sched_ext: Fixes for v6.12-rc7

- The fair sched class currently has a bug where its balance() returns true
   telling the sched core that it has tasks to run but then NULL from
   pick_task(). This makes sched core call sched_ext's pick_task() without
   preceding balance() which can lead to stalls in partial mode. For now,
   work around by detecting the condition and forcing the CPU to go through
   another scheduling cycle.
 
 - Add a missing newline to an error message and fix drgn introspection tool
   which went out of sync.
 -----BEGIN PGP SIGNATURE-----
 
 iIQEABYKACwWIQTfIjM1kS57o3GsC/uxYfJx3gVYGQUCZzI8sw4cdGpAa2VybmVs
 Lm9yZwAKCRCxYfJx3gVYGb5KAP40b/o6TyAFDG+Hn6GxyxQT7rcAUMXsdB2bcEpg
 /IjmzQEAwbHU5KP5vQXV6XHv+2V7Rs7u6ZqFtDnL88N0A9hf3wk=
 =7hL8
 -----END PGP SIGNATURE-----

Merge tag 'sched_ext-for-6.12-rc7-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/sched_ext

Pull sched_ext fixes from Tejun Heo:

 - The fair sched class currently has a bug where its balance() returns
   true telling the sched core that it has tasks to run but then NULL
   from pick_task(). This makes sched core call sched_ext's pick_task()
   without preceding balance() which can lead to stalls in partial mode.

   For now, work around by detecting the condition and forcing the CPU
   to go through another scheduling cycle.

 - Add a missing newline to an error message and fix drgn introspection
   tool which went out of sync.

* tag 'sched_ext-for-6.12-rc7-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/sched_ext:
  sched_ext: Handle cases where pick_task_scx() is called without preceding balance_scx()
  sched_ext: Update scx_show_state.py to match scx_ops_bypass_depth's new type
  sched_ext: Add a missing newline at the end of an error message

kernel/sched/core.c

@@ -5920,12 +5920,15 @@ static void prev_balance(struct rq *rq, struct task_struct *prev,
 
 #ifdef CONFIG_SCHED_CLASS_EXT
 	/*
-	 * SCX requires a balance() call before every pick_next_task() including
-	 * when waking up from SCHED_IDLE. If @start_class is below SCX, start
-	 * from SCX instead.
+	 * SCX requires a balance() call before every pick_task() including when
+	 * waking up from SCHED_IDLE. If @start_class is below SCX, start from
+	 * SCX instead. Also, set a flag to detect missing balance() call.
 	 */
-	if (scx_enabled() && sched_class_above(&ext_sched_class, start_class))
-		start_class = &ext_sched_class;
+	if (scx_enabled()) {
+		rq->scx.flags |= SCX_RQ_BAL_PENDING;
+		if (sched_class_above(&ext_sched_class, start_class))
+			start_class = &ext_sched_class;
+	}
 #endif
 
 	/*

kernel/sched/ext.c

@@ -2634,7 +2634,7 @@ static int balance_one(struct rq *rq, struct task_struct *prev)
 
 	lockdep_assert_rq_held(rq);
 	rq->scx.flags |= SCX_RQ_IN_BALANCE;
-	rq->scx.flags &= ~SCX_RQ_BAL_KEEP;
+	rq->scx.flags &= ~(SCX_RQ_BAL_PENDING | SCX_RQ_BAL_KEEP);
 
 	if (static_branch_unlikely(&scx_ops_cpu_preempt) &&
 	    unlikely(rq->scx.cpu_released)) {
@@ -2948,12 +2948,11 @@ static struct task_struct *pick_task_scx(struct rq *rq)
 {
 	struct task_struct *prev = rq->curr;
 	struct task_struct *p;
+	bool prev_on_scx = prev->sched_class == &ext_sched_class;
+	bool keep_prev = rq->scx.flags & SCX_RQ_BAL_KEEP;
+	bool kick_idle = false;
 
 	/*
-	 * If balance_scx() is telling us to keep running @prev, replenish slice
-	 * if necessary and keep running @prev. Otherwise, pop the first one
-	 * from the local DSQ.
-	 *
 	 * WORKAROUND:
 	 *
 	 * %SCX_RQ_BAL_KEEP should be set iff $prev is on SCX as it must just
@@ -2962,22 +2961,41 @@ static struct task_struct *pick_task_scx(struct rq *rq)
 	 * which then ends up calling pick_task_scx() without preceding
 	 * balance_scx().
 	 *
-	 * For now, ignore cases where $prev is not on SCX. This isn't great and
-	 * can theoretically lead to stalls. However, for switch_all cases, this
-	 * happens only while a BPF scheduler is being loaded or unloaded, and,
-	 * for partial cases, fair will likely keep triggering this CPU.
+	 * Keep running @prev if possible and avoid stalling from entering idle
+	 * without balancing.
 	 *
-	 * Once fair is fixed, restore WARN_ON_ONCE().
+	 * Once fair is fixed, remove the workaround and trigger WARN_ON_ONCE()
+	 * if pick_task_scx() is called without preceding balance_scx().
 	 */
-	if ((rq->scx.flags & SCX_RQ_BAL_KEEP) &&
-	    prev->sched_class == &ext_sched_class) {
+	if (unlikely(rq->scx.flags & SCX_RQ_BAL_PENDING)) {
+		if (prev_on_scx) {
+			keep_prev = true;
+		} else {
+			keep_prev = false;
+			kick_idle = true;
+		}
+	} else if (unlikely(keep_prev && !prev_on_scx)) {
+		/* only allowed during transitions */
+		WARN_ON_ONCE(scx_ops_enable_state() == SCX_OPS_ENABLED);
+		keep_prev = false;
+	}
+
+	/*
+	 * If balance_scx() is telling us to keep running @prev, replenish slice
+	 * if necessary and keep running @prev. Otherwise, pop the first one
+	 * from the local DSQ.
+	 */
+	if (keep_prev) {
 		p = prev;
 		if (!p->scx.slice)
 			p->scx.slice = SCX_SLICE_DFL;
 	} else {
 		p = first_local_task(rq);
-		if (!p)
+		if (!p) {
+			if (kick_idle)
+				scx_bpf_kick_cpu(cpu_of(rq), SCX_KICK_IDLE);
 			return NULL;
+		}
 
 		if (unlikely(!p->scx.slice)) {
 			if (!scx_rq_bypassing(rq) && !scx_warned_zero_slice) {
@@ -4979,7 +4997,7 @@ static int scx_ops_enable(struct sched_ext_ops *ops, struct bpf_link *link)
 
 	if (!cpumask_equal(housekeeping_cpumask(HK_TYPE_DOMAIN),
 			   cpu_possible_mask)) {
-		pr_err("sched_ext: Not compatible with \"isolcpus=\" domain isolation");
+		pr_err("sched_ext: Not compatible with \"isolcpus=\" domain isolation\n");
 		return -EINVAL;
 	}
 

kernel/sched/sched.h

@@ -751,8 +751,9 @@ enum scx_rq_flags {
 	 */
 	SCX_RQ_ONLINE		= 1 << 0,
 	SCX_RQ_CAN_STOP_TICK	= 1 << 1,
-	SCX_RQ_BAL_KEEP		= 1 << 2, /* balance decided to keep current */
-	SCX_RQ_BYPASSING	= 1 << 3,
+	SCX_RQ_BAL_PENDING	= 1 << 2, /* balance hasn't run yet */
+	SCX_RQ_BAL_KEEP		= 1 << 3, /* balance decided to keep current */
+	SCX_RQ_BYPASSING	= 1 << 4,
 
 	SCX_RQ_IN_WAKEUP	= 1 << 16,
 	SCX_RQ_IN_BALANCE	= 1 << 17,

tools/sched_ext/scx_show_state.py

@@ -35,6 +35,6 @@ print(f'enabled       : {read_static_key("__scx_ops_enabled")}')
 print(f'switching_all : {read_int("scx_switching_all")}')
 print(f'switched_all  : {read_static_key("__scx_switched_all")}')
 print(f'enable_state  : {ops_state_str(enable_state)} ({enable_state})')
-print(f'bypass_depth  : {read_atomic("scx_ops_bypass_depth")}')
+print(f'bypass_depth  : {prog["scx_ops_bypass_depth"].value_()}')
 print(f'nr_rejected   : {read_atomic("scx_nr_rejected")}')
 print(f'enable_seq    : {read_atomic("scx_enable_seq")}')