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, 143 insertions, 0 deletions

diff --git a/drivers/usb/host/xhci-debugfs.h b/drivers/usb/host/xhci-debugfs.h
new file mode 100644
index 000000000..7c074b4be
--- /dev/null
+++ b/drivers/usb/host/xhci-debugfs.h
@@ -0,0 +1,143 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * xhci-debugfs.h - xHCI debugfs interface
+ *
+ * Copyright (C) 2017 Intel Corporation
+ *
+ * Author: Lu Baolu <baolu.lu@linux.intel.com>
+ */
+
+#ifndef __LINUX_XHCI_DEBUGFS_H
+#define __LINUX_XHCI_DEBUGFS_H
+
+#include <linux/debugfs.h>
+
+#define DEBUGFS_NAMELEN 32
+
+#define REG_CAPLENGTH					0x00
+#define REG_HCSPARAMS1					0x04
+#define REG_HCSPARAMS2					0x08
+#define REG_HCSPARAMS3					0x0c
+#define REG_HCCPARAMS1					0x10
+#define REG_DOORBELLOFF					0x14
+#define REG_RUNTIMEOFF					0x18
+#define REG_HCCPARAMS2					0x1c
+
+#define	REG_USBCMD					0x00
+#define REG_USBSTS					0x04
+#define REG_PAGESIZE					0x08
+#define REG_DNCTRL					0x14
+#define REG_CRCR					0x18
+#define REG_DCBAAP_LOW					0x30
+#define REG_DCBAAP_HIGH					0x34
+#define REG_CONFIG					0x38
+
+#define REG_MFINDEX					0x00
+#define REG_IR0_IMAN					0x20
+#define REG_IR0_IMOD					0x24
+#define REG_IR0_ERSTSZ					0x28
+#define REG_IR0_ERSTBA_LOW				0x30
+#define REG_IR0_ERSTBA_HIGH				0x34
+#define REG_IR0_ERDP_LOW				0x38
+#define REG_IR0_ERDP_HIGH				0x3c
+
+#define REG_EXTCAP_USBLEGSUP				0x00
+#define REG_EXTCAP_USBLEGCTLSTS				0x04
+
+#define REG_EXTCAP_REVISION				0x00
+#define REG_EXTCAP_NAME					0x04
+#define REG_EXTCAP_PORTINFO				0x08
+#define REG_EXTCAP_PORTTYPE				0x0c
+#define REG_EXTCAP_MANTISSA1				0x10
+#define REG_EXTCAP_MANTISSA2				0x14
+#define REG_EXTCAP_MANTISSA3				0x18
+#define REG_EXTCAP_MANTISSA4				0x1c
+#define REG_EXTCAP_MANTISSA5				0x20
+#define REG_EXTCAP_MANTISSA6				0x24
+
+#define REG_EXTCAP_DBC_CAPABILITY			0x00
+#define REG_EXTCAP_DBC_DOORBELL				0x04
+#define REG_EXTCAP_DBC_ERSTSIZE				0x08
+#define REG_EXTCAP_DBC_ERST_LOW				0x10
+#define REG_EXTCAP_DBC_ERST_HIGH			0x14
+#define REG_EXTCAP_DBC_ERDP_LOW				0x18
+#define REG_EXTCAP_DBC_ERDP_HIGH			0x1c
+#define REG_EXTCAP_DBC_CONTROL				0x20
+#define REG_EXTCAP_DBC_STATUS				0x24
+#define REG_EXTCAP_DBC_PORTSC				0x28
+#define REG_EXTCAP_DBC_CONT_LOW				0x30
+#define REG_EXTCAP_DBC_CONT_HIGH			0x34
+#define REG_EXTCAP_DBC_DEVINFO1				0x38
+#define REG_EXTCAP_DBC_DEVINFO2				0x3c
+
+#define dump_register(nm)				\
+{							\
+	.name	= __stringify(nm),			\
+	.offset	= REG_ ##nm,				\
+}
+
+struct xhci_regset {
+	char			name[DEBUGFS_NAMELEN];
+	struct debugfs_regset32	regset;
+	size_t			nregs;
+	struct list_head	list;
+};
+
+struct xhci_file_map {
+	const char		*name;
+	int			(*show)(struct seq_file *s, void *unused);
+};
+
+struct xhci_ep_priv {
+	char			name[DEBUGFS_NAMELEN];
+	struct dentry		*root;
+	struct xhci_stream_info *stream_info;
+	struct xhci_ring	*show_ring;
+	unsigned int		stream_id;
+};
+
+struct xhci_slot_priv {
+	char			name[DEBUGFS_NAMELEN];
+	struct dentry		*root;
+	struct xhci_ep_priv	*eps[31];
+	struct xhci_virt_device	*dev;
+};
+
+#ifdef CONFIG_DEBUG_FS
+void xhci_debugfs_init(struct xhci_hcd *xhci);
+void xhci_debugfs_exit(struct xhci_hcd *xhci);
+void __init xhci_debugfs_create_root(void);
+void __exit xhci_debugfs_remove_root(void);
+void xhci_debugfs_create_slot(struct xhci_hcd *xhci, int slot_id);
+void xhci_debugfs_remove_slot(struct xhci_hcd *xhci, int slot_id);
+void xhci_debugfs_create_endpoint(struct xhci_hcd *xhci,
+				  struct xhci_virt_device *virt_dev,
+				  int ep_index);
+void xhci_debugfs_remove_endpoint(struct xhci_hcd *xhci,
+				  struct xhci_virt_device *virt_dev,
+				  int ep_index);
+void xhci_debugfs_create_stream_files(struct xhci_hcd *xhci,
+				      struct xhci_virt_device *virt_dev,
+				      int ep_index);
+#else
+static inline void xhci_debugfs_init(struct xhci_hcd *xhci) { }
+static inline void xhci_debugfs_exit(struct xhci_hcd *xhci) { }
+static inline void __init xhci_debugfs_create_root(void) { }
+static inline void __exit xhci_debugfs_remove_root(void) { }
+static inline void xhci_debugfs_create_slot(struct xhci_hcd *x, int s) { }
+static inline void xhci_debugfs_remove_slot(struct xhci_hcd *x, int s) { }
+static inline void
+xhci_debugfs_create_endpoint(struct xhci_hcd *xhci,
+			     struct xhci_virt_device *virt_dev,
+			     int ep_index) { }
+static inline void
+xhci_debugfs_remove_endpoint(struct xhci_hcd *xhci,
+			     struct xhci_virt_device *virt_dev,
+			     int ep_index) { }
+static inline void
+xhci_debugfs_create_stream_files(struct xhci_hcd *xhci,
+				 struct xhci_virt_device *virt_dev,
+				 int ep_index) { }
+#endif /* CONFIG_DEBUG_FS */
+
+#endif /* __LINUX_XHCI_DEBUGFS_H */