overlay: workaround libmount failure to remount,ro
PATCH 1/1 overlay: workaround libmount failure to remount,ro
libmount v1.41 calls several unneeded fsconfig() calls to reconfigure lowerdir/upperdir when user requests only -o remount,ro.
Those calls fail because overlayfs does not allow making any config changes with new mount api, besides MS_RDONLY.
force mount(8) to use mount(2) to remount ro/rw to workaround this issue, by setting LIBMOUNT_FORCE_MOUNT2=always.
--- common/overlay | 4 +++- tests/overlay/035 | 2 +- 2 files changed, 4 insertions(+), 2 deletions(-)
common/overlay +3 -1
@@ -127,7 +127,9 @@ _overlay_base_scratch_mount() _overlay_scratch_mount() { if echo "$*" | grep -q remount; then - $MOUNT_PROG $SCRATCH_MNT $* + # force mount(8) to use mount(2), to workaround libmount v1.41 + # failed fsconfig() calls to reconfigure lowerdir/upperdir + LIBMOUNT_FORCE_MOUNT2=always $MOUNT_PROG $SCRATCH_MNT $* return fi
tests/overlay/035 +1 -1
@@ -42,7 +42,7 @@ mkdir -p $lowerdir1 $lowerdir2 $upperdir $workdir # Verify that overlay is mounted read-only and that it cannot be remounted rw. _overlay_scratch_mount_opts -o"lowerdir=$lowerdir2:$lowerdir1" touch $SCRATCH_MNT/foo 2>&1 | _filter_scratch -$MOUNT_PROG -o remount,rw $SCRATCH_MNT 2>&1 | _filter_ro_mount +_scratch_remount rw 2>&1 | _filter_ro_mount $UMOUNT_PROG $SCRATCH_MNT # Make workdir immutable to prevent workdir re-create on mount
Signature · 1 line
2.34.1
Quoted 6 lines · Amir Goldstein
libmount v1.41 calls several unneeded
fsconfig()calls to reconfigure lowerdir/upperdir when user requests only -o remount,ro.Those calls fail because overlayfs does not allow making any config changes with new mount api, besides
MS_RDONLY.
Quoted 6 lines
force mount(8) to use mount(2) to remount ro/rw to workaround this issue, by setting LIBMOUNT_FORCE_MOUNT2=always.
Reported-by André Almeida <[email protected]>Cc Karel Zak <[email protected]>
Is my bug report (a year ago) gotten fixed?
https://lore.kernel.org/linux-fsdevel/20241026180741.cfqm6oqp3frvasfm@dell-per750-06-vm-08.rhts.eng.pek2.redhat.com/
If this kernel fix works, do we still need this workaround?
Thanks,
Zorro
Quoted 37 lines
Signed-off-by Amir Goldstein <[email protected]>--- common/overlay | 4 +++- tests/overlay/035 | 2 +- 2 files changed, 4 insertions(+), 2 deletions(-)common/overlay +3
-1@@ -127,7 +127,9 @@ _overlay_base_scratch_mount() _overlay_scratch_mount() { if echo "$*" | grep -q remount; then - $MOUNT_PROG $SCRATCH_MNT $* + # force mount(8) to use mount(2), to workaround libmount v1.41 + # failed fsconfig() calls to reconfigure lowerdir/upperdir + LIBMOUNT_FORCE_MOUNT2=always $MOUNT_PROG $SCRATCH_MNT $* return fitests/overlay/035 +1
-1@@ -42,7 +42,7 @@ mkdir -p $lowerdir1 $lowerdir2 $upperdir $workdir # Verify that overlay is mounted read-only and that it cannot be remounted rw. _overlay_scratch_mount_opts -o"lowerdir=$lowerdir2:$lowerdir1" touch $SCRATCH_MNT/foo 2>&1 | _filter_scratch -$MOUNT_PROG -o remount,rw $SCRATCH_MNT 2>&1 | _filter_ro_mount +_scratch_remount rw 2>&1 | _filter_ro_mount $UMOUNT_PROG $SCRATCH_MNT # Make workdir immutable to prevent workdir re-create on mountSignature · 1 line
2.34.1
Quoted 11 lines · Zorro Lang
Quoted 9 lines · Amir Goldstein
libmount v1.41 calls several unneeded
fsconfig()calls to reconfigure lowerdir/upperdir when user requests only -o remount,ro.Those calls fail because overlayfs does not allow making any config changes with new mount api, besides
MS_RDONLY.force mount(8) to use mount(2) to remount ro/rw to workaround this issue, by setting LIBMOUNT_FORCE_MOUNT2=always.
Quoted 7 lines
Quoted 3 lines
Reported-by André Almeida <[email protected]>Cc Karel Zak <[email protected]>Is my bug report (a year ago) gotten fixed?
https://lore.kernel.org/linux-fsdevel/20241026180741.cfqm6oqp3frvasfm@dell-per750-06-vm-08.rhts.eng.pek2.redhat.com/
Please see v2 (this is v1):
https://lore.kernel.org/fstests/[email protected]/
The suggested workaround is exactly the same as suggested one year ago
Quoted 2 lines
If this kernel fix works, do we still need this workaround?
Which kernel fix?
There was no agreement on a kernel fix (yet).
Thanks,
Amir.