Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S262326AbVCPKT4 (ORCPT ); Wed, 16 Mar 2005 05:19:56 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S262331AbVCPKT4 (ORCPT ); Wed, 16 Mar 2005 05:19:56 -0500 Received: from mx2.elte.hu ([157.181.151.9]:58517 "EHLO mx2.elte.hu") by vger.kernel.org with ESMTP id S262326AbVCPKTx (ORCPT ); Wed, 16 Mar 2005 05:19:53 -0500 Date: Wed, 16 Mar 2005 11:19:06 +0100 From: Ingo Molnar To: Andrew Morton Cc: rostedt@goodmis.org, rlrevell@joe-job.com, linux-kernel@vger.kernel.org Subject: Re: [patch 0/3] j_state_lock, j_list_lock, remove-bitlocks Message-ID: <20050316101906.GA17328@elte.hu> References: <20050315120053.GA4686@elte.hu> <20050315133540.GB4686@elte.hu> <20050316085029.GA11414@elte.hu> <20050316011510.2a3bdfdb.akpm@osdl.org> <20050316095155.GA15080@elte.hu> <20050316020408.434cc620.akpm@osdl.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20050316020408.434cc620.akpm@osdl.org> User-Agent: Mutt/1.4.1i X-ELTE-SpamVersion: MailScanner 4.31.6-itk1 (ELTE 1.2) SpamAssassin 2.63 ClamAV 0.73 X-ELTE-VirusStatus: clean X-ELTE-SpamCheck: no X-ELTE-SpamCheck-Details: score=-4.9, required 5.9, autolearn=not spam, BAYES_00 -4.90 X-ELTE-SpamLevel: X-ELTE-SpamScore: -4 Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 1224 Lines: 29 * Andrew Morton wrote: > > > There's a little lock ranking diagram in jbd.h which tells us that > > > these locks nest inside j_list_lock and j_state_lock. So I guess > > > you'll need to turn those into semaphores. > > > > indeed. I did this (see the three followup patches, against BK-curr), > > and it builds/boots/works just fine on an ext3 box. Do we want to try > > this in -mm? > > ooh, I'd rather not. I spent an intense three days removing all the > sleeping locks from ext3 (and three months debugging the result). > Ended up gaining 1000% on 16-way. > > Putting them back in will really hurt the SMP performance. seems like turning the bitlocks into spinlocks is the best option then. We'd need one lock in buffer_head (j_state_lock, renamed to something more sensible like b_private_lock), and one lock in journal_head (j_list_lock) i guess. How much would the +4/+8 bytes size increase in buffer_head [on SMP] be frowned upon? Ingo - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/