[PATCH v3 3/3] sched_ext: Fix cpu_released while changing sched policy of the running task

[liuwenfang <liuwenfang@honor.com>]

Priority inheritance policy should be cared, eg., one SCX task can be
boosted to REAL-TIME while holding a mutex lock, and then restored while
releasing the lock. The value cpu_released is fixed when changing the
sched class of the running task.

Signed-off-by: Wenfang Liu liuwenfang@honor.com
---
 kernel/sched/ext.c | 21 ++++++++++++++++++++-
 1 file changed, 20 insertions(+), 1 deletion(-)

diff --git a/kernel/sched/ext.c b/kernel/sched/ext.c
index ddf4bd060..e3b9032e2 100644
--- a/kernel/sched/ext.c
+++ b/kernel/sched/ext.c
@@ -3899,11 +3899,30 @@ static void switching_to_scx(struct rq *rq, struct task_struct *p)
 
 static void switched_from_scx(struct rq *rq, struct task_struct *p)
 {
+	if (task_current(rq, p))
+		switch_class(rq, p);
+
 	scx_ops_disable_task(p);
 }
 
 static void wakeup_preempt_scx(struct rq *rq, struct task_struct *p,int wake_flags) {}
-static void switched_to_scx(struct rq *rq, struct task_struct *p) {}
+static void switched_to_scx(struct rq *rq, struct task_struct *p)
+{
+	lockdep_assert_rq_held(rq);
+
+	if (static_branch_unlikely(&scx_ops_cpu_preempt) &&
+	    unlikely(rq->scx.cpu_released) && task_current(rq, p)) {
+		/*
+		 * If the previous sched_class for the current CPU was not SCX,
+		 * notify the BPF scheduler that it again has control of the
+		 * core. This callback complements ->cpu_release(), which is
+		 * emitted in switch_class().
+		 */
+		if (SCX_HAS_OP(cpu_acquire))
+			SCX_CALL_OP(SCX_KF_REST, cpu_acquire, rq, cpu_of(rq), NULL);
+		rq->scx.cpu_released = false;
+	}
+}
 
 int scx_check_setscheduler(struct task_struct *p, int policy)
 {
-- 
2.17.1