Re: [PATCH v2] fstests: btrfs/150 regression test for reading compressed data

[Liu Bo <bo.li.liu@oracle.com>]

On Tue, Sep 26, 2017 at 05:02:36PM +0800, Eryu Guan wrote:
> On Fri, Sep 22, 2017 at 05:21:27PM -0600, Liu Bo wrote:
> > We had a bug in btrfs compression code which could end up with a
> > kernel panic.
> > 
> > This is adding a regression test for the bug and I've also sent a
> > kernel patch to fix the bug.
> > 
> > The patch is "Btrfs: fix kernel oops while reading compressed data".
> > 
> > Signed-off-by: Liu Bo <bo.li.liu@oracle.com>
> 
> Hmm, I can't reproduce the panic with 4.13 kernel, which doesn't have
> the fix applied. Can you please help confirm if it panics on your test
> environment?
>

Yes, it is reproducible on my box, hrm...I'll be running it more times
to double check.

> > ---
> > v2: - Fix ambiguous copyright.
> >     - Use /proc/$pid/make-it-fail to specify IO failure
> >     - Use bash -c to run test only when pid is odd.
> >     - Add test to dangerous group.
> > 
> >  tests/btrfs/150     | 103 ++++++++++++++++++++++++++++++++++++++++++++++++++++
> >  tests/btrfs/150.out |   3 ++
> >  tests/btrfs/group   |   1 +
> >  3 files changed, 107 insertions(+)
> >  create mode 100755 tests/btrfs/150
> >  create mode 100644 tests/btrfs/150.out
> > 
> > diff --git a/tests/btrfs/150 b/tests/btrfs/150
> > new file mode 100755
> > index 0000000..8891c38
> > --- /dev/null
> > +++ b/tests/btrfs/150
> > @@ -0,0 +1,103 @@
> > +#! /bin/bash
> > +# FS QA Test btrfs/150
> > +#
> > +# This is a regression test which ends up with a kernel oops in btrfs.
> > +# It occurs when btrfs's read repair happens while reading a compressed
> > +# extent.
> > +# The patch to fix it is
> > +#	Btrfs: fix kernel oops while reading compressed data
> > +#
> > +#-----------------------------------------------------------------------
> > +# Copyright (c) 2017 Oracle.  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 -f $tmp.*
> > +}
> > +
> > +# get standard environment, filters and checks
> > +. ./common/rc
> > +. ./common/filter
> > +
> > +# remove previous $seqres.full before test
> > +rm -f $seqres.full
> > +
> > +# real QA test starts here
> > +
> > +# Modify as appropriate.
> > +_supported_fs btrfs
> > +_supported_os Linux
> > +_require_scratch
> > +_require_fail_make_request
> > +_require_scratch_dev_pool 2 
> 
> Trailing whitespace in above line.
> 
> > +
> > +SYSFS_BDEV=`_sysfs_dev $SCRATCH_DEV`
> > +enable_io_failure()
> > +{
> > +        echo 100 > $DEBUGFS_MNT/fail_make_request/probability
> > +        echo 1000 > $DEBUGFS_MNT/fail_make_request/times
> > +        echo 0 > $DEBUGFS_MNT/fail_make_request/verbose
> > +        echo 1 > $SYSFS_BDEV/make-it-fail
> > +}
> > +
> > +disable_io_failure()
> > +{
> > +        echo 0 > $SYSFS_BDEV/make-it-fail
> > +        echo 0 > $DEBUGFS_MNT/fail_make_request/probability
> > +        echo 0 > $DEBUGFS_MNT/fail_make_request/times
> > +}
> > +
> > +_scratch_pool_mkfs "-d raid1 -b 1G" >> $seqres.full 2>&1
> > +
> > +# It doesn't matter which compression algorithm we use.
> > +_scratch_mount -ocompress
> > +
> > +# Create a file with all data being compressed
> > +$XFS_IO_PROG -f -c "pwrite -W 0 8K" $SCRATCH_MNT/foobar | _filter_xfs_io
> > +
> > +# Raid1 consists of two copies and btrfs decides which copy to read by reader's
> > +# %pid.  Now we inject errors to copy #1 and copy #0 is good.  We want to read
> > +# the bad copy to trigger read-repair.
> > +while [[ -z $result ]]; do
> > +	# invalidate the page cache
> > +	$XFS_IO_PROG -f -c "fadvise -d 0 8K" $SCRATCH_MNT/foobar
> 
> Does 'echo 3 > /proc/sys/vm/drop_caches' work?
>

Yes, it works, drop_caches is system-wide, while here I'm just
dropping caches on this single inode.

Or are you implying that it's 'fadvise' that makes the test fail to
show oops?

thanks,

-liubo

> Thanks,
> Eryu
> 
> > +
> > +	enable_io_failure
> > +
> > +	result=$(bash -c "
> > +	if [ \$((\$\$ % 2)) == 1 ]; then
> > +		echo 1 > /proc/\$\$/make-it-fail
> > +		exec $XFS_IO_PROG -c \"pread 0 8K\" \$SCRATCH_MNT/foobar
> > +	fi")
> > +
> > +	disable_io_failure
> > +done
> > +
> > +# success, all done
> > +status=0
> > +exit
> > diff --git a/tests/btrfs/150.out b/tests/btrfs/150.out
> > new file mode 100644
> > index 0000000..c492c24
> > --- /dev/null
> > +++ b/tests/btrfs/150.out
> > @@ -0,0 +1,3 @@
> > +QA output created by 150
> > +wrote 8192/8192 bytes at offset 0
> > +XXX Bytes, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
> > diff --git a/tests/btrfs/group b/tests/btrfs/group
> > index 70c3f05..e73bb1b 100644
> > --- a/tests/btrfs/group
> > +++ b/tests/btrfs/group
> > @@ -152,3 +152,4 @@
> >  147 auto quick send
> >  148 auto quick rw
> >  149 auto quick send compress
> > +150 auto quick dangerous
> > -- 
> > 2.5.0
> > 
> > --
> > To unsubscribe from this list: send the line "unsubscribe fstests" in
> > the body of a message to majordomo@vger.kernel.org
> > More majordomo info at  http://vger.kernel.org/majordomo-info.html