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, 133 insertions, 0 deletions

diff --git a/samples/livepatch/livepatch-shadow-fix2.c b/samples/livepatch/livepatch-shadow-fix2.c
new file mode 100644
index 000000000..b99122cb2
--- /dev/null
+++ b/samples/livepatch/livepatch-shadow-fix2.c
@@ -0,0 +1,133 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (C) 2017 Joe Lawrence <joe.lawrence@redhat.com>
+ */
+
+/*
+ * livepatch-shadow-fix2.c - Shadow variables, livepatch demo
+ *
+ * Purpose
+ * -------
+ *
+ * Adds functionality to livepatch-shadow-mod's in-flight data
+ * structures through a shadow variable.  The livepatch patches a
+ * routine that periodically inspects data structures, incrementing a
+ * per-data-structure counter, creating the counter if needed.
+ *
+ *
+ * Usage
+ * -----
+ *
+ * This module is not intended to be standalone.  See the "Usage"
+ * section of livepatch-shadow-mod.c.
+ */
+
+#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
+
+#include <linux/module.h>
+#include <linux/kernel.h>
+#include <linux/livepatch.h>
+#include <linux/slab.h>
+
+/* Shadow variable enums */
+#define SV_LEAK		1
+#define SV_COUNTER	2
+
+struct dummy {
+	struct list_head list;
+	unsigned long jiffies_expire;
+};
+
+static bool livepatch_fix2_dummy_check(struct dummy *d, unsigned long jiffies)
+{
+	int *shadow_count;
+
+	/*
+	 * Patch: handle in-flight dummy structures, if they do not
+	 * already have a SV_COUNTER shadow variable, then attach a
+	 * new one.
+	 */
+	shadow_count = klp_shadow_get_or_alloc(d, SV_COUNTER,
+				sizeof(*shadow_count), GFP_NOWAIT,
+				NULL, NULL);
+	if (shadow_count)
+		*shadow_count += 1;
+
+	return time_after(jiffies, d->jiffies_expire);
+}
+
+static void livepatch_fix2_dummy_leak_dtor(void *obj, void *shadow_data)
+{
+	void *d = obj;
+	int **shadow_leak = shadow_data;
+
+	pr_info("%s: dummy @ %p, prevented leak @ %p\n",
+			 __func__, d, *shadow_leak);
+	kfree(*shadow_leak);
+}
+
+static void livepatch_fix2_dummy_free(struct dummy *d)
+{
+	int **shadow_leak;
+	int *shadow_count;
+
+	/* Patch: copy the memory leak patch from the fix1 module. */
+	shadow_leak = klp_shadow_get(d, SV_LEAK);
+	if (shadow_leak)
+		klp_shadow_free(d, SV_LEAK, livepatch_fix2_dummy_leak_dtor);
+	else
+		pr_info("%s: dummy @ %p leaked!\n", __func__, d);
+
+	/*
+	 * Patch: fetch the SV_COUNTER shadow variable and display
+	 * the final count.  Detach the shadow variable.
+	 */
+	shadow_count = klp_shadow_get(d, SV_COUNTER);
+	if (shadow_count) {
+		pr_info("%s: dummy @ %p, check counter = %d\n",
+			__func__, d, *shadow_count);
+		klp_shadow_free(d, SV_COUNTER, NULL);
+	}
+
+	kfree(d);
+}
+
+static struct klp_func funcs[] = {
+	{
+		.old_name = "dummy_check",
+		.new_func = livepatch_fix2_dummy_check,
+	},
+	{
+		.old_name = "dummy_free",
+		.new_func = livepatch_fix2_dummy_free,
+	}, { }
+};
+
+static struct klp_object objs[] = {
+	{
+		.name = "livepatch_shadow_mod",
+		.funcs = funcs,
+	}, { }
+};
+
+static struct klp_patch patch = {
+	.mod = THIS_MODULE,
+	.objs = objs,
+};
+
+static int livepatch_shadow_fix2_init(void)
+{
+	return klp_enable_patch(&patch);
+}
+
+static void livepatch_shadow_fix2_exit(void)
+{
+	/* Cleanup any existing SV_COUNTER shadow variables */
+	klp_shadow_free_all(SV_COUNTER, NULL);
+}
+
+module_init(livepatch_shadow_fix2_init);
+module_exit(livepatch_shadow_fix2_exit);
+MODULE_DESCRIPTION("Live patching demo for shadow variables");
+MODULE_LICENSE("GPL");
+MODULE_INFO(livepatch, "Y");