Merge tag 'sched_ext-for-6.15' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/sched_ext

Pull sched_ext updates from Tejun Heo:

 - Add mechanism to count and report internal events. This significantly
   improves visibility on subtle corner conditions.

 - The default idle CPU selection logic is revamped and improved in
   multiple ways including being made topology aware.

 - sched_ext was disabling ttwu_queue for simplicity, which can be
   costly when hardware topology is more complex. Implement
   SCX_OPS_ALLOWED_QUEUED_WAKEUP so that BPF schedulers can selectively
   enable ttwu_queue.

 - tools/sched_ext updates to improve compatibility among others.

 - Other misc updates and fixes.

 - sched_ext/for-6.14-fixes were pulled a few times to receive
   prerequisite fixes and resolve conflicts.

* tag 'sched_ext-for-6.15' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/sched_ext: (42 commits)
  sched_ext: idle: Refactor scx_select_cpu_dfl()
  sched_ext: idle: Honor idle flags in the built-in idle selection policy
  sched_ext: Skip per-CPU tasks in scx_bpf_reenqueue_local()
  sched_ext: Add trace point to track sched_ext core events
  sched_ext: Change the event type from u64 to s64
  sched_ext: Documentation: add task lifecycle summary
  tools/sched_ext: Provide a compatible helper for scx_bpf_events()
  selftests/sched_ext: Add NUMA-aware scheduler test
  tools/sched_ext: Provide consistent access to scx flags
  sched_ext: idle: Fix scx_bpf_pick_any_cpu_node() behavior
  sched_ext: idle: Introduce scx_bpf_nr_node_ids()
  sched_ext: idle: Introduce node-aware idle cpu kfunc helpers
  sched_ext: idle: Per-node idle cpumasks
  sched_ext: idle: Introduce SCX_OPS_BUILTIN_IDLE_PER_NODE
  sched_ext: idle: Make idle static keys private
  sched/topology: Introduce for_each_node_numadist() iterator
  mm/numa: Introduce nearest_node_nodemask()
  nodemask: numa: reorganize inclusion path
  nodemask: add nodes_copy()
  tools/sched_ext: Sync with scx repo
  ...

1 files changed, 22 insertions, 1 deletions

diff --git a/tools/sched_ext/scx_qmap.bpf.c b/tools/sched_ext/scx_qmap.bpf.c
index 3a20bb0c014a..26c40ca4f36c 100644
--- a/tools/sched_ext/scx_qmap.bpf.c
+++ b/tools/sched_ext/scx_qmap.bpf.c
@@ -231,7 +231,7 @@ void BPF_STRUCT_OPS(qmap_enqueue, struct task_struct *p, u64 enq_flags)
 	}
 
 	/* if select_cpu() wasn't called, try direct dispatch */
-	if (!(enq_flags & SCX_ENQ_CPU_SELECTED) &&
+	if (!__COMPAT_is_enq_cpu_selected(enq_flags) &&
 	    (cpu = pick_direct_dispatch_cpu(p, scx_bpf_task_cpu(p))) >= 0) {
 		__sync_fetch_and_add(&nr_ddsp_from_enq, 1);
 		scx_bpf_dsq_insert(p, SCX_DSQ_LOCAL_ON | cpu, slice_ns, enq_flags);
@@ -763,6 +763,8 @@ static void dump_shared_dsq(void)
 
 static int monitor_timerfn(void *map, int *key, struct bpf_timer *timer)
 {
+	struct scx_event_stats events;
+
 	bpf_rcu_read_lock();
 	dispatch_highpri(true);
 	bpf_rcu_read_unlock();
@@ -772,6 +774,25 @@ static int monitor_timerfn(void *map, int *key, struct bpf_timer *timer)
 	if (print_shared_dsq)
 		dump_shared_dsq();
 
+	__COMPAT_scx_bpf_events(&events, sizeof(events));
+
+	bpf_printk("%35s: %lld", "SCX_EV_SELECT_CPU_FALLBACK",
+		   scx_read_event(&events, SCX_EV_SELECT_CPU_FALLBACK));
+	bpf_printk("%35s: %lld", "SCX_EV_DISPATCH_LOCAL_DSQ_OFFLINE",
+		   scx_read_event(&events, SCX_EV_DISPATCH_LOCAL_DSQ_OFFLINE));
+	bpf_printk("%35s: %lld", "SCX_EV_DISPATCH_KEEP_LAST",
+		   scx_read_event(&events, SCX_EV_DISPATCH_KEEP_LAST));
+	bpf_printk("%35s: %lld", "SCX_EV_ENQ_SKIP_EXITING",
+		   scx_read_event(&events, SCX_EV_ENQ_SKIP_EXITING));
+	bpf_printk("%35s: %lld", "SCX_EV_ENQ_SLICE_DFL",
+		   scx_read_event(&events, SCX_EV_ENQ_SLICE_DFL));
+	bpf_printk("%35s: %lld", "SCX_EV_BYPASS_DURATION",
+		   scx_read_event(&events, SCX_EV_BYPASS_DURATION));
+	bpf_printk("%35s: %lld", "SCX_EV_BYPASS_DISPATCH",
+		   scx_read_event(&events, SCX_EV_BYPASS_DISPATCH));
+	bpf_printk("%35s: %lld", "SCX_EV_BYPASS_ACTIVATE",
+		   scx_read_event(&events, SCX_EV_BYPASS_ACTIVATE));
+
 	bpf_timer_start(timer, ONE_SEC_IN_NS, 0);
 	return 0;
 }