Home page logo
/

oss-sec logo oss-sec mailing list archives

Re: Linux kernel futex local privilege escalation (CVE-2014-3153)
From: John Johansen <john.johansen () canonical com>
Date: Thu, 05 Jun 2014 18:28:28 -0700

On 06/05/2014 04:25 PM, Phil Turnbull wrote:
On 05/06/14 10:45, Solar Designer wrote:
I've attached patches by Thomas Gleixner (four e-mails, in mbox format),
as well as back-ports of those by John Johansen of Canonical, who wrote:

---
For anyone who is interested I've attached back ports of the patches to

  3.13 - minor conflicts in patch 4. It has applied cleanly back to 3.2
and
  2.6.32 - conflict is in patches 3, and 4
---

Hi,

We are currently preparing Ksplice updates for our supported kernels and spotted
something odd in the 2.6.32 backport.

In patches-2.6.32.tgz:patches/0003-futex-Always-cleanup-owner-tid-in-unlock_pi.patch
there is this change (ignoring whitespace changes):

        curval = cmpxchg_futex_value_locked(uaddr, uval, newval);
-
-               if (curval == -EFAULT)
+       if (curval)
                ret = -EFAULT;

which seems to change the behaviour of the function.

The purpose of the return value of cmpxchg_futex_value_locked changed in

37a9d912b24f96a0591 "futex: Sanitize cmpxchg_futex_value_locked API"

which is not included in 2.6.32. This patch changes the return value to a
status code, but in 2.6.32 the return value is the value of the futex or
-EFAULT. With this backported patch, any futex with a non-zero value will
return -EFAULT.

Can you please clarify whether this change was intentional or is an error
introduced in the backport?

gah, that is indeed an error in the patch, but I could swear I remember
fixing this.

The corrected patch is below

---

From d6b2a11dab074c1a28b497525a61dc448d7073d2 Mon Sep 17 00:00:00 2001
From: Thomas Gleixner <tglx () linutronix de>
Date: Tue, 3 Jun 2014 12:27:07 +0000
Subject: [PATCH 3/4] futex: Always cleanup owner tid in unlock_pi

If the owner died bit is set at futex_unlock_pi, we currently do not
cleanup the user space futex. So the owner TID of the current owner
(the unlocker) persists. That's observable inconsistant state,
especially when the ownership of the pi state got transferred.

Clean it up unconditionally.

Backport to 2.6.32:
  conflicts: kernel/futex.c

Signed-off-by: Thomas Gleixner <tglx () linutronix de>
Signed-off-by: John Johansen <john.johansen () canonical com>
Cc: Kees Cook <keescook () chromium org>
Cc: Will Drewry <wad () chromium org>
Cc: Darren Hart <dvhart () linux intel com>
Cc: stable () vger kernel org
---
 kernel/futex.c | 43 +++++++++++++++++++------------------------
 1 file changed, 19 insertions(+), 24 deletions(-)

diff --git a/kernel/futex.c b/kernel/futex.c
index 179f08f..7be53af 100644
--- a/kernel/futex.c
+++ b/kernel/futex.c
@@ -811,6 +811,7 @@ static int wake_futex_pi(u32 __user *uaddr, u32 uval, struct futex_q *this)
        struct task_struct *new_owner;
        struct futex_pi_state *pi_state = this->pi_state;
        u32 curval, newval;
+       int ret = 0;
 
        if (!pi_state)
                return -EINVAL;
@@ -835,25 +836,20 @@ static int wake_futex_pi(u32 __user *uaddr, u32 uval, struct futex_q *this)
                new_owner = this->task;
 
        /*
-        * We pass it to the next owner. (The WAITERS bit is always
-        * kept enabled while there is PI state around. We must also
-        * preserve the owner died bit.)
+        * We pass it to the next owner. The WAITERS bit is always
+        * kept enabled while there is PI state around. We cleanup the
+        * owner died bit, because we are the owner.
         */
-       if (!(uval & FUTEX_OWNER_DIED)) {
-               int ret = 0;
-
-               newval = FUTEX_WAITERS | task_pid_vnr(new_owner);
-
-               curval = cmpxchg_futex_value_locked(uaddr, uval, newval);
+       newval = FUTEX_WAITERS | task_pid_vnr(new_owner);
 
-               if (curval == -EFAULT)
-                       ret = -EFAULT;
-               else if (curval != uval)
-                       ret = -EINVAL;
-               if (ret) {
-                       spin_unlock(&pi_state->pi_mutex.wait_lock);
-                       return ret;
-               }
+       curval = cmpxchg_futex_value_locked(uaddr, uval, newval);
+       if (curval == -EFAULT)
+               ret = -EFAULT;
+       else if (curval != uval)
+               ret = -EINVAL;
+       if (ret) {
+               spin_unlock(&pi_state->pi_mutex.wait_lock);
+               return ret;
        }
 
        spin_lock_irq(&pi_state->owner->pi_lock);
@@ -2107,9 +2103,10 @@ retry:
        /*
         * To avoid races, try to do the TID -> 0 atomic transition
         * again. If it succeeds then we can return without waking
-        * anyone else up:
+        * anyone else up. We only try this if neither the waiters nor
+        * the owner died bit are set.
         */
-       if (!(uval & FUTEX_OWNER_DIED))
+       if (!(uval & ~FUTEX_TID_MASK))
                uval = cmpxchg_futex_value_locked(uaddr, task_pid_vnr(current), 0);
 
 
@@ -2144,11 +2141,9 @@ retry:
        /*
         * No waiters - kernel unlocks the futex:
         */
-       if (!(uval & FUTEX_OWNER_DIED)) {
-               ret = unlock_futex_pi(uaddr, uval);
-               if (ret == -EFAULT)
-                       goto pi_faulted;
-       }
+       ret = unlock_futex_pi(uaddr, uval);
+       if (ret == -EFAULT)
+               goto pi_faulted;
 
 out_unlock:
        spin_unlock(&hb->lock);
-- 
2.0.0


  By Date           By Thread  

Current thread:
[ Nmap | Sec Tools | Mailing Lists | Site News | About/Contact | Advertising | Privacy ]
AlienVault