Re: more efficient current_is_keventd macro? [was Re: [lhcs-devel] Re: Kthread_create() never returns when called from worker_thread]

[Srivatsa Vaddagiri <vatsa@in.ibm.com>]

On Mon, Mar 08, 2004 at 02:36:58PM -0800, Andrew Morton wrote:
> Is racy in the presence of preemption.  Please replace smp_processor_id()
> with get_cpu(), stick a put_cpu() at the end, avoid having two function
> return points, test it and send me the diff?

Hi Andrew,
	I had considered preemption and had thought using just
smp_processor_id() should be safe, since anyway the per-cpu kevent
thread is bound to its cpu alone.

Patch below is booted/tested against 2.6.4-rc2-mm1 on a 4-way x86 SMP box.
For testing, I basically called call_usermodehelper from inside a 
work function (activated using schedule_work) and checked that 
current_is_keventd() macro was returning true.

--- workqueue.c.org	2004-03-09 11:34:30.000000000 +0530
+++ workqueue.c	2004-03-09 19:06:51.000000000 +0530
@@ -374,16 +374,17 @@
 int current_is_keventd(void)
 {
 	struct cpu_workqueue_struct *cwq;
-	int cpu;
+	int cpu = smp_processor_id();
+	int ret = 0;
 
 	BUG_ON(!keventd_wq);
 
-	for_each_cpu(cpu) {
-		cwq = keventd_wq->cpu_wq + cpu;
-		if (current == cwq->thread)
-			return 1;
-	}
-	return 0;
+	cwq = keventd_wq->cpu_wq + cpu;
+	if (current == cwq->thread)
+		ret = 1;
+	
+	return ret;
+
 }
 
 #ifdef CONFIG_HOTPLUG_CPU

-- 


Thanks and Regards,
Srivatsa Vaddagiri,
Linux Technology Center,
IBM Software Labs,
Bangalore, INDIA - 560017