sched/cache: Keep nr_pref_llc_running in the runnable domain, to fix LLC mis-scheduling buggrafted

alb_break_llc() decides whether to break LLC preference during active
load balance. It does so by testing that every runnable fair task on the
source rq prefers its LLC:

	env->src_rq->nr_pref_llc_running == env->src_rq->cfs.h_nr_runnable

But the two counters cover different sets. nr_pref_llc_running is updated
in account_llc_enqueue()/account_llc_dequeue(), next to cfs_rq->nr_queued,
so it follows queued tasks. h_nr_runnable is updated in set_delayed()/
clear_delayed() and drops delay-dequeued tasks.

So under DELAY_DEQUEUE, a preferring task that goes to sleep stays counted
in nr_pref_llc_running while h_nr_runnable falls. The equality then breaks,
alb_break_llc() returns false, and active balance is free to pull a task
off its preferred LLC. Active balance only moves runnable tasks, and this
is the only LLC check it consults: once the stopper runs, LBF_ACTIVE_LB
skips the per-task test in can_migrate_task(). The runnable set is the one
we want.

Fix it on the counter side. A task should be counted in
nr_pref_llc_running exactly while it is both queued on its preferred LLC
(pref_llc_queued) and runnable (!sched_delayed). Define that membership
once in task_pref_llc_runnable(), and adjust the counter only through
pref_llc_running_inc()/pref_llc_running_dec() from the four sites that
change either input: account_llc_enqueue(), account_llc_dequeue(),
set_delayed() and clear_delayed(). Gating every update on the same
predicate keeps the delay, wake and dequeue paths from double-counting
or underflowing; see the comments at those sites for the ordering.

nr_llc_running and sd->llc_counts are not touched and stay on queued
semantics.

Fixes: 714059f79ff0 ("sched/cache: Handle moving single tasks to/from their preferred LLC")
Closes: https://lore.kernel.org/lkml/20260827135000.735138-1-zhanxusheng@xiaomi.com/
Reported-by: Zhan Xusheng <zhanxusheng@xiaomi.com>
Suggested-by: Chen Yu <yu.c.chen@intel.com>
Signed-off-by: Tim Chen <tim.c.chen@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Reviewed-by: Kayra Cizmeci <kayracizmeci@gmail.com>
Cc: <stable@kernel.org> # v7.2.x
Link: https://patch.msgid.link/06af61afedac32e6477f57feb4d658f6c411c3af.1790035273.git.tim.c.chen@linux.intel.com

1 files changed, 72 insertions, 0 deletions

diff --git a/include/net/ncsi.h b/include/net/ncsi.h
new file mode 100644
index 000000000..e88635820
--- /dev/null
+++ b/include/net/ncsi.h
@@ -0,0 +1,72 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef __NET_NCSI_H
+#define __NET_NCSI_H
+
+#include <linux/types.h>
+
+/*
+ * The NCSI device states seen from external. More NCSI device states are
+ * only visible internally (in net/ncsi/internal.h). When the NCSI device
+ * is registered, it's in ncsi_dev_state_registered state. The state
+ * ncsi_dev_state_start is used to drive to choose active package and
+ * channel. After that, its state is changed to ncsi_dev_state_functional.
+ *
+ * The state ncsi_dev_state_stop helps to shut down the currently active
+ * package and channel while ncsi_dev_state_config helps to reconfigure
+ * them.
+ */
+enum {
+	ncsi_dev_state_registered	= 0x0000,
+	ncsi_dev_state_functional	= 0x0100,
+	ncsi_dev_state_probe		= 0x0200,
+	ncsi_dev_state_config		= 0x0300,
+	ncsi_dev_state_suspend		= 0x0400,
+};
+
+struct ncsi_dev {
+	int               state;
+	int		  link_up;
+	struct net_device *dev;
+	void		  (*handler)(struct ncsi_dev *ndev);
+};
+
+#ifdef CONFIG_NET_NCSI
+int ncsi_vlan_rx_add_vid(struct net_device *dev, __be16 proto, u16 vid);
+int ncsi_vlan_rx_kill_vid(struct net_device *dev, __be16 proto, u16 vid);
+struct ncsi_dev *ncsi_register_dev(struct net_device *dev,
+				   void (*notifier)(struct ncsi_dev *nd));
+int ncsi_start_dev(struct ncsi_dev *nd);
+void ncsi_stop_dev(struct ncsi_dev *nd);
+void ncsi_unregister_dev(struct ncsi_dev *nd);
+#else /* !CONFIG_NET_NCSI */
+static inline int ncsi_vlan_rx_add_vid(struct net_device *dev, __be16 proto, u16 vid)
+{
+	return -EINVAL;
+}
+
+static inline int ncsi_vlan_rx_kill_vid(struct net_device *dev, __be16 proto, u16 vid)
+{
+	return -EINVAL;
+}
+
+static inline struct ncsi_dev *ncsi_register_dev(struct net_device *dev,
+					void (*notifier)(struct ncsi_dev *nd))
+{
+	return NULL;
+}
+
+static inline int ncsi_start_dev(struct ncsi_dev *nd)
+{
+	return -ENOTTY;
+}
+
+static inline void ncsi_stop_dev(struct ncsi_dev *nd)
+{
+}
+
+static inline void ncsi_unregister_dev(struct ncsi_dev *nd)
+{
+}
+#endif /* CONFIG_NET_NCSI */
+
+#endif /* __NET_NCSI_H */