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, 162 insertions, 0 deletions

diff --git a/arch/mips/lib/bitops.c b/arch/mips/lib/bitops.c
new file mode 100644
index 000000000..00aee98e9
--- /dev/null
+++ b/arch/mips/lib/bitops.c
@@ -0,0 +1,162 @@
+/*
+ * 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) 1994-1997, 99, 2000, 06, 07 Ralf Baechle (ralf@linux-mips.org)
+ * Copyright (c) 1999, 2000  Silicon Graphics, Inc.
+ */
+#include <linux/bitops.h>
+#include <linux/bits.h>
+#include <linux/irqflags.h>
+#include <linux/export.h>
+
+
+/**
+ * __mips_set_bit - Atomically set a bit in memory.  This is called by
+ * set_bit() if it cannot find a faster solution.
+ * @nr: the bit to set
+ * @addr: the address to start counting from
+ */
+void __mips_set_bit(unsigned long nr, volatile unsigned long *addr)
+{
+	volatile unsigned long *a = &addr[BIT_WORD(nr)];
+	unsigned int bit = nr % BITS_PER_LONG;
+	unsigned long mask;
+	unsigned long flags;
+
+	mask = 1UL << bit;
+	raw_local_irq_save(flags);
+	*a |= mask;
+	raw_local_irq_restore(flags);
+}
+EXPORT_SYMBOL(__mips_set_bit);
+
+
+/**
+ * __mips_clear_bit - Clears a bit in memory.  This is called by clear_bit() if
+ * it cannot find a faster solution.
+ * @nr: Bit to clear
+ * @addr: Address to start counting from
+ */
+void __mips_clear_bit(unsigned long nr, volatile unsigned long *addr)
+{
+	volatile unsigned long *a = &addr[BIT_WORD(nr)];
+	unsigned int bit = nr % BITS_PER_LONG;
+	unsigned long mask;
+	unsigned long flags;
+
+	mask = 1UL << bit;
+	raw_local_irq_save(flags);
+	*a &= ~mask;
+	raw_local_irq_restore(flags);
+}
+EXPORT_SYMBOL(__mips_clear_bit);
+
+
+/**
+ * __mips_change_bit - Toggle a bit in memory.	This is called by change_bit()
+ * if it cannot find a faster solution.
+ * @nr: Bit to change
+ * @addr: Address to start counting from
+ */
+void __mips_change_bit(unsigned long nr, volatile unsigned long *addr)
+{
+	volatile unsigned long *a = &addr[BIT_WORD(nr)];
+	unsigned int bit = nr % BITS_PER_LONG;
+	unsigned long mask;
+	unsigned long flags;
+
+	mask = 1UL << bit;
+	raw_local_irq_save(flags);
+	*a ^= mask;
+	raw_local_irq_restore(flags);
+}
+EXPORT_SYMBOL(__mips_change_bit);
+
+
+/**
+ * __mips_test_and_set_bit_lock - Set a bit and return its old value.  This is
+ * called by test_and_set_bit_lock() if it cannot find a faster solution.
+ * @nr: Bit to set
+ * @addr: Address to count from
+ */
+int __mips_test_and_set_bit_lock(unsigned long nr,
+				 volatile unsigned long *addr)
+{
+	volatile unsigned long *a = &addr[BIT_WORD(nr)];
+	unsigned int bit = nr % BITS_PER_LONG;
+	unsigned long mask;
+	unsigned long flags;
+	int res;
+
+	mask = 1UL << bit;
+	raw_local_irq_save(flags);
+	res = (mask & *a) != 0;
+	*a |= mask;
+	raw_local_irq_restore(flags);
+	return res;
+}
+EXPORT_SYMBOL(__mips_test_and_set_bit_lock);
+
+
+/**
+ * __mips_test_and_clear_bit - Clear a bit and return its old value.  This is
+ * called by test_and_clear_bit() if it cannot find a faster solution.
+ * @nr: Bit to clear
+ * @addr: Address to count from
+ */
+int __mips_test_and_clear_bit(unsigned long nr, volatile unsigned long *addr)
+{
+	volatile unsigned long *a = &addr[BIT_WORD(nr)];
+	unsigned int bit = nr % BITS_PER_LONG;
+	unsigned long mask;
+	unsigned long flags;
+	int res;
+
+	mask = 1UL << bit;
+	raw_local_irq_save(flags);
+	res = (mask & *a) != 0;
+	*a &= ~mask;
+	raw_local_irq_restore(flags);
+	return res;
+}
+EXPORT_SYMBOL(__mips_test_and_clear_bit);
+
+
+/**
+ * __mips_test_and_change_bit - Change a bit and return its old value.	This is
+ * called by test_and_change_bit() if it cannot find a faster solution.
+ * @nr: Bit to change
+ * @addr: Address to count from
+ */
+int __mips_test_and_change_bit(unsigned long nr, volatile unsigned long *addr)
+{
+	volatile unsigned long *a = &addr[BIT_WORD(nr)];
+	unsigned int bit = nr % BITS_PER_LONG;
+	unsigned long mask;
+	unsigned long flags;
+	int res;
+
+	mask = 1UL << bit;
+	raw_local_irq_save(flags);
+	res = (mask & *a) != 0;
+	*a ^= mask;
+	raw_local_irq_restore(flags);
+	return res;
+}
+EXPORT_SYMBOL(__mips_test_and_change_bit);
+
+bool __mips_xor_is_negative_byte(unsigned long mask,
+		volatile unsigned long *addr)
+{
+	unsigned long flags;
+	unsigned long data;
+
+	raw_local_irq_save(flags);
+	data = *addr;
+	*addr = data ^ mask;
+	raw_local_irq_restore(flags);
+
+	return (data & BIT(7)) != 0;
+}