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, 174 insertions, 0 deletions

diff --git a/drivers/clk/mediatek/reset.c b/drivers/clk/mediatek/reset.c
new file mode 100644
index 000000000..2e3303975
--- /dev/null
+++ b/drivers/clk/mediatek/reset.c
@@ -0,0 +1,174 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (c) 2014 MediaTek Inc.
+ */
+
+#include <linux/mfd/syscon.h>
+#include <linux/module.h>
+#include <linux/of.h>
+#include <linux/platform_device.h>
+#include <linux/regmap.h>
+#include <linux/slab.h>
+
+#include "reset.h"
+
+static inline struct mtk_clk_rst_data *to_mtk_clk_rst_data(struct reset_controller_dev *rcdev)
+{
+	return container_of(rcdev, struct mtk_clk_rst_data, rcdev);
+}
+
+static int mtk_reset_update(struct reset_controller_dev *rcdev,
+			    unsigned long id, bool deassert)
+{
+	struct mtk_clk_rst_data *data = to_mtk_clk_rst_data(rcdev);
+	unsigned int val = deassert ? 0 : ~0;
+
+	return regmap_update_bits(data->regmap,
+				  data->desc->rst_bank_ofs[id / RST_NR_PER_BANK],
+				  BIT(id % RST_NR_PER_BANK), val);
+}
+
+static int mtk_reset_assert(struct reset_controller_dev *rcdev,
+			    unsigned long id)
+{
+	return mtk_reset_update(rcdev, id, false);
+}
+
+static int mtk_reset_deassert(struct reset_controller_dev *rcdev,
+			      unsigned long id)
+{
+	return mtk_reset_update(rcdev, id, true);
+}
+
+static int mtk_reset(struct reset_controller_dev *rcdev, unsigned long id)
+{
+	int ret;
+
+	ret = mtk_reset_assert(rcdev, id);
+	if (ret)
+		return ret;
+
+	return mtk_reset_deassert(rcdev, id);
+}
+
+static int mtk_reset_update_set_clr(struct reset_controller_dev *rcdev,
+				    unsigned long id, bool deassert)
+{
+	struct mtk_clk_rst_data *data = to_mtk_clk_rst_data(rcdev);
+	unsigned int deassert_ofs = deassert ? 0x4 : 0;
+
+	return regmap_write(data->regmap,
+			    data->desc->rst_bank_ofs[id / RST_NR_PER_BANK] +
+			    deassert_ofs,
+			    BIT(id % RST_NR_PER_BANK));
+}
+
+static int mtk_reset_assert_set_clr(struct reset_controller_dev *rcdev,
+				    unsigned long id)
+{
+	return mtk_reset_update_set_clr(rcdev, id, false);
+}
+
+static int mtk_reset_deassert_set_clr(struct reset_controller_dev *rcdev,
+				      unsigned long id)
+{
+	return mtk_reset_update_set_clr(rcdev, id, true);
+}
+
+static int mtk_reset_set_clr(struct reset_controller_dev *rcdev,
+			     unsigned long id)
+{
+	int ret;
+
+	ret = mtk_reset_assert_set_clr(rcdev, id);
+	if (ret)
+		return ret;
+	return mtk_reset_deassert_set_clr(rcdev, id);
+}
+
+static const struct reset_control_ops mtk_reset_ops = {
+	.assert = mtk_reset_assert,
+	.deassert = mtk_reset_deassert,
+	.reset = mtk_reset,
+};
+
+static const struct reset_control_ops mtk_reset_ops_set_clr = {
+	.assert = mtk_reset_assert_set_clr,
+	.deassert = mtk_reset_deassert_set_clr,
+	.reset = mtk_reset_set_clr,
+};
+
+static int reset_xlate(struct reset_controller_dev *rcdev,
+		       const struct of_phandle_args *reset_spec)
+{
+	struct mtk_clk_rst_data *data = to_mtk_clk_rst_data(rcdev);
+
+	if (reset_spec->args[0] >= rcdev->nr_resets ||
+	    reset_spec->args[0] >= data->desc->rst_idx_map_nr)
+		return -EINVAL;
+
+	return data->desc->rst_idx_map[reset_spec->args[0]];
+}
+
+int mtk_register_reset_controller_with_dev(struct device *dev,
+					   const struct mtk_clk_rst_desc *desc)
+{
+	struct device_node *np = dev->of_node;
+	struct regmap *regmap;
+	const struct reset_control_ops *rcops = NULL;
+	struct mtk_clk_rst_data *data;
+	int ret;
+
+	if (!desc) {
+		dev_err(dev, "mtk clock reset desc is NULL\n");
+		return -EINVAL;
+	}
+
+	switch (desc->version) {
+	case MTK_RST_SIMPLE:
+		rcops = &mtk_reset_ops;
+		break;
+	case MTK_RST_SET_CLR:
+		rcops = &mtk_reset_ops_set_clr;
+		break;
+	default:
+		dev_err(dev, "Unknown reset version %d\n", desc->version);
+		return -EINVAL;
+	}
+
+	regmap = device_node_to_regmap(np);
+	if (IS_ERR(regmap)) {
+		dev_err(dev, "Cannot find regmap %pe\n", regmap);
+		return PTR_ERR(regmap);
+	}
+
+	data = devm_kzalloc(dev, sizeof(*data), GFP_KERNEL);
+	if (!data)
+		return -ENOMEM;
+
+	data->desc = desc;
+	data->regmap = regmap;
+	data->rcdev.owner = THIS_MODULE;
+	data->rcdev.ops = rcops;
+	data->rcdev.of_node = np;
+	data->rcdev.dev = dev;
+
+	if (data->desc->rst_idx_map_nr > 0) {
+		data->rcdev.of_reset_n_cells = 1;
+		data->rcdev.nr_resets = desc->rst_idx_map_nr;
+		data->rcdev.of_xlate = reset_xlate;
+	} else {
+		data->rcdev.nr_resets = desc->rst_bank_nr * RST_NR_PER_BANK;
+	}
+
+	ret = devm_reset_controller_register(dev, &data->rcdev);
+	if (ret) {
+		dev_err(dev, "could not register reset controller: %d\n", ret);
+		return ret;
+	}
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mtk_register_reset_controller_with_dev);
+
+MODULE_LICENSE("GPL");