Re: VM-related Oops: 2.4.15pre1

["Jeff V. Merkey" <jmerkey@vger.timpanogas.org>]

On Sun, Nov 18, 2001 at 06:04:16PM -0800, Linus Torvalds wrote:
> 
> On Sun, 18 Nov 2001, Horst von Brand wrote:
> > Linus Torvalds <torvalds@transmeta.com> said:
> > > And nope, not really. It does use plain stores to page->flags, and I agree
> > > that it is ugly, but if the page was locked before calling it, all the
> > > stores will be with the PG_lock bit set - and even plain stores _are_
> > > documented to be atomic on x86 (and on all other reasonable architectures
> > > too).
> >
> > Even unaligned stores?
> 
> Actually, even unaligned stores (which page->flags is NOT) are atomic,
> even if Intel strongly discourages them (for performance reasons if no
> others) and there tends to be documentation that doesn't guarantee it.
> 
> 		Linus
> 

This is true.  They Generate what's called a "split lock" bus transaction
where the bus will hold LOCK# low across the several clock cycles to 
complete the write.  They are **VERY** heavy, BTW, and really cause 
nasty performance hits.

Jeff




> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/