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/include/asm-generic/bitops/lock.h b/include/asm-generic/bitops/lock.h
new file mode 100644
index 000000000..ffb73b612
--- /dev/null
+++ b/include/asm-generic/bitops/lock.h
@@ -0,0 +1,82 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef _ASM_GENERIC_BITOPS_LOCK_H_
+#define _ASM_GENERIC_BITOPS_LOCK_H_
+
+#include <linux/atomic.h>
+#include <linux/compiler.h>
+#include <asm/barrier.h>
+
+/**
+ * arch_test_and_set_bit_lock - Set a bit and return its old value, for lock
+ * @nr: Bit to set
+ * @addr: Address to count from
+ *
+ * This operation is atomic and provides acquire barrier semantics if
+ * the returned value is 0.
+ * It can be used to implement bit locks.
+ */
+static __always_inline int
+arch_test_and_set_bit_lock(unsigned int nr, volatile unsigned long *addr)
+{
+	long old;
+	unsigned long mask = BIT_MASK(nr);
+
+	addr += BIT_WORD(nr);
+	if (READ_ONCE(*addr) & mask)
+		return 1;
+
+	old = raw_atomic_long_fetch_or_acquire(mask, (atomic_long_t *)addr);
+	return !!(old & mask);
+}
+
+
+/**
+ * arch_clear_bit_unlock - Clear a bit in memory, for unlock
+ * @nr: the bit to set
+ * @addr: the address to start counting from
+ *
+ * This operation is atomic and provides release barrier semantics.
+ */
+static __always_inline void
+arch_clear_bit_unlock(unsigned int nr, volatile unsigned long *addr)
+{
+	addr += BIT_WORD(nr);
+	raw_atomic_long_fetch_andnot_release(BIT_MASK(nr), (atomic_long_t *)addr);
+}
+
+/**
+ * arch___clear_bit_unlock - Clear a bit in memory, for unlock
+ * @nr: the bit to set
+ * @addr: the address to start counting from
+ *
+ * A weaker form of clear_bit_unlock() as used by __bit_lock_unlock(). If all
+ * the bits in the word are protected by this lock some archs can use weaker
+ * ops to safely unlock.
+ *
+ * See for example x86's implementation.
+ */
+static inline void
+arch___clear_bit_unlock(unsigned int nr, volatile unsigned long *addr)
+{
+	unsigned long old;
+
+	addr += BIT_WORD(nr);
+	old = READ_ONCE(*addr);
+	old &= ~BIT_MASK(nr);
+	raw_atomic_long_set_release((atomic_long_t *)addr, old);
+}
+
+#ifndef arch_xor_unlock_is_negative_byte
+static inline bool arch_xor_unlock_is_negative_byte(unsigned long mask,
+		volatile unsigned long *p)
+{
+	long old;
+
+	old = raw_atomic_long_fetch_xor_release(mask, (atomic_long_t *)p);
+	return !!(old & BIT(7));
+}
+#endif
+
+#include <asm-generic/bitops/instrumented-lock.h>
+
+#endif /* _ASM_GENERIC_BITOPS_LOCK_H_ */