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, 130 insertions, 0 deletions

diff --git a/mm/memcontrol-v1.h b/mm/memcontrol-v1.h
new file mode 100644
index 000000000..1e394269c
--- /dev/null
+++ b/mm/memcontrol-v1.h
@@ -0,0 +1,130 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later */
+
+#ifndef __MM_MEMCONTROL_V1_H
+#define __MM_MEMCONTROL_V1_H
+
+#include <linux/cgroup-defs.h>
+#include <linux/memcontrol.h>
+
+/* Cgroup v1 and v2 common declarations */
+
+/*
+ * Iteration constructs for visiting all cgroups (under a tree).  If
+ * loops are exited prematurely (break), mem_cgroup_iter_break() must
+ * be used for reference counting.
+ */
+#define for_each_mem_cgroup_tree(iter, root)		\
+	for (iter = mem_cgroup_iter(root, NULL, NULL);	\
+	     iter != NULL;				\
+	     iter = mem_cgroup_iter(root, iter, NULL))
+
+void drain_all_stock(struct mem_cgroup *root_memcg);
+
+int memory_stat_show(struct seq_file *m, void *v);
+
+struct mem_cgroup *mem_cgroup_private_id_get_online(struct mem_cgroup *memcg,
+						    unsigned int n);
+
+/* Cgroup v1-specific declarations */
+#ifdef CONFIG_MEMCG_V1
+
+/* Whether legacy memory+swap accounting is active */
+static inline bool do_memsw_account(void)
+{
+	return !cgroup_subsys_on_dfl(memory_cgrp_subsys);
+}
+
+unsigned long memcg_events_local(struct mem_cgroup *memcg, int event);
+unsigned long memcg_page_state_local(struct mem_cgroup *memcg, int idx);
+unsigned long memcg_page_state_local_output(struct mem_cgroup *memcg, int item);
+bool memcg1_alloc_events(struct mem_cgroup *memcg);
+void memcg1_free_events(struct mem_cgroup *memcg);
+
+void memcg1_memcg_init(struct mem_cgroup *memcg);
+void memcg1_remove_from_trees(struct mem_cgroup *memcg);
+
+static inline void memcg1_soft_limit_reset(struct mem_cgroup *memcg)
+{
+	WRITE_ONCE(memcg->soft_limit, PAGE_COUNTER_MAX);
+}
+
+struct cgroup_taskset;
+void memcg1_css_offline(struct mem_cgroup *memcg);
+
+/* for encoding cft->private value on file */
+enum res_type {
+	_MEM,
+	_MEMSWAP,
+	_KMEM,
+	_TCP,
+};
+
+bool memcg1_oom_prepare(struct mem_cgroup *memcg, bool *locked);
+void memcg1_oom_finish(struct mem_cgroup *memcg, bool locked);
+void memcg1_oom_recover(struct mem_cgroup *memcg);
+
+void memcg1_commit_charge(struct folio *folio, struct mem_cgroup *memcg);
+void memcg1_uncharge_batch(struct mem_cgroup *memcg, unsigned long pgpgout,
+			   unsigned long nr_memory, int nid);
+
+void memcg1_stat_format(struct mem_cgroup *memcg, struct seq_buf *s);
+void reparent_memcg1_state_local(struct mem_cgroup *memcg, struct mem_cgroup *parent);
+void reparent_memcg1_lruvec_state_local(struct mem_cgroup *memcg, struct mem_cgroup *parent);
+
+void reparent_memcg_state_local(struct mem_cgroup *memcg,
+				struct mem_cgroup *parent, int idx);
+void reparent_memcg_lruvec_state_local(struct mem_cgroup *memcg,
+				       struct mem_cgroup *parent, int idx);
+
+void memcg1_account_kmem(struct mem_cgroup *memcg, int nr_pages);
+static inline bool memcg1_tcpmem_active(struct mem_cgroup *memcg)
+{
+	return memcg->tcpmem_active;
+}
+bool memcg1_charge_skmem(struct mem_cgroup *memcg, unsigned int nr_pages,
+			 gfp_t gfp_mask);
+static inline void memcg1_uncharge_skmem(struct mem_cgroup *memcg, unsigned int nr_pages)
+{
+	page_counter_uncharge(&memcg->tcpmem, nr_pages);
+}
+
+extern struct cftype memsw_files[];
+extern struct cftype mem_cgroup_legacy_files[];
+
+#else	/* CONFIG_MEMCG_V1 */
+
+static inline bool do_memsw_account(void) { return false; }
+static inline bool memcg1_alloc_events(struct mem_cgroup *memcg) { return true; }
+static inline void memcg1_free_events(struct mem_cgroup *memcg) {}
+
+static inline void memcg1_memcg_init(struct mem_cgroup *memcg) {}
+static inline void memcg1_remove_from_trees(struct mem_cgroup *memcg) {}
+static inline void memcg1_soft_limit_reset(struct mem_cgroup *memcg) {}
+static inline void memcg1_css_offline(struct mem_cgroup *memcg) {}
+
+static inline bool memcg1_oom_prepare(struct mem_cgroup *memcg, bool *locked)
+{
+	*locked = false;
+	return true;
+}
+static inline void memcg1_oom_finish(struct mem_cgroup *memcg, bool locked) {}
+static inline void memcg1_oom_recover(struct mem_cgroup *memcg) {}
+
+static inline void memcg1_commit_charge(struct folio *folio,
+					struct mem_cgroup *memcg) {}
+
+static inline void memcg1_uncharge_batch(struct mem_cgroup *memcg,
+					 unsigned long pgpgout,
+					 unsigned long nr_memory, int nid) {}
+
+static inline void memcg1_stat_format(struct mem_cgroup *memcg, struct seq_buf *s) {}
+
+static inline void memcg1_account_kmem(struct mem_cgroup *memcg, int nr_pages) {}
+static inline bool memcg1_tcpmem_active(struct mem_cgroup *memcg) { return false; }
+static inline bool memcg1_charge_skmem(struct mem_cgroup *memcg, unsigned int nr_pages,
+				       gfp_t gfp_mask) { return true; }
+static inline void memcg1_uncharge_skmem(struct mem_cgroup *memcg, unsigned int nr_pages) {}
+
+#endif	/* CONFIG_MEMCG_V1 */
+
+#endif	/* __MM_MEMCONTROL_V1_H */