vendor: update github.com/opencontainers/runc

when the systemd cgroup manager is used, controllers not handled by
systemd are created manually afterwards.
libcontainer didn't correctly cleanup these cgroups that were leaked
on cgroup v1.

Closes: https://github.com/kubernetes/kubernetes/issues/92766

Signed-off-by: Giuseppe Scrivano <gscrivan@redhat.com>
This commit is contained in:
Giuseppe Scrivano
2020-07-09 09:37:05 +02:00
parent 896da2253c
commit c655a5b636
44 changed files with 1122 additions and 722 deletions

View File

@@ -37,6 +37,9 @@ func canSkipEBPFError(cgroup *configs.Cgroup) bool {
}
func setDevices(dirPath string, cgroup *configs.Cgroup) error {
if cgroup.SkipDevices {
return nil
}
// XXX: This is currently a white-list (but all callers pass a blacklist of
// devices). This is bad for a whole variety of reasons, but will need
// to be fixed with co-ordinated effort with downstreams.