4.18-stable review patch. If anyone has any objections, please let me know.
------------------
From: Boqun Feng <boqun.feng@xxxxxxxxx>
commit fcc63543650150629c8a873cbef3578770acecd9 upstream.
Currently, the parallelized initialization of expedited grace periods uses
the workqueue associated with each rcu_node structure's ->grplo field.
This works fine unless that CPU is offline. This commit therefore uses
the CPU corresponding to the lowest-numbered online CPU, or just queues
the work on WORK_CPU_UNBOUND if there are no online CPUs corresponding
to this rcu_node structure.
Note that this patch uses cpu_is_offline() instead of the usual approach
of checking bits in the rcu_node structure's ->qsmaskinitnext field. This
is safe because preemption is disabled across both the cpu_is_offline()
check and the call to queue_work_on().
Signed-off-by: Boqun Feng <boqun.feng@xxxxxxxxx>
[ paulmck: Disable preemption to close offline race window. ]
Signed-off-by: Paul E. McKenney <paulmck@xxxxxxxxxxxxxxxxxx>
[ paulmck: Apply Peter Zijlstra feedback on CPU selection. ]
Tested-by: Aneesh Kumar K.V <aneesh.kumar@xxxxxxxxxxxxxxxxxx>
Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
---
kernel/rcu/tree_exp.h | 9 ++++++++-
1 file changed, 8 insertions(+), 1 deletion(-)
--- a/kernel/rcu/tree_exp.h
+++ b/kernel/rcu/tree_exp.h
@@ -472,6 +472,7 @@ retry_ipi:
static void sync_rcu_exp_select_cpus(struct rcu_state *rsp,
smp_call_func_t func)
{
+ int cpu;
struct rcu_node *rnp;
trace_rcu_exp_grace_period(rsp->name, rcu_exp_gp_seq_endval(rsp), TPS("reset"));
@@ -492,7 +493,13 @@ static void sync_rcu_exp_select_cpus(str
continue;
}
INIT_WORK(&rnp->rew.rew_work, sync_rcu_exp_select_node_cpus);
- queue_work_on(rnp->grplo, rcu_par_gp_wq, &rnp->rew.rew_work);
+ preempt_disable();
+ cpu = cpumask_next(rnp->grplo - 1, cpu_online_mask);
+ /* If all offline, queue the work on an unbound CPU. */
+ if (unlikely(cpu > rnp->grphi))
+ cpu = WORK_CPU_UNBOUND;
+ queue_work_on(cpu, rcu_par_gp_wq, &rnp->rew.rew_work);
+ preempt_enable();
rnp->exp_need_flush = true;
}
