Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id ; Tue, 26 Dec 2000 10:53:43 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id ; Tue, 26 Dec 2000 10:53:34 -0500 Received: from router-100M.swansea.linux.org.uk ([194.168.151.17]:18699 "EHLO the-village.bc.nu") by vger.kernel.org with ESMTP id ; Tue, 26 Dec 2000 10:53:25 -0500 Subject: Re: innd mmap bug in 2.4.0-test12 To: torvalds@transmeta.com (Linus Torvalds) Date: Mon, 25 Dec 2000 20:45:41 -0500 (EST) Cc: md@Linux.IT (Marco d'Itri), viro@math.psu.edu (Alexander Viro), linux-kernel@vger.kernel.org In-Reply-To: from "Linus Torvalds" at Dec 25, 2000 01:42:33 AM X-Mailer: ELM [version 2.5 PL3] MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Message-Id: From: Alan Cox Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org > The simple fix is along the lines of adding code to fsync() that walks the > inode page list and writes out dirty pages. > > The clever and clean fix is to split the inode page list into two lists, > one for dirty and one for clean pages, and only walk the dirty list. Like the patches that were floating around on the list for the past few months to make O_SYNC work. Could those be used for it ? > - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org Please read the FAQ at http://www.tux.org/lkml/