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, 59 insertions, 0 deletions

diff --git a/tools/testing/radix-tree/test.h b/tools/testing/radix-tree/test.h
new file mode 100644
index 000000000..7ef7067e9
--- /dev/null
+++ b/tools/testing/radix-tree/test.h
@@ -0,0 +1,59 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#include <linux/gfp.h>
+#include <linux/types.h>
+#include <linux/radix-tree.h>
+#include <linux/rcupdate.h>
+
+struct item {
+	struct rcu_head	rcu_head;
+	unsigned long index;
+	unsigned int order;
+};
+
+struct item *item_create(unsigned long index, unsigned int order);
+int item_insert(struct radix_tree_root *root, unsigned long index);
+void item_sanity(struct item *item, unsigned long index);
+void item_free(struct item *item, unsigned long index);
+int item_delete(struct radix_tree_root *root, unsigned long index);
+int item_delete_rcu(struct xarray *xa, unsigned long index);
+struct item *item_lookup(struct radix_tree_root *root, unsigned long index);
+
+void item_check_present(struct radix_tree_root *root, unsigned long index);
+void item_check_absent(struct radix_tree_root *root, unsigned long index);
+void item_gang_check_present(struct radix_tree_root *root,
+			unsigned long start, unsigned long nr,
+			int chunk, int hop);
+void item_full_scan(struct radix_tree_root *root, unsigned long start,
+			unsigned long nr, int chunk);
+void item_kill_tree(struct radix_tree_root *root);
+
+int tag_tagged_items(struct xarray *, unsigned long start, unsigned long end,
+		unsigned batch, xa_mark_t iftag, xa_mark_t thentag);
+
+void xarray_tests(void);
+void tag_check(void);
+void multiorder_checks(void);
+void iteration_test(unsigned order, unsigned duration);
+void iteration_test2(unsigned duration);
+void benchmark(void);
+void idr_checks(void);
+void ida_tests(void);
+
+struct item *
+item_tag_set(struct radix_tree_root *root, unsigned long index, int tag);
+struct item *
+item_tag_clear(struct radix_tree_root *root, unsigned long index, int tag);
+int item_tag_get(struct radix_tree_root *root, unsigned long index, int tag);
+void tree_verify_min_height(struct radix_tree_root *root, int maxindex);
+void verify_tag_consistency(struct radix_tree_root *root, unsigned int tag);
+
+extern int nr_allocated;
+
+/* Normally private parts of lib/radix-tree.c */
+struct radix_tree_node *entry_to_node(void *ptr);
+void radix_tree_dump(struct radix_tree_root *root);
+int root_tag_get(struct radix_tree_root *root, unsigned int tag);
+unsigned long node_maxindex(struct radix_tree_node *);
+unsigned long shift_maxindex(unsigned int shift);
+int radix_tree_cpu_dead(unsigned int cpu);
+extern struct radix_tree_preload radix_tree_preloads;