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, 184 insertions, 0 deletions

diff --git a/drivers/acpi/event.c b/drivers/acpi/event.c
new file mode 100644
index 000000000..4d840d2e7
--- /dev/null
+++ b/drivers/acpi/event.c
@@ -0,0 +1,184 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * event.c - exporting ACPI events via procfs
+ *
+ *  Copyright (C) 2001, 2002 Andy Grover <andrew.grover@intel.com>
+ *  Copyright (C) 2001, 2002 Paul Diefenbaugh <paul.s.diefenbaugh@intel.com>
+ *
+ */
+
+#define pr_fmt(fmt) "ACPI: " fmt
+
+#include <linux/spinlock.h>
+#include <linux/export.h>
+#include <linux/proc_fs.h>
+#include <linux/init.h>
+#include <linux/poll.h>
+#include <linux/gfp.h>
+#include <linux/acpi.h>
+#include <net/netlink.h>
+#include <net/genetlink.h>
+
+#include "internal.h"
+
+/* ACPI notifier chain */
+static BLOCKING_NOTIFIER_HEAD(acpi_chain_head);
+
+int acpi_notifier_call_chain(const char *device_class,
+			     const char *bus_id, u32 type, u32 data)
+{
+	struct acpi_bus_event event;
+
+	strscpy(event.device_class, device_class);
+	strscpy(event.bus_id, bus_id);
+	event.type = type;
+	event.data = data;
+	return (blocking_notifier_call_chain(&acpi_chain_head, 0, (void *)&event)
+			== NOTIFY_BAD) ? -EINVAL : 0;
+}
+EXPORT_SYMBOL(acpi_notifier_call_chain);
+
+int register_acpi_notifier(struct notifier_block *nb)
+{
+	return blocking_notifier_chain_register(&acpi_chain_head, nb);
+}
+EXPORT_SYMBOL(register_acpi_notifier);
+
+int unregister_acpi_notifier(struct notifier_block *nb)
+{
+	return blocking_notifier_chain_unregister(&acpi_chain_head, nb);
+}
+EXPORT_SYMBOL(unregister_acpi_notifier);
+
+#ifdef CONFIG_NET
+static unsigned int acpi_event_seqnum;
+struct acpi_genl_event {
+	acpi_device_class device_class;
+	char bus_id[15];
+	u32 type;
+	u32 data;
+};
+
+/* attributes of acpi_genl_family */
+enum {
+	ACPI_GENL_ATTR_UNSPEC,
+	ACPI_GENL_ATTR_EVENT,	/* ACPI event info needed by user space */
+	__ACPI_GENL_ATTR_MAX,
+};
+#define ACPI_GENL_ATTR_MAX (__ACPI_GENL_ATTR_MAX - 1)
+
+/* commands supported by the acpi_genl_family */
+enum {
+	ACPI_GENL_CMD_UNSPEC,
+	ACPI_GENL_CMD_EVENT,	/* kernel->user notifications for ACPI events */
+	__ACPI_GENL_CMD_MAX,
+};
+#define ACPI_GENL_CMD_MAX (__ACPI_GENL_CMD_MAX - 1)
+
+#define ACPI_GENL_FAMILY_NAME		"acpi_event"
+#define ACPI_GENL_VERSION		0x01
+#define ACPI_GENL_MCAST_GROUP_NAME 	"acpi_mc_group"
+
+static const struct genl_multicast_group acpi_event_mcgrps[] = {
+	{ .name = ACPI_GENL_MCAST_GROUP_NAME, },
+};
+
+static struct genl_family acpi_event_genl_family __ro_after_init = {
+	.module = THIS_MODULE,
+	.name = ACPI_GENL_FAMILY_NAME,
+	.version = ACPI_GENL_VERSION,
+	.maxattr = ACPI_GENL_ATTR_MAX,
+	.mcgrps = acpi_event_mcgrps,
+	.n_mcgrps = ARRAY_SIZE(acpi_event_mcgrps),
+};
+
+int acpi_bus_generate_netlink_event(const char *device_class,
+				      const char *bus_id,
+				      u8 type, int data)
+{
+	struct sk_buff *skb;
+	struct nlattr *attr;
+	struct acpi_genl_event *event;
+	void *msg_header;
+	int size;
+
+	/* allocate memory */
+	size = nla_total_size(sizeof(struct acpi_genl_event)) +
+	    nla_total_size(0);
+
+	skb = genlmsg_new(size, GFP_ATOMIC);
+	if (!skb)
+		return -ENOMEM;
+
+	/* add the genetlink message header */
+	msg_header = genlmsg_put(skb, 0, acpi_event_seqnum++,
+				 &acpi_event_genl_family, 0,
+				 ACPI_GENL_CMD_EVENT);
+	if (!msg_header) {
+		nlmsg_free(skb);
+		return -ENOMEM;
+	}
+
+	/* fill the data */
+	attr =
+	    nla_reserve(skb, ACPI_GENL_ATTR_EVENT,
+			sizeof(struct acpi_genl_event));
+	if (!attr) {
+		nlmsg_free(skb);
+		return -EINVAL;
+	}
+
+	event = nla_data(attr);
+	memset(event, 0, sizeof(struct acpi_genl_event));
+
+	strscpy(event->device_class, device_class, sizeof(event->device_class));
+	strscpy(event->bus_id, bus_id, sizeof(event->bus_id));
+	event->type = type;
+	event->data = data;
+
+	/* send multicast genetlink message */
+	genlmsg_end(skb, msg_header);
+
+	genlmsg_multicast(&acpi_event_genl_family, skb, 0, 0, GFP_ATOMIC);
+	return 0;
+}
+
+EXPORT_SYMBOL(acpi_bus_generate_netlink_event);
+
+static int __init acpi_event_genetlink_init(void)
+{
+	return genl_register_family(&acpi_event_genl_family);
+}
+
+#else
+int acpi_bus_generate_netlink_event(const char *device_class,
+				      const char *bus_id,
+				      u8 type, int data)
+{
+	return 0;
+}
+
+EXPORT_SYMBOL(acpi_bus_generate_netlink_event);
+
+static int acpi_event_genetlink_init(void)
+{
+	return -ENODEV;
+}
+#endif
+
+static int __init acpi_event_init(void)
+{
+	int error;
+
+	if (acpi_disabled)
+		return 0;
+
+	/* create genetlink for acpi event */
+	error = acpi_event_genetlink_init();
+	if (error)
+		pr_warn("Failed to create genetlink family for ACPI event\n");
+
+	return 0;
+}
+
+fs_initcall(acpi_event_init);