On Tue, Feb 08, 2011 at 12:29:57AM +0100, Rafael J. Wysocki wrote:
> From: Rafael J. Wysocki <rjw@xxxxxxx>
>
> If thermal polling is enabled, which for example is the case for
> ACPI thermal zones with the _TZP object defined, the thermal driver
> uses delayed work items for this purpose. Unfortunately, since
> they are queued up using schedule_delayed_work(), the work function
> may be executed during system suspend or resume, which is not
> desirable.
>
> To prevent that from happening, use a freezable workqueue for
> queuing up delayed work items in the thermal driver.
>
> Signed-off-by: Rafael J. Wysocki <rjw@xxxxxxx>
> ---
> drivers/thermal/thermal_sys.c | 22 ++++++++++++++++++++--
> 1 file changed, 20 insertions(+), 2 deletions(-)
>
> Index: linux-2.6/drivers/thermal/thermal_sys.c
> ===================================================================
> --- linux-2.6.orig/drivers/thermal/thermal_sys.c
> +++ linux-2.6/drivers/thermal/thermal_sys.c
> @@ -62,6 +62,20 @@ static DEFINE_MUTEX(thermal_list_lock);
>
> static unsigned int thermal_event_seqnum;
>
> +static struct workqueue_struct *thermal_wq;
> +
> +static int __init thermal_start_workqueue(void)
> +{
> + thermal_wq = alloc_workqueue("thermal", WQ_FREEZEABLE, 0);
Should probably be unbound as well.
FWIW, I would not mind if we had a global freezeable workqueue already
predefined. I could switch input_polldev and vmw_balloon to it and there
probably could be more users...
Thanks.
--
Dmitry
--
To unsubscribe from this list: send the line "unsubscribe linux-acpi" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
