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, 141 insertions, 0 deletions

diff --git a/drivers/net/thunderbolt/trace.h b/drivers/net/thunderbolt/trace.h
new file mode 100644
index 000000000..9626eadae
--- /dev/null
+++ b/drivers/net/thunderbolt/trace.h
@@ -0,0 +1,141 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * Tracepoints for Thunderbolt/USB4 networking driver
+ *
+ * Copyright (C) 2023, Intel Corporation
+ * Author: Mika Westerberg <mika.westerberg@linux.intel.com>
+ */
+
+#undef TRACE_SYSTEM
+#define TRACE_SYSTEM thunderbolt_net
+
+#if !defined(__TRACE_THUNDERBOLT_NET_H) || defined(TRACE_HEADER_MULTI_READ)
+#define __TRACE_THUNDERBOLT_NET_H
+
+#include <linux/dma-direction.h>
+#include <linux/skbuff.h>
+#include <linux/tracepoint.h>
+
+#define DMA_DATA_DIRECTION_NAMES			\
+	{ DMA_BIDIRECTIONAL, "DMA_BIDIRECTIONAL" },	\
+	{ DMA_TO_DEVICE, "DMA_TO_DEVICE" },		\
+	{ DMA_FROM_DEVICE, "DMA_FROM_DEVICE" },		\
+	{ DMA_NONE, "DMA_NONE" }
+
+DECLARE_EVENT_CLASS(tbnet_frame,
+	TP_PROTO(unsigned int index, const void *page, dma_addr_t phys,
+		 enum dma_data_direction dir),
+	TP_ARGS(index, page, phys, dir),
+	TP_STRUCT__entry(
+		__field(unsigned int, index)
+		__field(const void *, page)
+		__field(dma_addr_t, phys)
+		__field(enum dma_data_direction, dir)
+	),
+	TP_fast_assign(
+		__entry->index = index;
+		__entry->page = page;
+		__entry->phys = phys;
+		__entry->dir = dir;
+	),
+	TP_printk("index=%u page=%p phys=%pad dir=%s",
+		  __entry->index, __entry->page, &__entry->phys,
+		__print_symbolic(__entry->dir, DMA_DATA_DIRECTION_NAMES))
+);
+
+DEFINE_EVENT(tbnet_frame, tbnet_alloc_rx_frame,
+	TP_PROTO(unsigned int index, const void *page, dma_addr_t phys,
+		 enum dma_data_direction dir),
+	TP_ARGS(index, page, phys, dir)
+);
+
+DEFINE_EVENT(tbnet_frame, tbnet_alloc_tx_frame,
+	TP_PROTO(unsigned int index, const void *page, dma_addr_t phys,
+		 enum dma_data_direction dir),
+	TP_ARGS(index, page, phys, dir)
+);
+
+DEFINE_EVENT(tbnet_frame, tbnet_free_frame,
+	TP_PROTO(unsigned int index, const void *page, dma_addr_t phys,
+		 enum dma_data_direction dir),
+	TP_ARGS(index, page, phys, dir)
+);
+
+DECLARE_EVENT_CLASS(tbnet_ip_frame,
+	TP_PROTO(__le32 size, __le16 id, __le16 index, __le32 count),
+	TP_ARGS(size, id, index, count),
+	TP_STRUCT__entry(
+		__field(u32, size)
+		__field(u16, id)
+		__field(u16, index)
+		__field(u32, count)
+	),
+	TP_fast_assign(
+		__entry->size = le32_to_cpu(size);
+		__entry->id = le16_to_cpu(id);
+		__entry->index = le16_to_cpu(index);
+		__entry->count = le32_to_cpu(count);
+	),
+	TP_printk("id=%u size=%u index=%u count=%u",
+		  __entry->id, __entry->size, __entry->index, __entry->count)
+);
+
+DEFINE_EVENT(tbnet_ip_frame, tbnet_rx_ip_frame,
+	TP_PROTO(__le32 size, __le16 id, __le16 index, __le32 count),
+	TP_ARGS(size, id, index, count)
+);
+
+DEFINE_EVENT(tbnet_ip_frame, tbnet_invalid_rx_ip_frame,
+	TP_PROTO(__le32 size, __le16 id, __le16 index, __le32 count),
+	TP_ARGS(size, id, index, count)
+);
+
+DEFINE_EVENT(tbnet_ip_frame, tbnet_tx_ip_frame,
+	TP_PROTO(__le32 size, __le16 id, __le16 index, __le32 count),
+	TP_ARGS(size, id, index, count)
+);
+
+DECLARE_EVENT_CLASS(tbnet_skb,
+	TP_PROTO(const struct sk_buff *skb),
+	TP_ARGS(skb),
+	TP_STRUCT__entry(
+		__field(const void *, addr)
+		__field(unsigned int, len)
+		__field(unsigned int, data_len)
+		__field(unsigned int, nr_frags)
+	),
+	TP_fast_assign(
+		__entry->addr = skb;
+		__entry->len = skb->len;
+		__entry->data_len = skb->data_len;
+		__entry->nr_frags = skb_shinfo(skb)->nr_frags;
+	),
+	TP_printk("skb=%p len=%u data_len=%u nr_frags=%u",
+		  __entry->addr, __entry->len, __entry->data_len,
+		  __entry->nr_frags)
+);
+
+DEFINE_EVENT(tbnet_skb, tbnet_rx_skb,
+	TP_PROTO(const struct sk_buff *skb),
+	TP_ARGS(skb)
+);
+
+DEFINE_EVENT(tbnet_skb, tbnet_tx_skb,
+	TP_PROTO(const struct sk_buff *skb),
+	TP_ARGS(skb)
+);
+
+DEFINE_EVENT(tbnet_skb, tbnet_consume_skb,
+	TP_PROTO(const struct sk_buff *skb),
+	TP_ARGS(skb)
+);
+
+#endif /* _TRACE_THUNDERBOLT_NET_H */
+
+#undef TRACE_INCLUDE_PATH
+#define TRACE_INCLUDE_PATH .
+
+#undef TRACE_INCLUDE_FILE
+#define TRACE_INCLUDE_FILE trace
+
+#include <trace/define_trace.h>