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, 219 insertions, 0 deletions

diff --git a/net/mac802154/mib.c b/net/mac802154/mib.c
new file mode 100644
index 000000000..81666e1d7
--- /dev/null
+++ b/net/mac802154/mib.c
@@ -0,0 +1,219 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright 2007-2012 Siemens AG
+ *
+ * Written by:
+ * Dmitry Eremin-Solenikov <dbaryshkov@gmail.com>
+ * Sergey Lapin <slapin@ossfans.org>
+ * Maxim Gorbachyov <maxim.gorbachev@siemens.com>
+ * Alexander Smirnov <alex.bluesman.smirnov@gmail.com>
+ */
+
+#include <linux/if_arp.h>
+
+#include <net/mac802154.h>
+#include <net/ieee802154_netdev.h>
+#include <net/cfg802154.h>
+
+#include "ieee802154_i.h"
+#include "driver-ops.h"
+
+void mac802154_dev_set_page_channel(struct net_device *dev, u8 page, u8 chan)
+{
+	struct ieee802154_sub_if_data *sdata = IEEE802154_DEV_TO_SUB_IF(dev);
+	struct ieee802154_local *local = sdata->local;
+	int res;
+
+	ASSERT_RTNL();
+
+	BUG_ON(dev->type != ARPHRD_IEEE802154);
+
+	res = drv_set_channel(local, page, chan);
+	if (res) {
+		pr_debug("set_channel failed\n");
+	} else {
+		local->phy->current_channel = chan;
+		local->phy->current_page = page;
+	}
+}
+
+int mac802154_get_params(struct net_device *dev,
+			 struct ieee802154_llsec_params *params)
+{
+	struct ieee802154_sub_if_data *sdata = IEEE802154_DEV_TO_SUB_IF(dev);
+	int res;
+
+	BUG_ON(dev->type != ARPHRD_IEEE802154);
+
+	mutex_lock(&sdata->sec_mtx);
+	res = mac802154_llsec_get_params(&sdata->sec, params);
+	mutex_unlock(&sdata->sec_mtx);
+
+	return res;
+}
+
+int mac802154_set_params(struct net_device *dev,
+			 const struct ieee802154_llsec_params *params,
+			 int changed)
+{
+	struct ieee802154_sub_if_data *sdata = IEEE802154_DEV_TO_SUB_IF(dev);
+	int res;
+
+	BUG_ON(dev->type != ARPHRD_IEEE802154);
+
+	mutex_lock(&sdata->sec_mtx);
+	res = mac802154_llsec_set_params(&sdata->sec, params, changed);
+	mutex_unlock(&sdata->sec_mtx);
+
+	return res;
+}
+
+int mac802154_add_key(struct net_device *dev,
+		      const struct ieee802154_llsec_key_id *id,
+		      const struct ieee802154_llsec_key *key)
+{
+	struct ieee802154_sub_if_data *sdata = IEEE802154_DEV_TO_SUB_IF(dev);
+	int res;
+
+	BUG_ON(dev->type != ARPHRD_IEEE802154);
+
+	mutex_lock(&sdata->sec_mtx);
+	res = mac802154_llsec_key_add(&sdata->sec, id, key);
+	mutex_unlock(&sdata->sec_mtx);
+
+	return res;
+}
+
+int mac802154_del_key(struct net_device *dev,
+		      const struct ieee802154_llsec_key_id *id)
+{
+	struct ieee802154_sub_if_data *sdata = IEEE802154_DEV_TO_SUB_IF(dev);
+	int res;
+
+	BUG_ON(dev->type != ARPHRD_IEEE802154);
+
+	mutex_lock(&sdata->sec_mtx);
+	res = mac802154_llsec_key_del(&sdata->sec, id);
+	mutex_unlock(&sdata->sec_mtx);
+
+	return res;
+}
+
+int mac802154_add_dev(struct net_device *dev,
+		      const struct ieee802154_llsec_device *llsec_dev)
+{
+	struct ieee802154_sub_if_data *sdata = IEEE802154_DEV_TO_SUB_IF(dev);
+	int res;
+
+	BUG_ON(dev->type != ARPHRD_IEEE802154);
+
+	mutex_lock(&sdata->sec_mtx);
+	res = mac802154_llsec_dev_add(&sdata->sec, llsec_dev);
+	mutex_unlock(&sdata->sec_mtx);
+
+	return res;
+}
+
+int mac802154_del_dev(struct net_device *dev, __le64 dev_addr)
+{
+	struct ieee802154_sub_if_data *sdata = IEEE802154_DEV_TO_SUB_IF(dev);
+	int res;
+
+	BUG_ON(dev->type != ARPHRD_IEEE802154);
+
+	mutex_lock(&sdata->sec_mtx);
+	res = mac802154_llsec_dev_del(&sdata->sec, dev_addr);
+	mutex_unlock(&sdata->sec_mtx);
+
+	return res;
+}
+
+int mac802154_add_devkey(struct net_device *dev,
+			 __le64 device_addr,
+			 const struct ieee802154_llsec_device_key *key)
+{
+	struct ieee802154_sub_if_data *sdata = IEEE802154_DEV_TO_SUB_IF(dev);
+	int res;
+
+	BUG_ON(dev->type != ARPHRD_IEEE802154);
+
+	mutex_lock(&sdata->sec_mtx);
+	res = mac802154_llsec_devkey_add(&sdata->sec, device_addr, key);
+	mutex_unlock(&sdata->sec_mtx);
+
+	return res;
+}
+
+int mac802154_del_devkey(struct net_device *dev,
+			 __le64 device_addr,
+			 const struct ieee802154_llsec_device_key *key)
+{
+	struct ieee802154_sub_if_data *sdata = IEEE802154_DEV_TO_SUB_IF(dev);
+	int res;
+
+	BUG_ON(dev->type != ARPHRD_IEEE802154);
+
+	mutex_lock(&sdata->sec_mtx);
+	res = mac802154_llsec_devkey_del(&sdata->sec, device_addr, key);
+	mutex_unlock(&sdata->sec_mtx);
+
+	return res;
+}
+
+int mac802154_add_seclevel(struct net_device *dev,
+			   const struct ieee802154_llsec_seclevel *sl)
+{
+	struct ieee802154_sub_if_data *sdata = IEEE802154_DEV_TO_SUB_IF(dev);
+	int res;
+
+	BUG_ON(dev->type != ARPHRD_IEEE802154);
+
+	mutex_lock(&sdata->sec_mtx);
+	res = mac802154_llsec_seclevel_add(&sdata->sec, sl);
+	mutex_unlock(&sdata->sec_mtx);
+
+	return res;
+}
+
+int mac802154_del_seclevel(struct net_device *dev,
+			   const struct ieee802154_llsec_seclevel *sl)
+{
+	struct ieee802154_sub_if_data *sdata = IEEE802154_DEV_TO_SUB_IF(dev);
+	int res;
+
+	BUG_ON(dev->type != ARPHRD_IEEE802154);
+
+	mutex_lock(&sdata->sec_mtx);
+	res = mac802154_llsec_seclevel_del(&sdata->sec, sl);
+	mutex_unlock(&sdata->sec_mtx);
+
+	return res;
+}
+
+void mac802154_lock_table(struct net_device *dev)
+{
+	struct ieee802154_sub_if_data *sdata = IEEE802154_DEV_TO_SUB_IF(dev);
+
+	BUG_ON(dev->type != ARPHRD_IEEE802154);
+
+	mutex_lock(&sdata->sec_mtx);
+}
+
+void mac802154_get_table(struct net_device *dev,
+			 struct ieee802154_llsec_table **t)
+{
+	struct ieee802154_sub_if_data *sdata = IEEE802154_DEV_TO_SUB_IF(dev);
+
+	BUG_ON(dev->type != ARPHRD_IEEE802154);
+
+	*t = &sdata->sec.table;
+}
+
+void mac802154_unlock_table(struct net_device *dev)
+{
+	struct ieee802154_sub_if_data *sdata = IEEE802154_DEV_TO_SUB_IF(dev);
+
+	BUG_ON(dev->type != ARPHRD_IEEE802154);
+
+	mutex_unlock(&sdata->sec_mtx);
+}