On Wed, 8 Jul 1998 21:03:50 +0200 (CEST), Rik van Riel
<H.H.vanRiel@phys.uu.nl> said:
> It's been in the kernel since 2.1.90, or thereabouts...
> [vmscan.c:do_try_to_free_page()]
> Andrea has clearly pointed out that it doesn't work on small
> memory machines.
The mechanism there is just to preferentially victimise the cache as
soon as it exceeds a certain size, which is a rather different thing.
Forcing a preferential cache scan on _every_ cache allocation would have
the additional effect that by the time the cache grew that large, we'd
already have aged it significantly (it's the lack of page aging while
the cache is growing which kills us). However, to get decent
performance from the allocator, we'd really want to use some explicit
cache LRU rather than the physical page scan we currently have.
--Stephen
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu