Messages in this thread | | | Date | Wed, 1 Sep 1999 13:41:22 -0700 | From | "David S. Miller" <> | Subject | Re: set_current_state |
| |
Date: Wed, 1 Sep 1999 16:17:07 -0400 From: "Theodore Y. Ts'o" <tytso@mit.edu>
Are there any circumstances where we should keep the old usage?
Fundamentally it only matters when you are going into a "sleep until condition" polling loop such that:
add_wait_queue(...); for(;;) { set_current_state(...); if (some asynchronous state) break; schedule(); } remove_wait_queue(...);
The idea being that you wish the task state to propagate into the view of all cpus in the system before other cpus can potentially cause the asynronous event to occur _and_ check your task state while performing the wake_up(...)
What we're trying to prevent here is the asynchronous state test load not bypassing the store of the new task state. If the cpu reorders these, the wakeup event can be missed.
Later, David S. Miller davem@redhat.com
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |