Re: BUG(): sched.c: Line 944 - 2.5.35

[Robert Love]

On Mon, 2002-09-16 at 12:36, Adam J. Richter wrote:

> 	When I see this problem at boot, preempt_count() returns 0x4000000
> (PREEMPT_ACTIVE) and kernel_locked() returns 0.
> 
> 	I don't understand the semantics of PREEMPT_ACTIVE to know
> whether to
> 
> 	(1) change the test back to using in_interrupt instead of in_atomic, or
> 	(2) change the definition of in_atomic(), or
> 	(3) look for a bug somewhere else.

There are two problems: First, PREEMPT_ACTIVE is indeed set on entry to
schedule() from preempt_schedule() so we need to check for that too. 
Second, the BUG() is catching a bit of issues... you want something
like:

-	if (unlikely(in_atomic()))
-		BUG();
+	if (unlikely(in_atomic() && preempt_count() != PREEMPT_ACTIVE)) {
+		printk(KERN_ERR "schedule() called while non-atomic!\n");
+		show_stack(NULL);
+	}

I will send a patch to Linus.

> 	However, I know experimentally that changing the test back to
> using in_interrupt() results in a possibly unrelated BUG() at line 279
> of rmap.c:

This is unrelated.

	Robert Love