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, 118 insertions, 0 deletions

diff --git a/arch/mips/econet/init.c b/arch/mips/econet/init.c
new file mode 100644
index 000000000..ae45b8788
--- /dev/null
+++ b/arch/mips/econet/init.c
@@ -0,0 +1,118 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * EcoNet setup code
+ *
+ * Copyright (C) 2025 Caleb James DeLisle <cjd@cjdns.fr>
+ */
+
+#include <linux/init.h>
+#include <linux/of_clk.h>
+#include <linux/irqchip.h>
+
+#include <asm/addrspace.h>
+#include <asm/io.h>
+#include <asm/bootinfo.h>
+#include <asm/time.h>
+#include <asm/prom.h>
+#include <asm/smp-ops.h>
+#include <asm/reboot.h>
+
+#define CR_AHB_RSTCR		((void __iomem *)CKSEG1ADDR(0x1fb00040))
+#define RESET			BIT(31)
+
+#define UART_BASE		CKSEG1ADDR(0x1fbf0003)
+#define UART_REG_SHIFT		2
+
+static void hw_reset(char *command)
+{
+	iowrite32(RESET, CR_AHB_RSTCR);
+}
+
+/*
+ * vsmp_init_secondary expects either GIC or cascading interrupt configuration.
+ * The EN751221 is a 34Kc with VEIC, but not GIC compatible. The cascading
+ * configuration enables lines 6 and 7 for performance counters, but when this
+ * is done on the EN751221 intc with VEIC enabled, it causes the whole intc to
+ * stop sending interrupts.
+ * Only unmask lines 0 and 1 (software interrupts) in init_secondary.
+ */
+#ifdef CONFIG_MIPS_MT_SMP
+extern const struct plat_smp_ops vsmp_smp_ops;
+static struct plat_smp_ops en75_smp_ops __ro_after_init;
+
+static void en751221_init_secondary(void)
+{
+	write_c0_status((read_c0_status() & ~ST0_IM) |
+		(STATUSF_IP0 | STATUSF_IP1));
+}
+
+static int __init en751221_register_vsmp_smp_ops(void)
+{
+	if (!cpu_has_mipsmt)
+		return -ENODEV;
+
+	en75_smp_ops = vsmp_smp_ops;
+	en75_smp_ops.init_secondary = en751221_init_secondary;
+	register_smp_ops(&en75_smp_ops);
+	return 0;
+}
+#else
+static int __init en751221_register_vsmp_smp_ops(void)
+{
+	return -ENODEV;
+}
+#endif /* CONFIG_MIPS_MT_SMP */
+
+/* 1. Bring up early printk. */
+void __init prom_init(void)
+{
+	setup_8250_early_printk_port(UART_BASE, UART_REG_SHIFT, 0);
+	_machine_restart = hw_reset;
+}
+
+/* 2. Parse the DT and find memory */
+void __init plat_mem_setup(void)
+{
+	void *dtb;
+
+	set_io_port_base(KSEG1);
+
+	dtb = get_fdt();
+	if (!dtb)
+		panic("no dtb found");
+
+	__dt_setup_arch(dtb);
+
+	early_init_dt_scan_memory();
+}
+
+/* 3. Overload __weak device_tree_init(), add SMP ops */
+void __init device_tree_init(void)
+{
+	unflatten_and_copy_device_tree();
+
+	/* EN751221 dual-vpe */
+	if (!en751221_register_vsmp_smp_ops())
+		return;
+
+	/* EN751221 with MT_SMP disabled */
+	register_up_smp_ops();
+}
+
+const char *get_system_type(void)
+{
+	return "EcoNet-EN75xx";
+}
+
+/* 4. Initialize the IRQ subsystem */
+void __init arch_init_irq(void)
+{
+	irqchip_init();
+}
+
+/* 5. Timers */
+void __init plat_time_init(void)
+{
+	of_clk_init(NULL);
+	timer_probe();
+}