[PATCHSET sched_ext/for-7.1] sched_ext: Overhaul DSQ reenqueue infrastructure

[Tejun Heo]

scx_bpf_reenqueue_local() currently only supports the current CPU's local
DSQ and must be called from ops.cpu_release(). This patchset overhauls the
reenqueue infrastructure to support flexible, remote, and user DSQ reenqueue
operations through the new scx_bpf_dsq_reenq() kfunc.

The patchset:

- Refactors per-node data structures and DSQ lookup helpers as preparation.

- Converts the deferred reenqueue mechanism from lockless llist to a
  spinlock-protected regular list to support more complex list operations.

- Introduces scx_bpf_dsq_reenq() which can target any local DSQ including
  remote CPUs via SCX_DSQ_LOCAL_ON | cpu, and user-defined DSQs.

- Adds per-CPU data to DSQs and cursor-based iteration helpers to support
  user DSQ reenqueue with proper multi-rq locking.

- Adds reenqueue flags plumbing and a lockless fast-path optimization.

- Simplifies task state handling and adds SCX_TASK_REENQ_REASON flags so BPF
  schedulers can distinguish why a task is being reenqueued.

scx_bpf_reenqueue_local() is reimplemented as a wrapper around
scx_bpf_dsq_reenq(SCX_DSQ_LOCAL, 0) and may be deprecated in the future.

Based on sched_ext/for-7.1 (4f8b122848db).

Tejun Heo (15):
  sched_ext: Relocate scx_bpf_task_cgroup() and its BTF_ID to the end of kfunc section
  sched_ext: Wrap global DSQs in per-node structure
  sched_ext: Factor out pnode allocation and deallocation into helpers
  sched_ext: Change find_global_dsq() to take CPU number instead of task
  sched_ext: Relocate reenq_local() and run_deferred()
  sched_ext: Convert deferred_reenq_locals from llist to regular list
  sched_ext: Wrap deferred_reenq_local_node into a struct
  sched_ext: Introduce scx_bpf_dsq_reenq() for remote local DSQ reenqueue
  sched_ext: Add reenq_flags plumbing to scx_bpf_dsq_reenq()
  sched_ext: Add per-CPU data to DSQs
  sched_ext: Factor out nldsq_cursor_next_task() and nldsq_cursor_lost_task()
  sched_ext: Implement scx_bpf_dsq_reenq() for user DSQs
  sched_ext: Optimize schedule_dsq_reenq() with lockless fast path
  sched_ext: Simplify task state handling
  sched_ext: Add SCX_TASK_REENQ_REASON flags

Git tree:
  git://git.kernel.org/pub/scm/linux/kernel/git/tj/sched_ext.git scx-reenq

 include/linux/sched/ext.h                |  58 +-
 kernel/sched/ext.c                       | 882 ++++++++++++++++++++++---------
 kernel/sched/ext_internal.h              |  32 +-
 kernel/sched/sched.h                     |   4 +-
 tools/sched_ext/include/scx/compat.bpf.h |  21 +
 tools/sched_ext/scx_qmap.bpf.c           |  68 ++-
 tools/sched_ext/scx_qmap.c               |   5 +-
 7 files changed, 777 insertions(+), 293 deletions(-)

--
tejun