Linux kernel mirror (for testing)
git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
kernel
os
linux
sched_ext: Documentation: Update sched-ext.rst
- Remove CONFIG_PAHOLE_HAS_BTF_TAG from required config list
- Document ext_idle.c as the built-in idle CPU selection policy
- Add descriptions for example schedulers in tools/sched_ext/
Signed-off-by: Cheng-Yang Chou <yphbchou0911@gmail.com>
Signed-off-by: Tejun Heo <tj@kernel.org>
authored by
Cheng-Yang Chou
and committed by
Tejun Heo
2a0596d5
1dde5025
+25
-2
+25
-2
Documentation/scheduler/sched-ext.rst
+25
-2
Documentation/scheduler/sched-ext.rst
···
43
43
CONFIG_DEBUG_INFO_BTF=y
44
44
CONFIG_BPF_JIT_ALWAYS_ON=y
45
45
CONFIG_BPF_JIT_DEFAULT_ON=y
46
-
CONFIG_PAHOLE_HAS_BTF_TAG=y
47
46
48
47
sched_ext is used only when the BPF scheduler is loaded and running.
49
48
···
345
346
The functions prefixed with ``scx_bpf_`` can be called from the BPF
346
347
scheduler.
347
348
349
+
* ``kernel/sched/ext_idle.c`` contains the built-in idle CPU selection policy.
350
+
348
351
* ``tools/sched_ext/`` hosts example BPF scheduler implementations.
349
352
350
353
* ``scx_simple[.bpf].c``: Minimal global FIFO scheduler example using a
···
355
354
* ``scx_qmap[.bpf].c``: A multi-level FIFO scheduler supporting five
356
355
levels of priority implemented with ``BPF_MAP_TYPE_QUEUE``.
357
356
357
+
* ``scx_central[.bpf].c``: A central FIFO scheduler where all scheduling
358
+
decisions are made on one CPU, demonstrating ``LOCAL_ON`` dispatching,
359
+
tickless operation, and kthread preemption.
360
+
361
+
* ``scx_cpu0[.bpf].c``: A scheduler that queues all tasks to a shared DSQ
362
+
and only dispatches them on CPU0 in FIFO order. Useful for testing bypass
363
+
behavior.
364
+
365
+
* ``scx_flatcg[.bpf].c``: A flattened cgroup hierarchy scheduler
366
+
implementing hierarchical weight-based cgroup CPU control by compounding
367
+
each cgroup's share at every level into a single flat scheduling layer.
368
+
369
+
* ``scx_pair[.bpf].c``: A core-scheduling example that always makes
370
+
sibling CPU pairs execute tasks from the same CPU cgroup.
371
+
372
+
* ``scx_sdt[.bpf].c``: A variation of ``scx_simple`` demonstrating BPF
373
+
arena memory management for per-task data.
374
+
375
+
* ``scx_userland[.bpf].c``: A minimal scheduler demonstrating user space
376
+
scheduling. Tasks with CPU affinity are direct-dispatched in FIFO order;
377
+
all others are scheduled in user space by a simple vruntime scheduler.
378
+
358
379
ABI Instability
359
380
===============
360
381
361
382
The APIs provided by sched_ext to BPF schedulers programs have no stability
362
383
guarantees. This includes the ops table callbacks and constants defined in
363
384
``include/linux/sched/ext.h``, as well as the ``scx_bpf_`` kfuncs defined in
364
-
``kernel/sched/ext.c``.
385
+
``kernel/sched/ext.c`` and ``kernel/sched/ext_idle.c``.
365
386
366
387
While we will attempt to provide a relatively stable API surface when
367
388
possible, they are subject to change without warning between kernel