From: "Darrick J. Wong" <darrick.wong@oracle.com>
To: Eryu Guan <guaneryu@gmail.com>
Cc: linux-xfs@vger.kernel.org, linux-fsdevel@vger.kernel.org,
fstests@vger.kernel.org
Subject: Re: [PATCH 6/6] generic: test reflink side effects
Date: Mon, 15 Oct 2018 11:57:37 -0700 [thread overview]
Message-ID: <20181015185737.GC6703@magnolia> (raw)
In-Reply-To: <20181014154039.GQ17817@desktop>
On Sun, Oct 14, 2018 at 11:40:39PM +0800, Eryu Guan wrote:
> On Thu, Oct 04, 2018 at 05:19:45PM -0700, Darrick J. Wong wrote:
> > From: Darrick J. Wong <darrick.wong@oracle.com>
> >
> > Test that ctime gets updated and suid is cleared when we reflink.
> > Ensure we can't reflink about RLIMIT_FSIZE.
> >
> > Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
> > ---
> > tests/generic/934 | 63 ++++++++++++++++++++++++++++++++++++++
> > tests/generic/934.out | 2 +
> > tests/generic/935 | 48 +++++++++++++++++++++++++++++
> > tests/generic/935.out | 2 +
> > tests/generic/936 | 70 ++++++++++++++++++++++++++++++++++++++++++
> > tests/generic/936.out | 5 +++
> > tests/generic/937 | 81 +++++++++++++++++++++++++++++++++++++++++++++++++
> > tests/generic/937.out | 12 +++++++
> > tests/generic/group | 4 ++
> > 9 files changed, 287 insertions(+)
> > create mode 100755 tests/generic/934
> > create mode 100644 tests/generic/934.out
> > create mode 100755 tests/generic/935
> > create mode 100644 tests/generic/935.out
> > create mode 100755 tests/generic/936
> > create mode 100644 tests/generic/936.out
> > create mode 100755 tests/generic/937
> > create mode 100644 tests/generic/937.out
> >
> >
> > diff --git a/tests/generic/934 b/tests/generic/934
> > new file mode 100755
> > index 00000000..ec7ad092
> > --- /dev/null
> > +++ b/tests/generic/934
> > @@ -0,0 +1,63 @@
> > +#! /bin/bash
> > +# SPDX-License-Identifier: GPL-2.0
> > +# Copyright (c) 2018 Oracle, Inc. All Rights Reserved.
> > +#
> > +# FS QA Test No. 934
> > +#
> > +# Ensure that ctime is updated and capabilities are cleared when reflinking.
> > +#
> > +seq=`basename $0`
> > +seqres=$RESULT_DIR/$seq
> > +echo "QA output created by $seq"
> > +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
> > +. ./common/reflink
> > +
> > +# real QA test starts here
> > +_supported_fs generic
> > +_supported_os Linux
> > +_require_scratch_reflink
> > +_require_command "$GETCAP_PROG" getcap
> > +_require_command "$SETCAP_PROG" setcap
> > +
> > +rm -f $seqres.full
> > +
> > +_scratch_mkfs >>$seqres.full 2>&1
> > +_scratch_mount
> > +
> > +$XFS_IO_PROG -f -c "pwrite -S 0x18 0 1m" $SCRATCH_MNT/foo >>$seqres.full
> > +$XFS_IO_PROG -f -c "pwrite -S 0x20 0 1m" $SCRATCH_MNT/bar >>$seqres.full
> > +
> > +$SETCAP_PROG cap_setgid,cap_setuid+ep $SCRATCH_MNT/bar
> > +
> > +before_cap="$($GETCAP_PROG -v $SCRATCH_MNT/bar)"
> > +before_ctime="$(stat -c '%z' $SCRATCH_MNT/bar)"
> > +
> > +sleep 1
> > +
> > +$XFS_IO_PROG -c "reflink $SCRATCH_MNT/foo" $SCRATCH_MNT/bar >> $seqres.full 2>&1
> > +#trace-cmd record -p function_graph -l 'file_remove_privs' -g 'file_remove_privs' $XFS_IO_PROG -c "reflink $SCRATCH_MNT/foo" $SCRATCH_MNT/bar >> $seqres.full 2>&1
>
> Debug trace-cmd command?
Oops, I'll remove that.
> > +
> > +after_cap="$($GETCAP_PROG -v $SCRATCH_MNT/bar)"
> > +after_ctime="$(stat -c '%z' $SCRATCH_MNT/bar)"
> > +
> > +echo "$before_cap $before_ctime" >> $seqres.full
> > +echo "$after_cap $after_ctime" >> $seqres.full
> > +
> > +test "$before_cap" != "$after_cap" || echo "Expected capabilities to drop."
> > +test "$before_ctime" != "$after_ctime" || echo "Expected ctime to change."
> > +
> > +echo Silence is golden.
> > +status=0
> > +exit
> > diff --git a/tests/generic/934.out b/tests/generic/934.out
> > new file mode 100644
> > index 00000000..98f0bb3f
> > --- /dev/null
> > +++ b/tests/generic/934.out
> > @@ -0,0 +1,2 @@
> > +QA output created by 934
> > +Silence is golden.
> > diff --git a/tests/generic/935 b/tests/generic/935
> > new file mode 100755
> > index 00000000..3b0eaa95
> > --- /dev/null
> > +++ b/tests/generic/935
> > @@ -0,0 +1,48 @@
> > +#! /bin/bash
> > +# SPDX-License-Identifier: GPL-2.0
> > +# Copyright (c) 2018 Oracle, Inc. All Rights Reserved.
> > +#
> > +# FS QA Test No. 935
> > +#
> > +# Ensure that file size resource limits are respected when reflinking.
> > +#
> > +seq=`basename $0`
> > +seqres=$RESULT_DIR/$seq
> > +echo "QA output created by $seq"
> > +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
> > +. ./common/reflink
> > +
> > +# real QA test starts here
> > +_supported_fs generic
> > +_supported_os Linux
> > +_require_scratch_reflink
>
> _require_user
>
> > +
> > +rm -f $seqres.full
> > +
> > +_scratch_mkfs >>$seqres.full 2>&1
> > +_scratch_mount
> > +
> > +chmod a+rwx $SCRATCH_MNT
> > +$XFS_IO_PROG -f -c "pwrite -S 0x18 0 1m" $SCRATCH_MNT/foo >>$seqres.full
> > +su -s/bin/bash - $qa_user -c "ulimit -f 64 ; $XFS_IO_PROG -f -c \"reflink $SCRATCH_MNT/foo\" $SCRATCH_MNT/bar" >> $seqres.full 2>&1
> > +
> > +sz="$(stat -c '%s' $SCRATCH_MNT/bar)"
> > +if [ "$sz" -ne 0 ] && [ "$sz" -ne 65536 ]; then
> > + echo "Oddball file size $sz??"
> > +fi
> > +
> > +echo Silence is golden.
> > +status=0
> > +exit
> > diff --git a/tests/generic/935.out b/tests/generic/935.out
> > new file mode 100644
> > index 00000000..adb5e92a
> > --- /dev/null
> > +++ b/tests/generic/935.out
> > @@ -0,0 +1,2 @@
> > +QA output created by 935
> > +Silence is golden.
> > diff --git a/tests/generic/936 b/tests/generic/936
> > new file mode 100755
> > index 00000000..eacf9b5e
> > --- /dev/null
> > +++ b/tests/generic/936
> > @@ -0,0 +1,70 @@
> > +#! /bin/bash
> > +# SPDX-License-Identifier: GPL-2.0
> > +# Copyright (c) 2018 Oracle, Inc. All Rights Reserved.
> > +#
> > +# FS QA Test No. 936
> > +#
> > +# Ensure that reflinking into a file well beyond EOF zeroes everything between
> > +# the old EOF and the start of the newly linked chunk. This is an adaptation
> > +# of a reproducer script that Eric Sandeen formulated from a stale data
> > +# exposure bug uncovered by shared/010.
> > +#
> > +seq=`basename $0`
> > +seqres=$RESULT_DIR/$seq
> > +echo "QA output created by $seq"
> > +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
> > +. ./common/reflink
> > +
> > +# real QA test starts here
> > +_supported_fs generic
> > +_supported_os Linux
> > +_require_scratch_reflink
> > +
> > +rm -f $seqres.full
> > +
> > +# Fill disk with a well known pattern so that stale data exposure becomes much
> > +# more obvious.
> > +$XFS_IO_PROG -c "pwrite -S 0x58 -b 1m 0 300m" $SCRATCH_DEV >> $seqres.full
> > +_scratch_mkfs_sized $((300 * 1048576)) >>$seqres.full 2>&1
> > +_scratch_mount
> > +
> > +DONOR1=$SCRATCH_MNT/a
> > +TARGET=$SCRATCH_MNT/b
> > +blksz=65536
> > +
> > +$XFS_IO_PROG -f -c "pwrite -S 0x72 0 $blksz" $DONOR1 >> $seqres.full
> > +
> > +$XFS_IO_PROG -f \
> > + -c "falloc -k $((blksz*2)) $blksz" \
> > + -c "pwrite -S 0x57 $((blksz*16)) 8192" \
> > + -c "fdatasync" \
>
> Looks like an indention issue here.
>
> > + -c 'stat' \
> > + -c "reflink $DONOR1 0 $((blksz*17)) $blksz" \
> > + $TARGET >> $seqres.full
> > +
> > +od -tx1 -Ad -c $TARGET >> $seqres.full
> > +${FILEFRAG_PROG} -v $TARGET >> $seqres.full
>
> Seems like this filefrag is only here for debug purpose, and we could
> through away the stderr output if filefrag is not installed, so we don't
> have to _require_command for filefrag.
>
> > +md5sum $DONOR1 | _filter_scratch
> > +md5sum $TARGET | _filter_scratch
> > +
> > +_scratch_cycle_mount
> > +
> > +od -tx1 -Ad -c $TARGET >> $seqres.full
> > +${FILEFRAG_PROG} -v $TARGET >> $seqres.full
> > +md5sum $DONOR1 | _filter_scratch
> > +md5sum $TARGET | _filter_scratch
> > +
> > +status=0
> > +exit
> > diff --git a/tests/generic/936.out b/tests/generic/936.out
> > new file mode 100644
> > index 00000000..73685490
> > --- /dev/null
> > +++ b/tests/generic/936.out
> > @@ -0,0 +1,5 @@
> > +QA output created by 936
> > +34a4daf52c562253a2f6108c5123605a SCRATCH_MNT/a
> > +4fe4dfe3d2244bd380adcefdc5006e37 SCRATCH_MNT/b
> > +34a4daf52c562253a2f6108c5123605a SCRATCH_MNT/a
> > +4fe4dfe3d2244bd380adcefdc5006e37 SCRATCH_MNT/b
> > diff --git a/tests/generic/937 b/tests/generic/937
> > new file mode 100755
> > index 00000000..e3a94dbb
> > --- /dev/null
> > +++ b/tests/generic/937
> > @@ -0,0 +1,81 @@
> > +#! /bin/bash
> > +# SPDX-License-Identifier: GPL-2.0
> > +# Copyright (c) 2018, Oracle and/or its affiliates. All Rights Reserved.
> > +#
> > +# FS QA Test No. 937
> > +#
> > +# Ensuring that we cannot dedupe non-matching parts of files:
> > +# - Fail to dedupe non-identical parts of two different files
> > +# - Check that nothing changes in either file
> > +#
> > +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
> > +. ./common/reflink
> > +
> > +# real QA test starts here
> > +_supported_os Linux
> > +_require_test_dedupe
> > +
> > +rm -f $seqres.full
> > +
> > +testdir=$TEST_DIR/test-$seq
> > +rm -rf $testdir
> > +mkdir $testdir
> > +
> > +echo "Create the original files"
> > +blksz=65536
> > +_pwrite_byte 0x61 $((blksz * 2)) $((blksz * 6)) $testdir/file1 >> $seqres.full
> > +_pwrite_byte 0x61 $((blksz * 2)) $((blksz * 6)) $testdir/file2 >> $seqres.full
> > +_pwrite_byte 0x62 $(((blksz * 6) - 33)) 1 $testdir/file2 >> $seqres.full
> > +_test_cycle_mount
> > +
> > +md5sum $testdir/file1 | _filter_test_dir
> > +md5sum $testdir/file2 | _filter_test_dir
> > +
> > +_compare_range $testdir/file1 0 $testdir/file2 0 "$((blksz * 8))" \
> > + || echo "Files 1-2 do not match (intentional)"
> > +
> > +echo "(partial) dedupe the middle blocks together"
> > +free_before=$(stat -f -c '%a' $testdir)
> > +_dedupe_range $testdir/file1 $((blksz * 4)) $testdir/file2 \
> > + $((blksz * 4)) $((blksz * 2)) 2>&1 | _filter_xfs_io_numbers | _filter_dedupe_error
> > +_test_cycle_mount
> > +${FILEFRAG_PROG} -v $testdir/file1 >> $seqres.full
> > +${FILEFRAG_PROG} -v $testdir/file2 >> $seqres.full
> > +free_after=$(stat -f -c '%a' $testdir)
> > +echo "freesp changed by $free_before -> $free_after" >> $seqres.full
> > +
> > +echo "Compare sections"
> > +md5sum $testdir/file1 | _filter_test_dir
> > +md5sum $testdir/file2 | _filter_test_dir
> > +
> > +_compare_range $testdir/file1 0 $testdir/file2 0 $((blksz * 4)) \
> > + || echo "Start sections do not match (error)"
> > +
> > +_compare_range $testdir/file1 $((blksz * 4)) $testdir/file2 \
> > + $((blksz * 4)) $((blksz * 2)) \
> > + || echo "Middle sections do not match (intentional)"
> > +
> > +_compare_range $testdir/file1 $((blksz * 6)) $testdir/file2 \
> > + $((blksz * 6)) $((blksz * 2)) \
> > + || echo "End sections do not match (error)"
> > +
> > +# success, all done
> > +status=0
> > +exit
> > diff --git a/tests/generic/937.out b/tests/generic/937.out
> > new file mode 100644
> > index 00000000..496975d6
> > --- /dev/null
> > +++ b/tests/generic/937.out
> > @@ -0,0 +1,12 @@
> > +QA output created by 937
> > +Create the original files
> > +35ac8d7917305c385c30f3d82c30a8f6 TEST_DIR/test-937/file1
> > +39578c21e2cb9f6049b1cf7fc7be12a6 TEST_DIR/test-937/file2
> > +Files 1-2 do not match (intentional)
> > +(partial) dedupe the middle blocks together
> > +deduped XXXX/XXXX bytes at offset XXXX
> > +XXX Bytes, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
>
> I hit failure like:
>
> -deduped XXXX/XXXX bytes at offset XXXX
> -XXX Bytes, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
> +XFS_IOC_FILE_EXTENT_SAME: Extents did not match.
>
> I suspected "Extents did not match." is what we actually expected?
Yeah. I mistakenly thought that the dedupe ioctl could return
bytes_deduped < len, but the existing clients go berserk in shared/010
so ... I fixed the code and the test to follow the btrfs precedent.
--D
> Thanks,
> Eryu
>
> P.S.
> Patch 1-5 look fine to me, but drop this patch for now.
>
> > +Compare sections
> > +35ac8d7917305c385c30f3d82c30a8f6 TEST_DIR/test-937/file1
> > +39578c21e2cb9f6049b1cf7fc7be12a6 TEST_DIR/test-937/file2
> > +Middle sections do not match (intentional)
> > diff --git a/tests/generic/group b/tests/generic/group
> > index 06b062e1..ef08c0de 100644
> > --- a/tests/generic/group
> > +++ b/tests/generic/group
> > @@ -509,3 +509,7 @@
> > 504 auto quick locks
> > 505 shutdown auto quick metadata
> > 707 auto quick rw collapse zero
> > +934 auto quick clone
> > +935 auto quick clone
> > +936 auto quick clone
> > +937 auto quick dedupe clone
> >
prev parent reply other threads:[~2018-10-16 2:44 UTC|newest]
Thread overview: 10+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-10-05 0:19 [PATCH 0/6] fstests: fixes and new tests Darrick J. Wong
2018-10-05 0:19 ` [PATCH 1/6] generic: add bfoster's alternate reproducer Darrick J. Wong
2018-10-05 0:19 ` [PATCH 2/6] xfs: test mount time summary counter check and fix Darrick J. Wong
2018-10-05 0:19 ` [PATCH 3/6] xfs/189: skip 'barrier' mount option test on newer kernels Darrick J. Wong
2018-10-05 0:19 ` [PATCH 4/6] xfs/270: use _scratch_xfs_set_metadata_field to fuzz rocompat bits Darrick J. Wong
2018-10-05 6:54 ` Xiao Yang
2018-10-05 0:19 ` [PATCH 5/6] xfs/288: shut up verifier error Darrick J. Wong
2018-10-05 0:19 ` [PATCH 6/6] generic: test reflink side effects Darrick J. Wong
2018-10-14 15:40 ` Eryu Guan
2018-10-15 18:57 ` Darrick J. Wong [this message]
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=20181015185737.GC6703@magnolia \
--to=darrick.wong@oracle.com \
--cc=fstests@vger.kernel.org \
--cc=guaneryu@gmail.com \
--cc=linux-fsdevel@vger.kernel.org \
--cc=linux-xfs@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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).