Description of problem:

Restoring a checkpointed container fails with cgroupsv1, but then succeeds on the second try.

Version-Release number of selected component (if applicable):
podman-4.0.3-1.el9.x86_64
criu-3.15-13.el9.x86_64
crun-1.4.5-2.el9.x86_64
runc-1.0.3-4.el9.x86_64

How reproducible:
Always

Steps to Reproduce:
1. grubby --update-kernel=ALL --args="systemd.unified_cgroup_hierarchy=0"
2. reboot
3. podman run -dit --name swamped-crate busybox:latest sh
114ca3e855150cdb2595341011baaf906d83dfbb2c9d3bea5fb0098a799ef40c
4. podman container checkpoint swamped-crate
114ca3e855150cdb2595341011baaf906d83dfbb2c9d3bea5fb0098a799ef40c
5. podman container restore swamped-crate
Error: OCI runtime error: crun: CRIU restoring failed -52. Please check CRIU logfile /var/lib/containers/storage/overlay-containers/114ca3e855150cdb2595341011baaf906d83dfbb2c9d3bea5fb0098a799ef40c/userdata/restore.log
6. podman container restore swamped-crate
114ca3e855150cdb2595341011baaf906d83dfbb2c9d3bea5fb0098a799ef40c

Actual results:

"podman container restore" fails on the first try, then succeeds.

Expected results:

It should succeed on the first try.

Additional info:

The ultimate error is this:

(00.004812) cg: Restoring cgroup property value [9223372036854771712] to [memory/machine.slice/libpod-10150e8cd8438787ec0462945a871045d06c24574c81e1e4b7f1db70d69ac4a2.scope/memory.kmem.limit_in_bytes]
(00.004827) Error (criu/cgroup.c:1360): cg: Failed writing 9223372036854771712 to memory/machine.slice/libpod-10150e8cd8438787ec0462945a871045d06c24574c81e1e4b7f1db70d69ac4a2.scope/memory.kmem.limit_in_bytes: Operation not supported
(00.004833) Error (criu/cgroup.c:1629): cg: Restoring memory.kmem.limit_in_bytes special property failed
(00.004835) Error (criu/cgroup.c:1690): cg: Restoring special cpuset props failed!