cgroup-v1: Correct privileges check in release_agent writes

BugLink: https://bugs.launchpad.net/bugs/1963891
commit 467a726b754f474936980da793b4ff2ec3e382a7 upstream.

The idea is to check: a) the owning user_ns of cgroup_ns, b)
capabilities in init_user_ns.

The commit 24f600856418 ("cgroup-v1: Require capabilities to set
release_agent") got this wrong in the write handler of release_agent
since it checked user_ns of the opener (may be different from the owning
user_ns of cgroup_ns).
Secondly, to avoid possibly confused deputy, the capability of the
opener must be checked.

Fixes: 24f600856418 ("cgroup-v1: Require capabilities to set release_agent")
Cc: stable@vger.kernel.org
Link: https://lore.kernel.org/stable/20220216121142.GB30035@blackbody.suse.cz/
Signed-off-by: Michal Koutný <mkoutny@suse.com>
Reviewed-by: Masami Ichikawa(CIP) <masami.ichikawa@cybertrust.co.jp>
Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Paolo Pisati <paolo.pisati@canonical.com>

diff --git a/kernel/cgroup/cgroup-v1.c b/kernel/cgroup/cgroup-v1.c
index c59aa2c7749b2b0999942ab819b2e47319a2bdd0..58900dc92ac98cf818ff80b8b61283f345ee4028 100644 (file)
--- a/kernel/cgroup/cgroup-v1.c
+++ b/kernel/cgroup/cgroup-v1.c
@@ -549,6 +549,7 @@ static ssize_t cgroup_release_agent_write(struct kernfs_open_file *of,
                                          char *buf, size_t nbytes, loff_t off)
 {
        struct cgroup *cgrp;
+       struct cgroup_file_ctx *ctx;
 
        BUILD_BUG_ON(sizeof(cgrp->root->release_agent_path) < PATH_MAX);
 
@@ -556,8 +557,9 @@ static ssize_t cgroup_release_agent_write(struct kernfs_open_file *of,
         * Release agent gets called with all capabilities,
         * require capabilities to set release agent.
         */
-       if ((of->file->f_cred->user_ns != &init_user_ns) ||
-           !capable(CAP_SYS_ADMIN))
+       ctx = of->priv;
+       if ((ctx->ns->user_ns != &init_user_ns) ||
+           !file_ns_capable(of->file, &init_user_ns, CAP_SYS_ADMIN))
                return -EPERM;
 
        cgrp = cgroup_kn_lock_live(of->kn, false);