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, 115 insertions, 0 deletions

diff --git a/include/asm-generic/cmpxchg.h b/include/asm-generic/cmpxchg.h
new file mode 100644
index 000000000..848de25fc
--- /dev/null
+++ b/include/asm-generic/cmpxchg.h
@@ -0,0 +1,115 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * Generic UP xchg and cmpxchg using interrupt disablement.  Does not
+ * support SMP.
+ */
+
+#ifndef __ASM_GENERIC_CMPXCHG_H
+#define __ASM_GENERIC_CMPXCHG_H
+
+#ifdef CONFIG_SMP
+#error "Cannot use generic cmpxchg on SMP"
+#endif
+
+#include <linux/types.h>
+#include <linux/irqflags.h>
+
+/*
+ * This function doesn't exist, so you'll get a linker error if
+ * something tries to do an invalidly-sized xchg().
+ */
+extern void __generic_xchg_called_with_bad_pointer(void);
+
+static inline
+unsigned long __generic_xchg(unsigned long x, volatile void *ptr, int size)
+{
+	unsigned long ret, flags;
+
+	switch (size) {
+	case 1:
+#ifdef __xchg_u8
+		return __xchg_u8(x, ptr);
+#else
+		local_irq_save(flags);
+		ret = *(volatile u8 *)ptr;
+		*(volatile u8 *)ptr = (x & 0xffu);
+		local_irq_restore(flags);
+		return ret;
+#endif /* __xchg_u8 */
+
+	case 2:
+#ifdef __xchg_u16
+		return __xchg_u16(x, ptr);
+#else
+		local_irq_save(flags);
+		ret = *(volatile u16 *)ptr;
+		*(volatile u16 *)ptr = (x & 0xffffu);
+		local_irq_restore(flags);
+		return ret;
+#endif /* __xchg_u16 */
+
+	case 4:
+#ifdef __xchg_u32
+		return __xchg_u32(x, ptr);
+#else
+		local_irq_save(flags);
+		ret = *(volatile u32 *)ptr;
+		*(volatile u32 *)ptr = (x & 0xffffffffu);
+		local_irq_restore(flags);
+		return ret;
+#endif /* __xchg_u32 */
+
+#ifdef CONFIG_64BIT
+	case 8:
+#ifdef __xchg_u64
+		return __xchg_u64(x, ptr);
+#else
+		local_irq_save(flags);
+		ret = *(volatile u64 *)ptr;
+		*(volatile u64 *)ptr = x;
+		local_irq_restore(flags);
+		return ret;
+#endif /* __xchg_u64 */
+#endif /* CONFIG_64BIT */
+
+	default:
+		__generic_xchg_called_with_bad_pointer();
+		return x;
+	}
+}
+
+#define generic_xchg(ptr, x) ({							\
+	((__typeof__(*(ptr)))							\
+		__generic_xchg((unsigned long)(x), (ptr), sizeof(*(ptr))));	\
+})
+
+/*
+ * Atomic compare and exchange.
+ */
+#include <asm-generic/cmpxchg-local.h>
+
+#define generic_cmpxchg_local(ptr, o, n) ({					\
+	((__typeof__(*(ptr)))__generic_cmpxchg_local((ptr), (unsigned long)(o),	\
+			(unsigned long)(n), sizeof(*(ptr))));			\
+})
+
+#define generic_cmpxchg64_local(ptr, o, n) \
+	__generic_cmpxchg64_local((ptr), (o), (n))
+
+
+#ifndef arch_xchg
+#define arch_xchg		generic_xchg
+#endif
+
+#ifndef arch_cmpxchg_local
+#define arch_cmpxchg_local	generic_cmpxchg_local
+#endif
+
+#ifndef arch_cmpxchg64_local
+#define arch_cmpxchg64_local	generic_cmpxchg64_local
+#endif
+
+#define arch_cmpxchg		arch_cmpxchg_local
+#define arch_cmpxchg64		arch_cmpxchg64_local
+
+#endif /* __ASM_GENERIC_CMPXCHG_H */