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, 180 insertions, 0 deletions

diff --git a/include/uapi/asm-generic/socket.h b/include/uapi/asm-generic/socket.h
new file mode 100644
index 000000000..84ea7b929
--- /dev/null
+++ b/include/uapi/asm-generic/socket.h
@@ -0,0 +1,180 @@
+/* SPDX-License-Identifier: GPL-2.0 WITH Linux-syscall-note */
+#ifndef __ASM_GENERIC_SOCKET_H
+#define __ASM_GENERIC_SOCKET_H
+
+#include <linux/posix_types.h>
+#include <asm/sockios.h>
+
+/* For setsockopt(2) */
+#define SOL_SOCKET	1
+
+#define SO_DEBUG	1
+#define SO_REUSEADDR	2
+#define SO_TYPE		3
+#define SO_ERROR	4
+#define SO_DONTROUTE	5
+#define SO_BROADCAST	6
+#define SO_SNDBUF	7
+#define SO_RCVBUF	8
+#define SO_SNDBUFFORCE	32
+#define SO_RCVBUFFORCE	33
+#define SO_KEEPALIVE	9
+#define SO_OOBINLINE	10
+#define SO_NO_CHECK	11
+#define SO_PRIORITY	12
+#define SO_LINGER	13
+#define SO_BSDCOMPAT	14
+#define SO_REUSEPORT	15
+#ifndef SO_PASSCRED /* powerpc only differs in these */
+#define SO_PASSCRED	16
+#define SO_PEERCRED	17
+#define SO_RCVLOWAT	18
+#define SO_SNDLOWAT	19
+#define SO_RCVTIMEO_OLD	20
+#define SO_SNDTIMEO_OLD	21
+#endif
+
+/* Security levels - as per NRL IPv6 - don't actually do anything */
+#define SO_SECURITY_AUTHENTICATION		22
+#define SO_SECURITY_ENCRYPTION_TRANSPORT	23
+#define SO_SECURITY_ENCRYPTION_NETWORK		24
+
+#define SO_BINDTODEVICE	25
+
+/* Socket filtering */
+#define SO_ATTACH_FILTER	26
+#define SO_DETACH_FILTER	27
+#define SO_GET_FILTER		SO_ATTACH_FILTER
+
+#define SO_PEERNAME		28
+
+#define SO_ACCEPTCONN		30
+
+#define SO_PEERSEC		31
+#define SO_PASSSEC		34
+
+#define SO_MARK			36
+
+#define SO_PROTOCOL		38
+#define SO_DOMAIN		39
+
+#define SO_RXQ_OVFL             40
+
+#define SO_WIFI_STATUS		41
+#define SCM_WIFI_STATUS	SO_WIFI_STATUS
+#define SO_PEEK_OFF		42
+
+/* Instruct lower device to use last 4-bytes of skb data as FCS */
+#define SO_NOFCS		43
+
+#define SO_LOCK_FILTER		44
+
+#define SO_SELECT_ERR_QUEUE	45
+
+#define SO_BUSY_POLL		46
+
+#define SO_MAX_PACING_RATE	47
+
+#define SO_BPF_EXTENSIONS	48
+
+#define SO_INCOMING_CPU		49
+
+#define SO_ATTACH_BPF		50
+#define SO_DETACH_BPF		SO_DETACH_FILTER
+
+#define SO_ATTACH_REUSEPORT_CBPF	51
+#define SO_ATTACH_REUSEPORT_EBPF	52
+
+#define SO_CNX_ADVICE		53
+
+#define SCM_TIMESTAMPING_OPT_STATS	54
+
+#define SO_MEMINFO		55
+
+#define SO_INCOMING_NAPI_ID	56
+
+#define SO_COOKIE		57
+
+#define SCM_TIMESTAMPING_PKTINFO	58
+
+#define SO_PEERGROUPS		59
+
+#define SO_ZEROCOPY		60
+
+#define SO_TXTIME		61
+#define SCM_TXTIME		SO_TXTIME
+
+#define SO_BINDTOIFINDEX	62
+
+#define SO_TIMESTAMP_OLD        29
+#define SO_TIMESTAMPNS_OLD      35
+#define SO_TIMESTAMPING_OLD     37
+
+#define SO_TIMESTAMP_NEW        63
+#define SO_TIMESTAMPNS_NEW      64
+#define SO_TIMESTAMPING_NEW     65
+
+#define SO_RCVTIMEO_NEW         66
+#define SO_SNDTIMEO_NEW         67
+
+#define SO_DETACH_REUSEPORT_BPF 68
+
+#define SO_PREFER_BUSY_POLL	69
+#define SO_BUSY_POLL_BUDGET	70
+
+#define SO_NETNS_COOKIE		71
+
+#define SO_BUF_LOCK		72
+
+#define SO_RESERVE_MEM		73
+
+#define SO_TXREHASH		74
+
+#define SO_RCVMARK		75
+
+#define SO_PASSPIDFD		76
+#define SO_PEERPIDFD		77
+
+#define SO_DEVMEM_LINEAR	78
+#define SCM_DEVMEM_LINEAR	SO_DEVMEM_LINEAR
+#define SO_DEVMEM_DMABUF	79
+#define SCM_DEVMEM_DMABUF	SO_DEVMEM_DMABUF
+#define SO_DEVMEM_DONTNEED	80
+
+#define SCM_TS_OPT_ID		81
+
+#define SO_RCVPRIORITY		82
+
+#define SO_PASSRIGHTS		83
+
+#define SO_INQ			84
+#define SCM_INQ			SO_INQ
+
+#define SO_RIGHTS_NOTRUNC	85
+
+#if !defined(__KERNEL__)
+
+#if __BITS_PER_LONG == 64 || (defined(__x86_64__) && defined(__ILP32__))
+/* on 64-bit and x32, avoid the ?: operator */
+#define SO_TIMESTAMP		SO_TIMESTAMP_OLD
+#define SO_TIMESTAMPNS		SO_TIMESTAMPNS_OLD
+#define SO_TIMESTAMPING		SO_TIMESTAMPING_OLD
+
+#define SO_RCVTIMEO		SO_RCVTIMEO_OLD
+#define SO_SNDTIMEO		SO_SNDTIMEO_OLD
+#else
+#define SO_TIMESTAMP (sizeof(time_t) == sizeof(__kernel_long_t) ? SO_TIMESTAMP_OLD : SO_TIMESTAMP_NEW)
+#define SO_TIMESTAMPNS (sizeof(time_t) == sizeof(__kernel_long_t) ? SO_TIMESTAMPNS_OLD : SO_TIMESTAMPNS_NEW)
+#define SO_TIMESTAMPING (sizeof(time_t) == sizeof(__kernel_long_t) ? SO_TIMESTAMPING_OLD : SO_TIMESTAMPING_NEW)
+
+#define SO_RCVTIMEO (sizeof(time_t) == sizeof(__kernel_long_t) ? SO_RCVTIMEO_OLD : SO_RCVTIMEO_NEW)
+#define SO_SNDTIMEO (sizeof(time_t) == sizeof(__kernel_long_t) ? SO_SNDTIMEO_OLD : SO_SNDTIMEO_NEW)
+#endif
+
+#define SCM_TIMESTAMP           SO_TIMESTAMP
+#define SCM_TIMESTAMPNS         SO_TIMESTAMPNS
+#define SCM_TIMESTAMPING        SO_TIMESTAMPING
+
+#endif
+
+#endif /* __ASM_GENERIC_SOCKET_H */