During container start there might be a situation when not all cgroup hierarchies get virtualized by container manager (like vzctl). By virtualizing a cgroup hierarchy I mean creation of sub-directory within a particular mounted cgroup. When container starts it looks in css set of it's init process to list all affilated cgroups and perform actions on each. But non-virtualized cgroups will also be present in init's css_set and they should not be touched from inside of any non root ve.
Signed-off-by: Valeriy Vdovin <[email protected]> --- kernel/cgroup.c | 31 +++++++++++++++++++++++++++++++ 1 file changed, 31 insertions(+) diff --git a/kernel/cgroup.c b/kernel/cgroup.c index 3305032..aefe40b 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -4449,6 +4449,18 @@ static struct cftype *get_cftype_by_name(const char *name) } #ifdef CONFIG_VE +static inline bool is_virtualized_cgroup(struct cgroup *cgrp) +{ + lockdep_assert_held(&cgroup_mutex); + if (cgrp->root->subsys_mask) + return true; + + if (!strcmp(cgrp->root->name, "systemd")) + return true; + + return false; +} + int cgroup_mark_ve_roots(struct ve_struct *ve) { struct cgroup *cgrp, *tmp; @@ -4464,6 +4476,17 @@ int cgroup_mark_ve_roots(struct ve_struct *ve) mutex_lock(&cgroup_mutex); for_each_active_root(root) { cgrp = css_cgroup_from_root(ve->root_css_set, root); + + /* + * At container start, vzctl creates special cgroups to serve + * as virtualized cgroup roots. They are bind-mounted on top + * of original cgroup mount point in container namespace. But + * not all cgroup mounts undergo this procedure. We should + * skip cgroup mounts that are not virtualized. + */ + if (!is_virtualized_cgroup(cgrp)) + continue; + rcu_assign_pointer(cgrp->ve_owner, ve); set_bit(CGRP_VE_ROOT, &cgrp->flags); @@ -4513,6 +4536,14 @@ void cgroup_unmark_ve_roots(struct ve_struct *ve) mutex_lock(&cgroup_mutex); for_each_active_root(root) { cgrp = css_cgroup_from_root(ve->root_css_set, root); + + /* + * For this line see comments in + * cgroup_mark_ve_roots + */ + if (!is_virtualized_cgroup(cgrp)) + continue; + dget(cgrp->dentry); list_add_tail(&cgrp->cft_q_node, &pending); } -- 1.8.3.1 _______________________________________________ Devel mailing list [email protected] https://lists.openvz.org/mailman/listinfo/devel
