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, 51 insertions, 0 deletions

diff --git a/drivers/video/of_videomode.c b/drivers/video/of_videomode.c
new file mode 100644
index 000000000..e7d10ffd3
--- /dev/null
+++ b/drivers/video/of_videomode.c
@@ -0,0 +1,51 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * generic videomode helper
+ *
+ * Copyright (c) 2012 Steffen Trumtrar <s.trumtrar@pengutronix.de>, Pengutronix
+ */
+#include <linux/errno.h>
+#include <linux/export.h>
+#include <linux/of.h>
+#include <video/display_timing.h>
+#include <video/of_display_timing.h>
+#include <video/of_videomode.h>
+#include <video/videomode.h>
+
+/**
+ * of_get_videomode - get the videomode #<index> from devicetree
+ * @np: devicenode with the display_timings
+ * @vm: set to return value
+ * @index: index into list of display_timings
+ *	    (Set this to OF_USE_NATIVE_MODE to use whatever mode is
+ *	     specified as native mode in the DT.)
+ *
+ * DESCRIPTION:
+ * Get a list of all display timings and put the one
+ * specified by index into *vm. This function should only be used, if
+ * only one videomode is to be retrieved. A driver that needs to work
+ * with multiple/all videomodes should work with
+ * of_get_display_timings instead.
+ **/
+int of_get_videomode(struct device_node *np, struct videomode *vm,
+		     int index)
+{
+	struct display_timings *disp;
+	int ret;
+
+	disp = of_get_display_timings(np);
+	if (!disp) {
+		pr_err("%pOF: no timings specified\n", np);
+		return -EINVAL;
+	}
+
+	if (index == OF_USE_NATIVE_MODE)
+		index = disp->native_mode;
+
+	ret = videomode_from_timings(disp, vm, index);
+
+	display_timings_release(disp);
+
+	return ret;
+}
+EXPORT_SYMBOL_GPL(of_get_videomode);