linux-kernel - [PATCH] clone3: fix cgroup argument sanity check

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>] [thread-next>] [day] [month] [year] [list]

Message-ID: <20200412202533.GA29554@asgard.redhat.com>
Date:   Sun, 12 Apr 2020 22:25:33 +0200
From:   Eugene Syromiatnikov <esyr@...hat.com>
To:     linux-kernel@...r.kernel.org,
        Christian Brauner <christian@...uner.io>
Cc:     Andrew Morton <akpm@...ux-foundation.org>,
        Ingo Molnar <mingo@...nel.org>, Tejun Heo <tj@...nel.org>,
        Oleg Nesterov <oleg@...hat.com>,
        Johannes Weiner <hannes@...xchg.org>,
        Li Zefan <lizefan@...wei.com>,
        Peter Zijlstra <peterz@...radead.org>, cgroups@...r.kernel.org,
        "Dmitry V. Levin" <ldv@...linux.org>
Subject: [PATCH] clone3: fix cgroup argument sanity check

Checking that cgroup field value of struct clone_args is less than 0
is useless, as it is defined as unsigned 64-bit integer.  Moreover,
it doesn't catch the situations where its higher bits are lost during
the assignment to the cgroup field of the cgroup field of the internal
struct kernel_clone_args (where it is declared as signed 32-bit
integer), so it is still possible to pass garbage there.  A check
against INT_MAX solves both these issues.

Fixes: ef2c41cf38a7559b ("clone3: allow spawning processes into cgroups")
Signed-off-by: Eugene Syromiatnikov <esyr@...hat.com>
---
 kernel/fork.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/kernel/fork.c b/kernel/fork.c
index 4385f3d..b4f7775 100644
--- a/kernel/fork.c
+++ b/kernel/fork.c
@@ -2631,7 +2631,7 @@ noinline static int copy_clone_args_from_user(struct kernel_clone_args *kargs,
 		     !valid_signal(args.exit_signal)))
 		return -EINVAL;
 
-	if ((args.flags & CLONE_INTO_CGROUP) && args.cgroup < 0)
+	if ((args.flags & CLONE_INTO_CGROUP) && args.cgroup > INT_MAX)
 		return -EINVAL;
 
 	*kargs = (struct kernel_clone_args){
-- 
2.1.4