Messages in this thread | | | Date | Sun, 23 Nov 2008 12:48:50 +0900 | From | Tejun Heo <> | Subject | Re: poll: allow f_op->poll to sleep, take #3 |
| |
Brad Boyer wrote: > On Sun, Nov 23, 2008 at 12:05:53PM +0900, Tejun Heo wrote: >> I thought try_to_wake_up() was made static to avoid abuse but then again >> creating dummy waitqueue is an obvious abuse of waitqueue. What do >> other people think? I'll be happy to use try_to_wake_up() directly. > > Do you need all the extra arguments? The function wake_up_process() > is already a wrapper around try_to_wake_up() and is exported, but > it doesn't have any arguments other than the task_struct and uses > defaults for the other arguments. I'm not sure if anything in your > code would break by ignoring the other possible values instead of > passing them along from the arguments into the caller.
Hmmm... there was something which made wake_up_process() inappropriate. Ah, okay, it was @mode. We can add a WARN_ON() if @mode is an unexpected value and use a fixed one - TASK_INTERRUPTIBLE or TASK_ALL - but that's even hackier than the waitqueue hack.
Thanks.
-- tejun
| |