On Mon, Oct 4, 2010 at 11:28 AM, Vimal <j.vimal@xxxxxxxxx> wrote:
> Hi,
>
> I am trying to setup a workqueue to defer computation intensive work
> in a performance critical path. I read about the workqueue mechanism
> in the LDD3 book; however, the API seems to be different now.
>
> Specifically: the LDD3 book says that we can create a work_struct and
> initialise it as follows:
>
> INIT_WORK(struct work_struct *work, void (*fn)(void *), void *data);
>
> However, in 2.6.35, I am unable to see the "data" argument to the
> INIT_WORK macro. Also, the "fn" pointer definition has changed to:
> void (*fn)(struct work_struct *).
>
> I could get away with initialising the work_struct myself, by looking
> at the fields and setting "data" accordingly (work->data is of type
> atomic_long_t), but I think they are supposed to be opaque. Could
> someone clarify how I should go about doing this?
You have to use container_of() here.
For example
struct cpufreq_policy {
......
....
struct work_struct update;
........
};
static void handle_update(struct work_struct *work)
{
struct cpufreq_policy *policy =
container_of(work, struct cpufreq_policy, update);
.............
.............
}
module_init()
{
struct cpufreq_policy policy;
........
........
INIT_WORK(&policy, handle_update);
..........
}
Hope this helps.
Arun
>
> Or, is there a different way of implementing the requirement? (defer
> computation intensive work in a performance critical path.)
>
> Thanks,
> --
> Vimal
>
> --
> To unsubscribe from this list: send an email with
> "unsubscribe kernelnewbies" to ecartis@xxxxxxxxxxxx
> Please read the FAQ at http://kernelnewbies.org/FAQ
>
>
--
To unsubscribe from this list: send an email with
"unsubscribe kernelnewbies" to ecartis@xxxxxxxxxxxx
Please read the FAQ at http://kernelnewbies.org/FAQ
