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, 2 May 2002 07:10:32 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id ; Thu, 2 May 2002 07:10:31 -0400 Received: from rwcrmhc53.attbi.com ([204.127.198.39]:11458 "EHLO rwcrmhc53.attbi.com") by vger.kernel.org with ESMTP id ; Thu, 2 May 2002 07:10:30 -0400 From: "Guillaume Boissiere" To: Dave Hansen Date: Thu, 2 May 2002 07:10:14 -0400 MIME-Version: 1.0 Subject: Re: [STATUS 2.5] May 1, 2002 (BKL status) CC: linux-kernel@vger.kernel.org, kernel-janitor-discuss@lists.sourceforge.net Message-ID: <3CD0E656.5451.7D1A7885@localhost> In-Reply-To: <3CD0047B.4060605@us.ibm.com> X-mailer: Pegasus Mail for Windows (v4.01) Content-type: text/plain; charset=US-ASCII Content-transfer-encoding: 7BIT Content-description: Mail message body Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org > I may not be the leading BKL expert, but I play one on TV :) > > Perhaps one of the kernel-janitor people would like to assist me with > this (cc'ing that list). I'd be willing to keep a web page to list all > current BKL uses and keep track of them as they are removed/added > Perhaps a set of web pages which resemble the directory structure of the > kernel tree would be helpful?? > > Here's a good question for kernel-janitor, and anyone else who's > interested, what format describing BKL use would most encourage you to > go and remove it? We already have Rick Lindsley's Global spinlock list: > http://prdownloads.sourceforge.net/lse/locking_doc-2.4.16 . The BKL > use in there is somewhat dated, but might be a good start. > > I have some awk scripts that I use on each new kernel release to check > for new and removed uses of the BKL. I can adapt these to start > checking new BK changesets for BKL changes. I'm by no means a BKL expert, not even on TV :-) , but here is a suggestion. What about a listing sorted by how the BKL should be replaced, something like this: BKL --> dma_spin_lock - alpha/kernel/ptrace.c Held while doing a ptrace (John Doe) - drivers/block/rd.c Held while doing whatever (?) - ... BKL --> xprt_lock - ... BKL --> no lock required - ... BKL instances that will NOT be removed - ... The advantage is that it should be easier to group together instances that require the same type of work. Comments? -- Guillaume