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, 100 insertions, 0 deletions

diff --git a/include/net/raw.h b/include/net/raw.h
new file mode 100644
index 000000000..66c0ffead
--- /dev/null
+++ b/include/net/raw.h
@@ -0,0 +1,100 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later */
+/*
+ * INET		An implementation of the TCP/IP protocol suite for the LINUX
+ *		operating system.  INET is implemented using the  BSD Socket
+ *		interface as the means of communication with the user level.
+ *
+ *		Definitions for the RAW-IP module.
+ *
+ * Version:	@(#)raw.h	1.0.2	05/07/93
+ *
+ * Author:	Fred N. van Kempen, <waltje@uWalt.NL.Mugnet.ORG>
+ */
+#ifndef _RAW_H
+#define _RAW_H
+
+#include <net/inet_sock.h>
+#include <net/protocol.h>
+#include <net/netns/hash.h>
+#include <linux/hash.h>
+#include <linux/icmp.h>
+
+extern struct proto raw_prot;
+
+extern struct raw_hashinfo raw_v4_hashinfo;
+bool raw_v4_match(struct net *net, const struct sock *sk, unsigned short num,
+		  __be32 raddr, __be32 laddr, int dif, int sdif);
+
+int raw_abort(struct sock *sk, int err);
+void raw_icmp_error(struct sk_buff *, int, u32);
+int raw_local_deliver(struct sk_buff *, int);
+
+int raw_rcv(struct sock *, struct sk_buff *);
+
+#define RAW_HTABLE_LOG	8
+#define RAW_HTABLE_SIZE	(1U << RAW_HTABLE_LOG)
+
+struct raw_hashinfo {
+	spinlock_t lock;
+
+	struct hlist_head ht[RAW_HTABLE_SIZE] ____cacheline_aligned;
+};
+
+static inline u32 raw_hashfunc(const struct net *net, u32 proto)
+{
+	return hash_32(net_hash_mix(net) ^ proto, RAW_HTABLE_LOG);
+}
+
+static inline void raw_hashinfo_init(struct raw_hashinfo *hashinfo)
+{
+	int i;
+
+	spin_lock_init(&hashinfo->lock);
+	for (i = 0; i < RAW_HTABLE_SIZE; i++)
+		INIT_HLIST_HEAD(&hashinfo->ht[i]);
+}
+
+#ifdef CONFIG_PROC_FS
+int raw_proc_init(void);
+void raw_proc_exit(void);
+
+struct raw_iter_state {
+	struct seq_net_private p;
+	int bucket;
+};
+
+static inline struct raw_iter_state *raw_seq_private(struct seq_file *seq)
+{
+	return seq->private;
+}
+void *raw_seq_start(struct seq_file *seq, loff_t *pos);
+void *raw_seq_next(struct seq_file *seq, void *v, loff_t *pos);
+void raw_seq_stop(struct seq_file *seq, void *v);
+#endif
+
+int raw_hash_sk(struct sock *sk);
+void raw_unhash_sk(struct sock *sk);
+void raw_init(void);
+
+struct raw_sock {
+	/* inet_sock has to be the first member */
+	struct inet_sock   inet;
+	struct icmp_filter filter;
+	u32		   ipmr_table;
+	struct numa_drop_counters drop_counters;
+};
+
+#define raw_sk(ptr) container_of_const(ptr, struct raw_sock, inet.sk)
+
+static inline bool raw_sk_bound_dev_eq(struct net *net, int bound_dev_if,
+				       int dif, int sdif)
+{
+#if IS_ENABLED(CONFIG_NET_L3_MASTER_DEV)
+	return inet_bound_dev_eq(READ_ONCE(net->ipv4.sysctl_raw_l3mdev_accept),
+				 bound_dev_if, dif, sdif);
+#else
+	return inet_bound_dev_eq(true, bound_dev_if, dif, sdif);
+#endif
+}
+
+#endif	/* _RAW_H */