - Replace poweroff in all newly introduced variables and in text
>   with power_off or power-off as appropriate
> - Replace POWEROFF_PRIORITY_xxx with POWER_OFF_PRIORITY_xxx
> v2:
> - Use define to specify poweroff handler priority
> - Use dev_warn instead of dev_err
> 
>  drivers/staging/nvec/nvec.c | 21 ++++++++++++---------
>  drivers/staging/nvec/nvec.h |  2 ++
>  2 files changed, 14 insertions(+), 9 deletions(-)
> 
> diff --git a/drivers/staging/nvec/nvec.c b/drivers/staging/nvec/nvec.c
> index a93208a..acf2ea6 100644
> --- a/drivers/staging/nvec/nvec.c
> +++ b/drivers/staging/nvec/nvec.c
> @@ -33,6 +33,7 @@
>  #include <linux/mfd/core.h>
>  #include <linux/mutex.h>
>  #include <linux/notifier.h>
> +#include <linux/pm.h>
>  #include <linux/slab.h>
>  #include <linux/spinlock.h>
>  #include <linux/workqueue.h>
> @@ -80,8 +81,6 @@ enum nvec_sleep_subcmds {
>  #define LID_SWITCH BIT(1)
>  #define PWR_BUTTON BIT(15)
> 
> -static struct nvec_chip *nvec_power_handle;
> -
>  static const struct mfd_cell nvec_devices[] = {
>  	{
>  		.name = "nvec-kbd",
> @@ -759,12 +758,14 @@ static void nvec_disable_i2c_slave(struct nvec_chip
> *nvec) }
>  #endif
> 
> -static void nvec_power_off(void)
> +static void nvec_power_off(struct power_off_handler_block *this)
>  {
> +	struct nvec_chip *nvec = container_of(this, struct nvec_chip,
> +					      power_off_hb);
>  	char ap_pwr_down[] = { NVEC_SLEEP, AP_PWR_DOWN };
> 
> -	nvec_toggle_global_events(nvec_power_handle, false);
> -	nvec_write_async(nvec_power_handle, ap_pwr_down, 2);
> +	nvec_toggle_global_events(nvec, false);
> +	nvec_write_async(nvec, ap_pwr_down, 2);
>  }
> 
>  /*
> @@ -878,8 +879,11 @@ static int tegra_nvec_probe(struct platform_device
> *pdev) nvec->nvec_status_notifier.notifier_call = nvec_status_notifier;
> nvec_register_notifier(nvec, &nvec->nvec_status_notifier, 0);
> 
> -	nvec_power_handle = nvec;
> -	pm_power_off = nvec_power_off;
> +	nvec->power_off_hb.handler = nvec_power_off;
> +	nvec->power_off_hb.priority = POWER_OFF_PRIORITY_DEFAULT;
> +	ret = register_power_off_handler(&nvec->power_off_hb);
> +	if (ret)
> +		dev_warn(nvec->dev, "Failed to register power-off handler\n");
> 
>  	/* Get Firmware Version */
>  	msg = nvec_write_sync(nvec, get_firmware_version, 2);
> @@ -914,13 +918,12 @@ static int tegra_nvec_remove(struct platform_device
> *pdev) {
>  	struct nvec_chip *nvec = platform_get_drvdata(pdev);
> 
> +	unregister_power_off_handler(&nvec->power_off_hb);
>  	nvec_toggle_global_events(nvec, false);
>  	mfd_remove_devices(nvec->dev);
>  	nvec_unregister_notifier(nvec, &nvec->nvec_status_notifier);
>  	cancel_work_sync(&nvec->rx_work);
>  	cancel_work_sync(&nvec->tx_work);
> -	/* FIXME: needs check wether nvec is responsible for power off */
> -	pm_power_off = NULL;
> 
>  	return 0;
>  }
> diff --git a/drivers/staging/nvec/nvec.h b/drivers/staging/nvec/nvec.h
> index e271375..d4145da 100644
> --- a/drivers/staging/nvec/nvec.h
> +++ b/drivers/staging/nvec/nvec.h
> @@ -163,6 +163,8 @@ struct nvec_chip {
>  	struct nvec_msg *last_sync_msg;
> 
>  	int state;
> +
> +	struct power_off_handler_block power_off_hb;
>  };
> 
>  extern int nvec_write_async(struct nvec_chip *nvec, const unsigned char
> *data,

Re: [PATCH v5 33/48] staging: nvec: Register with kernel power-off handler

[Andreas Färber]

Am 09.11.2014 um 22:41 schrieb Marc Dietrich:
> Am Donnerstag 06 November 2014, 08:43:17 schrieb Guenter Roeck:
>> Register with kernel power-off handler instead of setting pm_power_off
>> directly. Register with default priority since we don't know any better.
>>
>> Cc: Julian Andres Klode <jak-4HMq4SXA452hPH1hqNUYSQ@public.gmane.org>
>> Cc: Marc Dietrich <marvin24-Mmb7MZpHnFY@public.gmane.org>
>> Cc: Greg Kroah-Hartman <gregkh-hQyY1W1yCW8ekmWlsbkhG0B+6BGkLq7r@public.gmane.org>
>> Acked-by: Greg Kroah-Hartman <gregkh-hQyY1W1yCW8ekmWlsbkhG0B+6BGkLq7r@public.gmane.org>
>> Signed-off-by: Guenter Roeck <linux-0h96xk9xTtrk1uMJSBkQmQ@public.gmane.org
> 
> ok, this version (V5) actually switches the device off in contrast to the 
> first implementation. Somehow I have to find out why it works at all without 
> irqs.
> 
> Acked-by: Marc Dietrich <mavin24-Mmb7MZpHnFY@public.gmane.org>

'r' missing in email? :)

Cheers,
Andreas

-- 
SUSE LINUX GmbH, Maxfeldstr. 5, 90409 Nürnberg, Germany
GF: Jeff Hawn, Jennifer Guild, Felix Imendörffer; HRB 21284 AG Nürnberg

Re: [PATCH v5 33/48] staging: nvec: Register with kernel power-off handler

[Guenter Roeck]

On 11/09/2014 01:41 PM, Marc Dietrich wrote:
> Am Donnerstag 06 November 2014, 08:43:17 schrieb Guenter Roeck:
>> Register with kernel power-off handler instead of setting pm_power_off
>> directly. Register with default priority since we don't know any better.
>>
>> Cc: Julian Andres Klode <jak@jak-linux.org>
>> Cc: Marc Dietrich <marvin24@gmx.de>
>> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
>> Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
>> Signed-off-by: Guenter Roeck <linux@roeck-us.net
>
> ok, this version (V5) actually switches the device off in contrast to the
> first implementation. Somehow I have to find out why it works at all without
> irqs.
>
> Acked-by: Marc Dietrich <mavin24@gmx.de>
>

Hi Marc,

The most recent version does not turn off interrupts as long as the
infrastructure (architecture code, ie machine_power_off) doesn't
turn it off.

Thanks a lot for (re-)testing.

Guenter