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/net/openvswitch/flow_table.h b/net/openvswitch/flow_table.h
new file mode 100644
index 000000000..f524dc3e4
--- /dev/null
+++ b/net/openvswitch/flow_table.h
@@ -0,0 +1,115 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/*
+ * Copyright (c) 2007-2013 Nicira, Inc.
+ */
+
+#ifndef FLOW_TABLE_H
+#define FLOW_TABLE_H 1
+
+#include <linux/kernel.h>
+#include <linux/netlink.h>
+#include <linux/openvswitch.h>
+#include <linux/spinlock.h>
+#include <linux/types.h>
+#include <linux/rcupdate.h>
+#include <linux/if_ether.h>
+#include <linux/in6.h>
+#include <linux/jiffies.h>
+#include <linux/time.h>
+
+#include <net/inet_ecn.h>
+#include <net/ip_tunnels.h>
+
+#include "flow.h"
+
+struct mask_cache_entry {
+	u32 skb_hash;
+	u32 mask_index;
+};
+
+struct mask_cache {
+	struct rcu_head rcu;
+	u32 cache_size;  /* Must be ^2 value. */
+	struct mask_cache_entry __percpu *mask_cache;
+};
+
+struct mask_count {
+	int index;
+	u64 counter;
+};
+
+struct mask_array_stats {
+	struct u64_stats_sync syncp;
+	u64 usage_cntrs[];
+};
+
+struct mask_array {
+	struct rcu_head rcu;
+	int count, max;
+	struct mask_array_stats __percpu *masks_usage_stats;
+	u64 *masks_usage_zero_cntr;
+	struct sw_flow_mask __rcu *masks[] __counted_by(max);
+};
+
+struct table_instance {
+	struct hlist_head *buckets;
+	unsigned int n_buckets;
+	struct rcu_head rcu;
+	int node_ver;
+	u32 hash_seed;
+};
+
+struct flow_table {
+	struct table_instance __rcu *ti;
+	struct table_instance __rcu *ufid_ti;
+	struct mask_cache __rcu *mask_cache;
+	struct mask_array __rcu *mask_array;
+	unsigned long last_rehash;
+	unsigned int count;
+	unsigned int ufid_count;
+};
+
+extern struct kmem_cache *flow_stats_cache;
+
+int ovs_flow_init(void);
+void ovs_flow_exit(void);
+
+struct sw_flow *ovs_flow_alloc(void);
+void ovs_flow_free(struct sw_flow *, bool deferred);
+
+int ovs_flow_tbl_init(struct flow_table *);
+int ovs_flow_tbl_count(const struct flow_table *table);
+void ovs_flow_tbl_destroy(struct flow_table *table);
+int ovs_flow_tbl_flush(struct flow_table *flow_table);
+
+int ovs_flow_tbl_insert(struct flow_table *table, struct sw_flow *flow,
+			const struct sw_flow_mask *mask);
+void ovs_flow_tbl_remove(struct flow_table *table, struct sw_flow *flow);
+int  ovs_flow_tbl_num_masks(const struct flow_table *table);
+u32  ovs_flow_tbl_masks_cache_size(const struct flow_table *table);
+int  ovs_flow_tbl_masks_cache_resize(struct flow_table *table, u32 size);
+struct sw_flow *ovs_flow_tbl_dump_next(struct table_instance *table,
+				       u32 *bucket, u32 *idx);
+struct sw_flow *ovs_flow_tbl_lookup_stats(struct flow_table *,
+					  const struct sw_flow_key *,
+					  u32 skb_hash,
+					  u32 *n_mask_hit,
+					  u32 *n_cache_hit);
+struct sw_flow *ovs_flow_tbl_lookup(struct flow_table *,
+				    const struct sw_flow_key *);
+struct sw_flow *ovs_flow_tbl_lookup_exact(struct flow_table *tbl,
+					  const struct sw_flow_match *match);
+struct sw_flow *ovs_flow_tbl_lookup_ufid(struct flow_table *,
+					 const struct sw_flow_id *);
+
+bool ovs_flow_cmp(const struct sw_flow *, const struct sw_flow_match *);
+
+void ovs_flow_mask_key(struct sw_flow_key *dst, const struct sw_flow_key *src,
+		       bool full, const struct sw_flow_mask *mask);
+
+void ovs_flow_masks_rebalance(struct flow_table *table);
+void table_instance_flow_flush(struct flow_table *table,
+			       struct table_instance *ti,
+			       struct table_instance *ufid_ti);
+
+#endif /* flow_table.h */