On Thu, May 14, 2026 at 09:33:10PM +0800, Z qiang wrote: > > > > From: "Paul E. McKenney" <[email protected]> > > > > Currently, rcutorture bypasses lazy RCU by using call_rcu_hurry(). > > This works, avoiding the dreaded rtort_pipe_count WARN(), but fails to > > fully test lazy RCU. The rtort_pipe_count WARN() splats because lazy RCU > > could delay the start of an RCU grace period for a full stutter period, > > which defaults to only three seconds. > > > > This commit therefore reverts the call_rcu_hurry() instances > > back to call_rcu(), but, in kernels built with CONFIG_RCU_LAZY=y, > > queues a workqueue handler just before the call to stutter_wait() in > > rcu_torture_writer(). This workqueue handler invokes rcu_barrier(), > > which motivates any lingering lazy callbacks, thus avoiding the splat. > > > > Questions for review: > > > > 1. Should we avoid queueing work for RCU implementations not > > supporting lazy callbacks? > > Hello, Paul > > maybe we can do this: > > rcu_ops = { > ... > .support_lazy = IS_ENABLED(CONFIG_RCU_LAZY), > }; > > and > > if (cur_ops->support_lazy ) > queue_work(..., &lazy_work); > > > > > 2. Should we avoid queueing work in kernels built with > > CONFIG_RCU_LAZY=y, but that were not booted with the > > rcutree.enable_rcu_lazy kernel boot parameter set? (Note that > > this requires some ugliness to access this parameter, and must > > also handle Tiny RCU.) > > > > 3. Does the rcu_torture_ops structure need a ->call_hurry() field, > > and if so, why? If not, why not? > > > > 4. Your additional questions here! > > > > Reported-by: Saravana Kannan <[email protected]> > > Signed-off-by: Paul E. McKenney <[email protected]> > > Signed-off-by: Uladzislau Rezki (Sony) <[email protected]> > > --- > > kernel/rcu/rcutorture.c | 21 ++++++++++++++++++--- > > 1 file changed, 18 insertions(+), 3 deletions(-) > > > > diff --git a/kernel/rcu/rcutorture.c b/kernel/rcu/rcutorture.c > > index 5f2848b828dc..91ba3160ba6a 100644 > > --- a/kernel/rcu/rcutorture.c > > +++ b/kernel/rcu/rcutorture.c > > @@ -572,7 +572,7 @@ static unsigned long rcu_no_completed(void) > > > > static void rcu_torture_deferred_free(struct rcu_torture *p) > > { > > - call_rcu_hurry(&p->rtort_rcu, rcu_torture_cb); > > + call_rcu(&p->rtort_rcu, rcu_torture_cb); > > } > > > > static void rcu_sync_torture_init(void) > > @@ -619,7 +619,7 @@ static struct rcu_torture_ops rcu_ops = { > > .poll_gp_state_exp = poll_state_synchronize_rcu, > > .cond_sync_exp = cond_synchronize_rcu_expedited, > > .cond_sync_exp_full = cond_synchronize_rcu_expedited_full, > > - .call = call_rcu_hurry, > > + .call = call_rcu, > > .cb_barrier = rcu_barrier, > > .fqs = rcu_force_quiescent_state, > > .gp_kthread_dbg = show_rcu_gp_kthreads, > > @@ -1145,7 +1145,7 @@ static void rcu_tasks_torture_deferred_free(struct > > rcu_torture *p) > > > > static void synchronize_rcu_mult_test(void) > > { > > - synchronize_rcu_mult(call_rcu_tasks, call_rcu_hurry); > > + synchronize_rcu_mult(call_rcu_tasks, call_rcu); > > } > > > > static struct rcu_torture_ops tasks_ops = { > > @@ -1631,6 +1631,17 @@ static void do_rtws_sync(struct torture_random_state > > *trsp, void (*sync)(void)) > > cpus_read_unlock(); > > } > > > > +/* > > + * Do an rcu_barrier() to motivate lazy callbacks during a stutter > > + * pause. Without this, we can get false-positives rtort_pipe_count > > + * splats. > > + */ > > +static void rcu_torture_writer_work(struct work_struct *work) > > +{ > > + if (cur_ops->cb_barrier) > > + cur_ops->cb_barrier(); > > +} > > + > > /* > > * RCU torture writer kthread. Repeatedly substitutes a new structure > > * for that pointed to by rcu_torture_current, freeing the old structure > > @@ -1651,6 +1662,7 @@ rcu_torture_writer(void *arg) > > int i; > > int idx; > > unsigned long j; > > + struct work_struct lazy_work; > > int oldnice = task_nice(current); > > struct rcu_gp_oldstate *rgo = NULL; > > int rgo_size = 0; > > @@ -1667,6 +1679,7 @@ rcu_torture_writer(void *arg) > > stallsdone += (stall_cpu_holdoff + stall_gp_kthread + > > stall_cpu + 60) * > > HZ * (stall_cpu_repeat + 1); > > VERBOSE_TOROUT_STRING("rcu_torture_writer task started"); > > + INIT_WORK_ONSTACK(&lazy_work, rcu_torture_writer_work); > > if (!can_expedite) > > pr_alert("%s" TORTURE_FLAG > > " GP expediting controlled from boot/sysfs for > > %s.\n", > > @@ -1895,6 +1908,8 @@ rcu_torture_writer(void *arg) > > !rcu_gp_is_normal(); > > } > > rcu_torture_writer_state = RTWS_STUTTER; > > + if (IS_ENABLED(CONFIG_RCU_LAZY)) > > + queue_work(system_percpu_wq, &lazy_work); > > > When the task ends, the lazy_work should be cancel and destroy: > > diff --git a/kernel/rcu/rcutorture.c b/kernel/rcu/rcutorture.c > index f593f8b794dd..5adf537ab410 100644 > --- a/kernel/rcu/rcutorture.c > +++ b/kernel/rcu/rcutorture.c > @@ -1682,7 +1682,6 @@ rcu_torture_writer(void *arg) > stallsdone += (stall_cpu_holdoff + stall_gp_kthread + stall_cpu + 60) * > HZ * (stall_cpu_repeat + 1); > VERBOSE_TOROUT_STRING("rcu_torture_writer task started"); > - INIT_WORK_ONSTACK(&lazy_work, rcu_torture_writer_work); > if (!can_expedite) > pr_alert("%s" TORTURE_FLAG > " GP expediting controlled from boot/sysfs for %s.\n", > @@ -1719,6 +1718,8 @@ rcu_torture_writer(void *arg) > pr_alert("%s" TORTURE_FLAG " Waited %lu jiffies for boot to complete.\n", > torture_type, jiffies - j); > > + INIT_WORK_ONSTACK(&lazy_work, rcu_torture_writer_work); > + > do { > rcu_torture_writer_state = RTWS_FIXED_DELAY; > torture_hrtimeout_us(500, 1000, &rand); > @@ -1943,6 +1944,9 @@ rcu_torture_writer(void *arg) > pr_alert("%s" TORTURE_FLAG > " Dynamic grace-period expediting was disabled.\n", > torture_type); > + if (IS_ENABLED(CONFIG_RCU_LAZY)) > + cancel_work_sync(&lazy_work); > + destroy_work_on_stack(&lazy_work); > kfree(ulo); > kfree(rgo); > rcu_torture_writer_state = RTWS_STOPPING; > I agree, we seem miss destroying the work via destroy_work_on_stack() and "sync" the lazy_work work via cancel_work_sync().
Paul, any thoughts? -- Uladzislau Rezki
