[CRIU] [PATCH v2 2/2] restore: correctly restore cgroup mounts inside a container

[Tycho Andersen]

Before the nsroot= mount option, we were just getting lucky because the
cgroup superblocks "matched" when inspecting them from userspace, so we
were actually getting a bind mount from the host when migrating from within
cgroup namespaces.

Instead, let's actually do a new (i.e. not a bind mount) for cgroup
namespaces. For this, we need two things:

1. to prepare the cgroup namespace (and thus the cgroups) before the mount
   ns, so when the mount() occurrs it is relative to the right cgroup path.

2. not reject cgroup filesystems with no root. A cgroup ns mount looks
   like:

	 223 222 0:22 /lxc/unpriv /sys/fs/cgroup/systemd rw,nosuid,nodev,noexec,relatime - cgroup cgroup rw,xattr,release_agent=/lib/systemd/systemd-cgroups-agent,name=systemd,nsroot=/lxc/unpriv

   i.e. it has /lxc/unpriv as its root, and thus doesn't look rooted to CRIU.
   Let's allow cgroup mounts to be unrooted so we can deal with this.

Signed-off-by: Tycho Andersen <tycho.andersen at canonical.com>
---
 criu/cr-restore.c | 18 +++++++++---------
 criu/mount.c      |  5 ++++-
 2 files changed, 13 insertions(+), 10 deletions(-)

diff --git a/criu/cr-restore.c b/criu/cr-restore.c
index ffd2f01..967d5fa 100644
--- a/criu/cr-restore.c
+++ b/criu/cr-restore.c
@@ -1608,6 +1608,15 @@ static int restore_task_with_children(void *_arg)
 			goto err;
 	}
 
+	/*
+	 * Call this _before_ forking to optimize cgroups
+	 * restore -- if all tasks live in one set of cgroups
+	 * we will only move the root one there, others will
+	 * just have it inherited.
+	 */
+	if (prepare_task_cgroup(current) < 0)
+		goto err;
+
 	/* Restore root task */
 	if (current->parent == NULL) {
 		if (restore_finish_stage(CR_STATE_RESTORE_NS) < 0)
@@ -1640,15 +1649,6 @@ static int restore_task_with_children(void *_arg)
 	if (prepare_mappings())
 		goto err;
 
-	/*
-	 * Call this _before_ forking to optimize cgroups
-	 * restore -- if all tasks live in one set of cgroups
-	 * we will only move the root one there, others will
-	 * just have it inherited.
-	 */
-	if (prepare_task_cgroup(current) < 0)
-		goto err;
-
 	if (prepare_sigactions() < 0)
 		goto err;
 
diff --git a/criu/mount.c b/criu/mount.c
index eb8d058..542f1d6 100644
--- a/criu/mount.c
+++ b/criu/mount.c
@@ -734,7 +734,7 @@ static int validate_mounts(struct mount_info *info, bool for_dump)
 					 */
 					ret = m->need_plugin ? 0 : -ENOTSUP;
 
-				if (ret < 0) {
+				if (ret < 0 && m->fstype->code != FSTYPE__CGROUP) {
 					if (ret == -ENOTSUP)
 						pr_err("%d:%s doesn't have a proper root mount\n",
 								m->mnt_id, m->mountpoint);
@@ -2453,6 +2453,9 @@ static bool can_mount_now(struct mount_info *mi)
 	if (mi->external)
 		goto shared;
 
+	if (mi->fstype->code == FSTYPE__CGROUP)
+		return true;
+
 	/*
 	 * We're the slave peer:
 	 *   - Make sure the master peer is already mounted
-- 
2.7.3