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
  ...

Author: torvalds

Documentation/scheduler/sched-ext.rst

@@ -294,6 +294,42 @@ dispatching, and must be dispatched to with ``scx_bpf_dsq_insert()``. See
 the function documentation and usage in ``tools/sched_ext/scx_simple.bpf.c``
 for more information.
 
+Task Lifecycle
+--------------
+
+The following pseudo-code summarizes the entire lifecycle of a task managed
+by a sched_ext scheduler:
+
+.. code-block:: c
+
+    ops.init_task();            /* A new task is created */
+    ops.enable();               /* Enable BPF scheduling for the task */
+
+    while (task in SCHED_EXT) {
+        if (task can migrate)
+            ops.select_cpu();   /* Called on wakeup (optimization) */
+
+        ops.runnable();         /* Task becomes ready to run */
+
+        while (task is runnable) {
+            if (task is not in a DSQ) {
+                ops.enqueue();  /* Task can be added to a DSQ */
+
+                /* A 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 */
+            ops.stopping();     /* Task stops running (time slice expires or wait) */
+        }
+
+        ops.quiescent();        /* Task releases its assigned CPU (wait) */
+    }
+
+    ops.disable();              /* Disable BPF scheduling for the task */
+    ops.exit_task();            /* Task is destroyed */
+
 Where to Look
 =============
 

MAINTAINERS

@@ -21196,8 +21196,7 @@ S:	Maintained
 W:	https://github.com/sched-ext/scx
 T:	git://git.kernel.org/pub/scm/linux/kernel/git/tj/sched_ext.git
 F:	include/linux/sched/ext.h
-F:	kernel/sched/ext.h
-F:	kernel/sched/ext.c
+F:	kernel/sched/ext*
 F:	tools/sched_ext/
 F:	tools/testing/selftests/sched_ext
 

include/linux/nodemask.h

@@ -94,7 +94,6 @@
 #include <linux/bitmap.h>
 #include <linux/minmax.h>
 #include <linux/nodemask_types.h>
-#include <linux/numa.h>
 #include <linux/random.h>
 
 extern nodemask_t _unused_nodemask_arg_;
@@ -191,6 +190,13 @@ static __always_inline void __nodes_andnot(nodemask_t *dstp, const nodemask_t *s
 	bitmap_andnot(dstp->bits, src1p->bits, src2p->bits, nbits);
 }
 
+#define nodes_copy(dst, src) __nodes_copy(&(dst), &(src), MAX_NUMNODES)
+static __always_inline void __nodes_copy(nodemask_t *dstp,
+					const nodemask_t *srcp, unsigned int nbits)
+{
+	bitmap_copy(dstp->bits, srcp->bits, nbits);
+}
+
 #define nodes_complement(dst, src) \
 			__nodes_complement(&(dst), &(src), MAX_NUMNODES)
 static __always_inline void __nodes_complement(nodemask_t *dstp,

include/linux/nodemask_types.h

@@ -3,7 +3,16 @@
 #define __LINUX_NODEMASK_TYPES_H
 
 #include <linux/bitops.h>
-#include <linux/numa.h>
+
+#ifdef CONFIG_NODES_SHIFT
+#define NODES_SHIFT     CONFIG_NODES_SHIFT
+#else
+#define NODES_SHIFT     0
+#endif
+
+#define MAX_NUMNODES    (1 << NODES_SHIFT)
+
+#define	NUMA_NO_NODE	(-1)
 
 typedef struct { DECLARE_BITMAP(bits, MAX_NUMNODES); } nodemask_t;
 

include/linux/numa.h

@@ -3,16 +3,8 @@
 #define _LINUX_NUMA_H
 #include <linux/init.h>
 #include <linux/types.h>
+#include <linux/nodemask.h>
 
-#ifdef CONFIG_NODES_SHIFT
-#define NODES_SHIFT     CONFIG_NODES_SHIFT
-#else
-#define NODES_SHIFT     0
-#endif
-
-#define MAX_NUMNODES    (1 << NODES_SHIFT)
-
-#define	NUMA_NO_NODE	(-1)
 #define	NUMA_NO_MEMBLK	(-1)
 
 static inline bool numa_valid_node(int nid)
@@ -39,6 +31,8 @@ void __init alloc_offline_node_data(int nid);
 /* Generic implementation available */
 int numa_nearest_node(int node, unsigned int state);
 
+int nearest_node_nodemask(int node, nodemask_t *mask);
+
 #ifndef memory_add_physaddr_to_nid
 int memory_add_physaddr_to_nid(u64 start);
 #endif
@@ -55,6 +49,11 @@ static inline int numa_nearest_node(int node, unsigned int state)
 	return NUMA_NO_NODE;
 }
 
+static inline int nearest_node_nodemask(int node, nodemask_t *mask)
+{
+	return NUMA_NO_NODE;
+}
+
 static inline int memory_add_physaddr_to_nid(u64 start)
 {
 	return 0;

include/linux/sched/ext.h

@@ -146,6 +146,7 @@ struct sched_ext_entity {
 	u32			weight;
 	s32			sticky_cpu;
 	s32			holding_cpu;
+	s32			selected_cpu;
 	u32			kf_mask;	/* see scx_kf_mask above */
 	struct task_struct	*kf_tasks[2];	/* see SCX_CALL_OP_TASK() */
 	atomic_long_t		ops_state;

include/linux/topology.h

@@ -261,6 +261,36 @@ sched_numa_hop_mask(unsigned int node, unsigned int hops)
 }
 #endif	/* CONFIG_NUMA */
 
+/**
+ * for_each_node_numadist() - iterate over nodes in increasing distance
+ *			      order, starting from a given node
+ * @node: the iteration variable and the starting node.
+ * @unvisited: a nodemask to keep track of the unvisited nodes.
+ *
+ * This macro iterates over NUMA node IDs in increasing distance from the
+ * starting @node and yields MAX_NUMNODES when all the nodes have been
+ * visited.
+ *
+ * Note that by the time the loop completes, the @unvisited nodemask will
+ * be fully cleared, unless the loop exits early.
+ *
+ * The difference between for_each_node() and for_each_node_numadist() is
+ * that the former allows to iterate over nodes in numerical order, whereas
+ * the latter iterates over nodes in increasing order of distance.
+ *
+ * This complexity of this iterator is O(N^2), where N represents the
+ * number of nodes, as each iteration involves scanning all nodes to
+ * find the one with the shortest distance.
+ *
+ * Requires rcu_lock to be held.
+ */
+#define for_each_node_numadist(node, unvisited)					\
+	for (int __start = (node),						\
+	     (node) = nearest_node_nodemask((__start), &(unvisited));		\
+	     (node) < MAX_NUMNODES;						\
+	     node_clear((node), (unvisited)),					\
+	     (node) = nearest_node_nodemask((__start), &(unvisited)))
+
 /**
  * for_each_numa_hop_mask - iterate over cpumasks of increasing NUMA distance
  *                          from a given node.

include/trace/events/sched_ext.h

@@ -26,6 +26,25 @@ TRACE_EVENT(sched_ext_dump,
 	)
 );
 
+TRACE_EVENT(sched_ext_event,
+	    TP_PROTO(const char *name, __s64 delta),
+	    TP_ARGS(name, delta),
+
+	TP_STRUCT__entry(
+		__string(name, name)
+		__field(	__s64,		delta		)
+	),
+
+	TP_fast_assign(
+		__assign_str(name);
+		__entry->delta		= delta;
+	),
+
+	TP_printk("name %s delta %lld",
+		  __get_str(name), __entry->delta
+	)
+);
+
 #endif /* _TRACE_SCHED_EXT_H */
 
 /* This part must be outside protection */

kernel/sched/build_policy.c

@@ -61,6 +61,7 @@
 
 #ifdef CONFIG_SCHED_CLASS_EXT
 # include "ext.c"
+# include "ext_idle.c"
 #endif
 
 #include "syscalls.c"

kernel/sched/core.c

@@ -3922,13 +3922,8 @@ bool cpus_share_resources(int this_cpu, int that_cpu)
 
 static inline bool ttwu_queue_cond(struct task_struct *p, int cpu)
 {
-	/*
-	 * The BPF scheduler may depend on select_task_rq() being invoked during
-	 * wakeups. In addition, @p may end up executing on a different CPU
-	 * regardless of what happens in the wakeup path making the ttwu_queue
-	 * optimization less meaningful. Skip if on SCX.
-	 */
-	if (task_on_scx(p))
+	/* See SCX_OPS_ALLOW_QUEUED_WAKEUP. */
+	if (!scx_allow_ttwu_queue(p))
 		return false;
 
 	/*