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, 96 insertions, 0 deletions

diff --git a/mm/cma.h b/mm/cma.h
new file mode 100644
index 000000000..ab6d39898
--- /dev/null
+++ b/mm/cma.h
@@ -0,0 +1,96 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef __MM_CMA_H__
+#define __MM_CMA_H__
+
+#include <linux/cma.h>
+#include <linux/debugfs.h>
+#include <linux/kobject.h>
+
+struct cma_kobject {
+	struct kobject kobj;
+	struct cma *cma;
+};
+
+/*
+ * Multi-range support. This can be useful if the size of the allocation
+ * is not expected to be larger than the alignment (like with hugetlb_cma),
+ * and the total amount of memory requested, while smaller than the total
+ * amount of memory available, is large enough that it doesn't fit in a
+ * single physical memory range because of memory holes.
+ *
+ * Fields:
+ *   @base_pfn: physical address of range
+ *   @early_pfn: first PFN not reserved through cma_reserve_early
+ *   @count: size of range
+ *   @bitmap: bitmap of allocated (1 << order_per_bit)-sized chunks.
+ */
+struct cma_memrange {
+	unsigned long base_pfn;
+	unsigned long count;
+	union {
+		unsigned long early_pfn;
+		unsigned long *bitmap;
+	};
+#ifdef CONFIG_CMA_DEBUGFS
+	struct debugfs_u32_array dfs_bitmap;
+#endif
+};
+#define CMA_MAX_RANGES 8
+
+struct cma {
+	unsigned long   count;
+	unsigned long	available_count;
+	unsigned int order_per_bit; /* Order of pages represented by one bit */
+	spinlock_t	lock;
+	struct mutex alloc_mutex;
+#ifdef CONFIG_CMA_DEBUGFS
+	struct hlist_head mem_head;
+	spinlock_t mem_head_lock;
+#endif
+	char name[CMA_MAX_NAME];
+	int nranges;
+	struct cma_memrange ranges[CMA_MAX_RANGES];
+#ifdef CONFIG_CMA_SYSFS
+	/* the number of CMA page successful allocations */
+	atomic64_t nr_pages_succeeded;
+	/* the number of CMA page allocation failures */
+	atomic64_t nr_pages_failed;
+	/* the number of CMA page released */
+	atomic64_t nr_pages_released;
+	/* kobject requires dynamic object */
+	struct cma_kobject *cma_kobj;
+#endif
+	unsigned long flags;
+	/* NUMA node (NUMA_NO_NODE if unspecified) */
+	int nid;
+};
+
+enum cma_flags {
+	CMA_RESERVE_PAGES_ON_ERROR,
+	CMA_ZONES_VALID,
+	CMA_ZONES_INVALID,
+	CMA_ACTIVATED,
+};
+
+extern struct cma cma_areas[MAX_CMA_AREAS];
+extern unsigned int cma_area_count;
+
+static inline unsigned long cma_bitmap_maxno(struct cma *cma,
+		struct cma_memrange *cmr)
+{
+	return cmr->count >> cma->order_per_bit;
+}
+
+#ifdef CONFIG_CMA_SYSFS
+void cma_sysfs_account_success_pages(struct cma *cma, unsigned long nr_pages);
+void cma_sysfs_account_fail_pages(struct cma *cma, unsigned long nr_pages);
+void cma_sysfs_account_release_pages(struct cma *cma, unsigned long nr_pages);
+#else
+static inline void cma_sysfs_account_success_pages(struct cma *cma,
+						   unsigned long nr_pages) {};
+static inline void cma_sysfs_account_fail_pages(struct cma *cma,
+						unsigned long nr_pages) {};
+static inline void cma_sysfs_account_release_pages(struct cma *cma,
+						   unsigned long nr_pages) {};
+#endif
+#endif