Messages in this thread | | | Date | Fri, 6 Aug 1999 13:22:01 -0400 (EDT) | From | Chuck Lever <> | Subject | Re: clustering page-ins |
| |
On Fri, 6 Aug 1999, Jamie Lokier wrote: > The only problem with pre-faulting is you don't get real faults to tell > you what the process is up to... but with your ultra-refined readahead > strategy, pre-faulting a cluster would be the final stroke of genius, > no? :-)
pre-faulting would need to not prevent a minor fault, so that we could track the application's progress through the file to determine that it is still streaming. but if a minor fault still occurs, even if it's mostly a nop, does that make anything more efficient?
> Chuck, if your read-ahead code is working could you send me the patch?
i have something i can send you now. it doesn't adjust the trigger point automatically yet (it still uses a trigger point a half cluster before the end of the read-ahead window), since i've been on vacation. i'll try to post it later today.
i've been testing with grep and mpg123, so i know my implementation is working. but i don't have anything real that gobbles data fast enough to really compare with the old filemap_nopage.
- Chuck Lever -- corporate: <chuckl@netscape.com> personal: <chucklever@netscape.net> or <cel@monkey.org>
The Linux Scalability project: http://www.citi.umich.edu/projects/linux-scalability/
- 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/
| |