linux-kernel - [PATCH v7 01/15] rcuref: Avoid false positive "imbalanced put" report.

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

Message-ID: <20250123202446.610203-2-bigeasy@linutronix.de>
Date: Thu, 23 Jan 2025 21:24:31 +0100
From: Sebastian Andrzej Siewior <bigeasy@...utronix.de>
To: linux-kernel@...r.kernel.org
Cc: André Almeida <andrealmeid@...lia.com>,
	Darren Hart <dvhart@...radead.org>,
	Davidlohr Bueso <dave@...olabs.net>,
	Ingo Molnar <mingo@...hat.com>,
	Juri Lelli <juri.lelli@...hat.com>,
	Peter Zijlstra <peterz@...radead.org>,
	Thomas Gleixner <tglx@...utronix.de>,
	Valentin Schneider <vschneid@...hat.com>,
	Waiman Long <longman@...hat.com>,
	kernel test robot <oliver.sang@...el.com>,
	Sebastian Andrzej Siewior <bigeasy@...utronix.de>
Subject: [PATCH v7 01/15] rcuref: Avoid false positive "imbalanced put" report.

From: Thomas Gleixner <tglx@...utronix.de>

The kernel test robot reported an "imbalanced put" which turned out to
be false positive. Consider the following race:

            ref  = 0 (via rcuref_init(ref, 1))
 T1                                      T2
 rcuref_put(ref)
 -> atomic_add_negative_release(-1, ref)                                         # ref -> 0xffffffff
 -> rcuref_put_slowpath(ref)
                                         rcuref_get(ref)
                                         -> atomic_add_negative_relaxed(1, &ref->refcnt)
                                           -> return true;                       # ref -> 0

                                         rcuref_put(ref)
                                         -> atomic_add_negative_release(-1, ref) # ref -> 0xffffffff
                                         -> rcuref_put_slowpath()

    -> cnt = atomic_read(&ref->refcnt);                                          # cnt -> 0xffffffff / RCUREF_NOREF
    -> atomic_try_cmpxchg_release(&ref->refcnt, &cnt, RCUREF_DEAD))              # ref -> 0xe0000000 / RCUREF_DEAD
       -> return true
                                           -> cnt = atomic_read(&ref->refcnt);   # cnt -> 0xe0000000 / RCUREF_DEAD
                                           -> if (cnt > RCUREF_RELEASED)         # 0xe0000000 > 0xc0000000
                                             -> WARN_ONCE(cnt >= RCUREF_RELEASED, "rcuref - imbalanced put()")

The problem is the additional read in the slow path (after it
decremented to RCUREF_NOREF) which can happen after the counter has been
marked RCUREF_DEAD.

Avoid the false positive by reusing the returning value from the
decrement. Now every "final" put uses RCUREF_NOREF in the slow path and
attempts the final cmpxchg() to RCUREF_DEAD.

Fixes: ee1ee6db07795 ("atomics: Provide rcuref - scalable reference counting")
Reported-by: kernel test robot <oliver.sang@...el.com>
Closes: https://lore.kernel.org/oe-lkp/202412311453.9d7636a2-lkp@intel.com
Debugged-by: Sebastian Andrzej Siewior <bigeasy@...utronix.de>
Reviewed-by: Sebastian Andrzej Siewior <bigeasy@...utronix.de>
Signed-off-by: Thomas Gleixner <tglx@...utronix.de>
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@...utronix.de>
---
 include/linux/rcuref.h | 9 ++++++---
 lib/rcuref.c           | 5 ++---
 2 files changed, 8 insertions(+), 6 deletions(-)

diff --git a/include/linux/rcuref.h b/include/linux/rcuref.h
index 2c8bfd0f1b6b3..6322d8c1c6b42 100644
--- a/include/linux/rcuref.h
+++ b/include/linux/rcuref.h
@@ -71,27 +71,30 @@ static inline __must_check bool rcuref_get(rcuref_t *ref)
 	return rcuref_get_slowpath(ref);
 }
 
-extern __must_check bool rcuref_put_slowpath(rcuref_t *ref);
+extern __must_check bool rcuref_put_slowpath(rcuref_t *ref, unsigned int cnt);
 
 /*
  * Internal helper. Do not invoke directly.
  */
 static __always_inline __must_check bool __rcuref_put(rcuref_t *ref)
 {
+	int cnt;
+
 	RCU_LOCKDEP_WARN(!rcu_read_lock_held() && preemptible(),
 			 "suspicious rcuref_put_rcusafe() usage");
 	/*
 	 * Unconditionally decrease the reference count. The saturation and
 	 * dead zones provide enough tolerance for this.
 	 */
-	if (likely(!atomic_add_negative_release(-1, &ref->refcnt)))
+	cnt = atomic_sub_return_release(1, &ref->refcnt);
+	if (likely(cnt >= 0))
 		return false;
 
 	/*
 	 * Handle the last reference drop and cases inside the saturation
 	 * and dead zones.
 	 */
-	return rcuref_put_slowpath(ref);
+	return rcuref_put_slowpath(ref, cnt);
 }
 
 /**
diff --git a/lib/rcuref.c b/lib/rcuref.c
index 97f300eca927c..5bd726b71e393 100644
--- a/lib/rcuref.c
+++ b/lib/rcuref.c
@@ -220,6 +220,7 @@ EXPORT_SYMBOL_GPL(rcuref_get_slowpath);
 /**
  * rcuref_put_slowpath - Slowpath of __rcuref_put()
  * @ref:	Pointer to the reference count
+ * @cnt:	The resulting value of the fastpath decrement
  *
  * Invoked when the reference count is outside of the valid zone.
  *
@@ -233,10 +234,8 @@ EXPORT_SYMBOL_GPL(rcuref_get_slowpath);
  *	with a concurrent get()/put() pair. Caller is not allowed to
  *	deconstruct the protected object.
  */
-bool rcuref_put_slowpath(rcuref_t *ref)
+bool rcuref_put_slowpath(rcuref_t *ref, unsigned int cnt)
 {
-	unsigned int cnt = atomic_read(&ref->refcnt);
-
 	/* Did this drop the last reference? */
 	if (likely(cnt == RCUREF_NOREF)) {
 		/*
-- 
2.47.2