linux-kernel - [PATCH 2/3] LSM: SafeSetID: Add setgroups() security policy handling

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [day] [month] [year] [list]

Date:   Thu, 16 Jun 2022 10:18:35 -0700
From:   Micah Morton <mortonm@...omium.org>
To:     linux-security-module@...r.kernel.org
Cc:     keescook@...omium.org, jmorris@...ei.org, serge@...lyn.com,
        linux-kernel@...r.kernel.org, Micah Morton <mortonm@...omium.org>
Subject: [PATCH 2/3] LSM: SafeSetID: Add setgroups() security policy handling

The SafeSetID LSM has functionality for restricting setuid()/setgid()
syscalls based on its configured security policies. This patch adds the
analogous functionality for the setgroups() syscall. Security policy
for the setgroups() syscall follows the same policies that are
installed on the system for setgid() syscalls.

Signed-off-by: Micah Morton <mortonm@...omium.org>
---
 security/safesetid/lsm.c | 39 ++++++++++++++++++++++++++++++---------
 1 file changed, 30 insertions(+), 9 deletions(-)

diff --git a/security/safesetid/lsm.c b/security/safesetid/lsm.c
index 963f4ad9cb66..e806739f7868 100644
--- a/security/safesetid/lsm.c
+++ b/security/safesetid/lsm.c
@@ -97,15 +97,9 @@ static int safesetid_security_capable(const struct cred *cred,
 		return 0;
 
 	/*
-	 * If CAP_SET{U/G}ID is currently used for a setid() syscall, we want to
-	 * let it go through here; the real security check happens later, in the
-	 * task_fix_set{u/g}id hook.
-         *
-         * NOTE:
-         * Until we add support for restricting setgroups() calls, GID security
-         * policies offer no meaningful security since we always return 0 here
-         * when called from within the setgroups() syscall and there is no
-         * additional hook later on to enforce security policies for setgroups().
+	 * If CAP_SET{U/G}ID is currently used for a setid or setgroups syscall, we
+	 * want to let it go through here; the real security check happens later, in
+	 * the task_fix_set{u/g}id or task_fix_setgroups hooks.
 	 */
 	if ((opts & CAP_OPT_INSETID) != 0)
 		return 0;
@@ -241,9 +235,36 @@ static int safesetid_task_fix_setgid(struct cred *new,
 	return -EACCES;
 }
 
+static int safesetid_task_fix_setgroups(struct cred *new, const struct cred *old)
+{
+	int i;
+
+	/* Do nothing if there are no setgid restrictions for our old RGID. */
+	if (setid_policy_lookup((kid_t){.gid = old->gid}, INVALID_ID, GID) == SIDPOL_DEFAULT)
+		return 0;
+
+	get_group_info(new->group_info);
+	for (i = 0; i < new->group_info->ngroups; i++) {
+		if (!id_permitted_for_cred(old, (kid_t){.gid = new->group_info->gid[i]}, GID)) {
+			put_group_info(new->group_info);
+			/*
+			 * Kill this process to avoid potential security vulnerabilities
+			 * that could arise from a missing allowlist entry preventing a
+			 * privileged process from dropping to a lesser-privileged one.
+			 */
+			force_sig(SIGKILL);
+			return -EACCES;
+		}
+	}
+
+	put_group_info(new->group_info);
+	return 0;
+}
+
 static struct security_hook_list safesetid_security_hooks[] = {
 	LSM_HOOK_INIT(task_fix_setuid, safesetid_task_fix_setuid),
 	LSM_HOOK_INIT(task_fix_setgid, safesetid_task_fix_setgid),
+	LSM_HOOK_INIT(task_fix_setgroups, safesetid_task_fix_setgroups),
 	LSM_HOOK_INIT(capable, safesetid_security_capable)
 };
 
-- 
2.36.1.476.g0c4daa206d-goog