From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jan Kara Subject: Re: [PATCH 1/6] ext4: cache extent hole in extent status tree for ext4_da_map_blocks() Date: Mon, 24 Nov 2014 10:34:34 +0100 Message-ID: <20141124093434.GC1888@quack.suse.cz> References: <1415961957-21621-1-git-send-email-jack@suse.cz> <1415961957-21621-2-git-send-email-jack@suse.cz> <20141123054335.GA4102@thunk.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: Jan Kara , linux-ext4@vger.kernel.org, Zheng Liu , Zheng Liu , Andreas Dilger To: Theodore Ts'o Return-path: Received: from cantor2.suse.de ([195.135.220.15]:41935 "EHLO mx2.suse.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751529AbaKYOUa (ORCPT ); Tue, 25 Nov 2014 09:20:30 -0500 Content-Disposition: inline In-Reply-To: <20141123054335.GA4102@thunk.org> Sender: linux-ext4-owner@vger.kernel.org List-ID: On Sun 23-11-14 00:43:35, Ted Tso wrote: > On Fri, Nov 14, 2014 at 11:45:52AM +0100, Jan Kara wrote: > > From: Zheng Liu > > > > Currently extent status tree doesn't cache extent hole when a write > > looks up in extent tree to make sure whether a block has been allocated > > or not. In this case, we don't put extent hole in extent cache because > > later this extent might be removed and a new delayed extent might be > > added back. But it will cause a defect when we do a lot of writes. If > > we don't put extent hole in extent cache, the following writes also need > > to access extent tree to look at whether or not a block has been > > allocated. It brings a cache miss. This commit fixes this defect. > > Also if the inode doesn't have any extent, this extent hole will be > > cached as well. > > > > Cc: "Theodore Ts'o" > > Cc: Andreas Dilger > > Signed-off-by: Zheng Liu > > Signed-off-by: Jan Kara > > This patch (either when applied alone, or as part of the entire patch > series) is causing an xfstests regression test failure: I saw this in my testing as well but when investigating I somehow decided it wasn't my bug and then forgot about it. Anyway, I'll check again and resolve it either way. Honza > FSTYP -- ext4 > PLATFORM -- Linux/i686 kvm-xfstests 3.18.0-rc3-00537-g932e3f9 > MKFS_OPTIONS -- -q /dev/vdc > MOUNT_OPTIONS -- -o acl,user_xattr -o block_validity /dev/vdc /vdc > > ext4/003 4s ... [00:14:39] [00:14:48] - output mismatch (see /results/results-4k/ext4/003.out.bad) > --- tests/ext4/003.out 2014-10-31 10:13:03.000000000 -0400 > +++ /results/results-4k/ext4/003.out.bad 2014-11-23 00:14:48.935587897 -0500 > @@ -1,3 +1,3 @@ > QA output created by 003 > -wrote 268435456/268435456 bytes at offset 0 > +wrote 241762304/268435456 bytes at offset 0 > XXX Bytes, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec) > ... > (Run 'diff -u tests/ext4/003.out /results/results-4k/ext4/003.out.bad' to see the entire diff) > > - Ted -- Jan Kara SUSE Labs, CR