Take user namespaces into account in policy checks

Change file related checks to use user namespaces and make
sure involved uids/gids are mappable in the current
namespace.

Note that checks without file ownership information will
still not take user namespaces into account, as some of
these should be handled via 'zfs allow' (otherwise root in a
user namespace could issue commands such as `zpool export`).

This also adds an initial user namespace regression test
for the setgid bit loss, with a user_ns_exec helper usable
in further tests.

Additionally, configure checks for the required user
namespace related features are added for:
  * ns_capable
  * kuid/kgid_has_mapping()
  * user_ns in cred_t

Reviewed-by: Brian Behlendorf <[email protected]>
Signed-off-by: Wolfgang Bumiller <[email protected]>
Closes #6800 
Closes #7270 

1 files changed, 59 insertions, 7 deletions

diff --git a/module/zfs/policy.c b/module/zfs/policy.c
index 03e8f748b..55c932747 100644
--- a/module/zfs/policy.c
+++ b/module/zfs/policy.c
@@ -42,19 +42,47 @@
  * all other cases this function must fail and return the passed err.
  */
 static int
-priv_policy(const cred_t *cr, int capability, boolean_t all, int err)
+priv_policy_ns(const cred_t *cr, int capability, boolean_t all, int err,
+    struct user_namespace *ns)
 {
 	ASSERT3S(all, ==, B_FALSE);
 
 	if (cr != CRED() && (cr != kcred))
 		return (err);
 
+#if defined(CONFIG_USER_NS) && defined(HAVE_NS_CAPABLE)
+	if (!(ns ? ns_capable(ns, capability) : capable(capability)))
+#else
 	if (!capable(capability))
+#endif
 		return (err);
 
 	return (0);
 }
 
+static int
+priv_policy(const cred_t *cr, int capability, boolean_t all, int err)
+{
+	return (priv_policy_ns(cr, capability, all, err, NULL));
+}
+
+static int
+priv_policy_user(const cred_t *cr, int capability, boolean_t all, int err)
+{
+	/*
+	 * All priv_policy_user checks are preceeded by kuid/kgid_has_mapping()
+	 * checks. If we cannot do them, we shouldn't be using ns_capable()
+	 * since we don't know whether the affected files are valid in our
+	 * namespace. Note that kuid_has_mapping() came after cred->user_ns, so
+	 * we shouldn't need to re-check for HAVE_CRED_USER_NS
+	 */
+#if defined(CONFIG_USER_NS) && defined(HAVE_KUID_HAS_MAPPING)
+	return (priv_policy_ns(cr, capability, all, err, cr->user_ns));
+#else
+	return (priv_policy_ns(cr, capability, all, err, NULL));
+#endif
+}
+
 /*
  * Checks for operations that are either client-only or are used by
  * both clients and servers.
@@ -102,10 +130,15 @@ secpolicy_vnode_any_access(const cred_t *cr, struct inode *ip, uid_t owner)
 	if (zpl_inode_owner_or_capable(ip))
 		return (0);
 
-	if (priv_policy(cr, CAP_DAC_OVERRIDE, B_FALSE, EPERM) == 0)
+#if defined(CONFIG_USER_NS) && defined(HAVE_KUID_HAS_MAPPING)
+	if (!kuid_has_mapping(cr->user_ns, SUID_TO_KUID(owner)))
+		return (EPERM);
+#endif
+
+	if (priv_policy_user(cr, CAP_DAC_OVERRIDE, B_FALSE, EPERM) == 0)
 		return (0);
 
-	if (priv_policy(cr, CAP_DAC_READ_SEARCH, B_FALSE, EPERM) == 0)
+	if (priv_policy_user(cr, CAP_DAC_READ_SEARCH, B_FALSE, EPERM) == 0)
 		return (0);
 
 	return (EPERM);
@@ -120,7 +153,12 @@ secpolicy_vnode_chown(const cred_t *cr, uid_t owner)
 	if (crgetfsuid(cr) == owner)
 		return (0);
 
-	return (priv_policy(cr, CAP_FOWNER, B_FALSE, EPERM));
+#if defined(CONFIG_USER_NS) && defined(HAVE_KUID_HAS_MAPPING)
+	if (!kuid_has_mapping(cr->user_ns, SUID_TO_KUID(owner)))
+		return (EPERM);
+#endif
+
+	return (priv_policy_user(cr, CAP_FOWNER, B_FALSE, EPERM));
 }
 
 /*
@@ -152,7 +190,12 @@ secpolicy_vnode_setdac(const cred_t *cr, uid_t owner)
 	if (crgetfsuid(cr) == owner)
 		return (0);
 
-	return (priv_policy(cr, CAP_FOWNER, B_FALSE, EPERM));
+#if defined(CONFIG_USER_NS) && defined(HAVE_KUID_HAS_MAPPING)
+	if (!kuid_has_mapping(cr->user_ns, SUID_TO_KUID(owner)))
+		return (EPERM);
+#endif
+
+	return (priv_policy_user(cr, CAP_FOWNER, B_FALSE, EPERM));
 }
 
 /*
@@ -175,8 +218,12 @@ secpolicy_vnode_setid_retain(const cred_t *cr, boolean_t issuidroot)
 int
 secpolicy_vnode_setids_setgids(const cred_t *cr, gid_t gid)
 {
+#if defined(CONFIG_USER_NS) && defined(HAVE_KUID_HAS_MAPPING)
+	if (!kgid_has_mapping(cr->user_ns, SGID_TO_KGID(gid)))
+		return (EPERM);
+#endif
 	if (crgetfsgid(cr) != gid && !groupmember(gid, cr))
-		return (priv_policy(cr, CAP_FSETID, B_FALSE, EPERM));
+		return (priv_policy_user(cr, CAP_FSETID, B_FALSE, EPERM));
 
 	return (0);
 }
@@ -222,7 +269,12 @@ secpolicy_vnode_setid_modify(const cred_t *cr, uid_t owner)
 	if (crgetfsuid(cr) == owner)
 		return (0);
 
-	return (priv_policy(cr, CAP_FSETID, B_FALSE, EPERM));
+#if defined(CONFIG_USER_NS) && defined(HAVE_KUID_HAS_MAPPING)
+	if (!kuid_has_mapping(cr->user_ns, SUID_TO_KUID(owner)))
+		return (EPERM);
+#endif
+
+	return (priv_policy_user(cr, CAP_FSETID, B_FALSE, EPERM));
 }
 
 /*