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, 108 insertions, 0 deletions

diff --git a/drivers/usb/chipidea/ulpi.c b/drivers/usb/chipidea/ulpi.c
new file mode 100644
index 000000000..dfec07e8a
--- /dev/null
+++ b/drivers/usb/chipidea/ulpi.c
@@ -0,0 +1,108 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (c) 2016 Linaro Ltd.
+ */
+
+#include <linux/device.h>
+#include <linux/usb/chipidea.h>
+#include <linux/ulpi/interface.h>
+
+#include "ci.h"
+
+#define ULPI_WAKEUP		BIT(31)
+#define ULPI_RUN		BIT(30)
+#define ULPI_WRITE		BIT(29)
+#define ULPI_SYNC_STATE		BIT(27)
+#define ULPI_ADDR(n)		((n) << 16)
+#define ULPI_DATA(n)		(n)
+
+static int ci_ulpi_wait(struct ci_hdrc *ci, u32 mask)
+{
+	unsigned long usec = 10000;
+
+	while (usec--) {
+		if (!hw_read(ci, OP_ULPI_VIEWPORT, mask))
+			return 0;
+
+		udelay(1);
+	}
+
+	return -ETIMEDOUT;
+}
+
+static int ci_ulpi_read(struct device *dev, u8 addr)
+{
+	struct ci_hdrc *ci = dev_get_drvdata(dev);
+	int ret;
+
+	hw_write(ci, OP_ULPI_VIEWPORT, 0xffffffff, ULPI_WRITE | ULPI_WAKEUP);
+	ret = ci_ulpi_wait(ci, ULPI_WAKEUP);
+	if (ret)
+		return ret;
+
+	hw_write(ci, OP_ULPI_VIEWPORT, 0xffffffff, ULPI_RUN | ULPI_ADDR(addr));
+	ret = ci_ulpi_wait(ci, ULPI_RUN);
+	if (ret)
+		return ret;
+
+	return hw_read(ci, OP_ULPI_VIEWPORT, GENMASK(15, 8)) >> 8;
+}
+
+static int ci_ulpi_write(struct device *dev, u8 addr, u8 val)
+{
+	struct ci_hdrc *ci = dev_get_drvdata(dev);
+	int ret;
+
+	hw_write(ci, OP_ULPI_VIEWPORT, 0xffffffff, ULPI_WRITE | ULPI_WAKEUP);
+	ret = ci_ulpi_wait(ci, ULPI_WAKEUP);
+	if (ret)
+		return ret;
+
+	hw_write(ci, OP_ULPI_VIEWPORT, 0xffffffff,
+		 ULPI_RUN | ULPI_WRITE | ULPI_ADDR(addr) | val);
+	return ci_ulpi_wait(ci, ULPI_RUN);
+}
+
+int ci_ulpi_init(struct ci_hdrc *ci)
+{
+	if (ci->platdata->phy_mode != USBPHY_INTERFACE_MODE_ULPI)
+		return 0;
+
+	/*
+	 * Set PORTSC correctly so we can read/write ULPI registers for
+	 * identification purposes
+	 */
+	hw_phymode_configure(ci);
+
+	ci->ulpi_ops.read = ci_ulpi_read;
+	ci->ulpi_ops.write = ci_ulpi_write;
+	ci->ulpi = ulpi_register_interface(ci->dev, &ci->ulpi_ops);
+	if (IS_ERR(ci->ulpi))
+		dev_err(ci->dev, "failed to register ULPI interface");
+
+	return PTR_ERR_OR_ZERO(ci->ulpi);
+}
+
+void ci_ulpi_exit(struct ci_hdrc *ci)
+{
+	if (ci->ulpi) {
+		ulpi_unregister_interface(ci->ulpi);
+		ci->ulpi = NULL;
+	}
+}
+
+int ci_ulpi_resume(struct ci_hdrc *ci)
+{
+	int cnt = 100000;
+
+	if (ci->platdata->phy_mode != USBPHY_INTERFACE_MODE_ULPI)
+		return 0;
+
+	while (cnt-- > 0) {
+		if (hw_read(ci, OP_ULPI_VIEWPORT, ULPI_SYNC_STATE))
+			return 0;
+		udelay(1);
+	}
+
+	return -ETIMEDOUT;
+}