[PATCH v10 29/40] sched/deadline: Start a period at a deboost out of the class
From: Mathieu Desnoyers
Date: Fri Nov 20 2026 - 15:46:05 EST
A task that leaves SCHED_DEADLINE keeps its runtime and deadline until
its 0-lag time. Commit a649f237db18 ("sched,dl: Fix sched class hopping
CBS hole") does this, so that leaving the class and coming back gains
nothing. A priority inheritance boost also keeps the runtime and
deadline of a task with parameters of its own.
But a boosted task is not throttled. When it runs out of runtime,
replenish_dl_entity() refills it from the donor's parameters. If the
task stays in the class at its deboost, "sched/deadline: Reset the
donor's runtime at a deboost in the class" starts a new period.
A task that leaves the class at its deboost is missed. switched_from_dl()
calls task_non_contending(), which computes the 0-lag time from the
runtime and deadline of the donor's refills, with the task's own
dl_runtime and dl_period. The root domain then keeps the task's
bandwidth in total_bw for too long, or releases it too early. Until it
is released, deadline tasks that would fit are refused.
Fix this by starting a period of its own in switched_from_dl(), as
dl_deboost() does. Share that code in dl_start_own_period(). The 0-lag
time of the new period is now, or has passed, so the bandwidth is
released at once. task_non_contending() now also releases it when the
0-lag time is exactly now, instead of arming a timer of 0ns.
A task can also block while boosted, after a refill. Its dequeue then
arms the inactive timer from the same mix. If the boost ends while the
task sleeps, switched_from_dl() finds that timer armed, and keeps it.
Fix this too: cancel the timer there, and compute the 0-lag time of the
new period instead.
The Sashiko AI review found this, on "sched/deadline: Reset the
deadline and the runtime with the parameters" of the first series. The
deboost_out_after_donor_refill case, added to deadline_bw_test later in
this series, tests it. Without this patch, the owner's bandwidth was
still in total_bw right after the unlock, in every run. The
deboost_out_asleep_after_refill case tests an owner asleep when the
boost ends. Without the cancel of the timer, its bandwidth was still in
total_bw after the boost ended, in every run.
Fixes: 2d3d891d3344 ("sched/deadline: Add SCHED_DEADLINE inheritance logic")
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@xxxxxxxxxxxx>
Cc: stable@xxxxxxxxxxxxxxx
---
kernel/sched/deadline.c | 49 +++++++++++++++++++++++++++++++++++------
1 file changed, 42 insertions(+), 7 deletions(-)
diff --git a/kernel/sched/deadline.c b/kernel/sched/deadline.c
index ee3a08ff3345..b1346be9e81d 100644
--- a/kernel/sched/deadline.c
+++ b/kernel/sched/deadline.c
@@ -83,6 +83,16 @@ static inline int on_dl_rq(struct sched_dl_entity *dl_se)
return !RB_EMPTY_NODE(&dl_se->rb_node);
}
+/*
+ * Start a period of the reservation of @dl_se now. The caller holds the rq
+ * lock and has updated the rq clock.
+ */
+static inline void dl_start_own_period(struct sched_dl_entity *dl_se)
+{
+ dl_se->deadline = rq_clock(rq_of_dl_se(dl_se)) + dl_se->dl_deadline;
+ dl_se->runtime = dl_se->dl_runtime;
+}
+
#ifdef CONFIG_RT_MUTEXES
static inline struct sched_dl_entity *pi_of(struct sched_dl_entity *dl_se)
{
@@ -119,11 +129,8 @@ static void task_non_contending(struct sched_dl_entity *dl_se, bool dl_task);
*/
void dl_deboost(struct sched_dl_entity *dl_se)
{
- struct rq *rq = rq_of_dl_se(dl_se);
-
if (dl_se->dl_donor_refill) {
- dl_se->deadline = rq_clock(rq) + dl_se->dl_deadline;
- dl_se->runtime = dl_se->dl_runtime;
+ dl_start_own_period(dl_se);
/*
* A task asleep since the refill has its inactive timer armed
* from the donor's runtime and deadline. The 0-lag time of the
@@ -826,10 +833,12 @@ static void task_non_contending(struct sched_dl_entity *dl_se, bool dl_task)
zerolag_time = dl_time_to_zerolag(dl_se, rq);
/*
- * If the "0-lag time" already passed, decrease the active
- * utilization now, instead of starting a timer
+ * If the "0-lag time" is now or already passed, decrease the active
+ * utilization now, instead of starting a timer. A task that leaves
+ * the class with a period of its own just started has a 0-lag time
+ * of now.
*/
- if ((zerolag_time < 0) || hrtimer_active(&dl_se->inactive_timer)) {
+ if ((zerolag_time <= 0) || hrtimer_active(&dl_se->inactive_timer)) {
if (dl_server(dl_se)) {
sub_running_bw(dl_se, dl_rq);
} else {
@@ -4662,6 +4671,19 @@ void dl_clear_root_domains(void)
static void switched_from_dl(struct rq *rq, struct task_struct *p)
{
+ bool own_period = p->dl.dl_runtime && p->dl.dl_donor_refill;
+
+ /*
+ * The boost of a task with parameters of its own ends, and the task
+ * leaves the class. If the boost refilled its runtime from the
+ * donor's parameters, its runtime and deadline are not those of its
+ * reservation. task_non_contending() would mix them with its own
+ * dl_runtime and dl_period. Start a period of its own instead, as
+ * dl_deboost() does. Its 0-lag time is then now, or has passed.
+ */
+ if (own_period)
+ dl_start_own_period(&p->dl);
+
/*
* task_non_contending() can start the "inactive timer" (if the 0-lag
* time is in the future). If the task switches back to dl before
@@ -4693,6 +4715,19 @@ static void switched_from_dl(struct rq *rq, struct task_struct *p)
if (p->dl.dl_non_contending)
sub_running_bw(&p->dl, &rq->dl);
sub_rq_bw(&p->dl, &rq->dl);
+ /*
+ * The task blocked while boosted, after a refill from the
+ * donor. The timer armed then mixes the donor's runtime and
+ * deadline with the task's own parameters. Cancel it, and
+ * work out the 0-lag time of the period just started below.
+ * If the timer handler is already running, it releases the
+ * bandwidth.
+ */
+ if (own_period && p->dl.dl_non_contending &&
+ hrtimer_try_to_cancel(&p->dl.inactive_timer) == 1) {
+ p->dl.dl_non_contending = 0;
+ put_task_struct(p);
+ }
/*
* Without an armed timer, nothing releases total_bw. Do as for
* a queued task: arm the timer if the 0-lag time is ahead, or
--
2.47.3