linux-kernel - [PATCH] user namespace: fix copy_user

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 for Android: free password hash cracker in your pocket

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

Message-ID: <20070719205757.GA15099@sergelap.austin.ibm.com>
Date:	Thu, 19 Jul 2007 15:57:57 -0500
From:	"Serge E. Hallyn" <serue@...ibm.com>
To:	Linus Torvalds <torvalds@...ux-foundation.org>
Cc:	lkml <linux-kernel@...r.kernel.org>
Subject: [PATCH] user namespace: fix copy_user_ns return value

As found by sukadev@...ibm.com, current -git, when compiled with
CONFIG_USER_NS=n, can be crashed by doing

#include <sched.h>
#include <sys/types.h>
#include <unistd.h>

int main()
{
	unshare(CLONE_NEWNS);
	setuid(1000);
}

as root.

The dummy version of copy_user_ns was returning NULL instead of
a valid namespace.  The following patch fixes the problem.
Please apply.

thanks,
-serge

>From 32f86740d27ef77160e438cd7dc4fcd5df159dd0 Mon Sep 17 00:00:00 2001
From: Serge E. Hallyn <serue@...ibm.com>
Date: Tue, 17 Jul 2007 15:28:17 -0400
Subject: [PATCH] user namespace: fix copy_user_ns return value

When a CONFIG_USER_NS=n and a user tries to unshare some
namespace other than the user namespace, the dummy
copy_user_ns returns NULL rather than the old_ns.  This
value then gets assigned to task->nsproxy->user_ns, so
that a subsequent setuid, which uses task->nsproxy->user_ns,
causes a NULL pointer deref.

Fix this by returning old_ns.

Signed-off-by: Serge E. Hallyn <serue@...ibm.com>
---
 include/linux/user_namespace.h |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/include/linux/user_namespace.h b/include/linux/user_namespace.h
index bb32057..1101b0c 100644
--- a/include/linux/user_namespace.h
+++ b/include/linux/user_namespace.h
@@ -49,7 +49,7 @@ static inline struct user_namespace *copy_user_ns(int flags,
 	if (flags & CLONE_NEWUSER)
 		return ERR_PTR(-EINVAL);

-	return NULL;
+	return old_ns;
 }

 static inline void put_user_ns(struct user_namespace *ns)
-- 
1.5.1.1.GIT

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/