On 2017/06/29 12:34, Eryu Guan wrote:
On Thu, Jun 29, 2017 at 12:01:25PM +0800, Xiao Yang wrote:
On 2017/06/28 18:13, Eryu Guan wrote:
On Tue, Jun 27, 2017 at 05:03:52PM +0800, Xiao Yang wrote:
This regression test reproduces a race between a direct
I/O read and a mapped write to a hole in a file. If the
race occurs, it will trigger a BUG_ON(). This XFS bug
has been fixed by:
04197b3 (xfs: don't BUG() on mixed direct and mapped I/O)
Signed-off-by: Xiao Yang<yangx.jy@xxxxxxxxxxxxxx>
---
tests/xfs/422 | 86 +++++++++++++++++++++++++++++++++++++++++++++++++++++++
tests/xfs/422.out | 2 ++
tests/xfs/group | 1 +
3 files changed, 89 insertions(+)
create mode 100755 tests/xfs/422
create mode 100644 tests/xfs/422.out
diff --git a/tests/xfs/422 b/tests/xfs/422
new file mode 100755
index 0000000..a57ffab
--- /dev/null
+++ b/tests/xfs/422
@@ -0,0 +1,86 @@
+#! /bin/bash
+# FS QA Test No. 422
+#
+# Regression test for commit:
+# 04197b3 ("xfs: don't BUG() on mixed direct and mapped I/O)
^^ missing ending " :)
+#
+# This case tests a race between a direct I/O read and a
+# mapped write to a hole in a file. If the race occurs,
+# it will trigger a BUG_ON().
+#
+#-----------------------------------------------------------------------
+# Copyright (c) 2017 Fujitsu. All Rights Reserved.
+# Author: Xiao Yang<yangx.jy@xxxxxxxxxxxxxx>
+#
+# 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()
+{
+ rm -rf $tmp.*
+}
+
+# get standard environment and checks
+. ./common/rc
+
+# real QA test starts here
+_supported_os Linux
+_supported_fs xfs
I see nothing xfs specific in this test, I think it can be made generic
easily.
Hi Eryu,
Agreed. It could be moved into generic.
+_require_scratch
+_disable_dmesg_check
Better to add comments on ignoring dmesg check result.
For xfs filesystem, this race triggers various warnings on some
distributions(RHEL7) and upstream kernel.
e.g,
1) This race triggers the following warnings on RHEL7.3GA
-----------------------------------------------------------------------------------------------------------------------------
[ 0.297084] WARNING: at arch/x86/mm/ioremap.c:101
__ioremap_caller+0x238/0x320()
I'm not sure about this one, but the other xfs or iomap warnings are
"known issues" when mixing direct I/O and mmap I/O on the same file.
We've filtered out these intentional warnings in generic/095, and Brian
has just sent a patch[1] to move this filter to common/xfs and added it
to generic/247. I think you can do the same to this new test, filter out
these known warnings so we won't miss any other potential bug.
I don't think disable dmesg check only because it triggers warnings is a
good idea :)
Hi Eryu
It sounds reasonable to me. I will filter out these known warnings
instead of disabling dmesg check.
Thanks,
Eryu
[1] [PATCH] generic/247: filter out expected XFS warnings for mixed mmap/direct I/O
[ 648.745481] WARNING: at fs/xfs/xfs_aops.c:1244
xfs_vm_releasepage+0xcb/0x100 [xfs]()
[ 648.745694] WARNING: at fs/xfs/xfs_file.c:377
xfs_file_aio_read+0x2dd/0x340 [xfs]()
-----------------------------------------------------------------------------------------------------------------------------
2) It triggers the following warnings on RHEL7.4Beta
----------------------------------------------------------------------------------------------------------------------------
[ 497.575319] WARNING: CPU: 0 PID: 26865 at fs/xfs/xfs_file.c:361
xfs_file_dio_aio_read+0x1ba/0x2a0 [xfs]
[ 497.577651] WARNING: CPU: 0 PID: 26865 at fs/xfs/xfs_aops.c:1285
__xfs_get_blocks+0x8ca/0x8e0 [xfs]
Could we add __xfs_get_blocks into filter_xfs_dmesg? :-)
This known warning was introduced by:
04197b3 (xfs: don't BUG() on mixed direct and mapped I/O)
It has been removed by:
acdda3a (xfs: use iomap_dio_rw)
Thanks,
Xiao Yang
----------------------------------------------------------------------------------------------------------------------------
3) It triggers the following warnings on upstream kernel
----------------------------------------------------------------------------------------------------------------------------
[ 57.946882] WARNING: CPU: 1 PID: 4053 at fs/iomap.c:890
iomap_dio_rw+0x37a/0x390
[ 57.947066] WARNING: CPU: 1 PID: 4053 at fs/iomap.c:766
iomap_dio_actor+0xca/0x370
----------------------------------------------------------------------------------------------------------------------------
It focus on the oops due to this race, and the warnings are different based
on different kernel.
So we may ignore dmesg check.
I will add comments on ignoring dmesg check result as you said.
Thanks,
Xiao Yang
Thanks,
Eryu
+_require_xfs_io_command "truncate"
+_require_xfs_io_command "fpunch"
+
+rm -f $seqres.full
+
+# format and mount
+_scratch_mkfs> $seqres.full 2>&1
+_scratch_mount>> $seqres.full 2>&1
+
+filesz=$((65536 * 2))
+
+# create a test file with a hole
+$XFS_IO_PROG -f -c "truncate $((filesz * 2))" $SCRATCH_MNT/file>> $seqres.full
+
+# run a background dio read to a hole in a loop
+while true; do
+ $XFS_IO_PROG -d -c "pread 0 $filesz" $SCRATCH_MNT/file> /dev/null 2>&1
+done&
+
+dread_pid=$!
+
+# run mapped write to the same hole as dio read
+for i in `seq 0 999`; do
+ $XFS_IO_PROG -c "mmap 0 $filesz" -c "mwrite 0 $filesz" $SCRATCH_MNT/file \
+ > /dev/null
+ $XFS_IO_PROG -c "fpunch 0 $filesz" $SCRATCH_MNT/file> /dev/null
+done
+
+kill -9 $dread_pid> /dev/null 2>&1
+wait $dread_pid> /dev/null 2>&1
+
+echo "Silence is golden"
+
+# success, all done
+status=0
+exit
diff --git a/tests/xfs/422.out b/tests/xfs/422.out
new file mode 100644
index 0000000..f70693f
--- /dev/null
+++ b/tests/xfs/422.out
@@ -0,0 +1,2 @@
+QA output created by 422
+Silence is golden
diff --git a/tests/xfs/group b/tests/xfs/group
index 2a3b950..f43ffcf 100644
--- a/tests/xfs/group
+++ b/tests/xfs/group
@@ -419,3 +419,4 @@
419 auto quick
420 auto quick clone dedupe
421 auto quick clone dedupe
+422 auto quick dangerous
--
1.8.3.1
--
To unsubscribe from this list: send the line "unsubscribe fstests" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
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@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
