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, 167 insertions, 0 deletions

diff --git a/drivers/clk/qcom/hfpll.c b/drivers/clk/qcom/hfpll.c
new file mode 100644
index 000000000..385964196
--- /dev/null
+++ b/drivers/clk/qcom/hfpll.c
@@ -0,0 +1,167 @@
+// SPDX-License-Identifier: GPL-2.0
+// Copyright (c) 2018, The Linux Foundation. All rights reserved.
+
+#include <linux/kernel.h>
+#include <linux/init.h>
+#include <linux/module.h>
+#include <linux/platform_device.h>
+#include <linux/of.h>
+#include <linux/of_device.h>
+#include <linux/clk.h>
+#include <linux/clk-provider.h>
+#include <linux/regmap.h>
+
+#include "clk-regmap.h"
+#include "clk-hfpll.h"
+
+static const struct hfpll_data qcs404 = {
+	.mode_reg = 0x00,
+	.l_reg = 0x04,
+	.m_reg = 0x08,
+	.n_reg = 0x0c,
+	.user_reg = 0x10,
+	.config_reg = 0x14,
+	.config_val = 0x430405d,
+	.status_reg = 0x1c,
+	.lock_bit = 16,
+
+	.user_val = 0x8,
+	.user_vco_mask = 0x100000,
+	.low_vco_max_rate = 1248000000,
+	.min_rate = 537600000UL,
+	.max_rate = 2900000000UL,
+};
+
+static const struct hfpll_data msm8976_a53 = {
+	.mode_reg = 0x00,
+	.l_reg = 0x04,
+	.m_reg = 0x08,
+	.n_reg = 0x0c,
+	.user_reg = 0x10,
+	.config_reg = 0x14,
+	.config_val = 0x341600,
+	.status_reg = 0x1c,
+	.lock_bit = 16,
+
+	.l_val = 0x35,
+	.user_val = 0x109,
+	.min_rate = 902400000UL,
+	.max_rate = 1478400000UL,
+};
+
+static const struct hfpll_data msm8976_a72 = {
+	.mode_reg = 0x00,
+	.l_reg = 0x04,
+	.m_reg = 0x08,
+	.n_reg = 0x0c,
+	.user_reg = 0x10,
+	.config_reg = 0x14,
+	.config_val = 0x4e0405d,
+	.status_reg = 0x1c,
+	.lock_bit = 16,
+
+	.l_val = 0x3e,
+	.user_val = 0x100109,
+	.min_rate = 940800000UL,
+	.max_rate = 2016000000UL,
+};
+
+static const struct hfpll_data msm8976_cci = {
+	.mode_reg = 0x00,
+	.l_reg = 0x04,
+	.m_reg = 0x08,
+	.n_reg = 0x0c,
+	.user_reg = 0x10,
+	.config_reg = 0x14,
+	.config_val = 0x141400,
+	.status_reg = 0x1c,
+	.lock_bit = 16,
+
+	.l_val = 0x20,
+	.user_val = 0x100109,
+	.min_rate = 556800000UL,
+	.max_rate = 902400000UL,
+};
+
+static const struct of_device_id qcom_hfpll_match_table[] = {
+	{ .compatible = "qcom,msm8976-hfpll-a53", .data = &msm8976_a53 },
+	{ .compatible = "qcom,msm8976-hfpll-a72", .data = &msm8976_a72 },
+	{ .compatible = "qcom,msm8976-hfpll-cci", .data = &msm8976_cci },
+	{ .compatible = "qcom,qcs404-hfpll", .data = &qcs404 },
+	/* Deprecated in bindings */
+	{ .compatible = "qcom,hfpll", .data = &qcs404 },
+	{ }
+};
+MODULE_DEVICE_TABLE(of, qcom_hfpll_match_table);
+
+static const struct regmap_config hfpll_regmap_config = {
+	.reg_bits	= 32,
+	.reg_stride	= 4,
+	.val_bits	= 32,
+	.max_register	= 0x30,
+};
+
+static int qcom_hfpll_probe(struct platform_device *pdev)
+{
+	struct device *dev = &pdev->dev;
+	void __iomem *base;
+	struct regmap *regmap;
+	struct clk_hfpll *h;
+	struct clk_init_data init = {
+		.num_parents = 1,
+		.ops = &clk_ops_hfpll,
+		/*
+		 * rather than marking the clock critical and forcing the clock
+		 * to be always enabled, we make sure that the clock is not
+		 * disabled: the firmware remains responsible of enabling this
+		 * clock (for more info check the commit log)
+		 */
+		.flags = CLK_IGNORE_UNUSED,
+	};
+	int ret;
+	struct clk_parent_data pdata = { .index = 0 };
+
+	h = devm_kzalloc(dev, sizeof(*h), GFP_KERNEL);
+	if (!h)
+		return -ENOMEM;
+
+	base = devm_platform_get_and_ioremap_resource(pdev, 0, NULL);
+	if (IS_ERR(base))
+		return PTR_ERR(base);
+
+	regmap = devm_regmap_init_mmio(&pdev->dev, base, &hfpll_regmap_config);
+	if (IS_ERR(regmap))
+		return PTR_ERR(regmap);
+
+	if (of_property_read_string_index(dev->of_node, "clock-output-names",
+					  0, &init.name))
+		return -ENODEV;
+
+	init.parent_data = &pdata;
+
+	h->d = of_device_get_match_data(&pdev->dev);
+	h->clkr.hw.init = &init;
+	spin_lock_init(&h->lock);
+
+	ret = devm_clk_register_regmap(dev, &h->clkr);
+	if (ret) {
+		dev_err(dev, "failed to register regmap clock: %d\n", ret);
+		return ret;
+	}
+
+	return devm_of_clk_add_hw_provider(dev, of_clk_hw_simple_get,
+					   &h->clkr.hw);
+}
+
+static struct platform_driver qcom_hfpll_driver = {
+	.probe		= qcom_hfpll_probe,
+	.driver		= {
+		.name	= "qcom-hfpll",
+		.of_match_table = qcom_hfpll_match_table,
+	},
+};
+module_platform_driver(qcom_hfpll_driver);
+
+MODULE_DESCRIPTION("QCOM HFPLL Clock Driver");
+MODULE_LICENSE("GPL v2");
+MODULE_ALIAS("platform:qcom-hfpll");