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

4 files changed, 300 insertions, 0 deletions

diff --git a/arch/m68k/apollo/Makefile b/arch/m68k/apollo/Makefile
new file mode 100644
index 000000000..676c74b26
--- /dev/null
+++ b/arch/m68k/apollo/Makefile
@@ -0,0 +1,6 @@
+# SPDX-License-Identifier: GPL-2.0-only
+#
+# Makefile for Linux arch/m68k/apollo source directory
+#
+
+obj-y		:= config.o dn_ints.o
diff --git a/arch/m68k/apollo/apollo.h b/arch/m68k/apollo/apollo.h
new file mode 100644
index 000000000..1fe9d856d
--- /dev/null
+++ b/arch/m68k/apollo/apollo.h
@@ -0,0 +1,4 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+
+/* dn_ints.c */
+void dn_init_IRQ(void);
diff --git a/arch/m68k/apollo/config.c b/arch/m68k/apollo/config.c
new file mode 100644
index 000000000..e324c5f67
--- /dev/null
+++ b/arch/m68k/apollo/config.c
@@ -0,0 +1,240 @@
+// SPDX-License-Identifier: GPL-2.0
+#include <linux/init.h>
+#include <linux/types.h>
+#include <linux/kernel.h>
+#include <linux/mm.h>
+#include <linux/rtc.h>
+#include <linux/interrupt.h>
+
+#include <asm/setup.h>
+#include <asm/bootinfo.h>
+#include <asm/bootinfo-apollo.h>
+#include <asm/byteorder.h>
+#include <asm/apollohw.h>
+#include <asm/irq.h>
+#include <asm/machdep.h>
+#include <asm/config.h>
+
+#include "apollo.h"
+
+u_long sio01_physaddr;
+u_long sio23_physaddr;
+u_long rtc_physaddr;
+u_long pica_physaddr;
+u_long picb_physaddr;
+u_long cpuctrl_physaddr;
+u_long timer_physaddr;
+u_long apollo_model;
+
+extern void dn_sched_init(void);
+extern int dn_dummy_hwclk(int, struct rtc_time *);
+static void dn_dummy_reset(void);
+#ifdef CONFIG_HEARTBEAT
+static void dn_heartbeat(int on);
+#endif
+static irqreturn_t dn_timer_int(int irq,void *);
+static void dn_get_model(char *model);
+static const char *apollo_models[] = {
+	[APOLLO_DN3000-APOLLO_DN3000] = "DN3000 (Otter)",
+	[APOLLO_DN3010-APOLLO_DN3000] = "DN3010 (Otter)",
+	[APOLLO_DN3500-APOLLO_DN3000] = "DN3500 (Cougar II)",
+	[APOLLO_DN4000-APOLLO_DN3000] = "DN4000 (Mink)",
+	[APOLLO_DN4500-APOLLO_DN3000] = "DN4500 (Roadrunner)"
+};
+
+int __init apollo_parse_bootinfo(const struct bi_record *record)
+{
+	int unknown = 0;
+	const void *data = record->data;
+
+	switch (be16_to_cpu(record->tag)) {
+	case BI_APOLLO_MODEL:
+		apollo_model = be32_to_cpup(data);
+		break;
+
+	default:
+		 unknown=1;
+	}
+
+	return unknown;
+}
+
+static void __init dn_setup_model(void)
+{
+	pr_info("Apollo hardware found: [%s]\n",
+		apollo_models[apollo_model - APOLLO_DN3000]);
+
+	switch(apollo_model) {
+		case APOLLO_UNKNOWN:
+			panic("Unknown apollo model");
+			break;
+		case APOLLO_DN3000:
+		case APOLLO_DN3010:
+			sio01_physaddr=SAU8_SIO01_PHYSADDR;
+			rtc_physaddr=SAU8_RTC_PHYSADDR;
+			pica_physaddr=SAU8_PICA;
+			picb_physaddr=SAU8_PICB;
+			cpuctrl_physaddr=SAU8_CPUCTRL;
+			timer_physaddr=SAU8_TIMER;
+			break;
+		case APOLLO_DN4000:
+			sio01_physaddr=SAU7_SIO01_PHYSADDR;
+			sio23_physaddr=SAU7_SIO23_PHYSADDR;
+			rtc_physaddr=SAU7_RTC_PHYSADDR;
+			pica_physaddr=SAU7_PICA;
+			picb_physaddr=SAU7_PICB;
+			cpuctrl_physaddr=SAU7_CPUCTRL;
+			timer_physaddr=SAU7_TIMER;
+			break;
+		case APOLLO_DN4500:
+			panic("Apollo model not yet supported");
+			break;
+		case APOLLO_DN3500:
+			sio01_physaddr=SAU7_SIO01_PHYSADDR;
+			sio23_physaddr=SAU7_SIO23_PHYSADDR;
+			rtc_physaddr=SAU7_RTC_PHYSADDR;
+			pica_physaddr=SAU7_PICA;
+			picb_physaddr=SAU7_PICB;
+			cpuctrl_physaddr=SAU7_CPUCTRL;
+			timer_physaddr=SAU7_TIMER;
+			break;
+		default:
+			panic("Undefined apollo model");
+			break;
+	}
+
+
+}
+
+static void dn_serial_print(const char *str)
+{
+    while (*str) {
+        if (*str == '\n') {
+            sio01.rhrb_thrb = (unsigned char)'\r';
+            while (!(sio01.srb_csrb & 0x4))
+                ;
+        }
+        sio01.rhrb_thrb = (unsigned char)*str++;
+        while (!(sio01.srb_csrb & 0x4))
+            ;
+    }
+}
+
+void __init config_apollo(void)
+{
+	int i;
+
+	dn_setup_model();
+
+	mach_sched_init=dn_sched_init; /* */
+	mach_init_IRQ=dn_init_IRQ;
+	mach_hwclk           = dn_dummy_hwclk; /* */
+	mach_reset	     = dn_dummy_reset;  /* */
+#ifdef CONFIG_HEARTBEAT
+	mach_heartbeat = dn_heartbeat;
+#endif
+	mach_get_model       = dn_get_model;
+
+	cpuctrl=0xaa00;
+
+	/* clear DMA translation table */
+	for(i=0;i<0x400;i++)
+		addr_xlat_map[i]=0;
+
+}
+
+irqreturn_t dn_timer_int(int irq, void *dev_id)
+{
+	unsigned char *at = (unsigned char *)apollo_timer;
+
+	legacy_timer_tick(1);
+	timer_heartbeat();
+
+	READ_ONCE(*(at + 3));
+	READ_ONCE(*(at + 5));
+
+	return IRQ_HANDLED;
+}
+
+void dn_sched_init(void)
+{
+	/* program timer 1 */
+	*(volatile unsigned char *)(apollo_timer + 3) = 0x01;
+	*(volatile unsigned char *)(apollo_timer + 1) = 0x40;
+	*(volatile unsigned char *)(apollo_timer + 5) = 0x09;
+	*(volatile unsigned char *)(apollo_timer + 7) = 0xc4;
+
+	/* enable IRQ of PIC B */
+	*(volatile unsigned char *)(pica+1)&=(~8);
+
+#if 0
+	pr_info("*(0x10803) %02x\n",
+		*(volatile unsigned char *)(apollo_timer + 0x3));
+	pr_info("*(0x10803) %02x\n",
+		*(volatile unsigned char *)(apollo_timer + 0x3));
+#endif
+
+	if (request_irq(IRQ_APOLLO, dn_timer_int, 0, "time", NULL))
+		pr_err("Couldn't register timer interrupt\n");
+}
+
+int dn_dummy_hwclk(int op, struct rtc_time *t) {
+
+
+  if(!op) { /* read */
+    t->tm_sec=rtc->second;
+    t->tm_min=rtc->minute;
+    t->tm_hour=rtc->hours;
+    t->tm_mday=rtc->day_of_month;
+    t->tm_wday=rtc->day_of_week;
+    t->tm_mon = rtc->month - 1;
+    t->tm_year=rtc->year;
+    if (t->tm_year < 70)
+	t->tm_year += 100;
+  } else {
+    rtc->second=t->tm_sec;
+    rtc->minute=t->tm_min;
+    rtc->hours=t->tm_hour;
+    rtc->day_of_month=t->tm_mday;
+    if(t->tm_wday!=-1)
+      rtc->day_of_week=t->tm_wday;
+    rtc->month = t->tm_mon + 1;
+    rtc->year = t->tm_year % 100;
+  }
+
+  return 0;
+
+}
+
+static void dn_dummy_reset(void)
+{
+  dn_serial_print("The end !\n");
+
+  for(;;);
+
+}
+
+static void dn_get_model(char *model)
+{
+    strcpy(model, "Apollo ");
+    if (apollo_model >= APOLLO_DN3000 && apollo_model <= APOLLO_DN4500)
+        strcat(model, apollo_models[apollo_model - APOLLO_DN3000]);
+}
+
+#ifdef CONFIG_HEARTBEAT
+static int dn_cpuctrl=0xff00;
+
+static void dn_heartbeat(int on) {
+
+	if(on) {
+		dn_cpuctrl&=~0x100;
+		cpuctrl=dn_cpuctrl;
+	}
+	else {
+		dn_cpuctrl&=~0x100;
+		dn_cpuctrl|=0x100;
+		cpuctrl=dn_cpuctrl;
+	}
+}
+#endif
+
diff --git a/arch/m68k/apollo/dn_ints.c b/arch/m68k/apollo/dn_ints.c
new file mode 100644
index 000000000..ba96a92f8
--- /dev/null
+++ b/arch/m68k/apollo/dn_ints.c
@@ -0,0 +1,50 @@
+// SPDX-License-Identifier: GPL-2.0
+#include <linux/interrupt.h>
+#include <linux/irq.h>
+
+#include <asm/traps.h>
+#include <asm/apollohw.h>
+
+#include "apollo.h"
+
+static unsigned int apollo_irq_startup(struct irq_data *data)
+{
+	unsigned int irq = data->irq;
+
+	if (irq < 8)
+		*(volatile unsigned char *)(pica+1) &= ~(1 << irq);
+	else
+		*(volatile unsigned char *)(picb+1) &= ~(1 << (irq - 8));
+	return 0;
+}
+
+static void apollo_irq_shutdown(struct irq_data *data)
+{
+	unsigned int irq = data->irq;
+
+	if (irq < 8)
+		*(volatile unsigned char *)(pica+1) |= (1 << irq);
+	else
+		*(volatile unsigned char *)(picb+1) |= (1 << (irq - 8));
+}
+
+static void apollo_irq_eoi(struct irq_data *data)
+{
+	*(volatile unsigned char *)(pica) = 0x20;
+	*(volatile unsigned char *)(picb) = 0x20;
+}
+
+static struct irq_chip apollo_irq_chip = {
+	.name           = "apollo",
+	.irq_startup    = apollo_irq_startup,
+	.irq_shutdown   = apollo_irq_shutdown,
+	.irq_eoi	= apollo_irq_eoi,
+};
+
+
+void __init dn_init_IRQ(void)
+{
+	m68k_setup_user_interrupt(VEC_USER + 96, 16);
+	m68k_setup_irq_controller(&apollo_irq_chip, handle_fasteoi_irq,
+				  IRQ_APOLLO, 16);
+}