Re: [PATCH] i386 spinlocks should use the full 32 bits, not only 8bits

[Linus Torvalds]

On Fri, 21 Oct 2005, Ingo Molnar wrote:
> 
> the unlock is simple even in the preemption case

No it's not. It needs to decrement the preemption counter and test it.

See kernel/spinlock.c:

	void __lockfunc _spin_unlock(spinlock_t *lock)
	{
	        _raw_spin_unlock(lock);
	        preempt_enable();
	}
	EXPORT_SYMBOL(_spin_unlock);

and look at what "preempt_enable()" does.

In other words, with CONFIG_PREEMPT, your patch is WRONG. You made 
"spin_unlock()" just skip the preempt_enable.

In fact, with preemption, the _locking_ is the simpler part. Unlock is the 
complex one.


			Linus
-
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/