[PATCH 1/2] acpi: don't free non-existant backlight in acpi video module

[PATCH 1/2] acpi: don't free non-existant backlight in acpi video module

[keithp]

From: Keith Packard <keithp@keithp.com>

acpi_video_put_one_device was attempting to remove sysfs entries and
unregister a backlight device without first checking that said backlight
device structure had been created.

Signed-off-by: Keith Packard <keithp@keithp.com>
---
 drivers/acpi/video.c |    7 +++++--
 1 files changed, 5 insertions(+), 2 deletions(-)

diff --git a/drivers/acpi/video.c b/drivers/acpi/video.c
index 8851315..60ea984 100644
--- a/drivers/acpi/video.c
+++ b/drivers/acpi/video.c
@@ -2004,8 +2004,11 @@ static int acpi_video_bus_put_one_device(struct acpi_video_device *device)
 	status = acpi_remove_notify_handler(device->dev->handle,
 					    ACPI_DEVICE_NOTIFY,
 					    acpi_video_device_notify);
-	sysfs_remove_link(&device->backlight->dev.kobj, "device");
-	backlight_device_unregister(device->backlight);
+	if (device->backlight) {
+		sysfs_remove_link(&device->backlight->dev.kobj, "device");
+		backlight_device_unregister(device->backlight);
+		device->backlight = NULL;
+	}
 	if (device->cdev) {
 		sysfs_remove_link(&device->dev->dev.kobj,
 				  "thermal_cooling");
-- 
1.6.3.3

Re: [PATCH 1/2] acpi: don't free non-existant backlight in acpi video module

[Zhang Rui]

On Sun, 2009-07-12 at 14:47 +0800, keithp@keithp.com wrote:
> From: Keith Packard <keithp@keithp.com>
> 
> acpi_video_put_one_device was attempting to remove sysfs entries and
> unregister a backlight device without first checking that said backlight
> device structure had been created.
> 
> Signed-off-by: Keith Packard <keithp@keithp.com>

Acked-by: Zhang Rui <rui.zhang@intel.com>

> ---
>  drivers/acpi/video.c |    7 +++++--
>  1 files changed, 5 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/acpi/video.c b/drivers/acpi/video.c
> index 8851315..60ea984 100644
> --- a/drivers/acpi/video.c
> +++ b/drivers/acpi/video.c
> @@ -2004,8 +2004,11 @@ static int acpi_video_bus_put_one_device(struct acpi_video_device *device)
>  	status = acpi_remove_notify_handler(device->dev->handle,
>  					    ACPI_DEVICE_NOTIFY,
>  					    acpi_video_device_notify);
> -	sysfs_remove_link(&device->backlight->dev.kobj, "device");
> -	backlight_device_unregister(device->backlight);
> +	if (device->backlight) {
> +		sysfs_remove_link(&device->backlight->dev.kobj, "device");
> +		backlight_device_unregister(device->backlight);
> +		device->backlight = NULL;
> +	}
>  	if (device->cdev) {
>  		sysfs_remove_link(&device->dev->dev.kobj,
>  				  "thermal_cooling");

Re: [PATCH 1/2] acpi: don't free non-existant backlight in acpi video module

[Andrew Morton]

> acpi: don't free non-existant backlight in acpi video module

"existent" :)

On Sat, 11 Jul 2009 23:47:31 -0700
keithp@keithp.com wrote:

> From: Keith Packard <keithp@keithp.com>
> 
> acpi_video_put_one_device was attempting to remove sysfs entries and
> unregister a backlight device without first checking that said backlight
> device structure had been created.
> 
> Signed-off-by: Keith Packard <keithp@keithp.com>
> ---
>  drivers/acpi/video.c |    7 +++++--
>  1 files changed, 5 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/acpi/video.c b/drivers/acpi/video.c
> index 8851315..60ea984 100644
> --- a/drivers/acpi/video.c
> +++ b/drivers/acpi/video.c
> @@ -2004,8 +2004,11 @@ static int acpi_video_bus_put_one_device(struct acpi_video_device *device)
>  	status = acpi_remove_notify_handler(device->dev->handle,
>  					    ACPI_DEVICE_NOTIFY,
>  					    acpi_video_device_notify);
> -	sysfs_remove_link(&device->backlight->dev.kobj, "device");
> -	backlight_device_unregister(device->backlight);
> +	if (device->backlight) {
> +		sysfs_remove_link(&device->backlight->dev.kobj, "device");
> +		backlight_device_unregister(device->backlight);
> +		device->backlight = NULL;
> +	}
>  	if (device->cdev) {
>  		sysfs_remove_link(&device->dev->dev.kobj,
>  				  "thermal_cooling");

um, OK.

Under which circumstances was this observed?

For symmetry we could instead test acpi_video_backlight_support() here.

The patch assumes that someone initially zeroed device->backlight.  Is
that true and reliable?  If so, is the 

	memset(&device->cap, 0, sizeof(device->cap));

in acpi_video_device_find_cap() needed?


Where's [patch 2/2], btw?

Re: [PATCH 1/2] acpi: don't free non-existant backlight in acpi video module

[Zhang Rui]

On Tue, 2009-07-14 at 07:41 +0800, Andrew Morton wrote:
> > acpi: don't free non-existant backlight in acpi video module
> 
> "existent" :)
> 
> On Sat, 11 Jul 2009 23:47:31 -0700
> keithp@keithp.com wrote:
> 
> > From: Keith Packard <keithp@keithp.com>
> > 
> > acpi_video_put_one_device was attempting to remove sysfs entries and
> > unregister a backlight device without first checking that said backlight
> > device structure had been created.
> > 
> > Signed-off-by: Keith Packard <keithp@keithp.com>
> > ---
> >  drivers/acpi/video.c |    7 +++++--
> >  1 files changed, 5 insertions(+), 2 deletions(-)
> > 
> > diff --git a/drivers/acpi/video.c b/drivers/acpi/video.c
> > index 8851315..60ea984 100644
> > --- a/drivers/acpi/video.c
> > +++ b/drivers/acpi/video.c
> > @@ -2004,8 +2004,11 @@ static int acpi_video_bus_put_one_device(struct acpi_video_device *device)
> >  	status = acpi_remove_notify_handler(device->dev->handle,
> >  					    ACPI_DEVICE_NOTIFY,
> >  					    acpi_video_device_notify);
> > -	sysfs_remove_link(&device->backlight->dev.kobj, "device");
> > -	backlight_device_unregister(device->backlight);
> > +	if (device->backlight) {
> > +		sysfs_remove_link(&device->backlight->dev.kobj, "device");
> > +		backlight_device_unregister(device->backlight);
> > +		device->backlight = NULL;
> > +	}
> >  	if (device->cdev) {
> >  		sysfs_remove_link(&device->dev->dev.kobj,
> >  				  "thermal_cooling");
> 
> um, OK.
> 
> Under which circumstances was this observed?
> 
Only part of the ACPI video devices support backlight switching.
device->backlight is NULL for many ACPI video devices, like, CRT, TV,
DVI, etc.

> For symmetry we could instead test acpi_video_backlight_support() here.
> 
No.
device->backlight is NULL for the devices that declare to support
backlight control while provides INVALID ACPI backlight control methods.

> The patch assumes that someone initially zeroed device->backlight.  Is
> that true and reliable?

yes.

>   If so, is the 
> 
> 	memset(&device->cap, 0, sizeof(device->cap));
> 
> in acpi_video_device_find_cap() needed?
> 
right. IMO, we can remove this.
patch attached.

thanks,
rui


device->cap and video->cap is zerod initially so we don't need
to clear it again.

Signed-off-by: Zhang Rui <rui.zhang@intel.com>
---
 drivers/acpi/video.c |    4 ----
 1 file changed, 4 deletions(-)

Index: linux-2.6/drivers/acpi/video.c
===================================================================
--- linux-2.6.orig/drivers/acpi/video.c
+++ linux-2.6/drivers/acpi/video.c
@@ -934,9 +934,6 @@ static void acpi_video_device_find_cap(s
 {
 	acpi_handle h_dummy1;
 
-
-	memset(&device->cap, 0, sizeof(device->cap));
-
 	if (ACPI_SUCCESS(acpi_get_handle(device->dev->handle, "_ADR", &h_dummy1))) {
 		device->cap._ADR = 1;
 	}
@@ -1039,7 +1036,6 @@ static void acpi_video_bus_find_cap(stru
 {
 	acpi_handle h_dummy1;
 
-	memset(&video->cap, 0, sizeof(video->cap));
 	if (ACPI_SUCCESS(acpi_get_handle(video->device->handle, "_DOS", &h_dummy1))) {
 		video->cap._DOS = 1;
 	}