[PATCH] sched_ext: Documentation: Clarify time slice handling in task lifecycle

[PATCH] sched_ext: Documentation: Clarify time slice handling in task lifecycle

[Andrea Righi]

It is not always obvious how a task's time slice can be refilled, either
explicitly from ops.dispatch() or automatically by the sched_ext core,
to skip subsequent ops.enqueue() and ops.dispatch() calls. This
typically happens when the task is the only one running on a CPU.

To make this behavior easier to understand, update the task lifecycle
diagram to explicitly document how time slice handling works in such
cases.

Signed-off-by: Andrea Righi <arighi@nvidia.com>
---
 Documentation/scheduler/sched-ext.rst | 11 ++++++++---
 1 file changed, 8 insertions(+), 3 deletions(-)

diff --git a/Documentation/scheduler/sched-ext.rst b/Documentation/scheduler/sched-ext.rst
index a1869c38046ed..404fe6126a769 100644
--- a/Documentation/scheduler/sched-ext.rst
+++ b/Documentation/scheduler/sched-ext.rst
@@ -313,16 +313,21 @@ by a sched_ext scheduler:
         ops.runnable();         /* Task becomes ready to run */
 
         while (task is runnable) {
-            if (task is not in a DSQ) {
+            if (task is not in a DSQ && task->scx.slice == 0) {
                 ops.enqueue();  /* Task can be added to a DSQ */
 
-                /* A CPU becomes available */
+                /* Any usable CPU becomes available */
 
                 ops.dispatch(); /* Task is moved to a local DSQ */
             }
             ops.running();      /* Task starts running on its assigned CPU */
-            ops.tick();         /* Called every 1/HZ seconds */
+            while (task->scx.slice > 0 && task is runnable)
+                ops.tick();     /* Called every 1/HZ seconds */
             ops.stopping();     /* Task stops running (time slice expires or wait) */
+
+            /* Task's CPU becomes available */
+
+            ops.dispatch();     /* task->scx.slice can be refilled */
         }
 
         ops.quiescent();        /* Task releases its assigned CPU (wait) */
-- 
2.49.0

Re: [PATCH] sched_ext: Documentation: Clarify time slice handling in task lifecycle

[Tejun Heo]

On Mon, Jun 09, 2025 at 06:23:42PM +0200, Andrea Righi wrote:
> It is not always obvious how a task's time slice can be refilled, either
> explicitly from ops.dispatch() or automatically by the sched_ext core,
> to skip subsequent ops.enqueue() and ops.dispatch() calls. This
> typically happens when the task is the only one running on a CPU.
> 
> To make this behavior easier to understand, update the task lifecycle
> diagram to explicitly document how time slice handling works in such
> cases.
> 
> Signed-off-by: Andrea Righi <arighi@nvidia.com>

Applied to sched_ext/for-6.17.

Thanks.

-- 
tejun