Date: Mon, 13 Nov 1995 23:25:53 -0800 From: David Greenman <davidg@root.com> To: uhclem%nemesis@fw.ast.com (Frank Durda IV) Cc: simonm@dcs.gla.ac.uk, davidg@root.com, current@freebsd.org Subject: Re: Disk I/O that binds Message-ID: <199511140725.XAA00188@corbin.Root.COM> In-Reply-To: Your message of "Mon, 13 Nov 95 22:34:00 %2B0700." <m0tFD4o-000J7hC@nemesis.lonestar.org>
next in thread | previous in thread | raw e-mail | index | archive | help
Attached is the patch to disable disk block sorting. It's interesting - I've been doing some tests here with disksort disabled and I'm actually seeing a performance increase. I'm sure the algorithm is doing what it was written to do...but... My (SCSI) drive supports tagged queueing, but even with that disabled things seem to be a bit faster. Hmmm. -DG Index: ufs_disksubr.c =================================================================== RCS file: /home/ncvs/src/sys/ufs/ufs/ufs_disksubr.c,v retrieving revision 1.19 diff -c -r1.19 ufs_disksubr.c *** 1.19 1995/09/16 17:04:06 --- ufs_disksubr.c 1995/11/14 06:30:43 *************** *** 81,86 **** --- 81,87 ---- return; } + #if 0 /* * If we lie after the first (currently active) request, then we * must locate the second request list and add ourselves to it. *************** *** 135,140 **** --- 136,146 ---- * Neither a second list nor a larger request... we go at the end of * the first list, which is the same as the end of the whole schebang. */ + #endif + + bq = ap->b_actf; + while (bq->b_actf) + bq = bq->b_actf; insert: bp->b_actf = bq->b_actf; bq->b_actf = bp;
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199511140725.XAA00188>