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, 283 insertions, 0 deletions

diff --git a/include/net/dst_metadata.h b/include/net/dst_metadata.h
new file mode 100644
index 000000000..f45d1e316
--- /dev/null
+++ b/include/net/dst_metadata.h
@@ -0,0 +1,283 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef __NET_DST_METADATA_H
+#define __NET_DST_METADATA_H 1
+
+#include <linux/skbuff.h>
+#include <net/ip.h>
+#include <net/ip_tunnels.h>
+#include <net/macsec.h>
+#include <net/dst.h>
+
+enum metadata_type {
+	METADATA_IP_TUNNEL,
+	METADATA_HW_PORT_MUX,
+	METADATA_MACSEC,
+	METADATA_XFRM,
+};
+
+struct hw_port_info {
+	struct net_device *lower_dev;
+	u32 port_id;
+};
+
+struct macsec_info {
+	sci_t sci;
+};
+
+struct xfrm_md_info {
+	u32 if_id;
+	int link;
+	struct dst_entry *dst_orig;
+};
+
+struct metadata_dst {
+	struct dst_entry		dst;
+	enum metadata_type		type;
+	union {
+		struct ip_tunnel_info	tun_info;
+		struct hw_port_info	port_info;
+		struct macsec_info	macsec_info;
+		struct xfrm_md_info	xfrm_info;
+	} u;
+};
+
+static inline struct metadata_dst *skb_metadata_dst(const struct sk_buff *skb)
+{
+	struct metadata_dst *md_dst = (struct metadata_dst *) skb_dst(skb);
+
+	if (md_dst && md_dst->dst.flags & DST_METADATA)
+		return md_dst;
+
+	return NULL;
+}
+
+static inline struct ip_tunnel_info *
+skb_tunnel_info(const struct sk_buff *skb)
+{
+	struct metadata_dst *md_dst = skb_metadata_dst(skb);
+	struct dst_entry *dst;
+
+	if (md_dst && md_dst->type == METADATA_IP_TUNNEL)
+		return &md_dst->u.tun_info;
+
+	dst = skb_dst(skb);
+	if (dst && dst->lwtstate &&
+	    (dst->lwtstate->type == LWTUNNEL_ENCAP_IP ||
+	     dst->lwtstate->type == LWTUNNEL_ENCAP_IP6))
+		return lwt_tun_info(dst->lwtstate);
+
+	return NULL;
+}
+
+static inline struct xfrm_md_info *lwt_xfrm_info(struct lwtunnel_state *lwt)
+{
+	return (struct xfrm_md_info *)lwt->data;
+}
+
+static inline struct xfrm_md_info *skb_xfrm_md_info(const struct sk_buff *skb)
+{
+	struct metadata_dst *md_dst = skb_metadata_dst(skb);
+	struct dst_entry *dst;
+
+	if (md_dst && md_dst->type == METADATA_XFRM)
+		return &md_dst->u.xfrm_info;
+
+	dst = skb_dst(skb);
+	if (dst && dst->lwtstate &&
+	    dst->lwtstate->type == LWTUNNEL_ENCAP_XFRM)
+		return lwt_xfrm_info(dst->lwtstate);
+
+	return NULL;
+}
+
+static inline bool skb_valid_dst(const struct sk_buff *skb)
+{
+	struct dst_entry *dst = skb_dst(skb);
+
+	return dst && !(dst->flags & DST_METADATA);
+}
+
+static inline int skb_metadata_dst_cmp(const struct sk_buff *skb_a,
+				       const struct sk_buff *skb_b)
+{
+	const struct metadata_dst *a, *b;
+
+	if (!(skb_a->_skb_refdst | skb_b->_skb_refdst))
+		return 0;
+
+	a = (const struct metadata_dst *) skb_dst(skb_a);
+	b = (const struct metadata_dst *) skb_dst(skb_b);
+
+	if (!a != !b || a->type != b->type)
+		return 1;
+
+	switch (a->type) {
+	case METADATA_HW_PORT_MUX:
+		return memcmp(&a->u.port_info, &b->u.port_info,
+			      sizeof(a->u.port_info));
+	case METADATA_IP_TUNNEL:
+		return memcmp(&a->u.tun_info, &b->u.tun_info,
+			      sizeof(a->u.tun_info) +
+					 a->u.tun_info.options_len);
+	case METADATA_MACSEC:
+		return memcmp(&a->u.macsec_info, &b->u.macsec_info,
+			      sizeof(a->u.macsec_info));
+	case METADATA_XFRM:
+		return memcmp(&a->u.xfrm_info, &b->u.xfrm_info,
+			      sizeof(a->u.xfrm_info));
+	default:
+		return 1;
+	}
+}
+
+void metadata_dst_free(struct metadata_dst *);
+struct metadata_dst *metadata_dst_alloc(u8 optslen, enum metadata_type type,
+					gfp_t flags);
+void metadata_dst_free_percpu(struct metadata_dst __percpu *md_dst);
+struct metadata_dst __percpu *
+metadata_dst_alloc_percpu(u8 optslen, enum metadata_type type, gfp_t flags);
+
+static inline struct metadata_dst *tun_rx_dst(int md_size)
+{
+	struct metadata_dst *tun_dst;
+
+	tun_dst = metadata_dst_alloc(md_size, METADATA_IP_TUNNEL, GFP_ATOMIC);
+	if (!tun_dst)
+		return NULL;
+
+	tun_dst->u.tun_info.options_len = 0;
+	tun_dst->u.tun_info.mode = 0;
+	return tun_dst;
+}
+
+static inline struct metadata_dst *tun_dst_unclone(struct sk_buff *skb)
+{
+	struct metadata_dst *md_dst = skb_metadata_dst(skb);
+	int md_size;
+	struct metadata_dst *new_md;
+
+	if (!md_dst || md_dst->type != METADATA_IP_TUNNEL)
+		return ERR_PTR(-EINVAL);
+
+	md_size = md_dst->u.tun_info.options_len;
+	new_md = metadata_dst_alloc(md_size, METADATA_IP_TUNNEL, GFP_ATOMIC);
+	if (!new_md)
+		return ERR_PTR(-ENOMEM);
+
+	/* Copy in two stages to keep the __counted_by happy. */
+	new_md->u.tun_info = md_dst->u.tun_info;
+	memcpy(ip_tunnel_info_opts(&new_md->u.tun_info),
+	       ip_tunnel_info_opts(&md_dst->u.tun_info), md_size);
+
+#ifdef CONFIG_DST_CACHE
+	/* Unclone the dst cache if there is one */
+	if (new_md->u.tun_info.dst_cache.cache) {
+		int ret;
+
+		ret = dst_cache_init(&new_md->u.tun_info.dst_cache, GFP_ATOMIC);
+		if (ret) {
+			metadata_dst_free(new_md);
+			return ERR_PTR(ret);
+		}
+	}
+#endif
+
+	skb_dst_drop(skb);
+	skb_dst_set(skb, &new_md->dst);
+	return new_md;
+}
+
+static inline struct ip_tunnel_info *skb_tunnel_info_unclone(struct sk_buff *skb)
+{
+	struct metadata_dst *dst;
+
+	dst = tun_dst_unclone(skb);
+	if (IS_ERR(dst))
+		return NULL;
+
+	return &dst->u.tun_info;
+}
+
+static inline struct metadata_dst *__ip_tun_set_dst(__be32 saddr,
+						    __be32 daddr,
+						    __u8 tos, __u8 ttl,
+						    __be16 tp_dst,
+						    const unsigned long *flags,
+						    __be64 tunnel_id,
+						    int md_size)
+{
+	struct metadata_dst *tun_dst;
+
+	tun_dst = tun_rx_dst(md_size);
+	if (!tun_dst)
+		return NULL;
+
+	ip_tunnel_key_init(&tun_dst->u.tun_info.key,
+			   saddr, daddr, tos, ttl,
+			   0, 0, tp_dst, tunnel_id, flags);
+	return tun_dst;
+}
+
+static inline struct metadata_dst *ip_tun_rx_dst(struct sk_buff *skb,
+						 const unsigned long *flags,
+						 __be64 tunnel_id,
+						 int md_size)
+{
+	const struct iphdr *iph = ip_hdr(skb);
+	struct metadata_dst *tun_dst;
+
+	tun_dst = __ip_tun_set_dst(iph->saddr, iph->daddr, iph->tos, iph->ttl,
+				   0, flags, tunnel_id, md_size);
+
+	if (tun_dst && (iph->frag_off & htons(IP_DF)))
+		__set_bit(IP_TUNNEL_DONT_FRAGMENT_BIT,
+			  tun_dst->u.tun_info.key.tun_flags);
+	return tun_dst;
+}
+
+static inline struct metadata_dst *__ipv6_tun_set_dst(const struct in6_addr *saddr,
+						      const struct in6_addr *daddr,
+						      __u8 tos, __u8 ttl,
+						      __be16 tp_dst,
+						      __be32 label,
+						      const unsigned long *flags,
+						      __be64 tunnel_id,
+						      int md_size)
+{
+	struct metadata_dst *tun_dst;
+	struct ip_tunnel_info *info;
+
+	tun_dst = tun_rx_dst(md_size);
+	if (!tun_dst)
+		return NULL;
+
+	info = &tun_dst->u.tun_info;
+	info->mode = IP_TUNNEL_INFO_IPV6;
+	ip_tunnel_flags_copy(info->key.tun_flags, flags);
+	info->key.tun_id = tunnel_id;
+	info->key.tp_src = 0;
+	info->key.tp_dst = tp_dst;
+
+	info->key.u.ipv6.src = *saddr;
+	info->key.u.ipv6.dst = *daddr;
+
+	info->key.tos = tos;
+	info->key.ttl = ttl;
+	info->key.label = label;
+
+	return tun_dst;
+}
+
+static inline struct metadata_dst *ipv6_tun_rx_dst(struct sk_buff *skb,
+						   const unsigned long *flags,
+						   __be64 tunnel_id,
+						   int md_size)
+{
+	const struct ipv6hdr *ip6h = ipv6_hdr(skb);
+
+	return __ipv6_tun_set_dst(&ip6h->saddr, &ip6h->daddr,
+				  ipv6_get_dsfield(ip6h), ip6h->hop_limit,
+				  0, ip6_flowlabel(ip6h), flags, tunnel_id,
+				  md_size);
+}
+#endif /* __NET_DST_METADATA_H */