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, 127 insertions, 0 deletions

diff --git a/drivers/reset/reset-tn48m.c b/drivers/reset/reset-tn48m.c
new file mode 100644
index 000000000..177ee5127
--- /dev/null
+++ b/drivers/reset/reset-tn48m.c
@@ -0,0 +1,127 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Delta TN48M CPLD reset driver
+ *
+ * Copyright (C) 2021 Sartura Ltd.
+ *
+ * Author: Robert Marko <robert.marko@sartura.hr>
+ */
+
+#include <linux/device.h>
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/platform_device.h>
+#include <linux/regmap.h>
+#include <linux/reset-controller.h>
+
+#include <dt-bindings/reset/delta,tn48m-reset.h>
+
+#define TN48M_RESET_REG		0x10
+
+#define TN48M_RESET_TIMEOUT_US	125000
+#define TN48M_RESET_SLEEP_US	10
+
+struct tn48_reset_map {
+	u8 bit;
+};
+
+struct tn48_reset_data {
+	struct reset_controller_dev rcdev;
+	struct regmap *regmap;
+};
+
+static const struct tn48_reset_map tn48m_resets[] = {
+	[CPU_88F7040_RESET] = {0},
+	[CPU_88F6820_RESET] = {1},
+	[MAC_98DX3265_RESET] = {2},
+	[PHY_88E1680_RESET] = {4},
+	[PHY_88E1512_RESET] = {6},
+	[POE_RESET] = {7},
+};
+
+static inline struct tn48_reset_data *to_tn48_reset_data(
+			struct reset_controller_dev *rcdev)
+{
+	return container_of(rcdev, struct tn48_reset_data, rcdev);
+}
+
+static int tn48m_control_reset(struct reset_controller_dev *rcdev,
+			       unsigned long id)
+{
+	struct tn48_reset_data *data = to_tn48_reset_data(rcdev);
+	unsigned int val;
+
+	regmap_update_bits(data->regmap, TN48M_RESET_REG,
+			   BIT(tn48m_resets[id].bit), 0);
+
+	return regmap_read_poll_timeout(data->regmap,
+					TN48M_RESET_REG,
+					val,
+					val & BIT(tn48m_resets[id].bit),
+					TN48M_RESET_SLEEP_US,
+					TN48M_RESET_TIMEOUT_US);
+}
+
+static int tn48m_control_status(struct reset_controller_dev *rcdev,
+				unsigned long id)
+{
+	struct tn48_reset_data *data = to_tn48_reset_data(rcdev);
+	unsigned int regval;
+	int ret;
+
+	ret = regmap_read(data->regmap, TN48M_RESET_REG, &regval);
+	if (ret < 0)
+		return ret;
+
+	if (BIT(tn48m_resets[id].bit) & regval)
+		return 0;
+	else
+		return 1;
+}
+
+static const struct reset_control_ops tn48_reset_ops = {
+	.reset		= tn48m_control_reset,
+	.status		= tn48m_control_status,
+};
+
+static int tn48m_reset_probe(struct platform_device *pdev)
+{
+	struct tn48_reset_data *data;
+	struct regmap *regmap;
+
+	regmap = dev_get_regmap(pdev->dev.parent, NULL);
+	if (!regmap)
+		return -ENODEV;
+
+	data = devm_kzalloc(&pdev->dev, sizeof(*data), GFP_KERNEL);
+	if (!data)
+		return -ENOMEM;
+
+	data->regmap = regmap;
+
+	data->rcdev.owner = THIS_MODULE;
+	data->rcdev.ops = &tn48_reset_ops;
+	data->rcdev.nr_resets = ARRAY_SIZE(tn48m_resets);
+	data->rcdev.of_node = pdev->dev.of_node;
+
+	return devm_reset_controller_register(&pdev->dev, &data->rcdev);
+}
+
+static const struct of_device_id tn48m_reset_of_match[] = {
+	{ .compatible = "delta,tn48m-reset" },
+	{ }
+};
+MODULE_DEVICE_TABLE(of, tn48m_reset_of_match);
+
+static struct platform_driver tn48m_reset_driver = {
+	.driver = {
+		.name = "delta-tn48m-reset",
+		.of_match_table = tn48m_reset_of_match,
+	},
+	.probe = tn48m_reset_probe,
+};
+module_platform_driver(tn48m_reset_driver);
+
+MODULE_AUTHOR("Robert Marko <robert.marko@sartura.hr>");
+MODULE_DESCRIPTION("Delta TN48M CPLD reset driver");
+MODULE_LICENSE("GPL");