From owner-freebsd-hackers Tue Aug 22 12:02:18 1995 Return-Path: hackers-owner Received: (from majordom@localhost) by freefall.FreeBSD.org (8.6.11/8.6.6) id MAA10060 for hackers-outgoing; Tue, 22 Aug 1995 12:02:18 -0700 Received: from godzilla.zeta.org.au (godzilla.zeta.org.au [203.2.228.34]) by freefall.FreeBSD.org (8.6.11/8.6.6) with ESMTP id MAA10054 for ; Tue, 22 Aug 1995 12:02:14 -0700 Received: (from bde@localhost) by godzilla.zeta.org.au (8.6.9/8.6.9) id FAA19377; Wed, 23 Aug 1995 05:01:59 +1000 Date: Wed, 23 Aug 1995 05:01:59 +1000 From: Bruce Evans Message-Id: <199508221901.FAA19377@godzilla.zeta.org.au> To: bde@zeta.org.au, jdli@linux.csie.nctu.edu.tw Subject: Re: Making a FreeBSD NFS server Cc: freebsd-hackers@FreeBSD.ORG Sender: hackers-owner@FreeBSD.ORG Precedence: bulk >> No, I mean the fix for ufs in -current. >> > Do you mean that "-o async" in current is really async now ? The commit mail said exactly what was fixed: "Honor -async mount option when doing the inode update." The -async mount option still isn't honored for directory updates. For creat() there used to be 2 sync writes, one for the inode and one for the directory. Now there is one sync write for the directory, so creat() is about twice as fast with "-o async" as before. The other times in my benchmarks show that it is still about 4 times slower under FreeBSD-2.2-current than it would be if all writes were asynchronous. Bruce