| Message ID | 20240429115826.9171-1-ddiss@suse.de (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | [v2] tests: _fail on _scratch_mkfs_sized failure | expand |
On Mon, Apr 29, 2024 at 09:58:27PM +1000, David Disseldorp wrote: > If _scratch_mkfs_sized() fails, e.g. due to an FS not supporting the > provided size, tests may subsequently mount and run atop a previously > created (e.g. non-size-bound) filesystem. > This can lead to difficult to debug failures, or for some -ENOSPC > exercising tests, near infinite runtimes. Avoid this by renaming the > current function to _try_scratch_mkfs_sized() and _fail in the parent > _scratch_mkfs_sized() wrapper. > > Signed-off-by: David Disseldorp <ddiss@suse.de> > Suggested-by: Darrick J. Wong <djwong@kernel.org> > --- > Changes since v1: > - call _fail in a wrapper function, instead of adding failure handlers > to individual test _scratch_mkfs_sized() callers. As we call _fail in _scratch_mkfs_sized, so don't need to do _fail after mkfs fails : $ grep -rsn -A 1 _scratch_mkfs_sized tests/|grep -E "_fail " tests/btrfs/007-47- || _fail "size=$fsz mkfs failed" tests/btrfs/007-74- || _fail "size=$fsz mkfs failed" tests/btrfs/004-161- || _fail "size=$fsz mkfs failed" tests/ext4/021:27:_scratch_mkfs_sized $fssize >> $seqres.full 2>&1 || _fail "mkfs failed" tests/ext4/059-26- >>$seqres.full 2>&1 || _fail "mkfs failed" tests/generic/449-28-_scratch_mount || _fail "mount failed" tests/generic/171:45:_scratch_mkfs_sized $sz_bytes >> $seqres.full 2>&1 || _fail "mkfs failed" tests/generic/172:43:_scratch_mkfs_sized $fs_size >> $seqres.full 2>&1 || _fail "mkfs failed" tests/generic/173:45:_scratch_mkfs_sized $sz_bytes >> $seqres.full 2>&1 || _fail "mkfs failed" tests/generic/174:46:_scratch_mkfs_sized $sz_bytes >> $seqres.full 2>&1 || _fail "mkfs failed" tests/generic/735:28:_scratch_mkfs_sized $dev_size >>$seqres.full 2>&1 || _fail "mkfs failed" tests/generic/015-35- || _fail "mkfs failed" tests/generic/077:53:_scratch_mkfs_sized $fs_size >> $seqres.full 2>&1 || _fail "mkfs failed" tests/generic/083-54- || _fail "size=$fsz mkfs failed" tests/xfs/057-55- _fail "mkfs failed" tests/xfs/015:46:_scratch_mkfs_sized $((96 * 1024 * 1024)) > $tmp.mkfs.raw || _fail "mkfs failed" > > common/rc | 9 +++++++-- > 1 file changed, 7 insertions(+), 2 deletions(-) > > diff --git a/common/rc b/common/rc > index 56f1afb6..e95070dd 100644 > --- a/common/rc > +++ b/common/rc > @@ -968,8 +968,8 @@ _small_fs_size_mb() > } > > # Create fs of certain size on scratch device > -# _scratch_mkfs_sized <size in bytes> [optional blocksize] > -_scratch_mkfs_sized() > +# _try_scratch_mkfs_sized <size in bytes> [optional blocksize] > +_try_scratch_mkfs_sized() > { > local fssize=$1 > local blocksize=$2 > @@ -1111,6 +1111,11 @@ _scratch_mkfs_sized() > esac > } > > +_scratch_mkfs_sized() > +{ > + _try_scratch_mkfs_sized $* || _fail "_scratch_mkfs_sized failed" How about printing the "size" arguments? > +} > + > # Emulate an N-data-disk stripe w/ various stripe units > # _scratch_mkfs_geom <sunit bytes> <swidth multiplier> [optional blocksize] > _scratch_mkfs_geom() > -- > 2.35.3 > >
diff --git a/common/rc b/common/rc index 56f1afb6..e95070dd 100644 --- a/common/rc +++ b/common/rc @@ -968,8 +968,8 @@ _small_fs_size_mb() } # Create fs of certain size on scratch device -# _scratch_mkfs_sized <size in bytes> [optional blocksize] -_scratch_mkfs_sized() +# _try_scratch_mkfs_sized <size in bytes> [optional blocksize] +_try_scratch_mkfs_sized() { local fssize=$1 local blocksize=$2 @@ -1111,6 +1111,11 @@ _scratch_mkfs_sized() esac } +_scratch_mkfs_sized() +{ + _try_scratch_mkfs_sized $* || _fail "_scratch_mkfs_sized failed" +} + # Emulate an N-data-disk stripe w/ various stripe units # _scratch_mkfs_geom <sunit bytes> <swidth multiplier> [optional blocksize] _scratch_mkfs_geom()
If _scratch_mkfs_sized() fails, e.g. due to an FS not supporting the provided size, tests may subsequently mount and run atop a previously created (e.g. non-size-bound) filesystem. This can lead to difficult to debug failures, or for some -ENOSPC exercising tests, near infinite runtimes. Avoid this by renaming the current function to _try_scratch_mkfs_sized() and _fail in the parent _scratch_mkfs_sized() wrapper. Signed-off-by: David Disseldorp <ddiss@suse.de> Suggested-by: Darrick J. Wong <djwong@kernel.org> --- Changes since v1: - call _fail in a wrapper function, instead of adding failure handlers to individual test _scratch_mkfs_sized() callers. common/rc | 9 +++++++-- 1 file changed, 7 insertions(+), 2 deletions(-)