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, 62 insertions, 0 deletions

diff --git a/include/uapi/asm-generic/resource.h b/include/uapi/asm-generic/resource.h
new file mode 100644
index 000000000..f12db7a0d
--- /dev/null
+++ b/include/uapi/asm-generic/resource.h
@@ -0,0 +1,62 @@
+/* SPDX-License-Identifier: GPL-2.0 WITH Linux-syscall-note */
+#ifndef _UAPI_ASM_GENERIC_RESOURCE_H
+#define _UAPI_ASM_GENERIC_RESOURCE_H
+
+/*
+ * Resource limit IDs
+ *
+ * ( Compatibility detail: there are architectures that have
+ *   a different rlimit ID order in the 5-9 range and want
+ *   to keep that order for binary compatibility. The reasons
+ *   are historic and all new rlimits are identical across all
+ *   arches. If an arch has such special order for some rlimits
+ *   then it defines them prior including asm-generic/resource.h. )
+ */
+
+#define RLIMIT_CPU		0	/* CPU time in sec */
+#define RLIMIT_FSIZE		1	/* Maximum filesize */
+#define RLIMIT_DATA		2	/* max data size */
+#define RLIMIT_STACK		3	/* max stack size */
+#define RLIMIT_CORE		4	/* max core file size */
+
+#ifndef RLIMIT_RSS
+# define RLIMIT_RSS		5	/* max resident set size */
+#endif
+
+#ifndef RLIMIT_NPROC
+# define RLIMIT_NPROC		6	/* max number of processes */
+#endif
+
+#ifndef RLIMIT_NOFILE
+# define RLIMIT_NOFILE		7	/* max number of open files */
+#endif
+
+#ifndef RLIMIT_MEMLOCK
+# define RLIMIT_MEMLOCK		8	/* max locked-in-memory address space */
+#endif
+
+#ifndef RLIMIT_AS
+# define RLIMIT_AS		9	/* address space limit */
+#endif
+
+#define RLIMIT_LOCKS		10	/* maximum file locks held */
+#define RLIMIT_SIGPENDING	11	/* max number of pending signals */
+#define RLIMIT_MSGQUEUE		12	/* maximum bytes in POSIX mqueues */
+#define RLIMIT_NICE		13	/* max nice prio allowed to raise to
+					   0-39 for nice level 19 .. -20 */
+#define RLIMIT_RTPRIO		14	/* maximum realtime priority */
+#define RLIMIT_RTTIME		15	/* timeout for RT tasks in us */
+#define RLIM_NLIMITS		16
+
+/*
+ * SuS says limits have to be unsigned.
+ * Which makes a ton more sense anyway.
+ *
+ * Some architectures override this (for compatibility reasons):
+ */
+#ifndef RLIM_INFINITY
+# define RLIM_INFINITY		(~0UL)
+#endif
+
+
+#endif /* _UAPI_ASM_GENERIC_RESOURCE_H */