In RCU code, rcu_implicit_dynticks_qs() checks if a CPU is offline,
while being protected by a spinlock. At first, it appears as if we need to
use the get/put_online_cpus_atomic() APIs to properly synchronize with CPU
hotplug, once we get rid of stop_machine(). However, RCU has adequate
synchronization with CPU hotplug, making that unnecessary. But since the
locking details are non-trivial, it is hard to teach this to the rudimentary
hotplug locking validator.
So use the _nocheck() variants of the cpu accessor functions to prevent false-
positive warnings from the CPU hotplug debug code. Also, add a comment
explaining the hotplug synchronization design used in RCU, so that its easy
to see why it is justified to use the _nocheck() variants.
Cc: Dipankar Sarma <dipankar@xxxxxxxxxx>
Cc: "Paul E. McKenney" <paulmck@xxxxxxxxxxxxxxxxxx>
Signed-off-by: Srivatsa S. Bhat <srivatsa.bhat@xxxxxxxxxxxxxxxxxx>
---
kernel/rcutree.c | 12 +++++++++++-
1 file changed, 11 insertions(+), 1 deletion(-)
diff --git a/kernel/rcutree.c b/kernel/rcutree.c
index cf3adc6..ced28a45 100644
--- a/kernel/rcutree.c
+++ b/kernel/rcutree.c
@@ -794,7 +794,17 @@ static int rcu_implicit_dynticks_qs(struct rcu_data *rdp)
if (ULONG_CMP_GE(rdp->rsp->gp_start + 2, jiffies))
return 0; /* Grace period is not old enough. */
barrier();
- if (cpu_is_offline(rdp->cpu)) {
+
+ /*
+ * It is safe to use the _nocheck() version of cpu_is_offline() here
+ * (to avoid false-positive warnings from CPU hotplug debug code),
+ * because:
+ * 1. rcu_gp_init() holds off CPU hotplug operations during grace
+ * period initialization.
+ * 2. The current grace period has not ended yet.
+ * So it is safe to sample the offline state without synchronization.
+ */
+ if (cpu_is_offline_nocheck(rdp->cpu)) {
trace_rcu_fqs(rdp->rsp->name, rdp->gpnum, rdp->cpu, "ofl");
rdp->offline_fqs++;
return 1;
--
To unsubscribe from this list: send the line "unsubscribe linux-arch" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
