Re: Purpose of the mm/slab.c changes

From: Manfred Spraul (manfred@colorfullife.com)
Date: Sun Sep 09 2001 - 11:55:00 EST


Alan Cox wrote:
>
> > > doesn't matter which free page is used first/last.
> >
> > You're full of crap.
> > LIFO is obviously superior due to cache re-use.
>
> Interersting question however. On SMP without sufficient per CPU slab caches
> is tht still the case ?

Correct. SMP was perfect LIFO even without Andrea's changes.

I thought Andrea tried to reduce the fragmentation, therefore I wrote
"free is free".

But even for cache re-use his changes are not a big change: The main
fifo/lifo ordering on UP is mandated by the defragmentation property of
the slab allocator.

Afaics there is exactly one case where my code is not lifo and Andrea's
is: kmem_cache_free frees the last object in slab, each slab contains
more than one object, and there are no further partial slabs.
In all other cases Andrea just adds list_del();list_add() instead of
changes to the firstnotfull pointer.

full->partial is/was lifo,
partial->partial doesn't change the lists at all
partial->empty was fifo, is now lifo_if_no_partial_slab_exists

--
	Manfred
-
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/



This archive was generated by hypermail 2b29 : Sat Sep 15 2001 - 21:00:19 EST