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, 82 insertions, 0 deletions

diff --git a/arch/mips/lib/uncached.c b/arch/mips/lib/uncached.c
new file mode 100644
index 000000000..f80a67c09
--- /dev/null
+++ b/arch/mips/lib/uncached.c
@@ -0,0 +1,82 @@
+/*
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License.  See the file "COPYING" in the main directory of this archive
+ * for more details.
+ *
+ * Copyright (C) 2005 Thiemo Seufer
+ * Copyright (C) 2005  MIPS Technologies, Inc.	All rights reserved.
+ *	Author: Maciej W. Rozycki <macro@mips.com>
+ */
+
+
+#include <asm/addrspace.h>
+#include <asm/bug.h>
+#include <asm/cacheflush.h>
+
+#ifndef CKSEG2
+#define CKSEG2 CKSSEG
+#endif
+#ifndef TO_PHYS_MASK
+#define TO_PHYS_MASK -1
+#endif
+
+/*
+ * FUNC is executed in one of the uncached segments, depending on its
+ * original address as follows:
+ *
+ * 1. If the original address is in CKSEG0 or CKSEG1, then the uncached
+ *    segment used is CKSEG1.
+ * 2. If the original address is in XKPHYS, then the uncached segment
+ *    used is XKPHYS(2).
+ * 3. Otherwise it's a bug.
+ *
+ * The same remapping is done with the stack pointer.  Stack handling
+ * works because we don't handle stack arguments or more complex return
+ * values, so we can avoid sharing the same stack area between a cached
+ * and the uncached mode.
+ */
+unsigned long run_uncached(void *func)
+{
+	register long ret __asm__("$2");
+	long lfunc = (long)func, ufunc;
+	long usp;
+	long sp;
+
+	__asm__("move %0, $sp" : "=r" (sp));
+
+	if (sp >= (long)CKSEG0 && sp < (long)CKSEG2)
+		usp = CKSEG1ADDR(sp);
+#ifdef CONFIG_64BIT
+	else if ((long long)sp >= (long long)PHYS_TO_XKPHYS(0, 0) &&
+		 (long long)sp < (long long)PHYS_TO_XKPHYS(8, 0))
+		usp = PHYS_TO_XKPHYS(K_CALG_UNCACHED,
+				     XKPHYS_TO_PHYS((long long)sp));
+#endif
+	else {
+		BUG();
+		usp = sp;
+	}
+	if (lfunc >= (long)CKSEG0 && lfunc < (long)CKSEG2)
+		ufunc = CKSEG1ADDR(lfunc);
+#ifdef CONFIG_64BIT
+	else if ((long long)lfunc >= (long long)PHYS_TO_XKPHYS(0, 0) &&
+		 (long long)lfunc < (long long)PHYS_TO_XKPHYS(8, 0))
+		ufunc = PHYS_TO_XKPHYS(K_CALG_UNCACHED,
+				       XKPHYS_TO_PHYS((long long)lfunc));
+#endif
+	else {
+		BUG();
+		ufunc = lfunc;
+	}
+
+	__asm__ __volatile__ (
+		"	move	$16, $sp\n"
+		"	move	$sp, %1\n"
+		"	jalr	%2\n"
+		"	move	$sp, $16"
+		: "=r" (ret)
+		: "r" (usp), "r" (ufunc)
+		: "$16", "$31");
+
+	return ret;
+}