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, 169 insertions, 0 deletions

diff --git a/drivers/input/misc/ariel-pwrbutton.c b/drivers/input/misc/ariel-pwrbutton.c
new file mode 100644
index 000000000..f0e06ee60
--- /dev/null
+++ b/drivers/input/misc/ariel-pwrbutton.c
@@ -0,0 +1,169 @@
+// SPDX-License-Identifier: BSD-2-Clause OR GPL-2.0-or-later
+/*
+ * Dell Wyse 3020 a.k.a. "Ariel" Power Button Driver
+ *
+ * Copyright (C) 2020 Lubomir Rintel
+ */
+
+#include <linux/device.h>
+#include <linux/gfp.h>
+#include <linux/input.h>
+#include <linux/interrupt.h>
+#include <linux/module.h>
+#include <linux/spi/spi.h>
+
+#define RESP_COUNTER(response)	(response.header & 0x3)
+#define RESP_SIZE(response)	((response.header >> 2) & 0x3)
+#define RESP_TYPE(response)	((response.header >> 4) & 0xf)
+
+struct ec_input_response {
+	u8 reserved;
+	u8 header;
+	u8 data[3];
+} __packed;
+
+struct ariel_pwrbutton {
+	struct spi_device *client;
+	struct input_dev *input;
+	u8 msg_counter;
+};
+
+static int ec_input_read(struct ariel_pwrbutton *priv,
+			 struct ec_input_response *response)
+{
+	u8 read_request[] = { 0x00, 0x5a, 0xa5, 0x00, 0x00 };
+	struct spi_device *spi = priv->client;
+	struct spi_transfer t = {
+		.tx_buf = read_request,
+		.rx_buf = response,
+		.len = sizeof(read_request),
+	};
+
+	compiletime_assert(sizeof(read_request) == sizeof(*response),
+			   "SPI xfer request/response size mismatch");
+
+	return spi_sync_transfer(spi, &t, 1);
+}
+
+static irqreturn_t ec_input_interrupt(int irq, void *dev_id)
+{
+	struct ariel_pwrbutton *priv = dev_id;
+	struct spi_device *spi = priv->client;
+	struct ec_input_response response;
+	int error;
+	int i;
+
+	error = ec_input_read(priv, &response);
+	if (error < 0) {
+		dev_err(&spi->dev, "EC read failed: %d\n", error);
+		goto out;
+	}
+
+	if (priv->msg_counter == RESP_COUNTER(response)) {
+		dev_warn(&spi->dev, "No new data to read?\n");
+		goto out;
+	}
+
+	priv->msg_counter = RESP_COUNTER(response);
+
+	if (RESP_TYPE(response) != 0x3 && RESP_TYPE(response) != 0xc) {
+		dev_dbg(&spi->dev, "Ignoring message that's not kbd data\n");
+		goto out;
+	}
+
+	for (i = 0; i < RESP_SIZE(response); i++) {
+		switch (response.data[i]) {
+		case 0x74:
+			input_report_key(priv->input, KEY_POWER, 1);
+			input_sync(priv->input);
+			break;
+		case 0xf4:
+			input_report_key(priv->input, KEY_POWER, 0);
+			input_sync(priv->input);
+			break;
+		default:
+			dev_dbg(&spi->dev, "Unknown scan code: %02x\n",
+				response.data[i]);
+		}
+	}
+
+out:
+	return IRQ_HANDLED;
+}
+
+static int ariel_pwrbutton_probe(struct spi_device *spi)
+{
+	struct ec_input_response response;
+	struct ariel_pwrbutton *priv;
+	int error;
+
+	if (!spi->irq) {
+		dev_err(&spi->dev, "Missing IRQ.\n");
+		return -EINVAL;
+	}
+
+	priv = devm_kzalloc(&spi->dev, sizeof(*priv), GFP_KERNEL);
+	if (!priv)
+		return -ENOMEM;
+
+	priv->client = spi;
+	spi_set_drvdata(spi, priv);
+
+	priv->input = devm_input_allocate_device(&spi->dev);
+	if (!priv->input)
+		return -ENOMEM;
+	priv->input->name = "Power Button";
+	priv->input->dev.parent = &spi->dev;
+	input_set_capability(priv->input, EV_KEY, KEY_POWER);
+	error = input_register_device(priv->input);
+	if (error) {
+		dev_err(&spi->dev, "error registering input device: %d\n", error);
+		return error;
+	}
+
+	error = ec_input_read(priv, &response);
+	if (error < 0) {
+		dev_err(&spi->dev, "EC read failed: %d\n", error);
+		return error;
+	}
+	priv->msg_counter = RESP_COUNTER(response);
+
+	error = devm_request_threaded_irq(&spi->dev, spi->irq, NULL,
+					  ec_input_interrupt,
+					  IRQF_ONESHOT,
+					  "Ariel EC Input", priv);
+
+	if (error) {
+		dev_err(&spi->dev, "Failed to request IRQ %d: %d\n",
+			spi->irq, error);
+		return error;
+	}
+
+	return 0;
+}
+
+static const struct of_device_id ariel_pwrbutton_of_match[] = {
+	{ .compatible = "dell,wyse-ariel-ec-input" },
+	{ }
+};
+MODULE_DEVICE_TABLE(of, ariel_pwrbutton_of_match);
+
+static const struct spi_device_id ariel_pwrbutton_spi_ids[] = {
+	{ .name = "wyse-ariel-ec-input" },
+	{ }
+};
+MODULE_DEVICE_TABLE(spi, ariel_pwrbutton_spi_ids);
+
+static struct spi_driver ariel_pwrbutton_driver = {
+	.driver = {
+		.name = "dell-wyse-ariel-ec-input",
+		.of_match_table = ariel_pwrbutton_of_match,
+	},
+	.probe = ariel_pwrbutton_probe,
+	.id_table = ariel_pwrbutton_spi_ids,
+};
+module_spi_driver(ariel_pwrbutton_driver);
+
+MODULE_AUTHOR("Lubomir Rintel <lkundrak@v3.sk>");
+MODULE_DESCRIPTION("Dell Wyse 3020 Power Button Input Driver");
+MODULE_LICENSE("Dual BSD/GPL");