rcu: Prevent lockdep-RCU splats on lock acquisition/release
The rcu_cpu_starting() and rcu_report_dead() functions transition the
current CPU between online and offline state from an RCU perspective.
Unfortunately, this means that the rcu_cpu_starting() function's lock
acquisition and the rcu_report_dead() function's lock releases happen
while the CPU is offline from an RCU perspective, which can result in
lockdep-RCU splats about using RCU from an offline CPU. In reality,
aside from the splats, both transitions are safe because a new grace
period cannot start until these functions release their locks.
But the false-positive splats nevertheless need to be eliminated.
This commit therefore uses sequence-count-like synchronization to forgive
use of RCU while RCU thinks a CPU is offline across the full extent of
the rcu_cpu_starting() and rcu_report_dead() function's lock acquisitions
and releases.
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
diff --git a/kernel/rcu/tree.c b/kernel/rcu/tree.c
index 1d42909..5b06886 100644
--- a/kernel/rcu/tree.c
+++ b/kernel/rcu/tree.c
@@ -1152,13 +1152,15 @@
struct rcu_data *rdp;
struct rcu_node *rnp;
bool ret = false;
+ unsigned long seq;
if (in_nmi() || !rcu_scheduler_fully_active)
return true;
preempt_disable_notrace();
rdp = this_cpu_ptr(&rcu_data);
rnp = rdp->mynode;
- if (rdp->grpmask & rcu_rnp_online_cpus(rnp))
+ seq = READ_ONCE(rnp->ofl_seq) & ~0x1;
+ if (rdp->grpmask & rcu_rnp_online_cpus(rnp) || seq != READ_ONCE(rnp->ofl_seq))
ret = true;
preempt_enable_notrace();
return ret;
@@ -4065,6 +4067,8 @@
rnp = rdp->mynode;
mask = rdp->grpmask;
+ WRITE_ONCE(rnp->ofl_seq, rnp->ofl_seq + 1);
+ WARN_ON_ONCE(!(rnp->ofl_seq & 0x1));
raw_spin_lock_irqsave_rcu_node(rnp, flags);
WRITE_ONCE(rnp->qsmaskinitnext, rnp->qsmaskinitnext | mask);
newcpu = !(rnp->expmaskinitnext & mask);
@@ -4084,6 +4088,8 @@
} else {
raw_spin_unlock_irqrestore_rcu_node(rnp, flags);
}
+ WRITE_ONCE(rnp->ofl_seq, rnp->ofl_seq + 1);
+ WARN_ON_ONCE(rnp->ofl_seq & 0x1);
smp_mb(); /* Ensure RCU read-side usage follows above initialization. */
}
@@ -4111,6 +4117,8 @@
/* Remove outgoing CPU from mask in the leaf rcu_node structure. */
mask = rdp->grpmask;
+ WRITE_ONCE(rnp->ofl_seq, rnp->ofl_seq + 1);
+ WARN_ON_ONCE(!(rnp->ofl_seq & 0x1));
raw_spin_lock(&rcu_state.ofl_lock);
raw_spin_lock_irqsave_rcu_node(rnp, flags); /* Enforce GP memory-order guarantee. */
rdp->rcu_ofl_gp_seq = READ_ONCE(rcu_state.gp_seq);
@@ -4123,6 +4131,8 @@
WRITE_ONCE(rnp->qsmaskinitnext, rnp->qsmaskinitnext & ~mask);
raw_spin_unlock_irqrestore_rcu_node(rnp, flags);
raw_spin_unlock(&rcu_state.ofl_lock);
+ WRITE_ONCE(rnp->ofl_seq, rnp->ofl_seq + 1);
+ WARN_ON_ONCE(rnp->ofl_seq & 0x1);
rdp->cpu_started = false;
}
diff --git a/kernel/rcu/tree.h b/kernel/rcu/tree.h
index 805c9eb..7d802b6 100644
--- a/kernel/rcu/tree.h
+++ b/kernel/rcu/tree.h
@@ -57,6 +57,7 @@
/* beginning of each grace period. */
unsigned long qsmaskinitnext;
/* Online CPUs for next grace period. */
+ unsigned long ofl_seq; /* CPU-hotplug operation sequence count. */
unsigned long expmask; /* CPUs or groups that need to check in */
/* to allow the current expedited GP */
/* to complete. */