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, 78 insertions, 0 deletions

diff --git a/tools/testing/vsock/timeout.c b/tools/testing/vsock/timeout.c
new file mode 100644
index 000000000..1453d38e0
--- /dev/null
+++ b/tools/testing/vsock/timeout.c
@@ -0,0 +1,78 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/* Timeout API for single-threaded programs that use blocking
+ * syscalls (read/write/send/recv/connect/accept).
+ *
+ * Copyright (C) 2017 Red Hat, Inc.
+ *
+ * Author: Stefan Hajnoczi <stefanha@redhat.com>
+ */
+
+/* Use the following pattern:
+ *
+ *   timeout_begin(TIMEOUT);
+ *   do {
+ *       ret = accept(...);
+ *       timeout_check("accept");
+ *   } while (ret < 0 && ret == EINTR);
+ *   timeout_end();
+ */
+
+#include <stdlib.h>
+#include <stdbool.h>
+#include <unistd.h>
+#include <stdio.h>
+#include <time.h>
+#include "timeout.h"
+
+static volatile bool timeout;
+
+/* SIGALRM handler function.  Do not use sleep(2), alarm(2), or
+ * setitimer(2) while using this API - they may interfere with each
+ * other.
+ *
+ * If you need to sleep, please use timeout_sleep() provided by this API.
+ */
+void sigalrm(int signo)
+{
+	timeout = true;
+}
+
+/* Start a timeout.  Call timeout_check() to verify that the timeout hasn't
+ * expired.  timeout_end() must be called to stop the timeout.  Timeouts cannot
+ * be nested.
+ */
+void timeout_begin(unsigned int seconds)
+{
+	alarm(seconds);
+}
+
+/* Exit with an error message if the timeout has expired */
+void timeout_check(const char *operation)
+{
+	if (timeout) {
+		fprintf(stderr, "%s timed out\n", operation);
+		exit(EXIT_FAILURE);
+	}
+}
+
+/* Stop a timeout */
+void timeout_end(void)
+{
+	alarm(0);
+	timeout = false;
+}
+
+/* Sleep in a timeout section.
+ *
+ * nanosleep(2) can be used with this API since POSIX.1 explicitly
+ * specifies that it does not interact with signals.
+ */
+int timeout_usleep(useconds_t usec)
+{
+	struct timespec ts = {
+		.tv_sec = usec / 1000000,
+		.tv_nsec = (usec % 1000000) * 1000,
+	};
+
+	return nanosleep(&ts, NULL);
+}