Kernel test robot reported an "imbalanced put" in the rcuref_put() slow
path, which turned out to be a 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.
Prevent this by reusing the return value of the decrement. Now every "final"
put uses RCUREF_NOREF in the slow path and attempts the final cmpxchg() to
RCUREF_DEAD.
[ bigeasy: Add changelog ]
Fixes:
ee1ee6db07795 ("atomics: Provide rcuref - scalable reference counting")
Reported-by: kernel test robot <oliver.sang@intel.com>
Debugged-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Cc: stable@vger.kernel.org
Closes: https://lore.kernel.org/oe-lkp/
202412311453.
9d7636a2-lkp@intel.com
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);
}
/**
/**
* 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.
*
* 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)) {
/*