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, 113 insertions, 0 deletions

diff --git a/drivers/clk/sprd/common.c b/drivers/clk/sprd/common.c
new file mode 100644
index 000000000..0f2181617
--- /dev/null
+++ b/drivers/clk/sprd/common.c
@@ -0,0 +1,113 @@
+// SPDX-License-Identifier: GPL-2.0
+//
+// Spreadtrum clock infrastructure
+//
+// Copyright (C) 2017 Spreadtrum, Inc.
+// Author: Chunyan Zhang <chunyan.zhang@spreadtrum.com>
+
+#include <linux/mfd/syscon.h>
+#include <linux/module.h>
+#include <linux/of_address.h>
+#include <linux/platform_device.h>
+#include <linux/regmap.h>
+
+#include "common.h"
+
+static const struct regmap_config sprdclk_regmap_config = {
+	.reg_bits	= 32,
+	.reg_stride	= 4,
+	.val_bits	= 32,
+	.fast_io	= true,
+};
+
+static void sprd_clk_set_regmap(const struct sprd_clk_desc *desc,
+			 struct regmap *regmap)
+{
+	int i;
+	struct sprd_clk_common *cclk;
+
+	for (i = 0; i < desc->num_clk_clks; i++) {
+		cclk = desc->clk_clks[i];
+		if (!cclk)
+			continue;
+
+		cclk->regmap = regmap;
+	}
+}
+
+int sprd_clk_regmap_init(struct platform_device *pdev,
+			 const struct sprd_clk_desc *desc)
+{
+	void __iomem *base;
+	struct device *dev = &pdev->dev;
+	struct device_node *node = dev->of_node, *np;
+	struct regmap *regmap;
+	struct resource *res;
+	struct regmap_config reg_config = sprdclk_regmap_config;
+
+	if (of_property_present(node, "sprd,syscon")) {
+		regmap = syscon_regmap_lookup_by_phandle(node, "sprd,syscon");
+		if (IS_ERR(regmap)) {
+			pr_err("%s: failed to get syscon regmap\n", __func__);
+			return PTR_ERR(regmap);
+		}
+	} else if (of_device_is_compatible(np =	of_get_parent(node), "syscon") ||
+		   (of_node_put(np), 0)) {
+		regmap = device_node_to_regmap(np);
+		of_node_put(np);
+		if (IS_ERR(regmap)) {
+			dev_err(dev, "failed to get regmap from its parent.\n");
+			return PTR_ERR(regmap);
+		}
+	} else {
+		base = devm_platform_get_and_ioremap_resource(pdev, 0, &res);
+		if (IS_ERR(base))
+			return PTR_ERR(base);
+
+		reg_config.max_register = resource_size(res) - reg_config.reg_stride;
+
+		regmap = devm_regmap_init_mmio(&pdev->dev, base,
+					       &reg_config);
+		if (IS_ERR(regmap)) {
+			pr_err("failed to init regmap\n");
+			return PTR_ERR(regmap);
+		}
+	}
+
+	sprd_clk_set_regmap(desc, regmap);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(sprd_clk_regmap_init);
+
+int sprd_clk_probe(struct device *dev, struct clk_hw_onecell_data *clkhw)
+{
+	int i, ret;
+	struct clk_hw *hw;
+
+	for (i = 0; i < clkhw->num; i++) {
+		const char *name;
+
+		hw = clkhw->hws[i];
+		if (!hw)
+			continue;
+
+		name = hw->init->name;
+		ret = devm_clk_hw_register(dev, hw);
+		if (ret) {
+			dev_err(dev, "Couldn't register clock %d - %s\n",
+				i, name);
+			return ret;
+		}
+	}
+
+	ret = devm_of_clk_add_hw_provider(dev, of_clk_hw_onecell_get, clkhw);
+	if (ret)
+		dev_err(dev, "Failed to add clock provider\n");
+
+	return ret;
+}
+EXPORT_SYMBOL_GPL(sprd_clk_probe);
+
+MODULE_DESCRIPTION("Spreadtrum clock infrastructure");
+MODULE_LICENSE("GPL v2");