Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S262636AbVAFMBb (ORCPT ); Thu, 6 Jan 2005 07:01:31 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S262630AbVAFMBb (ORCPT ); Thu, 6 Jan 2005 07:01:31 -0500 Received: from ns.virtualhost.dk ([195.184.98.160]:57019 "EHLO virtualhost.dk") by vger.kernel.org with ESMTP id S262636AbVAFMBE (ORCPT ); Thu, 6 Jan 2005 07:01:04 -0500 Date: Thu, 6 Jan 2005 13:00:34 +0100 From: Jens Axboe To: Nick Piggin Cc: Andrew Morton , riel@redhat.com, marcelo.tosatti@cyclades.com, andrea@suse.de, linux-kernel@vger.kernel.org Subject: Re: memory barrier in ll_rw_blk.c (was Re: [PATCH][5/?] count writeback pages in nr_scanned) Message-ID: <20050106120034.GK17821@suse.de> References: <41DCB577.9000205@yahoo.com.au> <20050105202611.65eb82cf.akpm@osdl.org> <41DCC014.80007@yahoo.com.au> <20050105204706.0781d672.akpm@osdl.org> <41DCC4C6.8000205@yahoo.com.au> <20050106080649.GE17821@suse.de> <41DCF3EC.3090506@yahoo.com.au> <20050106083251.GH17821@suse.de> <41DCFC91.5060703@yahoo.com.au> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <41DCFC91.5060703@yahoo.com.au> Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 1880 Lines: 55 On Thu, Jan 06 2005, Nick Piggin wrote: > Jens Axboe wrote: > >On Thu, Jan 06 2005, Nick Piggin wrote: > > > > >>No that's right... but between the prepare_to_wait and the io_schedule, > >>get_request takes the lock and checks nr_requests. I think we are safe? > > > > > >It looks like it, yes you are right. But it looks to be needed a few > >lines further down instead, though :-) > > > >===== drivers/block/ll_rw_blk.c 1.281 vs edited ===== > >--- 1.281/drivers/block/ll_rw_blk.c 2004-12-01 09:13:57 +01:00 > >+++ edited/drivers/block/ll_rw_blk.c 2005-01-06 09:32:19 +01:00 > >@@ -1630,11 +1630,11 @@ > > if (rl->count[rw] < queue_congestion_off_threshold(q)) > > clear_queue_congested(q, rw); > > if (rl->count[rw]+1 <= q->nr_requests) { > >- smp_mb(); > > if (waitqueue_active(&rl->wait[rw])) > > wake_up(&rl->wait[rw]); > > blk_clear_queue_full(q, rw); > > } > >+ smp_mb(); > > if (unlikely(waitqueue_active(&rl->drain)) && > > !rl->count[READ] && !rl->count[WRITE]) > > wake_up(&rl->drain); > > > > Yes, looks like you're right there. > > Any point in doing it like this > > if (!rl->count[READ] && !rl->count[WRITE]) { > smb_mb(); > if (unlikely(waitqueue_active(...))) > wake_up() > } > > I wonder? I don't have any feeling of how memory barriers impact performance > on a very parallel system with CPUs that do lots of memory reordering like > POWER5. I had the same idea - I think it's a good idea, it's definitely an optimization worth doing. -- Jens Axboe - 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/