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, 175 insertions, 0 deletions

diff --git a/net/mac80211/key.h b/net/mac80211/key.h
new file mode 100644
index 000000000..826e4e938
--- /dev/null
+++ b/net/mac80211/key.h
@@ -0,0 +1,175 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/*
+ * Copyright 2002-2004, Instant802 Networks, Inc.
+ * Copyright 2005, Devicescape Software, Inc.
+ * Copyright (C) 2019, 2022-2023 Intel Corporation
+ */
+
+#ifndef IEEE80211_KEY_H
+#define IEEE80211_KEY_H
+
+#include <linux/types.h>
+#include <linux/list.h>
+#include <linux/crypto.h>
+#include <linux/rcupdate.h>
+#include <crypto/aes-cbc-macs.h>
+#include <crypto/arc4.h>
+#include <net/mac80211.h>
+
+#define NUM_DEFAULT_KEYS 4
+#define NUM_DEFAULT_MGMT_KEYS 2
+#define NUM_DEFAULT_BEACON_KEYS 2
+#define INVALID_PTK_KEYIDX 2 /* Keyidx always pointing to a NULL key for PTK */
+
+struct ieee80211_local;
+struct ieee80211_sub_if_data;
+struct ieee80211_link_data;
+struct sta_info;
+
+/**
+ * enum ieee80211_internal_key_flags - internal key flags
+ *
+ * @KEY_FLAG_UPLOADED_TO_HARDWARE: Indicates that this key is present
+ *	in the hardware for TX crypto hardware acceleration.
+ * @KEY_FLAG_TAINTED: Key is tainted and packets should be dropped.
+ */
+enum ieee80211_internal_key_flags {
+	KEY_FLAG_UPLOADED_TO_HARDWARE	= BIT(0),
+	KEY_FLAG_TAINTED		= BIT(1),
+};
+
+enum ieee80211_internal_tkip_state {
+	TKIP_STATE_NOT_INIT,
+	TKIP_STATE_PHASE1_DONE,
+	TKIP_STATE_PHASE1_HW_UPLOADED,
+};
+
+struct tkip_ctx {
+	u16 p1k[5];	/* p1k cache */
+	u32 p1k_iv32;	/* iv32 for which p1k computed */
+	enum ieee80211_internal_tkip_state state;
+};
+
+struct tkip_ctx_rx {
+	struct tkip_ctx ctx;
+	u32 iv32;	/* current iv32 */
+	u16 iv16;	/* current iv16 */
+};
+
+struct ieee80211_key {
+	struct ieee80211_local *local;
+	struct ieee80211_sub_if_data *sdata;
+	struct sta_info *sta;
+
+	/* for sdata list */
+	struct list_head list;
+
+	/* protected by key mutex */
+	unsigned int flags;
+
+	union {
+		struct {
+			/* protects tx context */
+			spinlock_t txlock;
+
+			/* last used TSC */
+			struct tkip_ctx tx;
+
+			/* last received RSC */
+			struct tkip_ctx_rx rx[IEEE80211_NUM_TIDS];
+
+			/* number of mic failures */
+			u32 mic_failures;
+		} tkip;
+		struct {
+			/*
+			 * Last received packet number. The first
+			 * IEEE80211_NUM_TIDS counters are used with Data
+			 * frames and the last counter is used with Robust
+			 * Management frames.
+			 */
+			u8 rx_pn[IEEE80211_NUM_TIDS + 1][IEEE80211_CCMP_PN_LEN];
+			struct crypto_aead *tfm;
+			u32 replays; /* dot11RSNAStatsCCMPReplays */
+		} ccmp;
+		struct {
+			u8 rx_pn[IEEE80211_CMAC_PN_LEN];
+			struct aes_cmac_key key;
+			u32 replays; /* dot11RSNAStatsCMACReplays */
+			u32 icverrors; /* dot11RSNAStatsCMACICVErrors */
+		} aes_cmac;
+		struct {
+			u8 rx_pn[IEEE80211_GMAC_PN_LEN];
+			struct crypto_aead *tfm;
+			u32 replays; /* dot11RSNAStatsCMACReplays */
+			u32 icverrors; /* dot11RSNAStatsCMACICVErrors */
+		} aes_gmac;
+		struct {
+			/* Last received packet number. The first
+			 * IEEE80211_NUM_TIDS counters are used with Data
+			 * frames and the last counter is used with Robust
+			 * Management frames.
+			 */
+			u8 rx_pn[IEEE80211_NUM_TIDS + 1][IEEE80211_GCMP_PN_LEN];
+			struct crypto_aead *tfm;
+			u32 replays; /* dot11RSNAStatsGCMPReplays */
+		} gcmp;
+		struct {
+			/* generic cipher scheme */
+			u8 rx_pn[IEEE80211_NUM_TIDS + 1][IEEE80211_MAX_PN_LEN];
+		} gen;
+	} u;
+
+#ifdef CONFIG_MAC80211_DEBUGFS
+	struct {
+		struct dentry *stalink;
+		struct dentry *dir;
+		int cnt;
+	} debugfs;
+#endif
+
+	unsigned int color;
+
+	/*
+	 * key config, must be last because it contains key
+	 * material as variable length member
+	 */
+	struct ieee80211_key_conf conf;
+};
+
+struct ieee80211_key *
+ieee80211_key_alloc(u32 cipher, int idx, size_t key_len,
+		    const u8 *key_data,
+		    size_t seq_len, const u8 *seq);
+/*
+ * Insert a key into data structures (sdata, sta if necessary)
+ * to make it used, free old key. On failure, also free the new key.
+ */
+int ieee80211_key_link(struct ieee80211_key *key,
+		       struct ieee80211_link_data *link,
+		       struct sta_info *sta);
+int ieee80211_set_tx_key(struct ieee80211_key *key);
+void ieee80211_key_free(struct ieee80211_key *key, bool delay_tailroom);
+void ieee80211_key_free_unused(struct ieee80211_key *key);
+void ieee80211_set_default_key(struct ieee80211_link_data *link, int idx,
+			       bool uni, bool multi);
+void ieee80211_set_default_mgmt_key(struct ieee80211_link_data *link,
+				    int idx);
+void ieee80211_set_default_beacon_key(struct ieee80211_link_data *link,
+				      int idx);
+void ieee80211_remove_link_keys(struct ieee80211_link_data *link,
+				struct list_head *keys);
+void ieee80211_free_key_list(struct ieee80211_local *local,
+			     struct list_head *keys);
+void ieee80211_free_keys(struct ieee80211_sub_if_data *sdata,
+			 bool force_synchronize);
+void ieee80211_free_sta_keys(struct ieee80211_local *local,
+			     struct sta_info *sta);
+void ieee80211_reenable_keys(struct ieee80211_sub_if_data *sdata);
+int ieee80211_key_switch_links(struct ieee80211_sub_if_data *sdata,
+			       unsigned long del_links_mask,
+			       unsigned long add_links_mask);
+void ieee80211_delayed_tailroom_dec(struct wiphy *wiphy,
+				    struct wiphy_work *wk);
+
+#endif /* IEEE80211_KEY_H */