Re: Possible memory ordering bug in page reclaim?

[Andrea Arcangeli]

On Sun, Oct 16, 2005 at 05:48:55AM +1000, Herbert Xu wrote:
> On Sat, Oct 15, 2005 at 06:00:18PM +0000, Andrea Arcangeli wrote:
> > 
> > Note that the barrier in atomic_add_negative is useless here because it
> > happens way too late, _after_ the count is decremented (not _before_)
> > so the decreased count could be already visible to the other cpu.
> 
> Could you please point me to an architecture that does this?

sure see alpha:

	__asm__ __volatile__(
	"1:	ldq_l %0,%1\n"
	"	addq %0,%3,%2\n"
	"	addq %0,%3,%0\n"
	"	stq_c %0,%1\n"
	"	beq %0,2f\n"
	"	mb\n"

the memory barrier is applied way after the write is visible to other
cpus, you can even get an irq before the mb and block there for some
usec.

>From a common code point of view, the barrier you mentioned in
atomic_add_negative is absolutely useless for this specific case
(setpagedirty+put_page)
-
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/