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, 107 insertions, 0 deletions

diff --git a/drivers/md/dm-path-selector.h b/drivers/md/dm-path-selector.h
new file mode 100644
index 000000000..7b2270532
--- /dev/null
+++ b/drivers/md/dm-path-selector.h
@@ -0,0 +1,107 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/*
+ * Copyright (C) 2003 Sistina Software.
+ * Copyright (C) 2004 Red Hat, Inc. All rights reserved.
+ *
+ * Module Author: Heinz Mauelshagen
+ *
+ * This file is released under the GPL.
+ *
+ * Path-Selector registration.
+ */
+
+#ifndef	DM_PATH_SELECTOR_H
+#define	DM_PATH_SELECTOR_H
+
+#include <linux/device-mapper.h>
+
+#include "dm-mpath.h"
+
+/*
+ * We provide an abstraction for the code that chooses which path
+ * to send some io down.
+ */
+struct path_selector_type;
+struct path_selector {
+	struct path_selector_type *type;
+	void *context;
+};
+
+/*
+ * If a path selector uses this flag, a high resolution timer is used
+ * (via ktime_get_ns) to account for IO start time in BIO-based mpath.
+ * This improves performance of some path selectors (i.e. HST), in
+ * exchange for slightly higher overhead when submitting the BIO.
+ * The extra cost is usually offset by improved path selection for
+ * some benchmarks.
+ *
+ * This has no effect for request-based mpath, since it already uses a
+ * higher precision timer by default.
+ */
+#define DM_PS_USE_HR_TIMER		0x00000001
+#define dm_ps_use_hr_timer(type)	((type)->features & DM_PS_USE_HR_TIMER)
+
+/* Information about a path selector type */
+struct path_selector_type {
+	char *name;
+	struct module *module;
+
+	unsigned int features;
+	unsigned int table_args;
+	unsigned int info_args;
+
+	/*
+	 * Constructs a path selector object, takes custom arguments
+	 */
+	int (*create)(struct path_selector *ps, unsigned int argc, char **argv);
+	void (*destroy)(struct path_selector *ps);
+
+	/*
+	 * Add an opaque path object, along with some selector specific
+	 * path args (eg, path priority).
+	 */
+	int (*add_path)(struct path_selector *ps, struct dm_path *path,
+			int argc, char **argv, char **error);
+
+	/*
+	 * Chooses a path for this io, if no paths are available then
+	 * NULL will be returned.
+	 */
+	struct dm_path *(*select_path)(struct path_selector *ps, size_t nr_bytes);
+
+	/*
+	 * Notify the selector that a path has failed.
+	 */
+	void (*fail_path)(struct path_selector *ps, struct dm_path *p);
+
+	/*
+	 * Ask selector to reinstate a path.
+	 */
+	int (*reinstate_path)(struct path_selector *ps, struct dm_path *p);
+
+	/*
+	 * Table content based on parameters added in ps_add_path_fn
+	 * or path selector status
+	 */
+	int (*status)(struct path_selector *ps, struct dm_path *path,
+		      status_type_t type, char *result, unsigned int maxlen);
+
+	int (*start_io)(struct path_selector *ps, struct dm_path *path,
+			size_t nr_bytes);
+	int (*end_io)(struct path_selector *ps, struct dm_path *path,
+		      size_t nr_bytes, u64 start_time);
+};
+
+/* Register a path selector */
+int dm_register_path_selector(struct path_selector_type *type);
+
+/* Unregister a path selector */
+void dm_unregister_path_selector(struct path_selector_type *type);
+
+/* Returns a registered path selector type */
+struct path_selector_type *dm_get_path_selector(const char *name);
+
+/* Releases a path selector  */
+void dm_put_path_selector(struct path_selector_type *pst);
+
+#endif