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, 72 insertions, 0 deletions

diff --git a/drivers/clk/mediatek/reset.h b/drivers/clk/mediatek/reset.h
new file mode 100644
index 000000000..562ffd290
--- /dev/null
+++ b/drivers/clk/mediatek/reset.h
@@ -0,0 +1,72 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/*
+ * Copyright (c) 2022 MediaTek Inc.
+ */
+
+#ifndef __DRV_CLK_MTK_RESET_H
+#define __DRV_CLK_MTK_RESET_H
+
+#include <linux/reset-controller.h>
+#include <linux/types.h>
+
+#define RST_NR_PER_BANK 32
+
+/* Infra global controller reset set register */
+#define INFRA_RST0_SET_OFFSET 0x120
+#define INFRA_RST1_SET_OFFSET 0x130
+#define INFRA_RST2_SET_OFFSET 0x140
+#define INFRA_RST3_SET_OFFSET 0x150
+#define INFRA_RST4_SET_OFFSET 0x730
+
+/**
+ * enum mtk_reset_version - Version of MediaTek clock reset controller.
+ * @MTK_RST_SIMPLE: Use the same registers for bit set and clear.
+ * @MTK_RST_SET_CLR: Use separate registers for bit set and clear.
+ * @MTK_RST_MAX: Total quantity of version for MediaTek clock reset controller.
+ */
+enum mtk_reset_version {
+	MTK_RST_SIMPLE = 0,
+	MTK_RST_SET_CLR,
+	MTK_RST_MAX,
+};
+
+/**
+ * struct mtk_clk_rst_desc - Description of MediaTek clock reset.
+ * @version: Reset version which is defined in enum mtk_reset_version.
+ * @rst_bank_ofs: Pointer to an array containing base offsets of the reset register.
+ * @rst_bank_nr: Quantity of reset bank.
+ * @rst_idx_map:Pointer to an array containing ids if input argument is index.
+ *		This array is not necessary if our input argument does not mean index.
+ * @rst_idx_map_nr: Quantity of reset index map.
+ */
+struct mtk_clk_rst_desc {
+	enum mtk_reset_version version;
+	u16 *rst_bank_ofs;
+	u32 rst_bank_nr;
+	u16 *rst_idx_map;
+	u32 rst_idx_map_nr;
+};
+
+/**
+ * struct mtk_clk_rst_data - Data of MediaTek clock reset controller.
+ * @regmap: Pointer to base address of reset register address.
+ * @rcdev: Reset controller device.
+ * @desc: Pointer to description of the reset controller.
+ */
+struct mtk_clk_rst_data {
+	struct regmap *regmap;
+	struct reset_controller_dev rcdev;
+	const struct mtk_clk_rst_desc *desc;
+};
+
+/**
+ * mtk_register_reset_controller - Register mediatek clock reset controller with device
+ * @np: Pointer to device.
+ * @desc: Constant pointer to description of clock reset.
+ *
+ * Return: 0 on success and errorno otherwise.
+ */
+int mtk_register_reset_controller_with_dev(struct device *dev,
+					   const struct mtk_clk_rst_desc *desc);
+
+#endif /* __DRV_CLK_MTK_RESET_H */