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, 133 insertions, 0 deletions

diff --git a/drivers/reset/reset-bcm6345.c b/drivers/reset/reset-bcm6345.c
new file mode 100644
index 000000000..441d9dd76
--- /dev/null
+++ b/drivers/reset/reset-bcm6345.c
@@ -0,0 +1,133 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * BCM6345 Reset Controller Driver
+ *
+ * Copyright (C) 2020 Álvaro Fernández Rojas <noltari@gmail.com>
+ */
+
+#include <linux/delay.h>
+#include <linux/init.h>
+#include <linux/io.h>
+#include <linux/platform_device.h>
+#include <linux/reset-controller.h>
+
+#define BCM6345_RESET_NUM		32
+#define BCM6345_RESET_SLEEP_MIN_US	10000
+#define BCM6345_RESET_SLEEP_MAX_US	20000
+
+struct bcm6345_reset {
+	struct reset_controller_dev rcdev;
+	void __iomem *base;
+	spinlock_t lock;
+};
+
+static inline struct bcm6345_reset *
+to_bcm6345_reset(struct reset_controller_dev *rcdev)
+{
+	return container_of(rcdev, struct bcm6345_reset, rcdev);
+}
+
+static int bcm6345_reset_update(struct reset_controller_dev *rcdev,
+				unsigned long id, bool assert)
+{
+	struct bcm6345_reset *bcm6345_reset = to_bcm6345_reset(rcdev);
+	unsigned long flags;
+	uint32_t val;
+
+	spin_lock_irqsave(&bcm6345_reset->lock, flags);
+	val = __raw_readl(bcm6345_reset->base);
+	if (assert)
+		val &= ~BIT(id);
+	else
+		val |= BIT(id);
+	__raw_writel(val, bcm6345_reset->base);
+	spin_unlock_irqrestore(&bcm6345_reset->lock, flags);
+
+	return 0;
+}
+
+static int bcm6345_reset_assert(struct reset_controller_dev *rcdev,
+				unsigned long id)
+{
+	return bcm6345_reset_update(rcdev, id, true);
+}
+
+static int bcm6345_reset_deassert(struct reset_controller_dev *rcdev,
+				  unsigned long id)
+{
+	return bcm6345_reset_update(rcdev, id, false);
+}
+
+static int bcm6345_reset_reset(struct reset_controller_dev *rcdev,
+			       unsigned long id)
+{
+	bcm6345_reset_update(rcdev, id, true);
+	usleep_range(BCM6345_RESET_SLEEP_MIN_US,
+		     BCM6345_RESET_SLEEP_MAX_US);
+
+	bcm6345_reset_update(rcdev, id, false);
+	/*
+	 * Ensure component is taken out reset state by sleeping also after
+	 * deasserting the reset. Otherwise, the component may not be ready
+	 * for operation.
+	 */
+	usleep_range(BCM6345_RESET_SLEEP_MIN_US,
+		     BCM6345_RESET_SLEEP_MAX_US);
+
+	return 0;
+}
+
+static int bcm6345_reset_status(struct reset_controller_dev *rcdev,
+				unsigned long id)
+{
+	struct bcm6345_reset *bcm6345_reset = to_bcm6345_reset(rcdev);
+
+	return !(__raw_readl(bcm6345_reset->base) & BIT(id));
+}
+
+static const struct reset_control_ops bcm6345_reset_ops = {
+	.assert = bcm6345_reset_assert,
+	.deassert = bcm6345_reset_deassert,
+	.reset = bcm6345_reset_reset,
+	.status = bcm6345_reset_status,
+};
+
+static int bcm6345_reset_probe(struct platform_device *pdev)
+{
+	struct bcm6345_reset *bcm6345_reset;
+
+	bcm6345_reset = devm_kzalloc(&pdev->dev,
+				     sizeof(*bcm6345_reset), GFP_KERNEL);
+	if (!bcm6345_reset)
+		return -ENOMEM;
+
+	bcm6345_reset->base = devm_platform_ioremap_resource(pdev, 0);
+	if (IS_ERR(bcm6345_reset->base))
+		return PTR_ERR(bcm6345_reset->base);
+
+	spin_lock_init(&bcm6345_reset->lock);
+	bcm6345_reset->rcdev.ops = &bcm6345_reset_ops;
+	bcm6345_reset->rcdev.owner = THIS_MODULE;
+	bcm6345_reset->rcdev.of_node = pdev->dev.of_node;
+	bcm6345_reset->rcdev.of_reset_n_cells = 1;
+	bcm6345_reset->rcdev.nr_resets = BCM6345_RESET_NUM;
+
+	return devm_reset_controller_register(&pdev->dev,
+					      &bcm6345_reset->rcdev);
+}
+
+static const struct of_device_id bcm6345_reset_of_match[] = {
+	{ .compatible = "brcm,bcm6345-reset" },
+	{ .compatible = "brcm,bcm63xx-ephy-ctrl" },
+	{ /* sentinel */ },
+};
+
+static struct platform_driver bcm6345_reset_driver = {
+	.probe = bcm6345_reset_probe,
+	.driver	= {
+		.name = "bcm6345-reset",
+		.of_match_table = bcm6345_reset_of_match,
+		.suppress_bind_attrs = true,
+	},
+};
+builtin_platform_driver(bcm6345_reset_driver);