Message ID | 1498554232-20777-1-git-send-email-yangx.jy@cn.fujitsu.com (mailing list archive) |
---|---|
State | New, archived |
Headers | show |
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@cn.fujitsu.com> > --- > 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@cn.fujitsu.com> > +# > +# 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. > +_require_scratch > +_disable_dmesg_check Better to add comments on ignoring dmesg check result. 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@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
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@cn.fujitsu.com> >> --- >> 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@cn.fujitsu.com> >> +# >> +# 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() [ 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] ---------------------------------------------------------------------------------------------------------------------------- 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@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
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@cn.fujitsu.com> > > > --- > > > 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@cn.fujitsu.com> > > > +# > > > +# 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 :) 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] > ---------------------------------------------------------------------------------------------------------------------------- > 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@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
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@cn.fujitsu.com> >>>> --- >>>> 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@cn.fujitsu.com> >>>> +# >>>> +# 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@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
On Thu, Jun 29, 2017 at 01:21:12PM +0800, Xiao Yang wrote: > 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@cn.fujitsu.com> > > > > > --- > > > > > 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@cn.fujitsu.com> > > > > > +# > > > > > +# 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) Please send another patch with detailed background information to add it to the filter and cc linux-xfs list, perhaps after both Brian's patch and this new test patch being merged & pushed out to upstream. Thanks a lot! Eryu -- 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
On 2017/06/29 13:39, Eryu Guan wrote: > On Thu, Jun 29, 2017 at 01:21:12PM +0800, Xiao Yang wrote: >> 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@cn.fujitsu.com> >>>>>> --- >>>>>> 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@cn.fujitsu.com> >>>>>> +# >>>>>> +# 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) > Please send another patch with detailed background information to add it > to the filter and cc linux-xfs list, perhaps after both Brian's patch > and this new test patch being merged& pushed out to upstream. Thanks a > lot! Hi Eryu Thanks for your suggestion, I will send another patch as you said. :-) Thanks, Xiao Yang. > Eryu > > > . > -- 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 --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) +# +# 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@cn.fujitsu.com> +# +# 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 +_require_scratch +_disable_dmesg_check +_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
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@cn.fujitsu.com> --- 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