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, 144 insertions, 0 deletions

diff --git a/drivers/acpi/ec_sys.c b/drivers/acpi/ec_sys.c
new file mode 100644
index 000000000..c074a0fae
--- /dev/null
+++ b/drivers/acpi/ec_sys.c
@@ -0,0 +1,144 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * ec_sys.c
+ *
+ * Copyright (C) 2010 SUSE Products GmbH/Novell
+ * Author:
+ *      Thomas Renninger <trenn@suse.de>
+ */
+
+#include <linux/kernel.h>
+#include <linux/acpi.h>
+#include <linux/debugfs.h>
+#include <linux/module.h>
+#include <linux/uaccess.h>
+#include "internal.h"
+
+MODULE_AUTHOR("Thomas Renninger <trenn@suse.de>");
+MODULE_DESCRIPTION("ACPI EC sysfs access driver");
+MODULE_LICENSE("GPL");
+
+static bool write_support;
+module_param_hw(write_support, bool, other, 0644);
+MODULE_PARM_DESC(write_support, "Dangerous, reboot and removal of battery may "
+		 "be needed.");
+
+#define EC_SPACE_SIZE 256
+
+static struct dentry *acpi_ec_debugfs_dir;
+
+static ssize_t acpi_ec_read_io(struct file *f, char __user *buf,
+			       size_t count, loff_t *off)
+{
+	/* Use this if support reading/writing multiple ECs exists in ec.c:
+	 * struct acpi_ec *ec = ((struct seq_file *)f->private_data)->private;
+	 */
+	unsigned int size = EC_SPACE_SIZE;
+	loff_t init_off = *off;
+	int err = 0;
+
+	if (*off >= size)
+		return 0;
+	if (*off + count >= size) {
+		size -= *off;
+		count = size;
+	} else
+		size = count;
+
+	while (size) {
+		u8 byte_read;
+		err = ec_read(*off, &byte_read);
+		if (err)
+			return err;
+		if (put_user(byte_read, buf + *off - init_off)) {
+			if (*off - init_off)
+				return *off - init_off; /* partial read */
+			return -EFAULT;
+		}
+		*off += 1;
+		size--;
+	}
+	return count;
+}
+
+static ssize_t acpi_ec_write_io(struct file *f, const char __user *buf,
+				size_t count, loff_t *off)
+{
+	/* Use this if support reading/writing multiple ECs exists in ec.c:
+	 * struct acpi_ec *ec = ((struct seq_file *)f->private_data)->private;
+	 */
+
+	unsigned int size = count;
+	loff_t init_off = *off;
+	int err = 0;
+
+	if (!write_support)
+		return -EINVAL;
+
+	if (*off >= EC_SPACE_SIZE)
+		return 0;
+	if (*off + count >= EC_SPACE_SIZE) {
+		size = EC_SPACE_SIZE - *off;
+		count = size;
+	}
+
+	while (size) {
+		u8 byte_write;
+		if (get_user(byte_write, buf + *off - init_off)) {
+			if (*off - init_off)
+				return *off - init_off; /* partial write */
+			return -EFAULT;
+		}
+		err = ec_write(*off, byte_write);
+		if (err)
+			return err;
+
+		*off += 1;
+		size--;
+	}
+	return count;
+}
+
+static const struct file_operations acpi_ec_io_ops = {
+	.owner = THIS_MODULE,
+	.open  = simple_open,
+	.read  = acpi_ec_read_io,
+	.write = acpi_ec_write_io,
+	.llseek = default_llseek,
+};
+
+static void acpi_ec_add_debugfs(struct acpi_ec *ec, unsigned int ec_device_count)
+{
+	struct dentry *dev_dir;
+	char name[64];
+	umode_t mode = 0400;
+
+	if (ec_device_count == 0)
+		acpi_ec_debugfs_dir = debugfs_create_dir("ec", NULL);
+
+	sprintf(name, "ec%u", ec_device_count);
+	dev_dir = debugfs_create_dir(name, acpi_ec_debugfs_dir);
+
+	debugfs_create_x32("gpe", 0444, dev_dir, &first_ec->gpe);
+	debugfs_create_bool("use_global_lock", 0444, dev_dir,
+			    &first_ec->global_lock);
+
+	if (write_support)
+		mode = 0600;
+	debugfs_create_file("io", mode, dev_dir, ec, &acpi_ec_io_ops);
+}
+
+static int __init acpi_ec_sys_init(void)
+{
+	if (first_ec)
+		acpi_ec_add_debugfs(first_ec, 0);
+	return 0;
+}
+
+static void __exit acpi_ec_sys_exit(void)
+{
+	debugfs_remove_recursive(acpi_ec_debugfs_dir);
+}
+
+module_init(acpi_ec_sys_init);
+module_exit(acpi_ec_sys_exit);