Re: [PATCH] BUG(): sched.c: Line 944

From: Linus Torvalds (
Date: Tue Sep 17 2002 - 00:56:11 EST

On 16 Sep 2002, Robert Love wrote:
> I was this -> <- close to celebrating. Not so fast, smarty.

You forget - I'm not only a smarty, I'm sick and twisted too.

> What about release_kernel_lock() ?
> It sees task->lock_depth>=0 and calls spin_unlock() on a lock that it
> does not hold.

We have a simple rule:
 - task->lock_depth = -1 means "no lock held"
 - task->lock_depth >= 0 means "BKL really held"

... but what does "task->lock_depth < -1" mean?

Yup: "validly nonpreemptable".

So then you have:

        #define kernel_locked() (current->lock_depth >= 0)

        #define in_atomic() (preempt_count() & ~PREEMPT_ACTIVE) != (current->lock_depth != -1))

and you're all set - just set lock_depth to -2 when you exit to show that
you don't hold the BKL, but that you are validly not preemtable.

I get the award for having the most disgusting ideas.

                Linus "but it works and is efficient" Torvalds

To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to
More majordomo info at
Please read the FAQ at

This archive was generated by hypermail 2b29 : Mon Sep 23 2002 - 22:00:18 EST