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, 221 insertions, 0 deletions

diff --git a/drivers/input/serio/apbps2.c b/drivers/input/serio/apbps2.c
new file mode 100644
index 000000000..5f21acdd4
--- /dev/null
+++ b/drivers/input/serio/apbps2.c
@@ -0,0 +1,221 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (C) 2013 Aeroflex Gaisler
+ *
+ * This driver supports the APBPS2 PS/2 core available in the GRLIB
+ * VHDL IP core library.
+ *
+ * Full documentation of the APBPS2 core can be found here:
+ * http://www.gaisler.com/products/grlib/grip.pdf
+ *
+ * See "Documentation/devicetree/bindings/input/ps2keyb-mouse-apbps2.txt" for
+ * information on open firmware properties.
+ *
+ * Contributors: Daniel Hellstrom <daniel@gaisler.com>
+ */
+#include <linux/platform_device.h>
+#include <linux/module.h>
+#include <linux/serio.h>
+#include <linux/errno.h>
+#include <linux/interrupt.h>
+#include <linux/of.h>
+#include <linux/of_irq.h>
+#include <linux/device.h>
+#include <linux/delay.h>
+#include <linux/err.h>
+#include <linux/slab.h>
+#include <linux/string.h>
+#include <linux/kernel.h>
+#include <linux/io.h>
+
+struct apbps2_regs {
+	u32 __iomem data;	/* 0x00 */
+	u32 __iomem status;	/* 0x04 */
+	u32 __iomem ctrl;	/* 0x08 */
+	u32 __iomem reload;	/* 0x0c */
+};
+
+#define APBPS2_STATUS_DR	(1<<0)
+#define APBPS2_STATUS_PE	(1<<1)
+#define APBPS2_STATUS_FE	(1<<2)
+#define APBPS2_STATUS_KI	(1<<3)
+#define APBPS2_STATUS_RF	(1<<4)
+#define APBPS2_STATUS_TF	(1<<5)
+#define APBPS2_STATUS_TCNT	(0x1f<<22)
+#define APBPS2_STATUS_RCNT	(0x1f<<27)
+
+#define APBPS2_CTRL_RE		(1<<0)
+#define APBPS2_CTRL_TE		(1<<1)
+#define APBPS2_CTRL_RI		(1<<2)
+#define APBPS2_CTRL_TI		(1<<3)
+
+struct apbps2_priv {
+	struct serio		*io;
+	struct apbps2_regs	__iomem *regs;
+};
+
+static int apbps2_idx;
+
+static irqreturn_t apbps2_isr(int irq, void *dev_id)
+{
+	struct apbps2_priv *priv = dev_id;
+	unsigned long status, data, rxflags;
+	irqreturn_t ret = IRQ_NONE;
+
+	while ((status = ioread32be(&priv->regs->status)) & APBPS2_STATUS_DR) {
+		data = ioread32be(&priv->regs->data);
+		rxflags = (status & APBPS2_STATUS_PE) ? SERIO_PARITY : 0;
+		rxflags |= (status & APBPS2_STATUS_FE) ? SERIO_FRAME : 0;
+
+		/* Clear error bits */
+		if (rxflags)
+			iowrite32be(0, &priv->regs->status);
+
+		serio_interrupt(priv->io, data, rxflags);
+
+		ret = IRQ_HANDLED;
+	}
+
+	return ret;
+}
+
+static int apbps2_write(struct serio *io, unsigned char val)
+{
+	struct apbps2_priv *priv = io->port_data;
+	unsigned int tleft = 10000; /* Timeout in 100ms */
+
+	/* Delay until PS/2 controller has room for more chars */
+	while ((ioread32be(&priv->regs->status) & APBPS2_STATUS_TF) && tleft--)
+		udelay(10);
+
+	if ((ioread32be(&priv->regs->status) & APBPS2_STATUS_TF) == 0) {
+		iowrite32be(val, &priv->regs->data);
+
+		iowrite32be(APBPS2_CTRL_RE | APBPS2_CTRL_RI | APBPS2_CTRL_TE,
+				&priv->regs->ctrl);
+		return 0;
+	}
+
+	return -ETIMEDOUT;
+}
+
+static int apbps2_open(struct serio *io)
+{
+	struct apbps2_priv *priv = io->port_data;
+	int limit;
+
+	/* Clear error flags */
+	iowrite32be(0, &priv->regs->status);
+
+	/* Clear old data if available (unlikely) */
+	limit = 1024;
+	while ((ioread32be(&priv->regs->status) & APBPS2_STATUS_DR) && --limit)
+		ioread32be(&priv->regs->data);
+
+	/* Enable receiver and its interrupt */
+	iowrite32be(APBPS2_CTRL_RE | APBPS2_CTRL_RI, &priv->regs->ctrl);
+
+	return 0;
+}
+
+static void apbps2_close(struct serio *io)
+{
+	struct apbps2_priv *priv = io->port_data;
+
+	/* Stop interrupts at PS/2 HW level */
+	iowrite32be(0, &priv->regs->ctrl);
+}
+
+/* Initialize one APBPS2 PS/2 core */
+static int apbps2_of_probe(struct platform_device *ofdev)
+{
+	struct apbps2_priv *priv;
+	int irq, err;
+	u32 freq_hz;
+
+	priv = devm_kzalloc(&ofdev->dev, sizeof(*priv), GFP_KERNEL);
+	if (!priv) {
+		dev_err(&ofdev->dev, "memory allocation failed\n");
+		return -ENOMEM;
+	}
+
+	/* Find device address */
+	priv->regs = devm_platform_ioremap_resource(ofdev, 0);
+	if (IS_ERR(priv->regs))
+		return PTR_ERR(priv->regs);
+
+	/* Reset hardware, disable interrupt */
+	iowrite32be(0, &priv->regs->ctrl);
+
+	/* IRQ */
+	irq = platform_get_irq(ofdev, 0);
+	if (irq < 0)
+		return irq;
+
+	err = devm_request_irq(&ofdev->dev, irq, apbps2_isr,
+				IRQF_SHARED, "apbps2", priv);
+	if (err) {
+		dev_err(&ofdev->dev, "request IRQ%d failed\n", irq);
+		return err;
+	}
+
+	/* Get core frequency */
+	if (of_property_read_u32(ofdev->dev.of_node, "freq", &freq_hz)) {
+		dev_err(&ofdev->dev, "unable to get core frequency\n");
+		return -EINVAL;
+	}
+
+	/* Set reload register to core freq in kHz/10 */
+	iowrite32be(freq_hz / 10000, &priv->regs->reload);
+
+	priv->io = kzalloc_obj(*priv->io);
+	if (!priv->io)
+		return -ENOMEM;
+
+	priv->io->id.type = SERIO_8042;
+	priv->io->open = apbps2_open;
+	priv->io->close = apbps2_close;
+	priv->io->write = apbps2_write;
+	priv->io->port_data = priv;
+	strscpy(priv->io->name, "APBPS2 PS/2", sizeof(priv->io->name));
+	snprintf(priv->io->phys, sizeof(priv->io->phys),
+		 "apbps2_%d", apbps2_idx++);
+
+	dev_info(&ofdev->dev, "irq = %d, base = 0x%p\n", irq, priv->regs);
+
+	serio_register_port(priv->io);
+
+	platform_set_drvdata(ofdev, priv);
+
+	return 0;
+}
+
+static void apbps2_of_remove(struct platform_device *of_dev)
+{
+	struct apbps2_priv *priv = platform_get_drvdata(of_dev);
+
+	serio_unregister_port(priv->io);
+}
+
+static const struct of_device_id apbps2_of_match[] = {
+	{ .name = "GAISLER_APBPS2", },
+	{ .name = "01_060", },
+	{}
+};
+
+MODULE_DEVICE_TABLE(of, apbps2_of_match);
+
+static struct platform_driver apbps2_of_driver = {
+	.driver = {
+		.name = "grlib-apbps2",
+		.of_match_table = apbps2_of_match,
+	},
+	.probe = apbps2_of_probe,
+	.remove = apbps2_of_remove,
+};
+
+module_platform_driver(apbps2_of_driver);
+
+MODULE_AUTHOR("Aeroflex Gaisler AB.");
+MODULE_DESCRIPTION("GRLIB APBPS2 PS/2 serial I/O");
+MODULE_LICENSE("GPL");