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, 142 insertions, 0 deletions

diff --git a/arch/powerpc/boot/ugecon.c b/arch/powerpc/boot/ugecon.c
new file mode 100644
index 000000000..938a38bd4
--- /dev/null
+++ b/arch/powerpc/boot/ugecon.c
@@ -0,0 +1,142 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * arch/powerpc/boot/ugecon.c
+ *
+ * USB Gecko bootwrapper console.
+ * Copyright (C) 2008-2009 The GameCube Linux Team
+ * Copyright (C) 2008,2009 Albert Herranz
+ */
+
+#include <stddef.h>
+#include "stdio.h"
+#include "types.h"
+#include "io.h"
+#include "ops.h"
+
+
+#define EXI_CLK_32MHZ           5
+
+#define EXI_CSR                 0x00
+#define   EXI_CSR_CLKMASK       (0x7<<4)
+#define     EXI_CSR_CLK_32MHZ   (EXI_CLK_32MHZ<<4)
+#define   EXI_CSR_CSMASK        (0x7<<7)
+#define     EXI_CSR_CS_0        (0x1<<7)  /* Chip Select 001 */
+
+#define EXI_CR                  0x0c
+#define   EXI_CR_TSTART         (1<<0)
+#define   EXI_CR_WRITE		(1<<2)
+#define   EXI_CR_READ_WRITE     (2<<2)
+#define   EXI_CR_TLEN(len)      (((len)-1)<<4)
+
+#define EXI_DATA                0x10
+
+
+/* virtual address base for input/output, retrieved from device tree */
+static void *ug_io_base;
+
+
+static u32 ug_io_transaction(u32 in)
+{
+	u32 *csr_reg = ug_io_base + EXI_CSR;
+	u32 *data_reg = ug_io_base + EXI_DATA;
+	u32 *cr_reg = ug_io_base + EXI_CR;
+	u32 csr, data, cr;
+
+	/* select */
+	csr = EXI_CSR_CLK_32MHZ | EXI_CSR_CS_0;
+	out_be32(csr_reg, csr);
+
+	/* read/write */
+	data = in;
+	out_be32(data_reg, data);
+	cr = EXI_CR_TLEN(2) | EXI_CR_READ_WRITE | EXI_CR_TSTART;
+	out_be32(cr_reg, cr);
+
+	while (in_be32(cr_reg) & EXI_CR_TSTART)
+		barrier();
+
+	/* deselect */
+	out_be32(csr_reg, 0);
+
+	data = in_be32(data_reg);
+	return data;
+}
+
+static int ug_is_txfifo_ready(void)
+{
+	return ug_io_transaction(0xc0000000) & 0x04000000;
+}
+
+static void ug_raw_putc(char ch)
+{
+	ug_io_transaction(0xb0000000 | (ch << 20));
+}
+
+static void ug_putc(char ch)
+{
+	int count = 16;
+
+	if (!ug_io_base)
+		return;
+
+	while (!ug_is_txfifo_ready() && count--)
+		barrier();
+	if (count >= 0)
+		ug_raw_putc(ch);
+}
+
+void ug_console_write(const char *buf, int len)
+{
+	char *b = (char *)buf;
+
+	while (len--) {
+		if (*b == '\n')
+			ug_putc('\r');
+		ug_putc(*b++);
+	}
+}
+
+static int ug_is_adapter_present(void)
+{
+	if (!ug_io_base)
+		return 0;
+	return ug_io_transaction(0x90000000) == 0x04700000;
+}
+
+static void *ug_grab_exi_io_base(void)
+{
+	u32 v;
+	void *devp;
+
+	devp = find_node_by_compatible(NULL, "nintendo,flipper-exi");
+	if (devp == NULL)
+		goto err_out;
+	if (getprop(devp, "virtual-reg", &v, sizeof(v)) != sizeof(v))
+		goto err_out;
+
+	return (void *)v;
+
+err_out:
+	return NULL;
+}
+
+void *ug_probe(void)
+{
+	void *exi_io_base;
+	int i;
+
+	exi_io_base = ug_grab_exi_io_base();
+	if (!exi_io_base)
+		return NULL;
+
+	/* look for a usbgecko on memcard slots A and B */
+	for (i = 0; i < 2; i++) {
+		ug_io_base = exi_io_base + 0x14 * i;
+		if (ug_is_adapter_present())
+			break;
+	}
+	if (i == 2)
+		ug_io_base = NULL;
+	return ug_io_base;
+}
+