From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from relay.sgi.com (relay3.corp.sgi.com [198.149.34.15]) by oss.sgi.com (Postfix) with ESMTP id 00F9829DF9 for ; Thu, 21 Nov 2013 12:08:34 -0600 (CST) Received: from cuda.sgi.com (cuda1.sgi.com [192.48.157.11]) by relay3.corp.sgi.com (Postfix) with ESMTP id 8B508AC00D for ; Thu, 21 Nov 2013 10:08:30 -0800 (PST) Received: from mailgw1.uni-kl.de (mailgw1.uni-kl.de [131.246.120.220]) by cuda.sgi.com with ESMTP id C8tfYBKOosBzzxxb (version=TLSv1 cipher=AES256-SHA bits=256 verify=NO) for ; Thu, 21 Nov 2013 10:08:29 -0800 (PST) Received: from itwm2.itwm.fhg.de (itwm2.itwm.fhg.de [131.246.191.3]) by mailgw1.uni-kl.de (8.14.3/8.14.3/Debian-9.4) with ESMTP id rALI8Qcm025665 (version=TLSv1/SSLv3 cipher=EDH-RSA-DES-CBC3-SHA bits=168 verify=NOT) for ; Thu, 21 Nov 2013 19:08:26 +0100 Message-ID: <528E4C19.70507@itwm.fraunhofer.de> Date: Thu, 21 Nov 2013 19:08:25 +0100 From: Bernd Schubert MIME-Version: 1.0 Subject: Re: bio_add_page rw mode check by merge_bvec_fn References: <528E4A50.7040609@itwm.fraunhofer.de> In-Reply-To: <528E4A50.7040609@itwm.fraunhofer.de> List-Id: XFS Filesystem from SGI List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset="us-ascii"; Format="flowed" Errors-To: xfs-bounces@oss.sgi.com Sender: xfs-bounces@oss.sgi.com To: "linux-fsdevel@vger.kernel.org" , xfs@oss.sgi.com, linux-ext4@vger.kernel.org Cc: NeilBrown , Jens Axboe , Kent Overstreet [CCing linux-fsdevel instead of accidental fhgfs cc] On 11/21/2013 07:00 PM, Bernd Schubert wrote: > Hi all, > > raid5_mergeable_bvec() (q->merge_bvec_fn) checks for the rw-type and > makes decisions based on that if the bio is mergable or not. But so far > this value is only initialized on calling submit_bio(), > at least not by ext4 and xfs. I have not checked other file system so far. > > Question 1: merge_bvec_fn is supposed to be removed, is is > still worth to fix the usage of merge_bvec_fn? > > Question 2: If it is still supposed to be fixed, how do want to have > it? So far I have two patches for xfs and ext4 to set the rw type > directly there, but maybe bio_add_page should get an int rw_type > parameter and set bio->bi_rw itself? > > > > diff --git a/fs/ext4/page-io.c b/fs/ext4/page-io.c > index d488f80..4cddc15 100644 > --- a/fs/ext4/page-io.c > +++ b/fs/ext4/page-io.c > @@ -370,6 +370,7 @@ static int io_submit_init_bio(struct ext4_io_submit *io, > bio->bi_bdev = bh->b_bdev; > bio->bi_end_io = ext4_end_bio; > bio->bi_private = ext4_get_io_end(io->io_end); > + bio->bi_rw |= io->io_op; > io->io_bio = bio; > io->io_next_block = bh->b_blocknr; > return 0; > > diff --git a/fs/xfs/xfs_aops.c b/fs/xfs/xfs_aops.c > index 71c8c9d..b48048f 100644 > --- a/fs/xfs/xfs_aops.c > +++ b/fs/xfs/xfs_aops.c > @@ -401,7 +401,8 @@ xfs_submit_ioend_bio( > > STATIC struct bio * > xfs_alloc_ioend_bio( > - struct buffer_head *bh) > + struct buffer_head *bh, > + struct writeback_control *wbc) > { > int nvecs = bio_get_nr_vecs(bh->b_bdev); > struct bio *bio = bio_alloc(GFP_NOIO, nvecs); > @@ -409,6 +410,7 @@ xfs_alloc_ioend_bio( > ASSERT(bio->bi_private == NULL); > bio->bi_sector = bh->b_blocknr * (bh->b_size >> 9); > bio->bi_bdev = bh->b_bdev; > + bio->bi_rw |= (wbc->sync_mode == WB_SYNC_ALL) ? WRITE_SYNC : WRITE; > return bio; > } > > @@ -511,7 +513,8 @@ xfs_submit_ioend( > > if (!bio) { > retry: > - bio = xfs_alloc_ioend_bio(bh); > + bio = xfs_alloc_ioend_bio(bh, wbc); > + > } else if (bh->b_blocknr != lastblock + 1) { > xfs_submit_ioend_bio(wbc, ioend, bio); > goto retry; > > > > Thanks, > Bernd > _______________________________________________ xfs mailing list xfs@oss.sgi.com http://oss.sgi.com/mailman/listinfo/xfs