From owner-freebsd-hackers@FreeBSD.ORG Fri May 16 22:22:33 2008 Return-Path: Delivered-To: freebsd-hackers@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 14785106564A for ; Fri, 16 May 2008 22:22:33 +0000 (UTC) (envelope-from gonzo@pbxpress.com) Received: from bugor.portaone.com (bugor.portaone.com [65.61.203.147]) by mx1.freebsd.org (Postfix) with ESMTP id EE9DC8FC14 for ; Fri, 16 May 2008 22:22:32 +0000 (UTC) (envelope-from gonzo@pbxpress.com) Received: from mail.pbxpress.com ([65.61.203.142] helo=leaf.pbxpress.com) by bugor.portaone.com (8.11.3/8.11.3) with ESMTP (TLSv1:AES256-SHA:256)id 1Jx7vr-000BAI-Ic for freebsd-hackers@freebsd.org; Fri, 16 May 2008 14:57:45 -0700 Received: from jeeves.bluezbox.com (82.193.105.77.ipnet.kiev.ua [82.193.105.77]) (authenticated bits=0) by leaf.pbxpress.com (8.13.3/8.13.3) with ESMTP id m4GM17oR029048 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO) for ; Fri, 16 May 2008 15:01:15 -0700 (PDT) (envelope-from gonzo@pbxpress.com) Message-ID: <482E034C.2080402@pbxpress.com> Date: Sat, 17 May 2008 00:57:32 +0300 From: Oleksandr Tymoshenko User-Agent: Thunderbird 2.0.0.6 (X11/20070907) MIME-Version: 1.0 To: freebsd-hackers@freebsd.org Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit X-Spam-Status: No, recieved from trusted server Subject: aio_write issues X-BeenThere: freebsd-hackers@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Technical Discussions relating to FreeBSD List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 16 May 2008 22:22:33 -0000 It looks like aio_write is out of sync with its man page(if it ever has been). man 2 aio_write states: If O_APPEND is set for iocb->aio_fildes, aio_write() operations append to the file in the same order as the calls were made. Though simple test[1] reveals that it's not true. Order is not preserved. In addition the rest of the paragraph contradicts to SUSv3 specification of aio_write (or I got that part of man page wrong, because function behaves just as should be conforming to SUSv3). man page: If O_APPEND is not set for the file descriptor, the write operation will occur at the absolute position from the beginning of the file plus iocb->aio_offset. excerpt from IEEE 1003.1: If O_APPEND is not set for the file descriptor aio_fildes, then the requested operation shall take place at the absolute position in the file as given by aio_offset, as if lseek() were called immediately prior to the operation with an offset equal to aio_offset and a whence equal to SEEK_SET. I'd like to know if I got things right in order to avoid fixing things that are not broken. [1] http://people.freebsd.org/~gonzo/aio.c -- gonzo