Messages in this thread | | | Date | Mon, 16 Nov 1998 09:23:08 -0800 (PST) | From | Tim Smith <> | Subject | Re: elevator algorithm bug in ll_rw_blk.c |
| |
On 16 Nov 1998, H. Peter Anvin wrote: > Do you really want to bother keeping track of them, especially since > modern drives have different numbers of sectors per track?
I'm not suggesting that all of that information should be used. I was just responding to the post that said it wasn't determinable on modern drives. We *can* get pretty much all the information to implement any fancy scheduling algorithm, at the cost of perhaps a few hours of probing the disk at formating time. Whether that is a good idea or not I am not qualified to judge.
> I think the current algorithm (a) works well, (b) avoids penalizing > part of the disk, and (c) is simple.
Rotational optimization might be worthwhile. On many modern drives, the rotational latency is larger than the seek latency for requests that aren't a large number of cylinders apart.
--Tim Smith
- 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/
| |