diff mbox

fstests: generic: Test reserved extent map search routine on deduped file

Message ID 1463028695-7806-1-git-send-email-quwenruo@cn.fujitsu.com (mailing list archive)
State New, archived
Headers show

Commit Message

Qu Wenruo May 12, 2016, 4:51 a.m. UTC 
For fully dedupe file, which means all its file exntents are pointing to
the same bytenr, btrfs can cause soft lockup when calling fiemap ioctl
on that file, like the following output:
------
CPU: 1 PID: 7500 Comm: xfs_io Not tainted 4.5.0-rc6+ #2
Hardware name: XXXXXXXXXXXXXXXXXXXXXXXXXXX
task: ffff880027681b40 ti: ffff8800276e0000 task.ti: ffff8800276e0000
RIP: 0010:[<ffffffffa02583e4>]  [<ffffffffa02583e4>]
__merge_refs+0x34/0x120 [btrfs]
RSP: 0018:ffff8800276e3c08  EFLAGS: 00000202
RAX: ffff8800269cc330 RBX: ffff8800269cdb18 RCX: 0000000000000007
RDX: 00000000000061b0 RSI: ffff8800269cc4c8 RDI: ffff8800276e3c88
RBP: ffff8800276e3c20 R08: 0000000000000000 R09: 0000000000000001
R10: 0000000000000000 R11: 0000000000000000 R12: ffff880026ea3cb0
R13: ffff8800276e3c88 R14: ffff880027132a50 R15: ffff880027430000
FS:  00007f10201df700(0000) GS:ffff88003fa00000(0000)
knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 00007f10201ec000 CR3: 0000000027603000 CR4: 00000000000406e0
Stack:
 0000000000000000 0000000000000000 0000000000000000 ffff8800276e3ce8
 ffffffffa0259f38 0000000000000005 ffff8800274c6870 ffff8800274c7d88
 0000000000c10000 0000000000000000 0000000000000001 0000000027431190
Call Trace:
 [<ffffffffa0259f38>] find_parent_nodes+0x448/0x740 [btrfs]
 [<ffffffffa025a4f2>] btrfs_check_shared+0x102/0x1b0 [btrfs]
 [<ffffffff811fdcad>] ? __might_fault+0x4d/0xa0
 [<ffffffffa021899c>] extent_fiemap+0x2ac/0x550 [btrfs]
 [<ffffffff811ce156>] ? __filemap_fdatawait_range+0x96/0x160
 [<ffffffffa01f8ee0>] ? btrfs_get_extent+0xb30/0xb30 [btrfs]
 [<ffffffffa01f5da5>] btrfs_fiemap+0x45/0x50 [btrfs]
 [<ffffffff81246bb8>] do_vfs_ioctl+0x498/0x670
 [<ffffffff81246e09>] SyS_ioctl+0x79/0x90
 [<ffffffff8184e997>] entry_SYSCALL_64_fastpath+0x12/0x6f
Code: 41 55 41 54 53 4c 8b 27 4c 39 e7 0f 84 e9 00 00 00 49 89 fd 49 8b
34 24 49 39 f5 48 8b 1e 75 17 e9 d5 00 00 00 49 39 dd 48 8b 03 <48> 89
de 0f 84 b9 00 00 00 48 89 c3 8b 46 2c 41 39 44 24 2c 75
------

Also btrfs will return wrong flag for all these extents, they should
have SHARED(0x2000) flag, while btrfs still consider them as exclusive
extents.

On the other hand, with unmerged xfs reflink patches, xfs can handle it
without problem, and for patched btrfs, it can also handle it well.

This test case will create a large fully deduped file to check if the fs
can handle the fiemap ioctl and return correct SHARED flag for any fs
which support reflink.

Reported-by: Tsutomu Itoh <t-itoh@jp.fujitsu.com>
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
---
 common/punch          | 17 +++++++++++
 tests/generic/352     | 82 +++++++++++++++++++++++++++++++++++++++++++++++++++
 tests/generic/352.out |  5 ++++
 tests/generic/group   |  1 +
 4 files changed, 105 insertions(+)
 create mode 100755 tests/generic/352
 create mode 100644 tests/generic/352.out

Comments

Darrick J. Wong May 12, 2016, 5:01 a.m. UTC | #1 
On Thu, May 12, 2016 at 12:51:35PM +0800, Qu Wenruo wrote:
> For fully dedupe file, which means all its file exntents are pointing to
> the same bytenr, btrfs can cause soft lockup when calling fiemap ioctl
> on that file, like the following output:
> ------
> CPU: 1 PID: 7500 Comm: xfs_io Not tainted 4.5.0-rc6+ #2
> Hardware name: XXXXXXXXXXXXXXXXXXXXXXXXXXX
> task: ffff880027681b40 ti: ffff8800276e0000 task.ti: ffff8800276e0000
> RIP: 0010:[<ffffffffa02583e4>]  [<ffffffffa02583e4>]
> __merge_refs+0x34/0x120 [btrfs]
> RSP: 0018:ffff8800276e3c08  EFLAGS: 00000202
> RAX: ffff8800269cc330 RBX: ffff8800269cdb18 RCX: 0000000000000007
> RDX: 00000000000061b0 RSI: ffff8800269cc4c8 RDI: ffff8800276e3c88
> RBP: ffff8800276e3c20 R08: 0000000000000000 R09: 0000000000000001
> R10: 0000000000000000 R11: 0000000000000000 R12: ffff880026ea3cb0
> R13: ffff8800276e3c88 R14: ffff880027132a50 R15: ffff880027430000
> FS:  00007f10201df700(0000) GS:ffff88003fa00000(0000)
> knlGS:0000000000000000
> CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> CR2: 00007f10201ec000 CR3: 0000000027603000 CR4: 00000000000406e0
> Stack:
>  0000000000000000 0000000000000000 0000000000000000 ffff8800276e3ce8
>  ffffffffa0259f38 0000000000000005 ffff8800274c6870 ffff8800274c7d88
>  0000000000c10000 0000000000000000 0000000000000001 0000000027431190
> Call Trace:
>  [<ffffffffa0259f38>] find_parent_nodes+0x448/0x740 [btrfs]
>  [<ffffffffa025a4f2>] btrfs_check_shared+0x102/0x1b0 [btrfs]
>  [<ffffffff811fdcad>] ? __might_fault+0x4d/0xa0
>  [<ffffffffa021899c>] extent_fiemap+0x2ac/0x550 [btrfs]
>  [<ffffffff811ce156>] ? __filemap_fdatawait_range+0x96/0x160
>  [<ffffffffa01f8ee0>] ? btrfs_get_extent+0xb30/0xb30 [btrfs]
>  [<ffffffffa01f5da5>] btrfs_fiemap+0x45/0x50 [btrfs]
>  [<ffffffff81246bb8>] do_vfs_ioctl+0x498/0x670
>  [<ffffffff81246e09>] SyS_ioctl+0x79/0x90
>  [<ffffffff8184e997>] entry_SYSCALL_64_fastpath+0x12/0x6f
> Code: 41 55 41 54 53 4c 8b 27 4c 39 e7 0f 84 e9 00 00 00 49 89 fd 49 8b
> 34 24 49 39 f5 48 8b 1e 75 17 e9 d5 00 00 00 49 39 dd 48 8b 03 <48> 89
> de 0f 84 b9 00 00 00 48 89 c3 8b 46 2c 41 39 44 24 2c 75
> ------
> 
> Also btrfs will return wrong flag for all these extents, they should
> have SHARED(0x2000) flag, while btrfs still consider them as exclusive
> extents.
> 
> On the other hand, with unmerged xfs reflink patches, xfs can handle it
> without problem, and for patched btrfs, it can also handle it well.
> 
> This test case will create a large fully deduped file to check if the fs
> can handle the fiemap ioctl and return correct SHARED flag for any fs
> which support reflink.
> 
> Reported-by: Tsutomu Itoh <t-itoh@jp.fujitsu.com>
> Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
> ---
>  common/punch          | 17 +++++++++++
>  tests/generic/352     | 82 +++++++++++++++++++++++++++++++++++++++++++++++++++
>  tests/generic/352.out |  5 ++++
>  tests/generic/group   |  1 +
>  4 files changed, 105 insertions(+)
>  create mode 100755 tests/generic/352
>  create mode 100644 tests/generic/352.out
> 
> diff --git a/common/punch b/common/punch
> index 43f04c2..44c6e1c 100644
> --- a/common/punch
> +++ b/common/punch
> @@ -218,6 +218,23 @@ _filter_fiemap()
>  	_coalesce_extents
>  }
>  
> +_filter_fiemap_flags()
> +{
> +	$AWK_PROG '
> +		$3 ~ /hole/ {
> +			print $1, $2, $3;
> +			next;
> +		}
> +		$5 ~ /0x[[:xdigit:]]*8[[:xdigit:]][[:xdigit:]]/ {
> +			print $1, $2, "unwritten";
> +			next;
> +		}
> +		$5 ~ /0x[[:xdigit:]]+/ {
> +			print $1, $2, $5;
> +		}' |
> +	_coalesce_extents
> +}
> +
>  # Filters fiemap output to only print the 
>  # file offset column and whether or not
>  # it is an extent or a hole
> diff --git a/tests/generic/352 b/tests/generic/352
> new file mode 100755
> index 0000000..83f24ec
> --- /dev/null
> +++ b/tests/generic/352
> @@ -0,0 +1,82 @@
> +#! /bin/bash
> +# FS QA Test 352
> +#
> +# Test fiemap ioctl on heavily deduped file
> +#
> +# This test case will check if extent backref search goes
> +# without problem and return correct SHARED flag.
> +# Which btrfs will soft lock up and return wrong shared flag.
> +#
> +#-----------------------------------------------------------------------
> +# Copyright (c) 2016 Fujitsu.  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
> +. ./common/reflink
> +. ./common/punch
> +
> +# remove previous $seqres.full before test
> +rm -f $seqres.full
> +
> +# real QA test starts here
> +
> +# Modify as appropriate.
> +_supported_fs generic
> +_supported_os Linux
> +_require_scratch_reflink
> +
> +_scratch_mkfs > /dev/null 2>&1
> +_scratch_mount
> +
> +nr=$((8192 * $LOAD_FACTOR))
> +blocksize=$((128 * 1024))
> +file="$SCRATCH_MNT/tmp"
> +
> +# write the initial block for later reflink
> +$XFS_IO_PROG -f -c "pwrite 0 $blocksize" -c "fsync" $file | _filter_xfs_io
> +
> +# use reflink to create the rest of the file, whose all extents are all
> +# pointing to the first extent
> +for i in $(seq 1 $nr); do
> +	$XFS_IO_PROG -c "reflink $file 0 $(($i * $blocksize)) $blocksize" \
> +		$file > /dev/null

FYI, there's a _reflink_range helper for this now.

(There's also a _pwrite_byte helper.)

> +done
> +
> +# then call fiemap on that file to test both the shared flag and if
> +# reserved extent mapping search will cause soft lockup
> +$XFS_IO_PROG -c "fiemap -v" $file | _filter_fiemap_flags

I think this test needs _require_fiemap or else it'll fail on things
that don't support fiemap, like NFS.

PS: How well does btrfs handle generic/17[56]?

--D

> +
> +# success, all done
> +status=0
> +exit
> diff --git a/tests/generic/352.out b/tests/generic/352.out
> new file mode 100644
> index 0000000..a87c507
> --- /dev/null
> +++ b/tests/generic/352.out
> @@ -0,0 +1,5 @@
> +QA output created by 352
> +wrote 131072/131072 bytes at offset 0
> +XXX Bytes, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
> +0: [0..2097151]: 0x2000
> +1: [2097152..2097407]: 0x2001
> diff --git a/tests/generic/group b/tests/generic/group
> index 36fb759..3f00386 100644
> --- a/tests/generic/group
> +++ b/tests/generic/group
> @@ -354,3 +354,4 @@
>  349 blockdev quick rw
>  350 blockdev quick rw
>  351 blockdev quick rw
> +352 auto clone
> -- 
> 2.8.2
> 
> 
> 
--
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
Darrick J. Wong May 12, 2016, 5:34 a.m. UTC | #2 
On Wed, May 11, 2016 at 10:01:18PM -0700, Darrick J. Wong wrote:
> On Thu, May 12, 2016 at 12:51:35PM +0800, Qu Wenruo wrote:
> > For fully dedupe file, which means all its file exntents are pointing to
> > the same bytenr, btrfs can cause soft lockup when calling fiemap ioctl
> > on that file, like the following output:
> > ------
> > CPU: 1 PID: 7500 Comm: xfs_io Not tainted 4.5.0-rc6+ #2
> > Hardware name: XXXXXXXXXXXXXXXXXXXXXXXXXXX
> > task: ffff880027681b40 ti: ffff8800276e0000 task.ti: ffff8800276e0000
> > RIP: 0010:[<ffffffffa02583e4>]  [<ffffffffa02583e4>]
> > __merge_refs+0x34/0x120 [btrfs]
> > RSP: 0018:ffff8800276e3c08  EFLAGS: 00000202
> > RAX: ffff8800269cc330 RBX: ffff8800269cdb18 RCX: 0000000000000007
> > RDX: 00000000000061b0 RSI: ffff8800269cc4c8 RDI: ffff8800276e3c88
> > RBP: ffff8800276e3c20 R08: 0000000000000000 R09: 0000000000000001
> > R10: 0000000000000000 R11: 0000000000000000 R12: ffff880026ea3cb0
> > R13: ffff8800276e3c88 R14: ffff880027132a50 R15: ffff880027430000
> > FS:  00007f10201df700(0000) GS:ffff88003fa00000(0000)
> > knlGS:0000000000000000
> > CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> > CR2: 00007f10201ec000 CR3: 0000000027603000 CR4: 00000000000406e0
> > Stack:
> >  0000000000000000 0000000000000000 0000000000000000 ffff8800276e3ce8
> >  ffffffffa0259f38 0000000000000005 ffff8800274c6870 ffff8800274c7d88
> >  0000000000c10000 0000000000000000 0000000000000001 0000000027431190
> > Call Trace:
> >  [<ffffffffa0259f38>] find_parent_nodes+0x448/0x740 [btrfs]
> >  [<ffffffffa025a4f2>] btrfs_check_shared+0x102/0x1b0 [btrfs]
> >  [<ffffffff811fdcad>] ? __might_fault+0x4d/0xa0
> >  [<ffffffffa021899c>] extent_fiemap+0x2ac/0x550 [btrfs]
> >  [<ffffffff811ce156>] ? __filemap_fdatawait_range+0x96/0x160
> >  [<ffffffffa01f8ee0>] ? btrfs_get_extent+0xb30/0xb30 [btrfs]
> >  [<ffffffffa01f5da5>] btrfs_fiemap+0x45/0x50 [btrfs]
> >  [<ffffffff81246bb8>] do_vfs_ioctl+0x498/0x670
> >  [<ffffffff81246e09>] SyS_ioctl+0x79/0x90
> >  [<ffffffff8184e997>] entry_SYSCALL_64_fastpath+0x12/0x6f
> > Code: 41 55 41 54 53 4c 8b 27 4c 39 e7 0f 84 e9 00 00 00 49 89 fd 49 8b
> > 34 24 49 39 f5 48 8b 1e 75 17 e9 d5 00 00 00 49 39 dd 48 8b 03 <48> 89
> > de 0f 84 b9 00 00 00 48 89 c3 8b 46 2c 41 39 44 24 2c 75
> > ------
> > 
> > Also btrfs will return wrong flag for all these extents, they should
> > have SHARED(0x2000) flag, while btrfs still consider them as exclusive
> > extents.
> > 
> > On the other hand, with unmerged xfs reflink patches, xfs can handle it
> > without problem, and for patched btrfs, it can also handle it well.
> > 
> > This test case will create a large fully deduped file to check if the fs
> > can handle the fiemap ioctl and return correct SHARED flag for any fs
> > which support reflink.
> > 
> > Reported-by: Tsutomu Itoh <t-itoh@jp.fujitsu.com>
> > Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
> > ---
> >  common/punch          | 17 +++++++++++
> >  tests/generic/352     | 82 +++++++++++++++++++++++++++++++++++++++++++++++++++
> >  tests/generic/352.out |  5 ++++
> >  tests/generic/group   |  1 +
> >  4 files changed, 105 insertions(+)
> >  create mode 100755 tests/generic/352
> >  create mode 100644 tests/generic/352.out
> > 
> > diff --git a/common/punch b/common/punch
> > index 43f04c2..44c6e1c 100644
> > --- a/common/punch
> > +++ b/common/punch
> > @@ -218,6 +218,23 @@ _filter_fiemap()
> >  	_coalesce_extents
> >  }
> >  
> > +_filter_fiemap_flags()
> > +{
> > +	$AWK_PROG '
> > +		$3 ~ /hole/ {
> > +			print $1, $2, $3;
> > +			next;
> > +		}
> > +		$5 ~ /0x[[:xdigit:]]*8[[:xdigit:]][[:xdigit:]]/ {
> > +			print $1, $2, "unwritten";
> > +			next;
> > +		}
> > +		$5 ~ /0x[[:xdigit:]]+/ {
> > +			print $1, $2, $5;
> > +		}' |
> > +	_coalesce_extents
> > +}
> > +
> >  # Filters fiemap output to only print the 
> >  # file offset column and whether or not
> >  # it is an extent or a hole
> > diff --git a/tests/generic/352 b/tests/generic/352
> > new file mode 100755
> > index 0000000..83f24ec
> > --- /dev/null
> > +++ b/tests/generic/352
> > @@ -0,0 +1,82 @@
> > +#! /bin/bash
> > +# FS QA Test 352
> > +#
> > +# Test fiemap ioctl on heavily deduped file
> > +#
> > +# This test case will check if extent backref search goes
> > +# without problem and return correct SHARED flag.
> > +# Which btrfs will soft lock up and return wrong shared flag.
> > +#
> > +#-----------------------------------------------------------------------
> > +# Copyright (c) 2016 Fujitsu.  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
> > +. ./common/reflink
> > +. ./common/punch
> > +
> > +# remove previous $seqres.full before test
> > +rm -f $seqres.full
> > +
> > +# real QA test starts here
> > +
> > +# Modify as appropriate.
> > +_supported_fs generic
> > +_supported_os Linux
> > +_require_scratch_reflink
> > +
> > +_scratch_mkfs > /dev/null 2>&1
> > +_scratch_mount
> > +
> > +nr=$((8192 * $LOAD_FACTOR))
> > +blocksize=$((128 * 1024))
> > +file="$SCRATCH_MNT/tmp"
> > +
> > +# write the initial block for later reflink
> > +$XFS_IO_PROG -f -c "pwrite 0 $blocksize" -c "fsync" $file | _filter_xfs_io
> > +
> > +# use reflink to create the rest of the file, whose all extents are all
> > +# pointing to the first extent
> > +for i in $(seq 1 $nr); do
> > +	$XFS_IO_PROG -c "reflink $file 0 $(($i * $blocksize)) $blocksize" \
> > +		$file > /dev/null
> 
> FYI, there's a _reflink_range helper for this now.
> 
> (There's also a _pwrite_byte helper.)
> 
> > +done
> > +
> > +# then call fiemap on that file to test both the shared flag and if
> > +# reserved extent mapping search will cause soft lockup
> > +$XFS_IO_PROG -c "fiemap -v" $file | _filter_fiemap_flags
> 
> I think this test needs _require_fiemap or else it'll fail on things
> that don't support fiemap, like NFS.
> 
> PS: How well does btrfs handle generic/17[56]?

Well, running filefrag -v while it's running will indeed cause soft lockup
problems.  Ulp.

<reboots vm, tries again>

So left alone, 175 finishes in ~90s.  But then it runs btrfsck.

I also see that btrfsck seems to be doing this over and over:

pread64(4, "...", 16384, 347668480) = 16384 <0.000047>
pread64(4, "...", 16384, 151470080) = 16384 <0.000081>
pread64(4, "...", 16384, 337936384) = 16384 <0.000054>
readahead(4, 347783168, 16384) = 0 <0.000064>
readahead(4, 347815936, 16384) = 0 <0.000058>
readahead(4, 347832320, 16384) = 0 <0.000045>
readahead(4, 347848704, 16384) = 0 <0.000045>
readahead(4, 347881472, 16384) = 0 <0.000044>
readahead(4, 347897856, 16384) = 0 <0.000043>
readahead(4, 347914240, 16384) = 0 <0.000044>
readahead(4, 347930624, 16384) = 0 <0.000047>
readahead(4, 347947008, 16384) = 0 <0.000044>
readahead(4, 347963392, 16384) = 0 <0.000044>
readahead(4, 347979776, 16384) = 0 <0.000043>
readahead(4, 348012544, 16384) = 0 <0.000047>
readahead(4, 348028928, 16384) = 0 <0.000045>
readahead(4, 348045312, 16384) = 0 <0.000048>
readahead(4, 348078080, 16384) = 0 <0.000048>
readahead(4, 348110848, 16384) = 0 <0.000045>
readahead(4, 348127232, 16384) = 0 <0.000044>

(It's been running 20min now, I'm going to kill it and go to bed.)

--D

> 
> --D
> 
> > +
> > +# success, all done
> > +status=0
> > +exit
> > diff --git a/tests/generic/352.out b/tests/generic/352.out
> > new file mode 100644
> > index 0000000..a87c507
> > --- /dev/null
> > +++ b/tests/generic/352.out
> > @@ -0,0 +1,5 @@
> > +QA output created by 352
> > +wrote 131072/131072 bytes at offset 0
> > +XXX Bytes, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
> > +0: [0..2097151]: 0x2000
> > +1: [2097152..2097407]: 0x2001
> > diff --git a/tests/generic/group b/tests/generic/group
> > index 36fb759..3f00386 100644
> > --- a/tests/generic/group
> > +++ b/tests/generic/group
> > @@ -354,3 +354,4 @@
> >  349 blockdev quick rw
> >  350 blockdev quick rw
> >  351 blockdev quick rw
> > +352 auto clone
> > -- 
> > 2.8.2
> > 
> > 
> > 
> --
> 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
--
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
Qu Wenruo May 12, 2016, 5:44 a.m. UTC | #3 
Darrick J. Wong wrote on 2016/05/11 22:01 -0700:
> On Thu, May 12, 2016 at 12:51:35PM +0800, Qu Wenruo wrote:
>> For fully dedupe file, which means all its file exntents are pointing to
>> the same bytenr, btrfs can cause soft lockup when calling fiemap ioctl
>> on that file, like the following output:
>> ------
>> CPU: 1 PID: 7500 Comm: xfs_io Not tainted 4.5.0-rc6+ #2
>> Hardware name: XXXXXXXXXXXXXXXXXXXXXXXXXXX
>> task: ffff880027681b40 ti: ffff8800276e0000 task.ti: ffff8800276e0000
>> RIP: 0010:[<ffffffffa02583e4>]  [<ffffffffa02583e4>]
>> __merge_refs+0x34/0x120 [btrfs]
>> RSP: 0018:ffff8800276e3c08  EFLAGS: 00000202
>> RAX: ffff8800269cc330 RBX: ffff8800269cdb18 RCX: 0000000000000007
>> RDX: 00000000000061b0 RSI: ffff8800269cc4c8 RDI: ffff8800276e3c88
>> RBP: ffff8800276e3c20 R08: 0000000000000000 R09: 0000000000000001
>> R10: 0000000000000000 R11: 0000000000000000 R12: ffff880026ea3cb0
>> R13: ffff8800276e3c88 R14: ffff880027132a50 R15: ffff880027430000
>> FS:  00007f10201df700(0000) GS:ffff88003fa00000(0000)
>> knlGS:0000000000000000
>> CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
>> CR2: 00007f10201ec000 CR3: 0000000027603000 CR4: 00000000000406e0
>> Stack:
>>  0000000000000000 0000000000000000 0000000000000000 ffff8800276e3ce8
>>  ffffffffa0259f38 0000000000000005 ffff8800274c6870 ffff8800274c7d88
>>  0000000000c10000 0000000000000000 0000000000000001 0000000027431190
>> Call Trace:
>>  [<ffffffffa0259f38>] find_parent_nodes+0x448/0x740 [btrfs]
>>  [<ffffffffa025a4f2>] btrfs_check_shared+0x102/0x1b0 [btrfs]
>>  [<ffffffff811fdcad>] ? __might_fault+0x4d/0xa0
>>  [<ffffffffa021899c>] extent_fiemap+0x2ac/0x550 [btrfs]
>>  [<ffffffff811ce156>] ? __filemap_fdatawait_range+0x96/0x160
>>  [<ffffffffa01f8ee0>] ? btrfs_get_extent+0xb30/0xb30 [btrfs]
>>  [<ffffffffa01f5da5>] btrfs_fiemap+0x45/0x50 [btrfs]
>>  [<ffffffff81246bb8>] do_vfs_ioctl+0x498/0x670
>>  [<ffffffff81246e09>] SyS_ioctl+0x79/0x90
>>  [<ffffffff8184e997>] entry_SYSCALL_64_fastpath+0x12/0x6f
>> Code: 41 55 41 54 53 4c 8b 27 4c 39 e7 0f 84 e9 00 00 00 49 89 fd 49 8b
>> 34 24 49 39 f5 48 8b 1e 75 17 e9 d5 00 00 00 49 39 dd 48 8b 03 <48> 89
>> de 0f 84 b9 00 00 00 48 89 c3 8b 46 2c 41 39 44 24 2c 75
>> ------
>>
>> Also btrfs will return wrong flag for all these extents, they should
>> have SHARED(0x2000) flag, while btrfs still consider them as exclusive
>> extents.
>>
>> On the other hand, with unmerged xfs reflink patches, xfs can handle it
>> without problem, and for patched btrfs, it can also handle it well.
>>
>> This test case will create a large fully deduped file to check if the fs
>> can handle the fiemap ioctl and return correct SHARED flag for any fs
>> which support reflink.
>>
>> Reported-by: Tsutomu Itoh <t-itoh@jp.fujitsu.com>
>> Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
>> ---
>>  common/punch          | 17 +++++++++++
>>  tests/generic/352     | 82 +++++++++++++++++++++++++++++++++++++++++++++++++++
>>  tests/generic/352.out |  5 ++++
>>  tests/generic/group   |  1 +
>>  4 files changed, 105 insertions(+)
>>  create mode 100755 tests/generic/352
>>  create mode 100644 tests/generic/352.out
>>
>> diff --git a/common/punch b/common/punch
>> index 43f04c2..44c6e1c 100644
>> --- a/common/punch
>> +++ b/common/punch
>> @@ -218,6 +218,23 @@ _filter_fiemap()
>>  	_coalesce_extents
>>  }
>>
>> +_filter_fiemap_flags()
>> +{
>> +	$AWK_PROG '
>> +		$3 ~ /hole/ {
>> +			print $1, $2, $3;
>> +			next;
>> +		}
>> +		$5 ~ /0x[[:xdigit:]]*8[[:xdigit:]][[:xdigit:]]/ {
>> +			print $1, $2, "unwritten";
>> +			next;
>> +		}
>> +		$5 ~ /0x[[:xdigit:]]+/ {
>> +			print $1, $2, $5;
>> +		}' |
>> +	_coalesce_extents
>> +}
>> +
>>  # Filters fiemap output to only print the
>>  # file offset column and whether or not
>>  # it is an extent or a hole
>> diff --git a/tests/generic/352 b/tests/generic/352
>> new file mode 100755
>> index 0000000..83f24ec
>> --- /dev/null
>> +++ b/tests/generic/352
>> @@ -0,0 +1,82 @@
>> +#! /bin/bash
>> +# FS QA Test 352
>> +#
>> +# Test fiemap ioctl on heavily deduped file
>> +#
>> +# This test case will check if extent backref search goes
>> +# without problem and return correct SHARED flag.
>> +# Which btrfs will soft lock up and return wrong shared flag.
>> +#
>> +#-----------------------------------------------------------------------
>> +# Copyright (c) 2016 Fujitsu.  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
>> +. ./common/reflink
>> +. ./common/punch
>> +
>> +# remove previous $seqres.full before test
>> +rm -f $seqres.full
>> +
>> +# real QA test starts here
>> +
>> +# Modify as appropriate.
>> +_supported_fs generic
>> +_supported_os Linux
>> +_require_scratch_reflink
>> +
>> +_scratch_mkfs > /dev/null 2>&1
>> +_scratch_mount
>> +
>> +nr=$((8192 * $LOAD_FACTOR))
>> +blocksize=$((128 * 1024))
>> +file="$SCRATCH_MNT/tmp"
>> +
>> +# write the initial block for later reflink
>> +$XFS_IO_PROG -f -c "pwrite 0 $blocksize" -c "fsync" $file | _filter_xfs_io
>> +
>> +# use reflink to create the rest of the file, whose all extents are all
>> +# pointing to the first extent
>> +for i in $(seq 1 $nr); do
>> +	$XFS_IO_PROG -c "reflink $file 0 $(($i * $blocksize)) $blocksize" \
>> +		$file > /dev/null
>
> FYI, there's a _reflink_range helper for this now.
>
> (There's also a _pwrite_byte helper.)

Thanks, I'll use the new helper in next version.

>
>> +done
>> +
>> +# then call fiemap on that file to test both the shared flag and if
>> +# reserved extent mapping search will cause soft lockup
>> +$XFS_IO_PROG -c "fiemap -v" $file | _filter_fiemap_flags
>
> I think this test needs _require_fiemap or else it'll fail on things
> that don't support fiemap, like NFS.
>
> PS: How well does btrfs handle generic/17[56]?

Handles well, but very slow. (I bet it's slower than xfs)
175 takes me about 10min(579s) to finish the test case, and even more 
time to umount and run btrfsck.

As long as the test cases only involve reflink operation, btrfs won't 
cause too many problems except the slow metadata operation speed.

BTW, during that test, even I'm testing btrfs only, xfs reports lockdep 
problem (the test machine is running xfs as root).
The kernel is for-dave-4.6 branch, and I'll report it in another mail.

Thanks,
Qu

>
> --D
>
>> +
>> +# success, all done
>> +status=0
>> +exit
>> diff --git a/tests/generic/352.out b/tests/generic/352.out
>> new file mode 100644
>> index 0000000..a87c507
>> --- /dev/null
>> +++ b/tests/generic/352.out
>> @@ -0,0 +1,5 @@
>> +QA output created by 352
>> +wrote 131072/131072 bytes at offset 0
>> +XXX Bytes, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
>> +0: [0..2097151]: 0x2000
>> +1: [2097152..2097407]: 0x2001
>> diff --git a/tests/generic/group b/tests/generic/group
>> index 36fb759..3f00386 100644
>> --- a/tests/generic/group
>> +++ b/tests/generic/group
>> @@ -354,3 +354,4 @@
>>  349 blockdev quick rw
>>  350 blockdev quick rw
>>  351 blockdev quick rw
>> +352 auto clone
>> --
>> 2.8.2
>>
>>
>>
>
>


--
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
Darrick J. Wong May 12, 2016, 5:55 a.m. UTC | #4 
On Thu, May 12, 2016 at 01:44:01PM +0800, Qu Wenruo wrote:
> 
> 
> Darrick J. Wong wrote on 2016/05/11 22:01 -0700:
> >On Thu, May 12, 2016 at 12:51:35PM +0800, Qu Wenruo wrote:
> >>For fully dedupe file, which means all its file exntents are pointing to
> >>the same bytenr, btrfs can cause soft lockup when calling fiemap ioctl
> >>on that file, like the following output:
> >>------
> >>CPU: 1 PID: 7500 Comm: xfs_io Not tainted 4.5.0-rc6+ #2
> >>Hardware name: XXXXXXXXXXXXXXXXXXXXXXXXXXX
> >>task: ffff880027681b40 ti: ffff8800276e0000 task.ti: ffff8800276e0000
> >>RIP: 0010:[<ffffffffa02583e4>]  [<ffffffffa02583e4>]
> >>__merge_refs+0x34/0x120 [btrfs]
> >>RSP: 0018:ffff8800276e3c08  EFLAGS: 00000202
> >>RAX: ffff8800269cc330 RBX: ffff8800269cdb18 RCX: 0000000000000007
> >>RDX: 00000000000061b0 RSI: ffff8800269cc4c8 RDI: ffff8800276e3c88
> >>RBP: ffff8800276e3c20 R08: 0000000000000000 R09: 0000000000000001
> >>R10: 0000000000000000 R11: 0000000000000000 R12: ffff880026ea3cb0
> >>R13: ffff8800276e3c88 R14: ffff880027132a50 R15: ffff880027430000
> >>FS:  00007f10201df700(0000) GS:ffff88003fa00000(0000)
> >>knlGS:0000000000000000
> >>CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> >>CR2: 00007f10201ec000 CR3: 0000000027603000 CR4: 00000000000406e0
> >>Stack:
> >> 0000000000000000 0000000000000000 0000000000000000 ffff8800276e3ce8
> >> ffffffffa0259f38 0000000000000005 ffff8800274c6870 ffff8800274c7d88
> >> 0000000000c10000 0000000000000000 0000000000000001 0000000027431190
> >>Call Trace:
> >> [<ffffffffa0259f38>] find_parent_nodes+0x448/0x740 [btrfs]
> >> [<ffffffffa025a4f2>] btrfs_check_shared+0x102/0x1b0 [btrfs]
> >> [<ffffffff811fdcad>] ? __might_fault+0x4d/0xa0
> >> [<ffffffffa021899c>] extent_fiemap+0x2ac/0x550 [btrfs]
> >> [<ffffffff811ce156>] ? __filemap_fdatawait_range+0x96/0x160
> >> [<ffffffffa01f8ee0>] ? btrfs_get_extent+0xb30/0xb30 [btrfs]
> >> [<ffffffffa01f5da5>] btrfs_fiemap+0x45/0x50 [btrfs]
> >> [<ffffffff81246bb8>] do_vfs_ioctl+0x498/0x670
> >> [<ffffffff81246e09>] SyS_ioctl+0x79/0x90
> >> [<ffffffff8184e997>] entry_SYSCALL_64_fastpath+0x12/0x6f
> >>Code: 41 55 41 54 53 4c 8b 27 4c 39 e7 0f 84 e9 00 00 00 49 89 fd 49 8b
> >>34 24 49 39 f5 48 8b 1e 75 17 e9 d5 00 00 00 49 39 dd 48 8b 03 <48> 89
> >>de 0f 84 b9 00 00 00 48 89 c3 8b 46 2c 41 39 44 24 2c 75
> >>------
> >>
> >>Also btrfs will return wrong flag for all these extents, they should
> >>have SHARED(0x2000) flag, while btrfs still consider them as exclusive
> >>extents.
> >>
> >>On the other hand, with unmerged xfs reflink patches, xfs can handle it
> >>without problem, and for patched btrfs, it can also handle it well.
> >>
> >>This test case will create a large fully deduped file to check if the fs
> >>can handle the fiemap ioctl and return correct SHARED flag for any fs
> >>which support reflink.
> >>
> >>Reported-by: Tsutomu Itoh <t-itoh@jp.fujitsu.com>
> >>Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
> >>---
> >> common/punch          | 17 +++++++++++
> >> tests/generic/352     | 82 +++++++++++++++++++++++++++++++++++++++++++++++++++
> >> tests/generic/352.out |  5 ++++
> >> tests/generic/group   |  1 +
> >> 4 files changed, 105 insertions(+)
> >> create mode 100755 tests/generic/352
> >> create mode 100644 tests/generic/352.out
> >>
> >>diff --git a/common/punch b/common/punch
> >>index 43f04c2..44c6e1c 100644
> >>--- a/common/punch
> >>+++ b/common/punch
> >>@@ -218,6 +218,23 @@ _filter_fiemap()
> >> 	_coalesce_extents
> >> }
> >>
> >>+_filter_fiemap_flags()
> >>+{
> >>+	$AWK_PROG '
> >>+		$3 ~ /hole/ {
> >>+			print $1, $2, $3;
> >>+			next;
> >>+		}
> >>+		$5 ~ /0x[[:xdigit:]]*8[[:xdigit:]][[:xdigit:]]/ {
> >>+			print $1, $2, "unwritten";
> >>+			next;
> >>+		}
> >>+		$5 ~ /0x[[:xdigit:]]+/ {
> >>+			print $1, $2, $5;
> >>+		}' |
> >>+	_coalesce_extents
> >>+}
> >>+
> >> # Filters fiemap output to only print the
> >> # file offset column and whether or not
> >> # it is an extent or a hole
> >>diff --git a/tests/generic/352 b/tests/generic/352
> >>new file mode 100755
> >>index 0000000..83f24ec
> >>--- /dev/null
> >>+++ b/tests/generic/352
> >>@@ -0,0 +1,82 @@
> >>+#! /bin/bash
> >>+# FS QA Test 352
> >>+#
> >>+# Test fiemap ioctl on heavily deduped file
> >>+#
> >>+# This test case will check if extent backref search goes
> >>+# without problem and return correct SHARED flag.
> >>+# Which btrfs will soft lock up and return wrong shared flag.
> >>+#
> >>+#-----------------------------------------------------------------------
> >>+# Copyright (c) 2016 Fujitsu.  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
> >>+. ./common/reflink
> >>+. ./common/punch
> >>+
> >>+# remove previous $seqres.full before test
> >>+rm -f $seqres.full
> >>+
> >>+# real QA test starts here
> >>+
> >>+# Modify as appropriate.
> >>+_supported_fs generic
> >>+_supported_os Linux
> >>+_require_scratch_reflink
> >>+
> >>+_scratch_mkfs > /dev/null 2>&1
> >>+_scratch_mount
> >>+
> >>+nr=$((8192 * $LOAD_FACTOR))
> >>+blocksize=$((128 * 1024))
> >>+file="$SCRATCH_MNT/tmp"
> >>+
> >>+# write the initial block for later reflink
> >>+$XFS_IO_PROG -f -c "pwrite 0 $blocksize" -c "fsync" $file | _filter_xfs_io
> >>+
> >>+# use reflink to create the rest of the file, whose all extents are all
> >>+# pointing to the first extent
> >>+for i in $(seq 1 $nr); do
> >>+	$XFS_IO_PROG -c "reflink $file 0 $(($i * $blocksize)) $blocksize" \
> >>+		$file > /dev/null
> >
> >FYI, there's a _reflink_range helper for this now.
> >
> >(There's also a _pwrite_byte helper.)
> 
> Thanks, I'll use the new helper in next version.
> 
> >
> >>+done
> >>+
> >>+# then call fiemap on that file to test both the shared flag and if
> >>+# reserved extent mapping search will cause soft lockup
> >>+$XFS_IO_PROG -c "fiemap -v" $file | _filter_fiemap_flags
> >
> >I think this test needs _require_fiemap or else it'll fail on things
> >that don't support fiemap, like NFS.
> >
> >PS: How well does btrfs handle generic/17[56]?
> 
> Handles well, but very slow. (I bet it's slower than xfs)
> 175 takes me about 10min(579s) to finish the test case, and even more time
> to umount and run btrfsck.

Yeah.  I don't know if your recent patches make that better? <shrug>

> As long as the test cases only involve reflink operation, btrfs won't cause
> too many problems except the slow metadata operation speed.
> 
> BTW, during that test, even I'm testing btrfs only, xfs reports lockdep
> problem (the test machine is running xfs as root).
> The kernel is for-dave-4.6 branch, and I'll report it in another mail.

FWIW for-dave-for-4.6 is ancient.  You might try #djwong-devel for newer
(and hopefully more stable?) code.

Wait, *lockdep*?  Yeah, XFS and lockdep have problems.

--D

> 
> Thanks,
> Qu
> 
> >
> >--D
> >
> >>+
> >>+# success, all done
> >>+status=0
> >>+exit
> >>diff --git a/tests/generic/352.out b/tests/generic/352.out
> >>new file mode 100644
> >>index 0000000..a87c507
> >>--- /dev/null
> >>+++ b/tests/generic/352.out
> >>@@ -0,0 +1,5 @@
> >>+QA output created by 352
> >>+wrote 131072/131072 bytes at offset 0
> >>+XXX Bytes, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
> >>+0: [0..2097151]: 0x2000
> >>+1: [2097152..2097407]: 0x2001
> >>diff --git a/tests/generic/group b/tests/generic/group
> >>index 36fb759..3f00386 100644
> >>--- a/tests/generic/group
> >>+++ b/tests/generic/group
> >>@@ -354,3 +354,4 @@
> >> 349 blockdev quick rw
> >> 350 blockdev quick rw
> >> 351 blockdev quick rw
> >>+352 auto clone
> >>--
> >>2.8.2
> >>
> >>
> >>
> >
> >
> 
> 
--
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
diff mbox

Patch

diff --git a/common/punch b/common/punch
index 43f04c2..44c6e1c 100644
--- a/common/punch
+++ b/common/punch
@@ -218,6 +218,23 @@  _filter_fiemap()
 	_coalesce_extents
 }
 
+_filter_fiemap_flags()
+{
+	$AWK_PROG '
+		$3 ~ /hole/ {
+			print $1, $2, $3;
+			next;
+		}
+		$5 ~ /0x[[:xdigit:]]*8[[:xdigit:]][[:xdigit:]]/ {
+			print $1, $2, "unwritten";
+			next;
+		}
+		$5 ~ /0x[[:xdigit:]]+/ {
+			print $1, $2, $5;
+		}' |
+	_coalesce_extents
+}
+
 # Filters fiemap output to only print the 
 # file offset column and whether or not
 # it is an extent or a hole
diff --git a/tests/generic/352 b/tests/generic/352
new file mode 100755
index 0000000..83f24ec
--- /dev/null
+++ b/tests/generic/352
@@ -0,0 +1,82 @@ 
+#! /bin/bash
+# FS QA Test 352
+#
+# Test fiemap ioctl on heavily deduped file
+#
+# This test case will check if extent backref search goes
+# without problem and return correct SHARED flag.
+# Which btrfs will soft lock up and return wrong shared flag.
+#
+#-----------------------------------------------------------------------
+# Copyright (c) 2016 Fujitsu.  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
+. ./common/reflink
+. ./common/punch
+
+# remove previous $seqres.full before test
+rm -f $seqres.full
+
+# real QA test starts here
+
+# Modify as appropriate.
+_supported_fs generic
+_supported_os Linux
+_require_scratch_reflink
+
+_scratch_mkfs > /dev/null 2>&1
+_scratch_mount
+
+nr=$((8192 * $LOAD_FACTOR))
+blocksize=$((128 * 1024))
+file="$SCRATCH_MNT/tmp"
+
+# write the initial block for later reflink
+$XFS_IO_PROG -f -c "pwrite 0 $blocksize" -c "fsync" $file | _filter_xfs_io
+
+# use reflink to create the rest of the file, whose all extents are all
+# pointing to the first extent
+for i in $(seq 1 $nr); do
+	$XFS_IO_PROG -c "reflink $file 0 $(($i * $blocksize)) $blocksize" \
+		$file > /dev/null
+done
+
+# then call fiemap on that file to test both the shared flag and if
+# reserved extent mapping search will cause soft lockup
+$XFS_IO_PROG -c "fiemap -v" $file | _filter_fiemap_flags
+
+# success, all done
+status=0
+exit
diff --git a/tests/generic/352.out b/tests/generic/352.out
new file mode 100644
index 0000000..a87c507
--- /dev/null
+++ b/tests/generic/352.out
@@ -0,0 +1,5 @@ 
+QA output created by 352
+wrote 131072/131072 bytes at offset 0
+XXX Bytes, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
+0: [0..2097151]: 0x2000
+1: [2097152..2097407]: 0x2001
diff --git a/tests/generic/group b/tests/generic/group
index 36fb759..3f00386 100644
--- a/tests/generic/group
+++ b/tests/generic/group
@@ -354,3 +354,4 @@ 
 349 blockdev quick rw
 350 blockdev quick rw
 351 blockdev quick rw
+352 auto clone