[PATCH] thermal_sys: check get_temp return value

[PATCH] thermal_sys: check get_temp return value

[Michael Brunner]

The return value of the get_temp function is not checked when doing a
thermal zone update. This may lead to a critical shutdown if get_temp
fails and the content of the temp variable is incorrectly set higher
than the critical trip point.
This has been observed on a system with incorrect ACPI implementation
where the corresponding methods were not serialized and therefore
sometimes triggered ACPI errors (AE_ALREADY_EXISTS). The following
critical shutdowns indicated a temperature of 2097 C, which was
obviously wrong.

The patch adds a return value check that jumps over all trip point
evaluations printing a warning if get_temp fails. The trip points are
evaluated again on the next polling interval with successful get_temp
execution.

Signed-off-by: Michael Brunner <mibru@gmx.de>

---

 drivers/thermal/thermal_sys.c |    9 ++++++++-
 1 files changed, 8 insertions(+), 1 deletions(-)

diff --git a/drivers/thermal/thermal_sys.c b/drivers/thermal/thermal_sys.c
index 0a69672..4e83c29 100644
--- a/drivers/thermal/thermal_sys.c
+++ b/drivers/thermal/thermal_sys.c
@@ -953,7 +953,12 @@ void thermal_zone_device_update(struct thermal_zone_device *tz)
 
 	mutex_lock(&tz->lock);
 
-	tz->ops->get_temp(tz, &temp);
+	if (tz->ops->get_temp(tz, &temp)) {
+		/* get_temp failed - retry it later */
+		printk(KERN_WARNING PREFIX "failed to read out thermal zone "
+		       "%d\n", tz->id);
+		goto leave;
+	}
 
 	for (count = 0; count < tz->trips; count++) {
 		tz->ops->get_trip_type(tz, count, &trip_type);
@@ -1005,6 +1010,8 @@ void thermal_zone_device_update(struct thermal_zone_device *tz)
 					    THERMAL_TRIPS_NONE);
 
 	tz->last_temperature = temp;
+
+      leave:
 	if (tz->passive)
 		thermal_zone_device_set_polling(tz, tz->passive_delay);
 	else if (tz->polling_delay)

Re: [PATCH] thermal_sys: check get_temp return value

[Zhang Rui]

On Sun, 2009-08-02 at 20:06 +0800, Michael Brunner wrote:
> The return value of the get_temp function is not checked when doing a
> thermal zone update. This may lead to a critical shutdown if get_temp
> fails and the content of the temp variable is incorrectly set higher
> than the critical trip point.
> This has been observed on a system with incorrect ACPI implementation
> where the corresponding methods were not serialized and therefore
> sometimes triggered ACPI errors (AE_ALREADY_EXISTS). The following
> critical shutdowns indicated a temperature of 2097 C, which was
> obviously wrong.
> 
> The patch adds a return value check that jumps over all trip point
> evaluations printing a warning if get_temp fails. The trip points are
> evaluated again on the next polling interval with successful get_temp
> execution.
> 
> Signed-off-by: Michael Brunner <mibru@gmx.de>
> 
good catch, Michael. :)

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

> ---
> 
>  drivers/thermal/thermal_sys.c |    9 ++++++++-
>  1 files changed, 8 insertions(+), 1 deletions(-)
> 
> diff --git a/drivers/thermal/thermal_sys.c b/drivers/thermal/thermal_sys.c
> index 0a69672..4e83c29 100644
> --- a/drivers/thermal/thermal_sys.c
> +++ b/drivers/thermal/thermal_sys.c
> @@ -953,7 +953,12 @@ void thermal_zone_device_update(struct thermal_zone_device *tz)
>  
>  	mutex_lock(&tz->lock);
>  
> -	tz->ops->get_temp(tz, &temp);
> +	if (tz->ops->get_temp(tz, &temp)) {
> +		/* get_temp failed - retry it later */
> +		printk(KERN_WARNING PREFIX "failed to read out thermal zone "
> +		       "%d\n", tz->id);
> +		goto leave;
> +	}
>  
>  	for (count = 0; count < tz->trips; count++) {
>  		tz->ops->get_trip_type(tz, count, &trip_type);
> @@ -1005,6 +1010,8 @@ void thermal_zone_device_update(struct thermal_zone_device *tz)
>  					    THERMAL_TRIPS_NONE);
>  
>  	tz->last_temperature = temp;
> +
> +      leave:
>  	if (tz->passive)
>  		thermal_zone_device_set_polling(tz, tz->passive_delay);
>  	else if (tz->polling_delay)