Re: Patch: linux-2.5.8-pre1/kernel/exit.c change caused BUG() at

[Linus Torvalds]

On 4 Apr 2002, Robert Love wrote:
> 
> Thanks for the CC.  I've been looking into this problem.  I am not too
> sure why we require protection from concurrency via preemption and not
> via SMP.  In other words, why are we SMP-safe but not preempt-safe here.
> 
> I don't really have an answer.

The answer is that preempt_schedule() illegally sets 

	current->state = TASK_RUNNING;

without asking the process whether that's ok. The SMP code never does 
anything like that.

		Linus