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, 55 insertions, 0 deletions

diff --git a/tools/perf/python/tracepoint.py b/tools/perf/python/tracepoint.py
new file mode 100755
index 000000000..15b0c8268
--- /dev/null
+++ b/tools/perf/python/tracepoint.py
@@ -0,0 +1,55 @@
+#! /usr/bin/env python
+# SPDX-License-Identifier: GPL-2.0
+# -*- python -*-
+# -*- coding: utf-8 -*-
+
+import perf
+
+def change_proctitle():
+    try:
+        import setproctitle
+        setproctitle.setproctitle("tracepoint.py")
+    except:
+        print("Install the setproctitle python package to help with top and friends")
+
+def main():
+    change_proctitle()
+    cpus    = perf.cpu_map()
+    threads = perf.thread_map(-1)
+    evlist = perf.parse_events("sched:sched_switch", cpus, threads)
+    # Disable tracking of mmaps and similar that are unnecessary.
+    for ev in evlist:
+        ev.tracking = False
+    # Configure evsels with default record options.
+    evlist.config()
+    # Simplify the sample_type and read_format of evsels
+    for ev in evlist:
+        ev.sample_type = ev.sample_type & ~perf.SAMPLE_IP
+        ev.read_format = 0
+
+    evlist.open()
+    evlist.mmap()
+    evlist.enable();
+
+    while True:
+        evlist.poll(timeout = -1)
+        for cpu in cpus:
+            event = evlist.read_on_cpu(cpu)
+            if not event:
+                continue
+
+            if not isinstance(event, perf.sample_event):
+                continue
+
+            print("time %u prev_comm=%s prev_pid=%d prev_prio=%d prev_state=0x%x ==> next_comm=%s next_pid=%d next_prio=%d" % (
+                   event.sample_time,
+                   event.prev_comm,
+                   event.prev_pid,
+                   event.prev_prio,
+                   event.prev_state,
+                   event.next_comm,
+                   event.next_pid,
+                   event.next_prio))
+
+if __name__ == '__main__':
+    main()