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, 129 insertions, 0 deletions

diff --git a/drivers/pinctrl/pinctrl-utils.c b/drivers/pinctrl/pinctrl-utils.c
new file mode 100644
index 000000000..b880e44b8
--- /dev/null
+++ b/drivers/pinctrl/pinctrl-utils.c
@@ -0,0 +1,129 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Utils functions to implement the pincontrol driver.
+ *
+ * Copyright (c) 2013, NVIDIA Corporation.
+ *
+ * Author: Laxman Dewangan <ldewangan@nvidia.com>
+ */
+#include <linux/array_size.h>
+#include <linux/device.h>
+#include <linux/export.h>
+#include <linux/of.h>
+#include <linux/slab.h>
+
+#include <linux/pinctrl/pinctrl.h>
+
+#include "core.h"
+#include "pinctrl-utils.h"
+
+int pinctrl_utils_reserve_map(struct pinctrl_dev *pctldev,
+		struct pinctrl_map **map, unsigned int *reserved_maps,
+		unsigned int *num_maps, unsigned int reserve)
+{
+	unsigned int old_num = *reserved_maps;
+	unsigned int new_num = *num_maps + reserve;
+	struct pinctrl_map *new_map;
+
+	if (old_num >= new_num)
+		return 0;
+
+	new_map = krealloc_array(*map, new_num, sizeof(*new_map), GFP_KERNEL);
+	if (!new_map) {
+		dev_err(pctldev->dev, "krealloc(map) failed\n");
+		return -ENOMEM;
+	}
+
+	memset(new_map + old_num, 0, (new_num - old_num) * sizeof(*new_map));
+
+	*map = new_map;
+	*reserved_maps = new_num;
+	return 0;
+}
+EXPORT_SYMBOL_GPL(pinctrl_utils_reserve_map);
+
+int pinctrl_utils_add_map_mux(struct pinctrl_dev *pctldev,
+		struct pinctrl_map **map, unsigned int *reserved_maps,
+		unsigned int *num_maps, const char *group,
+		const char *function)
+{
+	if (WARN_ON(*num_maps == *reserved_maps))
+		return -ENOSPC;
+
+	(*map)[*num_maps].type = PIN_MAP_TYPE_MUX_GROUP;
+	(*map)[*num_maps].data.mux.group = group;
+	(*map)[*num_maps].data.mux.function = function;
+	(*num_maps)++;
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(pinctrl_utils_add_map_mux);
+
+int pinctrl_utils_add_map_configs(struct pinctrl_dev *pctldev,
+		struct pinctrl_map **map, unsigned int *reserved_maps,
+		unsigned int *num_maps, const char *group,
+		unsigned long *configs, unsigned int num_configs,
+		enum pinctrl_map_type type)
+{
+	unsigned long *dup_configs;
+
+	if (WARN_ON(*num_maps == *reserved_maps))
+		return -ENOSPC;
+
+	dup_configs = kmemdup_array(configs, num_configs,
+				sizeof(*dup_configs), GFP_KERNEL);
+	if (!dup_configs)
+		return -ENOMEM;
+
+	(*map)[*num_maps].type = type;
+	(*map)[*num_maps].data.configs.group_or_pin = group;
+	(*map)[*num_maps].data.configs.configs = dup_configs;
+	(*map)[*num_maps].data.configs.num_configs = num_configs;
+	(*num_maps)++;
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(pinctrl_utils_add_map_configs);
+
+int pinctrl_utils_add_config(struct pinctrl_dev *pctldev,
+		unsigned long **configs, unsigned int *num_configs,
+		unsigned long config)
+{
+	unsigned int old_num = *num_configs;
+	unsigned int new_num = old_num + 1;
+	unsigned long *new_configs;
+
+	new_configs = krealloc(*configs, sizeof(*new_configs) * new_num,
+			       GFP_KERNEL);
+	if (!new_configs) {
+		dev_err(pctldev->dev, "krealloc(configs) failed\n");
+		return -ENOMEM;
+	}
+
+	new_configs[old_num] = config;
+
+	*configs = new_configs;
+	*num_configs = new_num;
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(pinctrl_utils_add_config);
+
+void pinctrl_utils_free_map(struct pinctrl_dev *pctldev,
+	      struct pinctrl_map *map, unsigned int num_maps)
+{
+	int i;
+
+	for (i = 0; i < num_maps; i++) {
+		switch (map[i].type) {
+		case PIN_MAP_TYPE_CONFIGS_GROUP:
+		case PIN_MAP_TYPE_CONFIGS_PIN:
+			kfree(map[i].data.configs.configs);
+			break;
+		default:
+			break;
+		}
+	}
+	kfree(map);
+}
+EXPORT_SYMBOL_GPL(pinctrl_utils_free_map);