All of lore.kernel.org
 help / color / mirror / Atom feed
From: bfields@fieldses.org (J. Bruce Fields)
To: Christoph Hellwig <hch@infradead.org>
Cc: linux-nfs@vger.kernel.org
Subject: Re: odd sillyrename bug in 4.6 / 4.7-rc
Date: Thu, 16 Jun 2016 13:29:52 -0400	[thread overview]
Message-ID: <20160616172952.GA17604@fieldses.org> (raw)
In-Reply-To: <20160616150146.GA14015@infradead.org>

On Thu, Jun 16, 2016 at 08:01:47AM -0700, Christoph Hellwig wrote:
> I've run into an odd bug when testing the clone / reflink functionality,
> and managed to isolate a testcase not relying on clones.  The issue is
> that after a lot of direct I/O hitting ENOSPC on a NFS mount I run
> into a silly rename that never gets cleaned up after cleaning up the
> directory.

I have a redhat bug for what sounds like the same behavior.  I spent
some time looking at it but I don't think I got anywere.

	https://bugzilla.redhat.com/show_bug.cgi?id=1132610

(But you probably can't see that right now; I'll see if I can open it.)

--b.

> 
> Note that so far I've only managed to reproduce it against a Linux
> server running XFS (ext4 and btrfs seem fine) despite looking like
> a client bug, and although once in a while a single run of the testcase
> cause the bug I usually need a few iterations.
> 
> All over it seems like some odd dentry refcount race where we see
> a spurious reference for one reason or another.
> 
> 
> Below is a patch to xfstests to add this reproducer:
> 
> 
> diff --git a/tests/generic/700 b/tests/generic/700
> new file mode 100755
> index 0000000..33505b0
> --- /dev/null
> +++ b/tests/generic/700
> @@ -0,0 +1,130 @@
> +#! /bin/bash
> +# FS QA Test No. 700
> +#
> +# Dumbed down Copy of the original version of generic/187 that reproduced
> +# NFS silly rename issues.
> +#-----------------------------------------------------------------------
> +# Copyright (c) 2015, Oracle and/or its affiliates.  All Rights Reserved.
> +#
> +# This program is free software; you can redistribute it and/or
> +# modify it under the terms of the GNU General Public License as
> +# published by the Free Software Foundation.
> +#
> +# This program is distributed in the hope that it would be useful,
> +# but WITHOUT ANY WARRANTY; without even the implied warranty of
> +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
> +# GNU General Public License for more details.
> +#
> +# You should have received a copy of the GNU General Public License
> +# along with this program; if not, write the Free Software Foundation,
> +# Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
> +#-----------------------------------------------------------------------
> +
> +seq=`basename $0`
> +seqres=$RESULT_DIR/$seq
> +echo "QA output created by $seq"
> +
> +here=`pwd`
> +tmp=/tmp/$$
> +status=1    # failure is the default!
> +trap "_cleanup; exit \$status" 0 1 2 3 15
> +
> +_cleanup()
> +{
> +    cd /
> +    rm -rf $tmp.* $testdir
> +}
> +
> +# get standard environment, filters and checks
> +. ./common/rc
> +. ./common/filter
> +
> +# real QA test starts here
> +_supported_os Linux
> +_require_xfs_io_command "falloc"
> +_require_xfs_io_command "fpunch"
> +test $FSTYP = "btrfs" && _notrun "Can't fragment free space on btrfs."
> +_require_odirect
> +
> +rm -f $seqres.full
> +
> +_fragment_freesp()
> +{
> +	file=$1
> +
> +	# consume nearly all available space (leave ~1MB)
> +	avail=`_get_available_space $SCRATCH_MNT`
> +	echo "$avail bytes left"
> +	filesize=$((avail - 1048576))
> +	$XFS_IO_PROG -fc "truncate $filesize" $file
> +
> +	chunks=20
> +	chunksizemb=$((filesize / chunks / 1048576))
> +	seq 1 $chunks | while read f; do
> +		echo "$((f * chunksizemb)) file size $f / 20"
> +		$XFS_IO_PROG -fc "falloc -k $(( (f - 1) * chunksizemb))m ${chunksizemb}m" $file
> +	done
> +
> +	chunks=100
> +	chunksizemb=$((filesize / chunks / 1048576))
> +	seq 80 $chunks | while read f; do
> +		echo "$((f * chunksizemb)) file size $f / $chunks"
> +		$XFS_IO_PROG -fc "falloc -k $(( (f - 1) * chunksizemb))m ${chunksizemb}m" $file
> +	done
> +
> +	filesizemb=$((filesize / 1048576))
> +	$XFS_IO_PROG -fc "falloc -k 0 ${filesizemb}m" $file
> +
> +	# Try again anyway
> +	avail=`_get_available_space $SCRATCH_MNT`
> +	$XFS_IO_PROG -fc "pwrite -S 0x65 0 $avail" ${file}.${i}
> +
> +	# Punch out whatever we need
> +	seq 1 $((nr * 4)) | while read f; do
> +		$XFS_IO_PROG -f -c "fpunch $((f * 2 * blksz)) $blksz" $file
> +	done
> +}
> +
> +echo "Format and mount"
> +_scratch_mkfs > $seqres.full 2>&1
> +_scratch_mount >> $seqres.full 2>&1
> +
> +testdir=$SCRATCH_MNT/test-$seq
> +mkdir $testdir
> +
> +echo "Create the original files"
> +blksz=65536
> +nr=1024
> +filesize=$((blksz * nr))
> +_pwrite_byte 0x61 0 $filesize $testdir/file1 >> $seqres.full
> +_pwrite_byte 0x62 0 $filesize $testdir/file2 >> $seqres.full
> +seq 0 2 $((nr-1)) | while read f; do
> +#	_reflink_range $testdir/file1 $((blksz * f)) $testdir/file3 $((blksz * f)) $blksz >> $seqres.full
> +	touch $testdir/file3
> +	_pwrite_byte 0x61 $((blksz * f)) $blksz $testdir/file3.chk >> $seqres.full
> +done
> +seq 1 2 $((nr-1)) | while read f; do
> +	touch $testdir/file3
> +#	_reflink_range $testdir/file2 $((blksz * f)) $testdir/file3 $((blksz * f)) $blksz >> $seqres.full
> +	_pwrite_byte 0x62 $((blksz * f)) $blksz $testdir/file3.chk >> $seqres.full
> +done
> +_scratch_cycle_mount
> +_fragment_freesp $testdir/bigfile >> $seqres.full 2>&1
> +filesize=$((blksz * nr))
> +_scratch_cycle_mount
> +
> +echo "Overwrite"
> +cowoff=$((filesize / 4))
> +cowsz=$((filesize / 2))
> +$XFS_IO_PROG -d -f -c "pwrite -S 0x63 -b $cowsz $cowoff $cowsz" $testdir/file3 >> $seqres.full
> +_pwrite_byte 0x63 $cowoff $cowsz $testdir/file3.chk >> $seqres.full
> +
> +# mount / unmount
> +_scratch_cycle_mount
> +
> +echo "Deleting"
> +rm -r $testdir
> +
> +# success, all done
> +status=0
> +exit
> diff --git a/tests/generic/700.out b/tests/generic/700.out
> new file mode 100644
> index 0000000..9f7ce0f
> --- /dev/null
> +++ b/tests/generic/700.out
> @@ -0,0 +1,5 @@
> +QA output created by 700
> +Format and mount
> +Create the original files
> +Overwrite
> +Deleting
> diff --git a/tests/generic/group b/tests/generic/group
> index 2bd40a1..46a796a 100644
> --- a/tests/generic/group
> +++ b/tests/generic/group
> @@ -354,6 +354,7 @@
>  349 blockdev quick rw
>  350 blockdev quick rw
>  351 blockdev quick rw
> +700 auto nfs
>  923 auto quick clone
>  924 auto quick clone
>  925 auto quick clone
> --
> To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

  reply	other threads:[~2016-06-16 17:29 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-06-16 15:01 odd sillyrename bug in 4.6 / 4.7-rc Christoph Hellwig
2016-06-16 17:29 ` J. Bruce Fields [this message]
2016-06-17 11:31   ` Christoph Hellwig
2016-06-17 13:52     ` J. Bruce Fields

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20160616172952.GA17604@fieldses.org \
    --to=bfields@fieldses.org \
    --cc=hch@infradead.org \
    --cc=linux-nfs@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.