[PATCH] rpc_delete_timer() cleanup

[Oleg Nesterov]

When rpc task receives a signal while sleeping on RPC_TASK_QUEUED bit,
__rpc_execute() calls rpc_wake_up_task(). If rpc timer starts on another
cpu at the same time it can set RPC_TASK_WAKEUP bit before rpc_wake_up_task()
called from __rpc_execute() does so. Now the main loop in __rpc_execute
restarted with RPC_TASK_QUEUED bit and with still executing tk_timer.

I don't see too bad in this (the task already has RPC_TASK_KILLED flag,
so ->tk_exit() should not call rpc_delay(), and it has TIF_SIGPENDING),
but still I think it was not intended to happen.

Signed-off-by: Oleg Nesterov <oleg@xxxxxxxxxx>

--- 2.6.13-rc1/net/sunrpc/sched.c~	Thu Jun 30 17:03:08 2005
+++ 2.6.13-rc1/net/sunrpc/sched.c	Thu Jun 30 17:03:25 2005
@@ -135,8 +135,6 @@ __rpc_add_timer(struct rpc_task *task, r
 static void
 rpc_delete_timer(struct rpc_task *task)
 {
-	if (RPC_IS_QUEUED(task))
-		return;
 	if (test_and_clear_bit(RPC_TASK_HAS_TIMER, &task->tk_runstate)) {
 		del_singleshot_timer_sync(&task->tk_timer);
 		dprintk("RPC: %4d deleting timer\n", task->tk_pid);
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/