Messages in this thread | | | Date | Thu, 18 Mar 1999 17:13:16 -0500 | From | Arvind Sankar <> | Subject | Re: disk head scheduling |
| |
On Thu, Mar 18, 1999 at 10:52:44PM +0100, Ingo Molnar wrote: > > On Thu, 18 Mar 1999, Arvind Sankar wrote: > > > > > Another point is that IN_ORDER seems to be called only for two requests > > > > on the same device, so no idea why it compares the device numbers. > > > > > > no, all requests (for all devices) are in a single 'queue'. (Per-major > > > device queues is candidate 2.3 feature, it's really simple) > > > > oh. I realized that it was at least per-major, but it's global, huh? > > the sorting (the elevator algorithm) is per-device. The queue is > technically one queue, but thats just an implementation matter. (and of > course it matters in the case when we hit the queue length limit)
ah. But it is per major-device linking. It would be nice to have per physical device linking, then we could do a two-way elevator algorithm.
-- arvind
- 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/
| |