From: Qu Wenruo <quwenruo@cn.fujitsu.com> To: <fstests@vger.kernel.org> Cc: <linux-btrfs@vger.kernel.org>, Filipe Manana <fdmanana@suse.com> Subject: [PATCH v2] btrfs: regression test for btrfs extent merge Date: Thu, 16 Oct 2014 17:08:58 +0800 [thread overview] Message-ID: <1413450538-9783-1-git-send-email-quwenruo@cn.fujitsu.com> (raw) The following kernel commit introduced the bug: 51f395ad btrfs: Use right extent length when inserting overlap extent map. When btrfs commit race with btrfs_get_extent(), merge_extent_mapping() may build up a new extent which length overflows and cause extent map insert fail, causing the caller get a -EEXIST error. This regression is fixed by the following patches: btrfs: Fix and enhance merge_extent_mapping() to insert best fitted extent map btrfs: Fix the wrong condition judgment about subset extent map Cc: Filipe Manana <fdmanana@suse.com> Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com> Reviewed-by: Filipe Manana <fdmanana@suse.com> Tested-by: Filipe Manana <fdmanana@suse.com> --- changlog: v2: Remove duplicated variant Use _require_fs_space helper Use temp file other than temp dir --- tests/btrfs/078 | 126 ++++++++++++++++++++++++++++++++++++++++++++++++++++ tests/btrfs/078.out | 2 + tests/btrfs/group | 1 + 3 files changed, 129 insertions(+) create mode 100755 tests/btrfs/078 create mode 100644 tests/btrfs/078.out diff --git a/tests/btrfs/078 b/tests/btrfs/078 new file mode 100755 index 0000000..78132c3 --- /dev/null +++ b/tests/btrfs/078 @@ -0,0 +1,126 @@ +#! /bin/bash +# FS QA Test No. btrfs/078 +# +# Do write along with fiemap ioctl. +# Regression test for the kernel comit: +# 51f395ad btrfs: Use right extent length when inserting overlap extent map. +# +# When calling fiemap(without SYNC flag) and btrfs fs is commiting, +# it will cause race condition and cause btrfs to generate a wrong extent +# whose len is overflow and fail to insert into the extent map tree, +# returning -EEXIST. +# +# Fixed by the following patches (not merged in mainline yet): +# btrfs: Fix and enhance merge_extent_mapping() to insert best fitted extent map +# btrfs: Fix the wrong condition judgment about subset extent map +# +#----------------------------------------------------------------------- +# Copyright (C) 2014 Fujitsu All Rights Reserved. +# Author: Qu Wenruo <quwenruo@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" + +tmp=/tmp/$$ +status=1 # failure is the default! +trap "_cleanup; exit \$status" 0 1 2 3 15 + +_cleanup() +{ + kill $dd_pid &> /dev/null + kill $fiemap_pid &> /dev/null + wait + rm -fr $testfile + rm -fr $tmp.* $tmp +} + +# get standard environment, filters and checks +. ./common/rc +. ./common/filter + +# real QA test starts here +_supported_fs btrfs +_supported_os Linux +_require_scratch +_need_to_be_root +# Since xfs_io's fiemap always use SYNC flag and can't be unset, +# we must use filefrag to call fiemap without SYNC flag. +_require_command "/usr/sbin/filefrag" +_require_xfs_io_command "falloc" + +filesize=$((10 * 1024 * 1024 * 1024)) #10G size +buffersize=$((1024 * 1024)) # 1M bs for dd +count=$(($filesize / $buffersize)) +testfile=$SCRATCH_MNT/testfile + +rm -f $seqres.full + +_scratch_mkfs >>$seqres.full 2>&1 +_scratch_mount +_require_fs_space $SCRATCH_MNT $(($filesize / 1024 / 1024 / 1024)) +$XFS_IO_PROG -f -c "falloc 0 $filesize" $testfile + +dd_work() { + out=$1 + dd if=/dev/zero of=$out bs=$buffersize count=$count \ + conv=notrunc &> /dev/null +} + +# There is a bug for e2fsprogs, at least in version 1.42.9, filefrag will +# leak the return value, so we can't judge return value only, +# but also to filter the output +_filter_error() { + # when filefrag fails FIEMAP ioctl, it will fall back to FIBMAP, + # which is not supported by btrfs and will report "FIBMAP: strerr()" + # However old e2fsprogs will use wrong errno EINVAL other than ENOTTY + # so only grep for "FIBMAP" for max compatibility. + grep "FIBMAP" +} + +fiemap_work() { + filename=$1 + while true; do + filefrag $filename 2> $tmp.output 1> /dev/null + ret=$? + err=`cat $tmp.output | _filter_error` + if [ $ret -ne 0 -o -n "$err" ]; then + kill $dd_pid + return 1 + fi + done +} + +dd_work $testfile & +dd_pid=$! +fiemap_work $testfile & +fiemap_pid=$! +wait $dd_pid +ddret=$? +kill $fiemap_pid &> /dev/null +wait $fiemap_pid + +if [ $ddret -ne 0 ]; then + echo "Extent merge bug detected" + status=1 + exit +else + echo "Silence is golden" + status=0 + exit +fi diff --git a/tests/btrfs/078.out b/tests/btrfs/078.out new file mode 100644 index 0000000..b8acea8 --- /dev/null +++ b/tests/btrfs/078.out @@ -0,0 +1,2 @@ +QA output created by 078 +Silence is golden diff --git a/tests/btrfs/group b/tests/btrfs/group index 9adf862..40e7430 100644 --- a/tests/btrfs/group +++ b/tests/btrfs/group @@ -80,3 +80,4 @@ 075 auto quick subvol 076 auto quick 077 auto quick +078 auto -- 2.1.2
WARNING: multiple messages have this Message-ID (diff)
From: Qu Wenruo <quwenruo@cn.fujitsu.com> To: fstests@vger.kernel.org Cc: linux-btrfs@vger.kernel.org, Filipe Manana <fdmanana@suse.com> Subject: [PATCH v2] btrfs: regression test for btrfs extent merge Date: Thu, 16 Oct 2014 17:08:58 +0800 [thread overview] Message-ID: <1413450538-9783-1-git-send-email-quwenruo@cn.fujitsu.com> (raw) The following kernel commit introduced the bug: 51f395ad btrfs: Use right extent length when inserting overlap extent map. When btrfs commit race with btrfs_get_extent(), merge_extent_mapping() may build up a new extent which length overflows and cause extent map insert fail, causing the caller get a -EEXIST error. This regression is fixed by the following patches: btrfs: Fix and enhance merge_extent_mapping() to insert best fitted extent map btrfs: Fix the wrong condition judgment about subset extent map Cc: Filipe Manana <fdmanana@suse.com> Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com> Reviewed-by: Filipe Manana <fdmanana@suse.com> Tested-by: Filipe Manana <fdmanana@suse.com> --- changlog: v2: Remove duplicated variant Use _require_fs_space helper Use temp file other than temp dir --- tests/btrfs/078 | 126 ++++++++++++++++++++++++++++++++++++++++++++++++++++ tests/btrfs/078.out | 2 + tests/btrfs/group | 1 + 3 files changed, 129 insertions(+) create mode 100755 tests/btrfs/078 create mode 100644 tests/btrfs/078.out diff --git a/tests/btrfs/078 b/tests/btrfs/078 new file mode 100755 index 0000000..78132c3 --- /dev/null +++ b/tests/btrfs/078 @@ -0,0 +1,126 @@ +#! /bin/bash +# FS QA Test No. btrfs/078 +# +# Do write along with fiemap ioctl. +# Regression test for the kernel comit: +# 51f395ad btrfs: Use right extent length when inserting overlap extent map. +# +# When calling fiemap(without SYNC flag) and btrfs fs is commiting, +# it will cause race condition and cause btrfs to generate a wrong extent +# whose len is overflow and fail to insert into the extent map tree, +# returning -EEXIST. +# +# Fixed by the following patches (not merged in mainline yet): +# btrfs: Fix and enhance merge_extent_mapping() to insert best fitted extent map +# btrfs: Fix the wrong condition judgment about subset extent map +# +#----------------------------------------------------------------------- +# Copyright (C) 2014 Fujitsu All Rights Reserved. +# Author: Qu Wenruo <quwenruo@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" + +tmp=/tmp/$$ +status=1 # failure is the default! +trap "_cleanup; exit \$status" 0 1 2 3 15 + +_cleanup() +{ + kill $dd_pid &> /dev/null + kill $fiemap_pid &> /dev/null + wait + rm -fr $testfile + rm -fr $tmp.* $tmp +} + +# get standard environment, filters and checks +. ./common/rc +. ./common/filter + +# real QA test starts here +_supported_fs btrfs +_supported_os Linux +_require_scratch +_need_to_be_root +# Since xfs_io's fiemap always use SYNC flag and can't be unset, +# we must use filefrag to call fiemap without SYNC flag. +_require_command "/usr/sbin/filefrag" +_require_xfs_io_command "falloc" + +filesize=$((10 * 1024 * 1024 * 1024)) #10G size +buffersize=$((1024 * 1024)) # 1M bs for dd +count=$(($filesize / $buffersize)) +testfile=$SCRATCH_MNT/testfile + +rm -f $seqres.full + +_scratch_mkfs >>$seqres.full 2>&1 +_scratch_mount +_require_fs_space $SCRATCH_MNT $(($filesize / 1024 / 1024 / 1024)) +$XFS_IO_PROG -f -c "falloc 0 $filesize" $testfile + +dd_work() { + out=$1 + dd if=/dev/zero of=$out bs=$buffersize count=$count \ + conv=notrunc &> /dev/null +} + +# There is a bug for e2fsprogs, at least in version 1.42.9, filefrag will +# leak the return value, so we can't judge return value only, +# but also to filter the output +_filter_error() { + # when filefrag fails FIEMAP ioctl, it will fall back to FIBMAP, + # which is not supported by btrfs and will report "FIBMAP: strerr()" + # However old e2fsprogs will use wrong errno EINVAL other than ENOTTY + # so only grep for "FIBMAP" for max compatibility. + grep "FIBMAP" +} + +fiemap_work() { + filename=$1 + while true; do + filefrag $filename 2> $tmp.output 1> /dev/null + ret=$? + err=`cat $tmp.output | _filter_error` + if [ $ret -ne 0 -o -n "$err" ]; then + kill $dd_pid + return 1 + fi + done +} + +dd_work $testfile & +dd_pid=$! +fiemap_work $testfile & +fiemap_pid=$! +wait $dd_pid +ddret=$? +kill $fiemap_pid &> /dev/null +wait $fiemap_pid + +if [ $ddret -ne 0 ]; then + echo "Extent merge bug detected" + status=1 + exit +else + echo "Silence is golden" + status=0 + exit +fi diff --git a/tests/btrfs/078.out b/tests/btrfs/078.out new file mode 100644 index 0000000..b8acea8 --- /dev/null +++ b/tests/btrfs/078.out @@ -0,0 +1,2 @@ +QA output created by 078 +Silence is golden diff --git a/tests/btrfs/group b/tests/btrfs/group index 9adf862..40e7430 100644 --- a/tests/btrfs/group +++ b/tests/btrfs/group @@ -80,3 +80,4 @@ 075 auto quick subvol 076 auto quick 077 auto quick +078 auto -- 2.1.2
next reply other threads:[~2014-10-16 9:09 UTC|newest] Thread overview: 2+ messages / expand[flat|nested] mbox.gz Atom feed top 2014-10-16 9:08 Qu Wenruo [this message] 2014-10-16 9:08 ` [PATCH v2] btrfs: regression test for btrfs extent merge Qu Wenruo
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=1413450538-9783-1-git-send-email-quwenruo@cn.fujitsu.com \ --to=quwenruo@cn.fujitsu.com \ --cc=fdmanana@suse.com \ --cc=fstests@vger.kernel.org \ --cc=linux-btrfs@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: linkBe sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.