[PATCHSET sched/urgent, sched_ext/for-6.12-fixes] sched/core, sched_ext: Add ENQUEUE_RQ_SELECTED to fix ->select_task_rq() skip detection in sched_ext schedulers

[Tejun Heo <tj@kernel.org>]

During ttwu, ->select_task_rq() can be skipped if only one CPU is allowed or
migration is disabled. sched_ext schedulers may perform operations such as
direct dispatch from ->select_task_rq() path and it is useful for them to
know whether ->select_task_rq() was skipped in the ->enqueue_task() path.

Currently, sched_ext schedulers are using ENQUEUE_WAKEUP for this purpose
and end up assuming incorrectly that ->select_task_rq() was called for tasks
that are bound to a single CPU or migration disabled.

This patchset adds ENQUEUE_RQ_SELECTED to indicate whether
->select_task_rq() was called to ->enqueue_task() and propagate that to
sched_ext schedulers so that they can reliably detect whether
->select_task_rq() was skipped.

Peter, please let me know how 0001-0002 should be routed. If they get
applied to sched/urgent (when it opens), I'll pull it into
sched_ext/for-6.12-fixes and apply 0003 on top. If you'd prefer, I can take
all three through sched_ext/for-6.12-fixes too.

This patchset contains the following patches:

 0001-sched-core-Make-select_task_rq-take-the-pointer-to-w.patch
 0002-sched-core-Add-ENQUEUE_RQ_SELECTED-to-indicate-wheth.patch
 0003-sched_ext-scx_qmap-Add-and-use-SCX_ENQ_CPU_SELECTED.patch

 0001-0002 are sched/core patches to add ENQUEUE_RQ_SELECTED.

 0003 makes sched_ext propagate the flag as SCX_ENQ_CPU_SELECTED and fix
 scx_qmap using the new flag.

and can also be found in the following git branch:

 git://git.kernel.org/pub/scm/linux/kernel/git/tj/sched_ext.git scx-ENQUEUE_RQ_SELECTED

diffstat follows. Thanks.

 kernel/sched/core.c            |   21 ++++++++++++++-------
 kernel/sched/ext.c             |    1 +
 kernel/sched/sched.h           |    3 +++
 tools/sched_ext/scx_qmap.bpf.c |    4 ++--
 4 files changed, 20 insertions(+), 9 deletions(-)

--
tejun