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, 67 insertions, 0 deletions

diff --git a/arch/powerpc/boot/util.S b/arch/powerpc/boot/util.S
new file mode 100644
index 000000000..6a92376da
--- /dev/null
+++ b/arch/powerpc/boot/util.S
@@ -0,0 +1,67 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later */
+/*
+ * Copied from <file:arch/powerpc/kernel/misc_32.S>
+ *
+ * This file contains miscellaneous low-level functions.
+ *    Copyright (C) 1995-1996 Gary Thomas (gdt@linuxppc.org)
+ *
+ * Largely rewritten by Cort Dougan (cort@cs.nmt.edu)
+ * and Paul Mackerras.
+ *
+ * kexec bits:
+ * Copyright (C) 2002-2003 Eric Biederman  <ebiederm@xmission.com>
+ * GameCube/ppc32 port Copyright (C) 2004 Albert Herranz
+ */
+#include "ppc_asm.h"
+
+#define SPRN_PVR        0x11F   /* Processor Version Register */
+
+	.text
+
+/* udelay needs to know the period of the
+ * timebase in nanoseconds.  This used to be hardcoded to be 60ns
+ * (period of 66MHz/4).  Now a variable is used that is initialized to
+ * 60 for backward compatibility, but it can be overridden as necessary
+ * with code something like this:
+ *    extern unsigned long timebase_period_ns;
+ *    timebase_period_ns = 1000000000 / bd->bi_tbfreq;
+ */
+	.data
+	.globl timebase_period_ns
+timebase_period_ns:
+	.long	60
+
+	.text
+/*
+ * Delay for a number of microseconds
+ */
+	.globl	udelay
+udelay:
+	mulli	r4,r3,1000	/* nanoseconds */
+	/*  Change r4 to be the number of ticks using:
+	 *	(nanoseconds + (timebase_period_ns - 1 )) / timebase_period_ns
+	 *  timebase_period_ns defaults to 60 (16.6MHz) */
+	mflr	r5
+	bcl	20,31,0f
+0:	mflr	r6
+	mtlr	r5
+	addis	r5,r6,(timebase_period_ns-0b)@ha
+	lwz	r5,(timebase_period_ns-0b)@l(r5)
+	add	r4,r4,r5
+	addi	r4,r4,-1
+	divw	r4,r4,r5	/* BUS ticks */
+1:	MFTBU(r5)
+	MFTBL(r6)
+	MFTBU(r7)
+	cmpw	0,r5,r7
+	bne	1b		/* Get [synced] base time */
+	addc	r9,r6,r4	/* Compute end time */
+	addze	r8,r5
+2:	MFTBU(r5)
+	cmpw	0,r5,r8
+	blt	2b
+	bgt	3f
+	MFTBL(r6)
+	cmpw	0,r6,r9
+	blt	2b
+3:	blr