RE: [patch, CFT] improved disk read latency

[Giuliano Pochini]

> +             if (__rq->elevator_sequence-- <= 0) {
> +                     /*
> +                      * OK, we've exceeded someone's latency limit.
> +                      * But we still continue to look for merges,
> +                      * because they're so much better than seeks.
> +                      */
> +                     merge_only = 1;

That's true for hard discs, but for most removables seek time is
small compared to write time. Think about a magneto-optical drive or
a ZIP. Seek is about 20-50ms while writing goes at 300-1000KiB/s.

Bye.