Messages in this thread | | | Date | Wed, 12 Jul 2000 02:52:26 +0200 (CEST) | From | Andrea Arcangeli <> | Subject | Re: [PATCH] 2.2.17pre7 VM enhancement Re: I/O performance on 2.4.0-test2 |
| |
On Wed, 12 Jul 2000, John Alvord wrote:
>One question that puzzles me... cache for disk files and cache for >program data will have very unlike characteristics. Executable program
Agreed. That is exactly what I'm trying say by telling that lru_cache and lru_mapped_cache have different implicit priorities and we can't threat them in the same way.
>enough, what algorithm is used to achieve an effective balance of >usage?
In 2.[234].x we basically first try to shrink the cache for disk and when we run low in cache for disk (so when we start to fail in shrinking it) we fallback shrinking the cache for program. That's sane algorithm even if currently it's not very smart in understanding when it's time to shrink the cache for programs and it's also not able to shrink it properly in some case.
Andrea
- 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/
| |