[PATCH sched_ext/for-7.1] sched_ext: Pass full dequeue flags to ops.quiescent()

Andrea Righi posted 1 patch 1 month ago
kernel/sched/ext.c | 20 ++++++++++----------
1 file changed, 10 insertions(+), 10 deletions(-)
[PATCH sched_ext/for-7.1] sched_ext: Pass full dequeue flags to ops.quiescent()
Posted by Andrea Righi 1 month ago
ops.quiescent() is invoked with the same deq_flags as ops.dequeue(), so
the BPF scheduler is able to distinguish sleep vs property changes in
both callbacks.

However, dequeue_task_scx() receives deq_flags as an int from the
sched_class interface, so SCX flags above bit 32 (%SCX_DEQ_SCHED_CHANGE)
are truncated. ops_dequeue() reconstructs the full u64 for ops.dequeue(),
but ops.quiescent() is still called with the original int and can never
see %SCX_DEQ_SCHED_CHANGE.

Fix this by constructing the full u64 deq_flags in dequeue_task_scx()
(renaming the int parameter to core_deq_flags) and passing the complete
flags to both ops_dequeue() and ops.quiescent().

Fixes: ebf1ccff79c43 ("sched_ext: Fix ops.dequeue() semantics")
Signed-off-by: Andrea Righi <arighi@nvidia.com>
---
 kernel/sched/ext.c | 20 ++++++++++----------
 1 file changed, 10 insertions(+), 10 deletions(-)

diff --git a/kernel/sched/ext.c b/kernel/sched/ext.c
index 863400f91543e..75e38f44dd7ac 100644
--- a/kernel/sched/ext.c
+++ b/kernel/sched/ext.c
@@ -1900,14 +1900,6 @@ static void ops_dequeue(struct rq *rq, struct task_struct *p, u64 deq_flags)
 {
 	struct scx_sched *sch = scx_task_sched(p);
 	unsigned long opss;
-	u64 op_deq_flags = deq_flags;
-
-	/*
-	 * Set %SCX_DEQ_SCHED_CHANGE when the dequeue is due to a property
-	 * change (not sleep or core-sched pick).
-	 */
-	if (!(op_deq_flags & (DEQUEUE_SLEEP | SCX_DEQ_CORE_SCHED_EXEC)))
-		op_deq_flags |= SCX_DEQ_SCHED_CHANGE;
 
 	/* dequeue is always temporary, don't reset runnable_at */
 	clr_task_runnable(p, false);
@@ -1964,12 +1956,20 @@ static void ops_dequeue(struct rq *rq, struct task_struct *p, u64 deq_flags)
 	 * NONE but the task may still have %SCX_TASK_IN_CUSTODY set until
 	 * it is enqueued on the destination.
 	 */
-	call_task_dequeue(sch, rq, p, op_deq_flags);
+	call_task_dequeue(sch, rq, p, deq_flags);
 }
 
-static bool dequeue_task_scx(struct rq *rq, struct task_struct *p, int deq_flags)
+static bool dequeue_task_scx(struct rq *rq, struct task_struct *p, int core_deq_flags)
 {
 	struct scx_sched *sch = scx_task_sched(p);
+	u64 deq_flags = core_deq_flags;
+
+	/*
+	 * Set %SCX_DEQ_SCHED_CHANGE when the dequeue is due to a property
+	 * change (not sleep or core-sched pick).
+	 */
+	if (!(deq_flags & (DEQUEUE_SLEEP | SCX_DEQ_CORE_SCHED_EXEC)))
+		deq_flags |= SCX_DEQ_SCHED_CHANGE;
 
 	if (!(p->scx.flags & SCX_TASK_QUEUED)) {
 		WARN_ON_ONCE(task_runnable(p));
-- 
2.53.0
Re: [PATCH sched_ext/for-7.1] sched_ext: Pass full dequeue flags to ops.quiescent()
Posted by Tejun Heo 1 month ago
Applied to sched_ext/for-7.1 w/ Fixes tag hash fixed to 12 chars.

Thanks.

--
tejun