aboutsummaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorTejun Heo <tj@kernel.org>2014-11-22 09:22:42 -0500
committerTejun Heo <tj@kernel.org>2014-11-23 12:36:06 -0500
commit4aab3b5b3ccf94fc907e66233e6ca4d8675759a6 (patch)
treec14d2a47ec7cd23e11525ae608c88433c392ed01
parentcac7f2429872d3733dc3f9915857b1691da2eb2f (diff)
percpu-ref: fix DEAD flag contamination of percpu pointer
While decoupling ATOMIC and DEAD flags, f47ad4578461 ("percpu_ref: decouple switching to percpu mode and reinit") updated __ref_is_percpu() so that it only tests ATOMIC flag to determine whether the ref is in percpu mode or not; however, while DEAD implies ATOMIC, the two flags are set separately during percpu_ref_kill() and if __ref_is_percpu() races percpu_ref_kill(), it may see DEAD w/o ATOMIC. Because __ref_is_percpu() returns @ref->percpu_count_ptr value verbatim as the percpu pointer after testing ATOMIC, the pointer may now be contaminated with the DEAD flag. This can be fixed by clearing the flag bits before returning the pointer which was the fix proposed by Shaohua; however, as DEAD implies ATOMIC, we can just test for both flags at once and avoid the explicit masking. Update __ref_is_percpu() so that it tests that both ATOMIC and DEAD are clear before returning @ref->percpu_count_ptr as the percpu pointer. Signed-off-by: Tejun Heo <tj@kernel.org> Reported-and-Reviewed-by: Shaohua Li <shli@kernel.org> Link: http://lkml.kernel.org/r/995deb699f5b873c45d667df4add3b06f73c2c25.1416638887.git.shli@kernel.org Fixes: f47ad4578461 ("percpu_ref: decouple switching to percpu mode and reinit")
-rw-r--r--include/linux/percpu-refcount.h8
1 files changed, 7 insertions, 1 deletions
diff --git a/include/linux/percpu-refcount.h b/include/linux/percpu-refcount.h
index d5c89e0dd0e6..51ce60c35f4c 100644
--- a/include/linux/percpu-refcount.h
+++ b/include/linux/percpu-refcount.h
@@ -133,7 +133,13 @@ static inline bool __ref_is_percpu(struct percpu_ref *ref,
/* paired with smp_store_release() in percpu_ref_reinit() */
smp_read_barrier_depends();
- if (unlikely(percpu_ptr & __PERCPU_REF_ATOMIC))
+ /*
+ * Theoretically, the following could test just ATOMIC; however,
+ * then we'd have to mask off DEAD separately as DEAD may be
+ * visible without ATOMIC if we race with percpu_ref_kill(). DEAD
+ * implies ATOMIC anyway. Test them together.
+ */
+ if (unlikely(percpu_ptr & __PERCPU_REF_ATOMIC_DEAD))
return false;
*percpu_countp = (unsigned long __percpu *)percpu_ptr;