[patch] I/O scheduler starvation

From: Jens Axboe (axboe@kernel.org)
Date: Tue May 29 2001 - 11:04:29 EST


Hi,

Going over this a bit, an deviant from the original design of
elevator_linus seems to have snuck in. Well, almost, I've been aware of
this for some time...

The idea was _not_ to account request merges, since they will (in by far
the most cases) incur the cost of a seek. Instead we just account seeks,
any seek (and please don't start a debate on short vs long seeks again,
I'm not listening). The attached patch corrects this, and also corrects
the initial sequence numbers (typically this would be sequence / 8
number of seeks allowed to pass a request now).

I'll do some formal timing on this later.

-- 
Jens Axboe


- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/



This archive was generated by hypermail 2b29 : Thu May 31 2001 - 21:00:40 EST