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, 156 insertions, 0 deletions

diff --git a/include/trace/events/intel-sst.h b/include/trace/events/intel-sst.h
new file mode 100644
index 000000000..0416e91f8
--- /dev/null
+++ b/include/trace/events/intel-sst.h
@@ -0,0 +1,156 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#undef TRACE_SYSTEM
+#define TRACE_SYSTEM intel-sst
+
+/*
+ * The TRACE_SYSTEM_VAR defaults to TRACE_SYSTEM, but must be a
+ * legitimate C variable. It is not exported to user space.
+ */
+#undef TRACE_SYSTEM_VAR
+#define TRACE_SYSTEM_VAR intel_sst
+
+#if !defined(_TRACE_INTEL_SST_H) || defined(TRACE_HEADER_MULTI_READ)
+#define _TRACE_INTEL_SST_H
+
+#include <linux/types.h>
+#include <linux/ktime.h>
+#include <linux/tracepoint.h>
+
+DECLARE_EVENT_CLASS(sst_ipc_msg,
+
+	TP_PROTO(unsigned int val),
+
+	TP_ARGS(val),
+
+	TP_STRUCT__entry(
+		__field(	unsigned int,	val		)
+	),
+
+	TP_fast_assign(
+		__entry->val = val;
+	),
+
+	TP_printk("0x%8.8x", (unsigned int)__entry->val)
+);
+
+DEFINE_EVENT(sst_ipc_msg, sst_ipc_msg_tx,
+
+	TP_PROTO(unsigned int val),
+
+	TP_ARGS(val)
+
+);
+
+DEFINE_EVENT(sst_ipc_msg, sst_ipc_msg_rx,
+
+	TP_PROTO(unsigned int val),
+
+	TP_ARGS(val)
+
+);
+
+DECLARE_EVENT_CLASS(sst_ipc_mailbox,
+
+	TP_PROTO(unsigned int offset, unsigned int val),
+
+	TP_ARGS(offset, val),
+
+	TP_STRUCT__entry(
+		__field(	unsigned int,	offset		)
+		__field(	unsigned int,	val		)
+	),
+
+	TP_fast_assign(
+		__entry->offset = offset;
+		__entry->val = val;
+	),
+
+	TP_printk(" 0x%4.4x = 0x%8.8x",
+		(unsigned int)__entry->offset, (unsigned int)__entry->val)
+);
+
+DEFINE_EVENT(sst_ipc_mailbox, sst_ipc_inbox_rdata,
+
+	TP_PROTO(unsigned int offset, unsigned int val),
+
+	TP_ARGS(offset, val)
+
+);
+
+DEFINE_EVENT(sst_ipc_mailbox, sst_ipc_inbox_wdata,
+
+	TP_PROTO(unsigned int offset, unsigned int val),
+
+	TP_ARGS(offset, val)
+
+);
+
+DEFINE_EVENT(sst_ipc_mailbox, sst_ipc_outbox_rdata,
+
+	TP_PROTO(unsigned int offset, unsigned int val),
+
+	TP_ARGS(offset, val)
+
+);
+
+DEFINE_EVENT(sst_ipc_mailbox, sst_ipc_outbox_wdata,
+
+	TP_PROTO(unsigned int offset, unsigned int val),
+
+	TP_ARGS(offset, val)
+
+);
+
+DECLARE_EVENT_CLASS(sst_ipc_mailbox_info,
+
+	TP_PROTO(unsigned int size),
+
+	TP_ARGS(size),
+
+	TP_STRUCT__entry(
+		__field(	unsigned int,	size		)
+	),
+
+	TP_fast_assign(
+		__entry->size = size;
+	),
+
+	TP_printk("Mailbox bytes 0x%8.8x", (unsigned int)__entry->size)
+);
+
+DEFINE_EVENT(sst_ipc_mailbox_info, sst_ipc_inbox_read,
+
+	TP_PROTO(unsigned int size),
+
+	TP_ARGS(size)
+
+);
+
+DEFINE_EVENT(sst_ipc_mailbox_info, sst_ipc_inbox_write,
+
+	TP_PROTO(unsigned int size),
+
+	TP_ARGS(size)
+
+);
+
+DEFINE_EVENT(sst_ipc_mailbox_info, sst_ipc_outbox_read,
+
+	TP_PROTO(unsigned int size),
+
+	TP_ARGS(size)
+
+);
+
+DEFINE_EVENT(sst_ipc_mailbox_info, sst_ipc_outbox_write,
+
+	TP_PROTO(unsigned int size),
+
+	TP_ARGS(size)
+
+);
+
+#endif /* _TRACE_SST_H */
+
+/* This part must be outside protection */
+#include <trace/define_trace.h>