Re: [patch] smptimers, old BH removal, tq-cleanup, 2.5.39

[Jeff Garzik]

Jeff Garzik wrote:
> 3) Maybe this is an improvement worth considering, maybe not:  I think 
> it would be useful to have "process context timers."  What I mean by 
> this is, make it easier to implement code that can be boiled down to:
>     run in process context
>     sleep for a little while
> Sure I can do this by implementing a timer that does nothing more than 
> call schedule_task(), but that seems a little redundant if done in 
> multiple places :)

To sum, I guess I want something to "run schedule_task() [at least] NN 
usecs from now."  schedule_task_future() ?