[PATCH] ACPI / scan: Make it clear that acpi_bus_trim() cannot fail

[PATCH] ACPI / scan: Make it clear that acpi_bus_trim() cannot fail

[Rafael J. Wysocki]

From: Rafael J. Wysocki <rafael.j.wysocki@intel.com>

Since acpi_bus_trim() cannot fail, change its definition to a void
function, so that its callers don't check the return value in vain
and update the callers.

Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
---

On top of current linux-pm.git/linux-next.

Thanks,
Rafael

---
 drivers/acpi/dock.c                |    8 ++------
 drivers/acpi/scan.c                |   12 +++---------
 drivers/pci/hotplug/acpiphp_glue.c |   10 +++-------
 include/acpi/acpi_bus.h            |    2 +-
 4 files changed, 9 insertions(+), 23 deletions(-)

Index: linux-pm/drivers/acpi/scan.c
===================================================================
--- linux-pm.orig/drivers/acpi/scan.c
+++ linux-pm/drivers/acpi/scan.c
@@ -127,13 +127,8 @@ void acpi_bus_hot_remove_device(void *co
 	ACPI_DEBUG_PRINT((ACPI_DB_INFO,
 		"Hot-removing device %s...\n", dev_name(&device->dev)));
 
-	if (acpi_bus_trim(device)) {
-		printk(KERN_ERR PREFIX
-				"Removing device failed\n");
-		goto err_out;
-	}
-
-	/* device has been freed */
+	acpi_bus_trim(device);
+	/* Device node has been released. */
 	device = NULL;
 
 	/* power off device */
@@ -1656,7 +1651,7 @@ static acpi_status acpi_bus_remove(acpi_
 	return AE_OK;
 }
 
-int acpi_bus_trim(struct acpi_device *start)
+void acpi_bus_trim(struct acpi_device *start)
 {
 	/*
 	 * Execute acpi_bus_device_detach() as a post-order callback to detach
@@ -1672,7 +1667,6 @@ int acpi_bus_trim(struct acpi_device *st
 	acpi_walk_namespace(ACPI_TYPE_ANY, start->handle, ACPI_UINT32_MAX, NULL,
 			    acpi_bus_remove, NULL, NULL);
 	acpi_bus_remove(start->handle, 0, NULL, NULL);
-	return 0;
 }
 EXPORT_SYMBOL_GPL(acpi_bus_trim);
 
Index: linux-pm/include/acpi/acpi_bus.h
===================================================================
--- linux-pm.orig/include/acpi/acpi_bus.h
+++ linux-pm/include/acpi/acpi_bus.h
@@ -386,7 +386,7 @@ int acpi_bus_register_driver(struct acpi
 void acpi_bus_unregister_driver(struct acpi_driver *driver);
 int acpi_bus_scan(acpi_handle handle);
 void acpi_bus_hot_remove_device(void *context);
-int acpi_bus_trim(struct acpi_device *start);
+void acpi_bus_trim(struct acpi_device *start);
 acpi_status acpi_bus_get_ejd(acpi_handle handle, acpi_handle * ejd);
 int acpi_match_device_ids(struct acpi_device *device,
 			  const struct acpi_device_id *ids);
Index: linux-pm/drivers/acpi/dock.c
===================================================================
--- linux-pm.orig/drivers/acpi/dock.c
+++ linux-pm/drivers/acpi/dock.c
@@ -336,13 +336,9 @@ static struct acpi_device * dock_create_
 static void dock_remove_acpi_device(acpi_handle handle)
 {
 	struct acpi_device *device;
-	int ret;
 
-	if (!acpi_bus_get_device(handle, &device)) {
-		ret = acpi_bus_trim(device);
-		if (ret)
-			pr_debug("error removing bus, %x\n", -ret);
-	}
+	if (!acpi_bus_get_device(handle, &device))
+		acpi_bus_trim(device);
 }
 
 /**
Index: linux-pm/drivers/pci/hotplug/acpiphp_glue.c
===================================================================
--- linux-pm.orig/drivers/pci/hotplug/acpiphp_glue.c
+++ linux-pm/drivers/pci/hotplug/acpiphp_glue.c
@@ -742,8 +742,7 @@ static int acpiphp_bus_add(struct acpiph
 		/* this shouldn't be in here, so remove
 		 * the bus then re-add it...
 		 */
-		ret_val = acpi_bus_trim(device);
-		dbg("acpi_bus_trim return %x\n", ret_val);
+		acpi_bus_trim(device);
 	}
 
 	ret_val = acpi_bus_scan(func->handle);
@@ -772,11 +771,8 @@ static int acpiphp_bus_trim(acpi_handle
 		return retval;
 	}
 
-	retval = acpi_bus_trim(device);
-	if (retval)
-		err("cannot remove from acpi list\n");
-
-	return retval;
+	acpi_bus_trim(device);
+	return 0;
 }
 
 static void acpiphp_set_acpi_region(struct acpiphp_slot *slot)

Re: [PATCH] ACPI / scan: Make it clear that acpi_bus_trim() cannot fail

[Yinghai Lu]

On Thu, Jan 24, 2013 at 2:56 PM, Rafael J. Wysocki <rjw@sisk.pl> wrote:
> From: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
>
> Since acpi_bus_trim() cannot fail, change its definition to a void
> function, so that its callers don't check the return value in vain
> and update the callers.
>
> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
> ---

Acked-by: Yinghai Lu <yinghai@kernel.org>

Re: [PATCH] ACPI / scan: Make it clear that acpi_bus_trim() cannot fail

[Yasuaki Ishimatsu]

2013/01/25 7:56, Rafael J. Wysocki wrote:
> From: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
>
> Since acpi_bus_trim() cannot fail, change its definition to a void
> function, so that its callers don't check the return value in vain
> and update the callers.
>
> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
> ---

Acked-by: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com>

>
> On top of current linux-pm.git/linux-next.
>
> Thanks,
> Rafael
>
> ---
>   drivers/acpi/dock.c                |    8 ++------
>   drivers/acpi/scan.c                |   12 +++---------
>   drivers/pci/hotplug/acpiphp_glue.c |   10 +++-------
>   include/acpi/acpi_bus.h            |    2 +-
>   4 files changed, 9 insertions(+), 23 deletions(-)
>
> Index: linux-pm/drivers/acpi/scan.c
> ===================================================================
> --- linux-pm.orig/drivers/acpi/scan.c
> +++ linux-pm/drivers/acpi/scan.c
> @@ -127,13 +127,8 @@ void acpi_bus_hot_remove_device(void *co
>   	ACPI_DEBUG_PRINT((ACPI_DB_INFO,
>   		"Hot-removing device %s...\n", dev_name(&device->dev)));
>
> -	if (acpi_bus_trim(device)) {
> -		printk(KERN_ERR PREFIX
> -				"Removing device failed\n");
> -		goto err_out;
> -	}
> -
> -	/* device has been freed */
> +	acpi_bus_trim(device);
> +	/* Device node has been released. */
>   	device = NULL;
>
>   	/* power off device */
> @@ -1656,7 +1651,7 @@ static acpi_status acpi_bus_remove(acpi_
>   	return AE_OK;
>   }
>
> -int acpi_bus_trim(struct acpi_device *start)
> +void acpi_bus_trim(struct acpi_device *start)
>   {
>   	/*
>   	 * Execute acpi_bus_device_detach() as a post-order callback to detach
> @@ -1672,7 +1667,6 @@ int acpi_bus_trim(struct acpi_device *st
>   	acpi_walk_namespace(ACPI_TYPE_ANY, start->handle, ACPI_UINT32_MAX, NULL,
>   			    acpi_bus_remove, NULL, NULL);
>   	acpi_bus_remove(start->handle, 0, NULL, NULL);
> -	return 0;
>   }
>   EXPORT_SYMBOL_GPL(acpi_bus_trim);
>
> Index: linux-pm/include/acpi/acpi_bus.h
> ===================================================================
> --- linux-pm.orig/include/acpi/acpi_bus.h
> +++ linux-pm/include/acpi/acpi_bus.h
> @@ -386,7 +386,7 @@ int acpi_bus_register_driver(struct acpi
>   void acpi_bus_unregister_driver(struct acpi_driver *driver);
>   int acpi_bus_scan(acpi_handle handle);
>   void acpi_bus_hot_remove_device(void *context);
> -int acpi_bus_trim(struct acpi_device *start);
> +void acpi_bus_trim(struct acpi_device *start);
>   acpi_status acpi_bus_get_ejd(acpi_handle handle, acpi_handle * ejd);
>   int acpi_match_device_ids(struct acpi_device *device,
>   			  const struct acpi_device_id *ids);
> Index: linux-pm/drivers/acpi/dock.c
> ===================================================================
> --- linux-pm.orig/drivers/acpi/dock.c
> +++ linux-pm/drivers/acpi/dock.c
> @@ -336,13 +336,9 @@ static struct acpi_device * dock_create_
>   static void dock_remove_acpi_device(acpi_handle handle)
>   {
>   	struct acpi_device *device;
> -	int ret;
>
> -	if (!acpi_bus_get_device(handle, &device)) {
> -		ret = acpi_bus_trim(device);
> -		if (ret)
> -			pr_debug("error removing bus, %x\n", -ret);
> -	}
> +	if (!acpi_bus_get_device(handle, &device))
> +		acpi_bus_trim(device);
>   }
>
>   /**
> Index: linux-pm/drivers/pci/hotplug/acpiphp_glue.c
> ===================================================================
> --- linux-pm.orig/drivers/pci/hotplug/acpiphp_glue.c
> +++ linux-pm/drivers/pci/hotplug/acpiphp_glue.c
> @@ -742,8 +742,7 @@ static int acpiphp_bus_add(struct acpiph
>   		/* this shouldn't be in here, so remove
>   		 * the bus then re-add it...
>   		 */
> -		ret_val = acpi_bus_trim(device);
> -		dbg("acpi_bus_trim return %x\n", ret_val);
> +		acpi_bus_trim(device);
>   	}
>
>   	ret_val = acpi_bus_scan(func->handle);
> @@ -772,11 +771,8 @@ static int acpiphp_bus_trim(acpi_handle
>   		return retval;
>   	}
>
> -	retval = acpi_bus_trim(device);
> -	if (retval)
> -		err("cannot remove from acpi list\n");
> -
> -	return retval;
> +	acpi_bus_trim(device);
> +	return 0;
>   }
>
>   static void acpiphp_set_acpi_region(struct acpiphp_slot *slot)
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-acpi" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>

Re: [PATCH] ACPI / scan: Make it clear that acpi_bus_trim() cannot fail

[Toshi Kani]

On Thu, 2013-01-24 at 23:56 +0100, Rafael J. Wysocki wrote:
> From: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
> 
> Since acpi_bus_trim() cannot fail, change its definition to a void
> function, so that its callers don't check the return value in vain
> and update the callers.
> 
> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>

Acked-by: Toshi Kani <toshi.kani@hp.com>

Thanks,
-Toshi

> ---
> 
> On top of current linux-pm.git/linux-next.
> 
> Thanks,
> Rafael
> 
> ---
>  drivers/acpi/dock.c                |    8 ++------
>  drivers/acpi/scan.c                |   12 +++---------
>  drivers/pci/hotplug/acpiphp_glue.c |   10 +++-------
>  include/acpi/acpi_bus.h            |    2 +-
>  4 files changed, 9 insertions(+), 23 deletions(-)
> 
> Index: linux-pm/drivers/acpi/scan.c
> ===================================================================
> --- linux-pm.orig/drivers/acpi/scan.c
> +++ linux-pm/drivers/acpi/scan.c
> @@ -127,13 +127,8 @@ void acpi_bus_hot_remove_device(void *co
>  	ACPI_DEBUG_PRINT((ACPI_DB_INFO,
>  		"Hot-removing device %s...\n", dev_name(&device->dev)));
>  
> -	if (acpi_bus_trim(device)) {
> -		printk(KERN_ERR PREFIX
> -				"Removing device failed\n");
> -		goto err_out;
> -	}
> -
> -	/* device has been freed */
> +	acpi_bus_trim(device);
> +	/* Device node has been released. */
>  	device = NULL;
>  
>  	/* power off device */
> @@ -1656,7 +1651,7 @@ static acpi_status acpi_bus_remove(acpi_
>  	return AE_OK;
>  }
>  
> -int acpi_bus_trim(struct acpi_device *start)
> +void acpi_bus_trim(struct acpi_device *start)
>  {
>  	/*
>  	 * Execute acpi_bus_device_detach() as a post-order callback to detach
> @@ -1672,7 +1667,6 @@ int acpi_bus_trim(struct acpi_device *st
>  	acpi_walk_namespace(ACPI_TYPE_ANY, start->handle, ACPI_UINT32_MAX, NULL,
>  			    acpi_bus_remove, NULL, NULL);
>  	acpi_bus_remove(start->handle, 0, NULL, NULL);
> -	return 0;
>  }
>  EXPORT_SYMBOL_GPL(acpi_bus_trim);
>  
> Index: linux-pm/include/acpi/acpi_bus.h
> ===================================================================
> --- linux-pm.orig/include/acpi/acpi_bus.h
> +++ linux-pm/include/acpi/acpi_bus.h
> @@ -386,7 +386,7 @@ int acpi_bus_register_driver(struct acpi
>  void acpi_bus_unregister_driver(struct acpi_driver *driver);
>  int acpi_bus_scan(acpi_handle handle);
>  void acpi_bus_hot_remove_device(void *context);
> -int acpi_bus_trim(struct acpi_device *start);
> +void acpi_bus_trim(struct acpi_device *start);
>  acpi_status acpi_bus_get_ejd(acpi_handle handle, acpi_handle * ejd);
>  int acpi_match_device_ids(struct acpi_device *device,
>  			  const struct acpi_device_id *ids);
> Index: linux-pm/drivers/acpi/dock.c
> ===================================================================
> --- linux-pm.orig/drivers/acpi/dock.c
> +++ linux-pm/drivers/acpi/dock.c
> @@ -336,13 +336,9 @@ static struct acpi_device * dock_create_
>  static void dock_remove_acpi_device(acpi_handle handle)
>  {
>  	struct acpi_device *device;
> -	int ret;
>  
> -	if (!acpi_bus_get_device(handle, &device)) {
> -		ret = acpi_bus_trim(device);
> -		if (ret)
> -			pr_debug("error removing bus, %x\n", -ret);
> -	}
> +	if (!acpi_bus_get_device(handle, &device))
> +		acpi_bus_trim(device);
>  }
>  
>  /**
> Index: linux-pm/drivers/pci/hotplug/acpiphp_glue.c
> ===================================================================
> --- linux-pm.orig/drivers/pci/hotplug/acpiphp_glue.c
> +++ linux-pm/drivers/pci/hotplug/acpiphp_glue.c
> @@ -742,8 +742,7 @@ static int acpiphp_bus_add(struct acpiph
>  		/* this shouldn't be in here, so remove
>  		 * the bus then re-add it...
>  		 */
> -		ret_val = acpi_bus_trim(device);
> -		dbg("acpi_bus_trim return %x\n", ret_val);
> +		acpi_bus_trim(device);
>  	}
>  
>  	ret_val = acpi_bus_scan(func->handle);
> @@ -772,11 +771,8 @@ static int acpiphp_bus_trim(acpi_handle
>  		return retval;
>  	}
>  
> -	retval = acpi_bus_trim(device);
> -	if (retval)
> -		err("cannot remove from acpi list\n");
> -
> -	return retval;
> +	acpi_bus_trim(device);
> +	return 0;
>  }
>  
>  static void acpiphp_set_acpi_region(struct acpiphp_slot *slot)
> 

Re: [PATCH] ACPI / scan: Make it clear that acpi_bus_trim() cannot fail

[Vasilis Liaskovitis]

Hi,

On Thu, Jan 24, 2013 at 11:56:50PM +0100, Rafael J. Wysocki wrote:
> From: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
> 
> Since acpi_bus_trim() cannot fail, change its definition to a void
> function, so that its callers don't check the return value in vain
> and update the callers.

I have missed a few patchsets/discussions in the last month and wanted to
ask a question related to this: Does the new always-succeed 2-pass
trim_device design guarantee safe memory hot-remove operations?

Afaict if memory offline fails now, the device is ejected (_EJ0) anyways
causing a panic. Tested in a VM with linux-next-20130207 and
linux-next-20130218 by doing an SCI-eject request on a hot-plugged dimm.

Are there more patches in development for safe memory hot-remove?

thanks,

- Vasilis

Re: [PATCH] ACPI / scan: Make it clear that acpi_bus_trim() cannot fail

[Rafael J. Wysocki]

On Monday, February 18, 2013 01:25:49 PM Vasilis Liaskovitis wrote:
> Hi,
> 
> On Thu, Jan 24, 2013 at 11:56:50PM +0100, Rafael J. Wysocki wrote:
> > From: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
> > 
> > Since acpi_bus_trim() cannot fail, change its definition to a void
> > function, so that its callers don't check the return value in vain
> > and update the callers.
> 
> I have missed a few patchsets/discussions in the last month and wanted to
> ask a question related to this: Does the new always-succeed 2-pass
> trim_device design guarantee safe memory hot-remove operations?

I doesn't by itself.  Nor it really can, because the .remove() callbacks of
device drivers are not allowed to fail.

> Afaict if memory offline fails now, the device is ejected (_EJ0) anyways
> causing a panic. Tested in a VM with linux-next-20130207 and
> linux-next-20130218 by doing an SCI-eject request on a hot-plugged dimm.
> 
> Are there more patches in development for safe memory hot-remove?

Yes, there are.  I sent a patch series yesterday introducing some safety
measures (you can disable memory hotplug from user space or disable
automatic ejection).  There's more to come still.

The plan is to introduce offline/online operations for memory modules (in
analogy with CPU core online/offline) that can be started by user space and
memory eject will only be possible after offline (i.e. when the memory module
is known to be not in use).

Thanks,
Rafael


-- 
I speak only for myself.
Rafael J. Wysocki, Intel Open Source Technology Center.

Re: [PATCH] ACPI / scan: Make it clear that acpi_bus_trim() cannot fail

[Vasilis Liaskovitis]

On Mon, Feb 18, 2013 at 02:17:02PM +0100, Rafael J. Wysocki wrote:
> On Monday, February 18, 2013 01:25:49 PM Vasilis Liaskovitis wrote:
> > Hi,
> > 
> > On Thu, Jan 24, 2013 at 11:56:50PM +0100, Rafael J. Wysocki wrote:
> > > From: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
> > > 
> > > Since acpi_bus_trim() cannot fail, change its definition to a void
> > > function, so that its callers don't check the return value in vain
> > > and update the callers.
> > 
> > I have missed a few patchsets/discussions in the last month and wanted to
> > ask a question related to this: Does the new always-succeed 2-pass
> > trim_device design guarantee safe memory hot-remove operations?
> 
> I doesn't by itself.  Nor it really can, because the .remove() callbacks of
> device drivers are not allowed to fail.

right.

> 
> > Afaict if memory offline fails now, the device is ejected (_EJ0) anyways
> > causing a panic. Tested in a VM with linux-next-20130207 and
> > linux-next-20130218 by doing an SCI-eject request on a hot-plugged dimm.
> > 
> > Are there more patches in development for safe memory hot-remove?
> 
> Yes, there are.  I sent a patch series yesterday introducing some safety
> measures (you can disable memory hotplug from user space or disable
> automatic ejection).  There's more to come still.
> 
> The plan is to introduce offline/online operations for memory modules (in
> analogy with CPU core online/offline) that can be started by user space and
> memory eject will only be possible after offline (i.e. when the memory module
> is known to be not in use).

ok, thanks for the info. I 'll test these and the follow ups.

thanks again,

- Vasilis