From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id ; Thu, 20 Sep 2001 14:22:06 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id ; Thu, 20 Sep 2001 14:21:46 -0400 Received: from [208.129.208.52] ([208.129.208.52]:48654 "EHLO xmailserver.org") by vger.kernel.org with ESMTP id ; Thu, 20 Sep 2001 14:21:42 -0400 Message-ID: X-Mailer: XFMail 1.5.0 on Linux X-Priority: 3 (Normal) Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 8bit MIME-Version: 1.0 In-Reply-To: <20010920010502.A7960@redhat.com> Date: Thu, 20 Sep 2001 11:25:13 -0700 (PDT) From: Davide Libenzi To: Benjamin LaHaise Subject: Re: [PATCH] /dev/epoll update ... Cc: linux-kernel@vger.kernel.org Cc: linux-kernel@vger.kernel.org, "Christopher K. St. John" Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On 20-Sep-2001 Benjamin LaHaise wrote: > On Wed, Sep 19, 2001 at 11:43:57PM -0500, Christopher K. St. John wrote: >> Sorry, bad editing, that should be: >> >> Assume a large but bursty current of low bandwidth >> high latency connections instead of a continuous steady >> flood of high bandwidth low latency connections. > > Isn't asynchronous io a better model for that case? The advantage /dev/epoll has compared to aio_* and RTsig is 1) multiple event delivery/system call 2) less user<->kernel memory moves The concept is very similar anyway coz you basically have to initiate the io-call and wait for an event. The difference is how events are collected. - Davide