| Message ID | 20220512175603.2833164-1-zlang@kernel.org (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | common/overlay: source base fs specific common file | expand |
On Thu, May 12, 2022 at 8:56 PM Zorro Lang <zlang@kernel.org> wrote: > > When tests overlayfs, sometimes we need the underlying fs specific > helpers, e.g. common/rc has: > _filesystem_timestamp_range $OVL_BASE_TEST_DEV $OVL_BASE_FSTYP > > So when we source common/overlay, better to source then OVL_BASE_FSTYP > too. > > Signed-off-by: Zorro Lang <zlang@kernel.org> > --- > > The generic/402 always fails on overlayfs as: > > QA output created by 402 > ./common/rc: line 2441: _xfs_timestamp_range: command not found > ./common/rc: line 2403: [: syntax error: '-1' unexpected > ./common/rc: line 2408: [: -le: unary operator expected > ./common/rc: line 2441: _xfs_timestamp_range: command not found > 0;0 != 1;1 > 0;0 != 1;1 > Silence is golden > > Due to if FSTYP is overlay, we only source common/overlay file. But we > still need the helpers of the underlying filesystem. So I think we need > to source the common/$OVL_BASE_FSTYP in common/overlay. What do you think? That makes sense. I don't know why I have never hit this. Maybe because I am not running with multi section config or because I am runing with kvm-xfstests runner. Thanks, Amir.
On Thu, May 12, 2022 at 09:11:34PM +0300, Amir Goldstein wrote: > On Thu, May 12, 2022 at 8:56 PM Zorro Lang <zlang@kernel.org> wrote: > > > > When tests overlayfs, sometimes we need the underlying fs specific > > helpers, e.g. common/rc has: > > _filesystem_timestamp_range $OVL_BASE_TEST_DEV $OVL_BASE_FSTYP > > > > So when we source common/overlay, better to source then OVL_BASE_FSTYP > > too. > > > > Signed-off-by: Zorro Lang <zlang@kernel.org> > > --- > > > > The generic/402 always fails on overlayfs as: > > > > QA output created by 402 > > ./common/rc: line 2441: _xfs_timestamp_range: command not found > > ./common/rc: line 2403: [: syntax error: '-1' unexpected > > ./common/rc: line 2408: [: -le: unary operator expected > > ./common/rc: line 2441: _xfs_timestamp_range: command not found > > 0;0 != 1;1 > > 0;0 != 1;1 > > Silence is golden > > > > Due to if FSTYP is overlay, we only source common/overlay file. But we > > still need the helpers of the underlying filesystem. So I think we need > > to source the common/$OVL_BASE_FSTYP in common/overlay. What do you think? > > That makes sense. I haven't given it enough test, hope it won't break something:) > I don't know why I have never hit this. > Maybe because I am not running with multi section config or > because I am runing with kvm-xfstests runner. Do you test overlay with XFS? For now, I think only xfs+overlay can reproduce this failure. Thanks, Zorro > > Thanks, > Amir. >
On Thu, May 12, 2022 at 9:29 PM Zorro Lang <zlang@kernel.org> wrote: > > On Thu, May 12, 2022 at 09:11:34PM +0300, Amir Goldstein wrote: > > On Thu, May 12, 2022 at 8:56 PM Zorro Lang <zlang@kernel.org> wrote: > > > > > > When tests overlayfs, sometimes we need the underlying fs specific > > > helpers, e.g. common/rc has: > > > _filesystem_timestamp_range $OVL_BASE_TEST_DEV $OVL_BASE_FSTYP > > > > > > So when we source common/overlay, better to source then OVL_BASE_FSTYP > > > too. > > > > > > Signed-off-by: Zorro Lang <zlang@kernel.org> > > > --- > > > > > > The generic/402 always fails on overlayfs as: > > > > > > QA output created by 402 > > > ./common/rc: line 2441: _xfs_timestamp_range: command not found > > > ./common/rc: line 2403: [: syntax error: '-1' unexpected > > > ./common/rc: line 2408: [: -le: unary operator expected > > > ./common/rc: line 2441: _xfs_timestamp_range: command not found > > > 0;0 != 1;1 > > > 0;0 != 1;1 > > > Silence is golden > > > > > > Due to if FSTYP is overlay, we only source common/overlay file. But we > > > still need the helpers of the underlying filesystem. So I think we need > > > to source the common/$OVL_BASE_FSTYP in common/overlay. What do you think? > > > > That makes sense. > > I haven't given it enough test, hope it won't break something:) > > > I don't know why I have never hit this. > > Maybe because I am not running with multi section config or > > because I am runing with kvm-xfstests runner. > > Do you test overlay with XFS? For now, I think only xfs+overlay can reproduce > this failure. I only test xfs+overlay. And Darrick also tests this regularly for xfs regression test, so it must be a different way in which we invoke the test. Thanks, Amir.
diff --git a/common/config b/common/config index efcf12e9..b4b2a8d4 100644 --- a/common/config +++ b/common/config @@ -481,6 +481,65 @@ _fsck_opts() esac } +_source_specific_fs() +{ + local fs=$1 + + if [ -z "$fs" ];then + fs=$FSTYP + fi + + case "$fs" in + xfs) + [ "$XFS_LOGPRINT_PROG" = "" ] && _fatal "xfs_logprint not found" + [ "$XFS_REPAIR_PROG" = "" ] && _fatal "xfs_repair not found" + [ "$XFS_DB_PROG" = "" ] && _fatal "xfs_db not found" + [ "$MKFS_XFS_PROG" = "" ] && _fatal "mkfs_xfs not found" + [ "$XFS_INFO_PROG" = "" ] && _fatal "xfs_info not found" + + . ./common/xfs + ;; + udf) + [ "$MKFS_UDF_PROG" = "" ] && _fatal "mkfs_udf/mkudffs not found" + ;; + btrfs) + [ "$MKFS_BTRFS_PROG" = "" ] && _fatal "mkfs.btrfs not found" + + . ./common/btrfs + ;; + ext4) + [ "$MKFS_EXT4_PROG" = "" ] && _fatal "mkfs.ext4 not found" + ;; + f2fs) + [ "$MKFS_F2FS_PROG" = "" ] && _fatal "mkfs.f2fs not found" + ;; + nfs) + . ./common/nfs + ;; + cifs) + ;; + 9p) + ;; + ceph) + . ./common/ceph + ;; + glusterfs) + ;; + overlay) + . ./common/overlay + ;; + reiser4) + [ "$MKFS_REISER4_PROG" = "" ] && _fatal "mkfs.reiser4 not found" + ;; + pvfs2) + ;; + ubifs) + [ "$UBIUPDATEVOL_PROG" = "" ] && _fatal "ubiupdatevol not found" + ;; + esac +} + + known_hosts() { [ "$HOST_CONFIG_DIR" ] || HOST_CONFIG_DIR=`pwd`/configs diff --git a/common/overlay b/common/overlay index fff67ba1..4708d393 100644 --- a/common/overlay +++ b/common/overlay @@ -12,6 +12,10 @@ export OVL_XATTR_NLINK="trusted.overlay.nlink" export OVL_XATTR_UPPER="trusted.overlay.upper" export OVL_XATTR_METACOPY="trusted.overlay.metacopy" +if [ -n "$OVL_BASE_FSTYP" ];then + _source_specific_fs $OVL_BASE_FSTYP +fi + # helper function to do the actual overlayfs mount operation _overlay_mount_dirs() { diff --git a/common/rc b/common/rc index e907e362..ebdeeb93 100644 --- a/common/rc +++ b/common/rc @@ -96,54 +96,7 @@ _log_err() umask 022 # check for correct setup and source the $FSTYP specific functions now -case "$FSTYP" in - xfs) - [ "$XFS_LOGPRINT_PROG" = "" ] && _fatal "xfs_logprint not found" - [ "$XFS_REPAIR_PROG" = "" ] && _fatal "xfs_repair not found" - [ "$XFS_DB_PROG" = "" ] && _fatal "xfs_db not found" - [ "$MKFS_XFS_PROG" = "" ] && _fatal "mkfs_xfs not found" - [ "$XFS_INFO_PROG" = "" ] && _fatal "xfs_info not found" - - . ./common/xfs - ;; - udf) - [ "$MKFS_UDF_PROG" = "" ] && _fatal "mkfs_udf/mkudffs not found" - ;; - btrfs) - [ "$MKFS_BTRFS_PROG" = "" ] && _fatal "mkfs.btrfs not found" - - . ./common/btrfs - ;; - ext4) - [ "$MKFS_EXT4_PROG" = "" ] && _fatal "mkfs.ext4 not found" - ;; - f2fs) - [ "$MKFS_F2FS_PROG" = "" ] && _fatal "mkfs.f2fs not found" - ;; - nfs) - . ./common/nfs - ;; - cifs) - ;; - 9p) - ;; - ceph) - . ./common/ceph - ;; - glusterfs) - ;; - overlay) - . ./common/overlay - ;; - reiser4) - [ "$MKFS_REISER4_PROG" = "" ] && _fatal "mkfs.reiser4 not found" - ;; - pvfs2) - ;; - ubifs) - [ "$UBIUPDATEVOL_PROG" = "" ] && _fatal "ubiupdatevol not found" - ;; -esac +_source_specific_fs $FSTYP if [ ! -z "$REPORT_LIST" ]; then . ./common/report
When tests overlayfs, sometimes we need the underlying fs specific helpers, e.g. common/rc has: _filesystem_timestamp_range $OVL_BASE_TEST_DEV $OVL_BASE_FSTYP So when we source common/overlay, better to source then OVL_BASE_FSTYP too. Signed-off-by: Zorro Lang <zlang@kernel.org> --- The generic/402 always fails on overlayfs as: QA output created by 402 ./common/rc: line 2441: _xfs_timestamp_range: command not found ./common/rc: line 2403: [: syntax error: '-1' unexpected ./common/rc: line 2408: [: -le: unary operator expected ./common/rc: line 2441: _xfs_timestamp_range: command not found 0;0 != 1;1 0;0 != 1;1 Silence is golden Due to if FSTYP is overlay, we only source common/overlay file. But we still need the helpers of the underlying filesystem. So I think we need to source the common/$OVL_BASE_FSTYP in common/overlay. What do you think? Thanks, Zorro common/config | 59 ++++++++++++++++++++++++++++++++++++++++++++++++++ common/overlay | 4 ++++ common/rc | 49 +---------------------------------------- 3 files changed, 64 insertions(+), 48 deletions(-)