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, 47 insertions, 0 deletions

diff --git a/drivers/mfd/intel-lpss.h b/drivers/mfd/intel-lpss.h
new file mode 100644
index 000000000..6f8f668f4
--- /dev/null
+++ b/drivers/mfd/intel-lpss.h
@@ -0,0 +1,47 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/*
+ * Intel LPSS core support.
+ *
+ * Copyright (C) 2015, Intel Corporation
+ *
+ * Authors: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
+ *          Mika Westerberg <mika.westerberg@linux.intel.com>
+ */
+
+#ifndef __MFD_INTEL_LPSS_H
+#define __MFD_INTEL_LPSS_H
+
+#include <linux/bits.h>
+#include <linux/pm.h>
+
+/*
+ * Some DSDTs have an unused GEXP ACPI device conflicting with I2C4 resources.
+ * Set to ignore resource conflicts with ACPI declared SystemMemory regions.
+ */
+#define QUIRK_IGNORE_RESOURCE_CONFLICTS BIT(0)
+/*
+ * Some devices have misconfigured clock divider due to a firmware bug.
+ * Set this to force the clock divider to 1:1 ratio.
+ */
+#define QUIRK_CLOCK_DIVIDER_UNITY		BIT(1)
+
+struct device;
+struct resource;
+struct software_node;
+
+struct intel_lpss_platform_info {
+	struct resource *mem;
+	int irq;
+	unsigned int quirks;
+	unsigned long clk_rate;
+	const char *clk_con_id;
+	const struct software_node *swnode;
+};
+
+int intel_lpss_probe(struct device *dev,
+		     const struct intel_lpss_platform_info *info);
+void intel_lpss_remove(struct device *dev);
+
+extern const struct dev_pm_ops intel_lpss_pm_ops;
+
+#endif /* __MFD_INTEL_LPSS_H */