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, 201 insertions, 0 deletions

diff --git a/drivers/rtc/rtc-test.c b/drivers/rtc/rtc-test.c
new file mode 100644
index 000000000..94f995feb
--- /dev/null
+++ b/drivers/rtc/rtc-test.c
@@ -0,0 +1,201 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * An RTC test device/driver
+ * Copyright (C) 2005 Tower Technologies
+ * Author: Alessandro Zummo <a.zummo@towertech.it>
+ */
+
+#include <linux/module.h>
+#include <linux/err.h>
+#include <linux/rtc.h>
+#include <linux/platform_device.h>
+
+#define MAX_RTC_TEST 3
+
+struct rtc_test_data {
+	struct rtc_device *rtc;
+	time64_t offset;
+	struct timer_list alarm;
+	bool alarm_en;
+};
+
+static struct platform_device *pdev[MAX_RTC_TEST];
+
+static int test_rtc_read_alarm(struct device *dev, struct rtc_wkalrm *alrm)
+{
+	struct rtc_test_data *rtd = dev_get_drvdata(dev);
+	time64_t alarm;
+
+	alarm = (rtd->alarm.expires - jiffies) / HZ;
+	alarm += ktime_get_real_seconds() + rtd->offset;
+
+	rtc_time64_to_tm(alarm, &alrm->time);
+	alrm->enabled = rtd->alarm_en;
+
+	return 0;
+}
+
+static int test_rtc_set_alarm(struct device *dev, struct rtc_wkalrm *alrm)
+{
+	struct rtc_test_data *rtd = dev_get_drvdata(dev);
+	ktime_t timeout;
+	u64 expires;
+
+	timeout = rtc_tm_to_time64(&alrm->time) - ktime_get_real_seconds();
+	timeout -= rtd->offset;
+
+	timer_delete(&rtd->alarm);
+
+	expires = jiffies + timeout * HZ;
+	if (expires > U32_MAX)
+		expires = U32_MAX;
+
+	rtd->alarm.expires = expires;
+
+	if (alrm->enabled)
+		add_timer(&rtd->alarm);
+
+	rtd->alarm_en = alrm->enabled;
+
+	return 0;
+}
+
+static int test_rtc_read_time(struct device *dev, struct rtc_time *tm)
+{
+	struct rtc_test_data *rtd = dev_get_drvdata(dev);
+
+	rtc_time64_to_tm(ktime_get_real_seconds() + rtd->offset, tm);
+
+	return 0;
+}
+
+static int test_rtc_set_time(struct device *dev, struct rtc_time *tm)
+{
+	struct rtc_test_data *rtd = dev_get_drvdata(dev);
+
+	rtd->offset = rtc_tm_to_time64(tm) - ktime_get_real_seconds();
+
+	return 0;
+}
+
+static int test_rtc_alarm_irq_enable(struct device *dev, unsigned int enable)
+{
+	struct rtc_test_data *rtd = dev_get_drvdata(dev);
+
+	rtd->alarm_en = enable;
+	if (enable)
+		add_timer(&rtd->alarm);
+	else
+		timer_delete(&rtd->alarm);
+
+	return 0;
+}
+
+static const struct rtc_class_ops test_rtc_ops_noalm = {
+	.read_time = test_rtc_read_time,
+	.set_time = test_rtc_set_time,
+	.alarm_irq_enable = test_rtc_alarm_irq_enable,
+};
+
+static const struct rtc_class_ops test_rtc_ops = {
+	.read_time = test_rtc_read_time,
+	.set_time = test_rtc_set_time,
+	.read_alarm = test_rtc_read_alarm,
+	.set_alarm = test_rtc_set_alarm,
+	.alarm_irq_enable = test_rtc_alarm_irq_enable,
+};
+
+static void test_rtc_alarm_handler(struct timer_list *t)
+{
+	struct rtc_test_data *rtd = timer_container_of(rtd, t, alarm);
+
+	rtc_update_irq(rtd->rtc, 1, RTC_AF | RTC_IRQF);
+}
+
+static int test_probe(struct platform_device *plat_dev)
+{
+	struct rtc_test_data *rtd;
+
+	rtd = devm_kzalloc(&plat_dev->dev, sizeof(*rtd), GFP_KERNEL);
+	if (!rtd)
+		return -ENOMEM;
+
+	platform_set_drvdata(plat_dev, rtd);
+
+	rtd->rtc = devm_rtc_allocate_device(&plat_dev->dev);
+	if (IS_ERR(rtd->rtc))
+		return PTR_ERR(rtd->rtc);
+
+	switch (plat_dev->id) {
+	case 0:
+		rtd->rtc->ops = &test_rtc_ops_noalm;
+		break;
+	default:
+		rtd->rtc->ops = &test_rtc_ops;
+		device_init_wakeup(&plat_dev->dev, true);
+	}
+
+	timer_setup(&rtd->alarm, test_rtc_alarm_handler, 0);
+	rtd->alarm.expires = 0;
+
+	return devm_rtc_register_device(rtd->rtc);
+}
+
+static struct platform_driver test_driver = {
+	.probe	= test_probe,
+	.driver = {
+		.name = "rtc-test",
+	},
+};
+
+static int __init test_init(void)
+{
+	int i, err;
+
+	err = platform_driver_register(&test_driver);
+	if (err)
+		return err;
+
+	err = -ENOMEM;
+	for (i = 0; i < MAX_RTC_TEST; i++) {
+		pdev[i] = platform_device_alloc("rtc-test", i);
+		if (!pdev[i])
+			goto exit_free_mem;
+	}
+
+	for (i = 0; i < MAX_RTC_TEST; i++) {
+		err = platform_device_add(pdev[i]);
+		if (err)
+			goto exit_device_del;
+	}
+
+	return 0;
+
+exit_device_del:
+	for (; i > 0; i--)
+		platform_device_del(pdev[i - 1]);
+
+exit_free_mem:
+	for (i = 0; i < MAX_RTC_TEST; i++)
+		platform_device_put(pdev[i]);
+
+	platform_driver_unregister(&test_driver);
+	return err;
+}
+
+static void __exit test_exit(void)
+{
+	int i;
+
+	for (i = 0; i < MAX_RTC_TEST; i++)
+		platform_device_unregister(pdev[i]);
+
+	platform_driver_unregister(&test_driver);
+}
+
+MODULE_AUTHOR("Alessandro Zummo <a.zummo@towertech.it>");
+MODULE_DESCRIPTION("RTC test driver/device");
+MODULE_LICENSE("GPL v2");
+
+module_init(test_init);
+module_exit(test_exit);