> In article <Pine.LNX.3.96.980819081756.10109G-100000@flashy.is.co.za>,
> Craig Schlenter <craig@is.co.za> wrote:
> >On Tue, 18 Aug 1998, Linus Torvalds wrote:
> >
> >[snip]
> >> [snip] 2.1.116 has a few patches that it
> >> _shouldn't_ have had and that made it into the final release by mistake
> >> (the page aging code shouldn't have been ifdeffed out), but it seems
> >[snip]
> >
> >Are you referring to this part of the patch?
>
> Yup, that's the thing.
>
> It reportedly speeds things up, but I really wanted to try some other
> alternatives instead, rather than just disable the page age checking
> altogether. It may be that eventually we just remove the code, but I
> hadn't meant for that part to actually be in 2.1.116.
>
Then what do you think about THIS alternative:
diff -urN --exclude-from=exclude linux-old/mm/filemap.c linux/mm/filemap.c
--- linux-2.1.115/mm/filemap.c Mon Aug 3 13:10:57 1998
+++ linux/mm/filemap.c Tue Aug 4 16:10:03 1998
@@ -171,6 +171,11 @@
touch_page(page);
break;
}
+ /* Age named pages aggresively, so page cache
+ * doesn't grow too fast. -zcalusic
+ */
+ age_page(page);
+ age_page(page);
age_page(page);
if (page->age)
break;
I did sent it to you on another occasion, but never really got your
comment back.
Triple aging preserves good page aging characteristics (swapping out
unused pages), but is much less aggressive than current aging wrt
situations when you do lots of I/O. Current implementation swaps too
much useful data out in such conditions, thus in fact slowing things
down, even though you have lots of memory to spare for the page cache.
Could I have your comment please?
TIA,
-- Posted by Zlatko Calusic E-mail: <Zlatko.Calusic@CARNet.hr> --------------------------------------------------------------------- Life would be easier if I had the source code.- 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.altern.org/andrebalsa/doc/lkml-faq.html