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, 131 insertions, 0 deletions

diff --git a/drivers/reset/reset-k210.c b/drivers/reset/reset-k210.c
new file mode 100644
index 000000000..e77e4cca3
--- /dev/null
+++ b/drivers/reset/reset-k210.c
@@ -0,0 +1,131 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (c) 2020 Western Digital Corporation or its affiliates.
+ */
+#include <linux/of.h>
+#include <linux/platform_device.h>
+#include <linux/reset-controller.h>
+#include <linux/delay.h>
+#include <linux/mfd/syscon.h>
+#include <linux/regmap.h>
+#include <soc/canaan/k210-sysctl.h>
+
+#include <dt-bindings/reset/k210-rst.h>
+
+#define K210_RST_MASK	0x27FFFFFF
+
+struct k210_rst {
+	struct regmap *map;
+	struct reset_controller_dev rcdev;
+};
+
+static inline struct k210_rst *
+to_k210_rst(struct reset_controller_dev *rcdev)
+{
+	return container_of(rcdev, struct k210_rst, rcdev);
+}
+
+static inline int k210_rst_assert(struct reset_controller_dev *rcdev,
+				  unsigned long id)
+{
+	struct k210_rst *ksr = to_k210_rst(rcdev);
+
+	return regmap_update_bits(ksr->map, K210_SYSCTL_PERI_RESET, BIT(id), 1);
+}
+
+static inline int k210_rst_deassert(struct reset_controller_dev *rcdev,
+				    unsigned long id)
+{
+	struct k210_rst *ksr = to_k210_rst(rcdev);
+
+	return regmap_update_bits(ksr->map, K210_SYSCTL_PERI_RESET, BIT(id), 0);
+}
+
+static int k210_rst_reset(struct reset_controller_dev *rcdev,
+			  unsigned long id)
+{
+	int ret;
+
+	ret = k210_rst_assert(rcdev, id);
+	if (ret == 0) {
+		udelay(10);
+		ret = k210_rst_deassert(rcdev, id);
+	}
+
+	return ret;
+}
+
+static int k210_rst_status(struct reset_controller_dev *rcdev,
+			   unsigned long id)
+{
+	struct k210_rst *ksr = to_k210_rst(rcdev);
+	u32 reg, bit = BIT(id);
+	int ret;
+
+	ret = regmap_read(ksr->map, K210_SYSCTL_PERI_RESET, &reg);
+	if (ret)
+		return ret;
+
+	return reg & bit;
+}
+
+static int k210_rst_xlate(struct reset_controller_dev *rcdev,
+			  const struct of_phandle_args *reset_spec)
+{
+	unsigned long id = reset_spec->args[0];
+
+	if (!(BIT(id) & K210_RST_MASK))
+		return -EINVAL;
+
+	return id;
+}
+
+static const struct reset_control_ops k210_rst_ops = {
+	.assert		= k210_rst_assert,
+	.deassert	= k210_rst_deassert,
+	.reset		= k210_rst_reset,
+	.status		= k210_rst_status,
+};
+
+static int k210_rst_probe(struct platform_device *pdev)
+{
+	struct device *dev = &pdev->dev;
+	struct device_node *parent_np;
+	struct k210_rst *ksr;
+
+	dev_info(dev, "K210 reset controller\n");
+
+	ksr = devm_kzalloc(dev, sizeof(*ksr), GFP_KERNEL);
+	if (!ksr)
+		return -ENOMEM;
+
+	parent_np = of_get_parent(dev->of_node);
+	ksr->map = syscon_node_to_regmap(parent_np);
+	of_node_put(parent_np);
+	if (IS_ERR(ksr->map))
+		return PTR_ERR(ksr->map);
+
+	ksr->rcdev.owner = THIS_MODULE;
+	ksr->rcdev.dev = dev;
+	ksr->rcdev.of_node = dev->of_node;
+	ksr->rcdev.ops = &k210_rst_ops;
+	ksr->rcdev.nr_resets = fls(K210_RST_MASK);
+	ksr->rcdev.of_reset_n_cells = 1;
+	ksr->rcdev.of_xlate = k210_rst_xlate;
+
+	return devm_reset_controller_register(dev, &ksr->rcdev);
+}
+
+static const struct of_device_id k210_rst_dt_ids[] = {
+	{ .compatible = "canaan,k210-rst" },
+	{ /* sentinel */ },
+};
+
+static struct platform_driver k210_rst_driver = {
+	.probe	= k210_rst_probe,
+	.driver = {
+		.name		= "k210-rst",
+		.of_match_table	= k210_rst_dt_ids,
+	},
+};
+builtin_platform_driver(k210_rst_driver);