Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1762606AbYBHLjD (ORCPT ); Fri, 8 Feb 2008 06:39:03 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1753066AbYBHLix (ORCPT ); Fri, 8 Feb 2008 06:38:53 -0500 Received: from brick.kernel.dk ([87.55.233.238]:4561 "EHLO kernel.dk" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752103AbYBHLiw (ORCPT ); Fri, 8 Feb 2008 06:38:52 -0500 Date: Fri, 8 Feb 2008 12:38:49 +0100 From: Jens Axboe To: Linus Torvalds Cc: Ingo Molnar , linux-kernel@vger.kernel.org, Alan.Brunelle@hp.com, arjan@linux.intel.com, dgc@sgi.com, npiggin@suse.de, Andrew Morton , Vegard Nossum , Pekka Enberg Subject: Re: [patch] block layer: kmemcheck fixes Message-ID: <20080208113847.GI15220@kernel.dk> References: <1202375945-29525-1-git-send-email-jens.axboe@oracle.com> <1202375945-29525-5-git-send-email-jens.axboe@oracle.com> <20080207100738.GB7716@elte.hu> <20080207101727.GE15220@kernel.dk> <20080207102534.GB16735@elte.hu> <20080207103136.GG15220@kernel.dk> <20080207104901.GF16735@elte.hu> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 3931 Lines: 132 On Thu, Feb 07 2008, Linus Torvalds wrote: > > > On Thu, 7 Feb 2008, Ingo Molnar wrote: > > INIT_HLIST_NODE(&rq->hash); > > RB_CLEAR_NODE(&rq->rb_node); > > - rq->ioprio = 0; > > - rq->buffer = NULL; > > - rq->ref_count = 1; > > - rq->q = q; > > - rq->special = NULL; > > - rq->data_len = 0; > > - rq->data = NULL; > > - rq->nr_phys_segments = 0; > > - rq->sense = NULL; > > - rq->end_io = NULL; > > - rq->end_io_data = NULL; > > - rq->completion_data = NULL; > > - rq->next_rq = NULL; > > + rq->completion_data = NULL; > > + /* rq->elevator_private */ > > + /* rq->elevator_private2 */ > > + /* rq->rq_disk */ > > + /* rq->start_time */ > > + rq->nr_phys_segments = 0; > > + /* rq->nr_hw_segments */ > > + rq->ioprio = 0; > > + rq->special = NULL; > > + rq->buffer = NULL; > ... > > Can we please just stop doing these one-by-one assignments, and just do > something like > > memset(rq, 0, sizeof(*rq)); > rq->q = q; > rq->ref_count = 1; > INIT_HLIST_NODE(&rq->hash); > RB_CLEAR_NODE(&rq->rb_node); > > instead? > > The memset() is likely faster and smaller than one-by-one assignments > anyway, even if the one-by-ones can avoid initializing some field or there > ends up being a double initialization.. Looked into this approach and we can't currently do that here, since some members of the request are being set from blk_alloc_request() and then from the io scheduler attaching private data to it. So we have to preserve ->cmd_flags and ->elevator_private and ->elevator_private2 at least. Now rq_init() is also used for stored requests, so we cannot just rely on clearing at allocation time. So I'd prefer being a little conservative here. The below reorders rq_init() a bit and clears some more members to be on the safer side, adding comments to why we cannot memset and an associated comment in blkdev.h. diff --git a/block/blk-core.c b/block/blk-core.c index 4afb39c..fba4ca7 100644 --- a/block/blk-core.c +++ b/block/blk-core.c @@ -102,27 +102,38 @@ struct backing_dev_info *blk_get_backing_dev_info(struct block_device *bdev) } EXPORT_SYMBOL(blk_get_backing_dev_info); +/* + * We can't just memset() the structure, since the allocation path + * already stored some information in the request. + */ void rq_init(struct request_queue *q, struct request *rq) { INIT_LIST_HEAD(&rq->queuelist); INIT_LIST_HEAD(&rq->donelist); - - rq->errors = 0; + rq->q = q; + rq->sector = rq->hard_sector = (sector_t) -1; + rq->nr_sectors = rq->hard_nr_sectors = 0; + rq->current_nr_sectors = rq->hard_cur_sectors = 0; rq->bio = rq->biotail = NULL; INIT_HLIST_NODE(&rq->hash); RB_CLEAR_NODE(&rq->rb_node); + rq->rq_disk = NULL; + rq->nr_phys_segments = 0; + rq->nr_hw_segments = 0; rq->ioprio = 0; + rq->special = NULL; rq->buffer = NULL; + rq->tag = -1; + rq->errors = 0; rq->ref_count = 1; - rq->q = q; - rq->special = NULL; + rq->cmd_len = 0; + memset(rq->cmd, 0, sizeof(rq->cmd)); rq->data_len = 0; + rq->sense_len = 0; rq->data = NULL; - rq->nr_phys_segments = 0; rq->sense = NULL; rq->end_io = NULL; rq->end_io_data = NULL; - rq->completion_data = NULL; rq->next_rq = NULL; } diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h index 90392a9..e1888cc 100644 --- a/include/linux/blkdev.h +++ b/include/linux/blkdev.h @@ -137,7 +137,9 @@ enum rq_flag_bits { #define BLK_MAX_CDB 16 /* - * try to put the fields that are referenced together in the same cacheline + * try to put the fields that are referenced together in the same cacheline. + * if you modify this structure, be sure to check block/blk-core.c:rq_init() + * as well! */ struct request { struct list_head queuelist; -- 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/