>>>>> "andrea" == Andrea Arcangeli <andrea@suse.de> writes:
andrea> On Tue, 11 Jul 2000, Stephen C. Tweedie wrote:
>> Hi,
>>
>> On Sun, Jul 09, 2000 at 10:31:46PM +0200, Andrea Arcangeli wrote:
>>>
>>> Think what happens if we shrink lru_mapped first.
>>
>> It's not supposed to work that way.
andrea> The object of this simple example is to show that the lrus have different
andrea> priorities. These priorities will probably change in function of the
andrea> workload of course but we can try to take care of that.
I agree with Stephen here, if my cache page is older than my mmaped vi
page, I want to unmap first the vi page.
andrea> I see what you plan to do. Fact is that I'm not convinced it's necessary
andrea> and I prefer to have a dynamic falling back algorithms between caches that
andrea> will avoid me to have additional lru lists and additional refile between
andrea> lrus. Also I will be able to say when I did progress because my progress
andrea> will _always_ correspond to a page freed (so I'll remove the unrobusteness
andrea> of the current swap_out completly).
Yes, but you have to find a _magic_ number for knowing when to free
for the maped pages/cache pages. That number comes for free with the
inactive list implementation and is based in the actual workload,
i.e. we don't need to guess.
Later, Juan.
-- In theory, practice and theory are the same, but in practice they are different -- Larry McVoy- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Sat Jul 15 2000 - 21:00:13 EST