[ Hey, I can have long discussions by myself. I don't need you guys to
answer me at all.
This must be what senility feels like. Linus "doddering fool" Torvalds ]
On Tue, 7 Aug 2001, Linus Torvalds wrote:
>
> We should really document the ranges clearly. Right now (with these
> changes), the ranges would be (with users in parenthesis):
>
> free (free + inactive_clean):
>
> low water mark: zone->pages_high (__alloc_pages, zone_free_shortage)
> high water mark: zone->pages_high*2 (zone_free_plenty)
That's subtly wrong, __alloc_pages really has "zone->pages_low", not
"pages_high" as the low water mark on when to start kswapd.
It does use "pages_high" too - but it is used as a "prefer this zone
because it isn't close to empty" marker. In short, it's really more of a
local high water mark.
So the free_shortage() logic really should be:
- we want zone to have "pages_high" on _average_, but it's a shortage
only if any zone is under the "pages_low" thing.
Which is actually exactly what you get if "total_free_shortage()" first
tests the global sums against "freepages.high", and then checks individual
zones with "zone_free_shortage()" against zone->pages_low. So I think I
happened to get that one right originally by pure luck.
Linus
-
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 : Wed Aug 15 2001 - 21:00:15 EST