From owner-freebsd-hackers Sat Apr 27 12:12:10 1996 Return-Path: owner-hackers Received: (from root@localhost) by freefall.freebsd.org (8.7.3/8.7.3) id MAA00780 for hackers-outgoing; Sat, 27 Apr 1996 12:12:10 -0700 (PDT) Received: from hda.com (ip31-max1-fitch.zipnet.net [199.232.245.31]) by freefall.freebsd.org (8.7.3/8.7.3) with SMTP id MAA00774 Sat, 27 Apr 1996 12:12:07 -0700 (PDT) Received: (from dufault@localhost) by hda.com (8.6.11/8.6.9) id PAA09603; Sat, 27 Apr 1996 15:12:47 -0400 From: Peter Dufault Message-Id: <199604271912.PAA09603@hda.com> Subject: Re: O_SYNC To: bde@zeta.org.au (Bruce Evans) Date: Sat, 27 Apr 1996 15:12:46 -0400 (EDT) Cc: freebsd-doc@FreeBSD.org, jmacd@CS.Berkeley.EDU, freebsd-hackers@FreeBSD.org In-Reply-To: <199604270258.MAA07041@godzilla.zeta.org.au> from "Bruce Evans" at Apr 27, 96 12:58:55 pm Reply-to: hdalog@zipnet.net X-Mailer: ELM [version 2.4 PL24] Content-Type: text Sender: owner-hackers@FreeBSD.org X-Loop: FreeBSD.org Precedence: bulk > > >Isn't O_SYNC the POSIX way to do things? Why does FreeBSD have > >O_FSYNC, and why isn't it documented? > > POSIX.1 doesn't specify any way to sync files. Perhaps FreeBSD > has O_FSYNC as a compatibility hack, and it isn't documented because > it doesn't do anything. The new copy of POSIX.1 (P1003.1b-1993) adds O_DSYNC, O_RSYNC and O_SYNC in the open call. O_DSYNC versus O_SYNC permits data-only integrity versus full file info integrity (access stamps, etc). I haven't spent the time to puzzle out exactly what O_RSYNC means - that once you read it it better also be on the disk? There is a feature test macro "_POSIX_SYNCHRONIZED_IO" that has to be defined for these to be present. Anyone working off a new spec better keep an eye out for the change bars- I didn't realize this was new until this mail. -- Temporarily via "hdalog@zipnet.net"... Peter Dufault Real-Time Machine Control and Simulation HD Associates, Inc. Voice: 508 433 6936 dufault@hda.com Fax: 508 433 5267