| Message ID | 87244a2006270f7f8a32680ce8ba7656e277a4bd.1441373012.git.zhaolei@cn.fujitsu.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
Zhao Lei wrote on 2015/09/04 21:23 +0800: > To check is btrfs-convert create bad filesystem with > leaf accross stripes. > > It is happened in progs version <=v4.1.2, and fixed by patch titled: > btrfs: convert: Avoid allocating metadata extent crossing stripe boundary > which was merged in v4.2. > > Notice thar this testcase can not report error in old version of > btrfs-progs, because "btrfs check" can't check this type of error > in those version, but we have another testcase in fsck-tests, to > check is "btrfs check" support this check. > > So, use above 2 testcase together can check out leaf-accross-stripes > bug in all version. > > Signed-off-by: Zhao Lei <zhaolei@cn.fujitsu.com> > --- > tests/misc-tests/008-leaf-accross-stripes/test.sh | 24 +++++++++++++++++++++++ > 1 file changed, 24 insertions(+) > create mode 100755 tests/misc-tests/008-leaf-accross-stripes/test.sh > > diff --git a/tests/misc-tests/008-leaf-accross-stripes/test.sh b/tests/misc-tests/008-leaf-accross-stripes/test.sh > new file mode 100755 > index 0000000..4801dce > --- /dev/null > +++ b/tests/misc-tests/008-leaf-accross-stripes/test.sh > @@ -0,0 +1,24 @@ > +#!/bin/bash > +# test btrfs subvolume run normally with more than one subvolume > +# > +# - btrfs subvolume must not loop indefinetelly > +# - btrfs subvolume return 0 in normal case > + > +source $TOP/tests/common > + > +check_prereq btrfs-convert > +check_prereq btrfs > + > +# In my test, it happened in 514M~560M, 737M~769M, 929M~917M, > +# and HAVE_ERROR=((size + 1) / 2) % 2 if size >= 970 > +# > +SIZE_FROM=514 > +SIZE_END=560 > +A_PRIME_NUM=17 > +for ((size = SIZE_FROM; size <= SIZE_END; size += A_PRIME_NUM)); do > + run_check truncate -s "$size"M "$IMAGE" > + run_check mkfs.ext4 -F "$IMAGE" > + run_check $TOP/btrfs-convert "$IMAGE" > + $TOP/btrfs check "$IMAGE" 2>&1 | grep "crossing stripe boundary" && > + _fail "leaf accross stripes in btrfs-convert" > +done > Maybe I'm wrong, but even with your script, I still can't reproduce the bug. After some search in git log, I found that, from 4.1 btrfs-convert won't allow mixed block group since the following commit: commit c9b73b702be71bbc1a3a815a745e6e4e57eadffc Author: David Sterba <dsterba@suse.cz> Date: Mon Mar 23 19:31:31 2015 +0100 btrfs-progs: convert: allow to set filesystem features So the new test may not be necessary and won't reproduce the bug. Thanks, Qu -- To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Hi, Qu Wenruo > -----Original Message----- > From: Qu Wenruo [mailto:quwenruo@cn.fujitsu.com] > Sent: Wednesday, September 09, 2015 9:46 AM > To: Zhao Lei <zhaolei@cn.fujitsu.com>; linux-btrfs@vger.kernel.org > Subject: Re: [PATCH 3/3] btrfs-progs: tests: Introduce > misc-tests/008-leaf-accross-stripes > > > > Zhao Lei wrote on 2015/09/04 21:23 +0800: > > To check is btrfs-convert create bad filesystem with leaf accross > > stripes. > > > > It is happened in progs version <=v4.1.2, and fixed by patch titled: > > btrfs: convert: Avoid allocating metadata extent crossing stripe > > boundary which was merged in v4.2. > > > > Notice thar this testcase can not report error in old version of > > btrfs-progs, because "btrfs check" can't check this type of error in > > those version, but we have another testcase in fsck-tests, to check is > > "btrfs check" support this check. > > > > So, use above 2 testcase together can check out leaf-accross-stripes > > bug in all version. > > > > Signed-off-by: Zhao Lei <zhaolei@cn.fujitsu.com> > > --- > > tests/misc-tests/008-leaf-accross-stripes/test.sh | 24 > +++++++++++++++++++++++ > > 1 file changed, 24 insertions(+) > > create mode 100755 tests/misc-tests/008-leaf-accross-stripes/test.sh > > > > diff --git a/tests/misc-tests/008-leaf-accross-stripes/test.sh > > b/tests/misc-tests/008-leaf-accross-stripes/test.sh > > new file mode 100755 > > index 0000000..4801dce > > --- /dev/null > > +++ b/tests/misc-tests/008-leaf-accross-stripes/test.sh > > @@ -0,0 +1,24 @@ > > +#!/bin/bash > > +# test btrfs subvolume run normally with more than one subvolume # # > > +- btrfs subvolume must not loop indefinetelly # - btrfs subvolume > > +return 0 in normal case > > + > > +source $TOP/tests/common > > + > > +check_prereq btrfs-convert > > +check_prereq btrfs > > + > > +# In my test, it happened in 514M~560M, 737M~769M, 929M~917M, # and > > +HAVE_ERROR=((size + 1) / 2) % 2 if size >= 970 # > > +SIZE_FROM=514 > > +SIZE_END=560 > > +A_PRIME_NUM=17 > > +for ((size = SIZE_FROM; size <= SIZE_END; size += A_PRIME_NUM)); do > > + run_check truncate -s "$size"M "$IMAGE" > > + run_check mkfs.ext4 -F "$IMAGE" > > + run_check $TOP/btrfs-convert "$IMAGE" > > + $TOP/btrfs check "$IMAGE" 2>&1 | grep "crossing stripe boundary" && > > + _fail "leaf accross stripes in btrfs-convert" > > +done > > > Maybe I'm wrong, but even with your script, I still can't reproduce the bug. > It is right case. In old version of progs, convert have above bug, but btrfs check can not report. In new version of progs, btrfs check can report the bug, but convert fixed the bug. It is noticed in patch's description. > After some search in git log, I found that, from 4.1 btrfs-convert won't allow > mixed block group since the following commit: > commit c9b73b702be71bbc1a3a815a745e6e4e57eadffc > Author: David Sterba <dsterba@suse.cz> > Date: Mon Mar 23 19:31:31 2015 +0100 > > btrfs-progs: convert: allow to set filesystem features > > So the new test may not be necessary and won't reproduce the bug. > Yes, it is fixed. But we found a bug, we need fix the bug and write a testcase for it simultaneously. Thanks Zhaolei > Thanks, > Qu -- To unsubscribe from this list: send the line "unsubscribe linux-btrfs" 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/misc-tests/008-leaf-accross-stripes/test.sh b/tests/misc-tests/008-leaf-accross-stripes/test.sh new file mode 100755 index 0000000..4801dce --- /dev/null +++ b/tests/misc-tests/008-leaf-accross-stripes/test.sh @@ -0,0 +1,24 @@ +#!/bin/bash +# test btrfs subvolume run normally with more than one subvolume +# +# - btrfs subvolume must not loop indefinetelly +# - btrfs subvolume return 0 in normal case + +source $TOP/tests/common + +check_prereq btrfs-convert +check_prereq btrfs + +# In my test, it happened in 514M~560M, 737M~769M, 929M~917M, +# and HAVE_ERROR=((size + 1) / 2) % 2 if size >= 970 +# +SIZE_FROM=514 +SIZE_END=560 +A_PRIME_NUM=17 +for ((size = SIZE_FROM; size <= SIZE_END; size += A_PRIME_NUM)); do + run_check truncate -s "$size"M "$IMAGE" + run_check mkfs.ext4 -F "$IMAGE" + run_check $TOP/btrfs-convert "$IMAGE" + $TOP/btrfs check "$IMAGE" 2>&1 | grep "crossing stripe boundary" && + _fail "leaf accross stripes in btrfs-convert" +done
To check is btrfs-convert create bad filesystem with leaf accross stripes. It is happened in progs version <=v4.1.2, and fixed by patch titled: btrfs: convert: Avoid allocating metadata extent crossing stripe boundary which was merged in v4.2. Notice thar this testcase can not report error in old version of btrfs-progs, because "btrfs check" can't check this type of error in those version, but we have another testcase in fsck-tests, to check is "btrfs check" support this check. So, use above 2 testcase together can check out leaf-accross-stripes bug in all version. Signed-off-by: Zhao Lei <zhaolei@cn.fujitsu.com> --- tests/misc-tests/008-leaf-accross-stripes/test.sh | 24 +++++++++++++++++++++++ 1 file changed, 24 insertions(+) create mode 100755 tests/misc-tests/008-leaf-accross-stripes/test.sh