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, 84 insertions, 0 deletions

diff --git a/drivers/scsi/sd_trace.h b/drivers/scsi/sd_trace.h
new file mode 100644
index 000000000..cba3c0b82
--- /dev/null
+++ b/drivers/scsi/sd_trace.h
@@ -0,0 +1,84 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * Copyright (C) 2022 Western Digital Corporation or its affiliates.
+ */
+#undef TRACE_SYSTEM
+#define TRACE_SYSTEM sd
+
+#undef TRACE_INCLUDE_FILE
+#define TRACE_INCLUDE_FILE sd_trace
+
+#if !defined(_SD_TRACE_H) || defined(TRACE_HEADER_MULTI_READ)
+#include <scsi/scsi_cmnd.h>
+#include <scsi/scsi_host.h>
+#include <linux/tracepoint.h>
+
+TRACE_EVENT(scsi_prepare_zone_append,
+
+	    TP_PROTO(struct scsi_cmnd *cmnd, sector_t lba,
+		     unsigned int wp_offset),
+
+	    TP_ARGS(cmnd, lba, wp_offset),
+
+	    TP_STRUCT__entry(
+		     __field( unsigned int, host_no )
+		     __field( unsigned int, channel )
+		     __field( unsigned int, id )
+		     __field( unsigned int, lun )
+		     __field( sector_t,     lba )
+		     __field( unsigned int, wp_offset )
+	    ),
+
+	    TP_fast_assign(
+		__entry->host_no	= cmnd->device->host->host_no;
+		__entry->channel	= cmnd->device->channel;
+		__entry->id		= cmnd->device->id;
+		__entry->lun		= cmnd->device->lun;
+		__entry->lba		= lba;
+		__entry->wp_offset	= wp_offset;
+	    ),
+
+	    TP_printk("host_no=%u, channel=%u id=%u lun=%u lba=%llu wp_offset=%u",
+		      __entry->host_no, __entry->channel, __entry->id,
+		      __entry->lun, __entry->lba, __entry->wp_offset)
+);
+
+TRACE_EVENT(scsi_zone_wp_update,
+
+	    TP_PROTO(struct scsi_cmnd *cmnd, sector_t rq_sector,
+		     unsigned int wp_offset, unsigned int good_bytes),
+
+	    TP_ARGS(cmnd, rq_sector, wp_offset, good_bytes),
+
+	    TP_STRUCT__entry(
+		     __field( unsigned int, host_no )
+		     __field( unsigned int, channel )
+		     __field( unsigned int, id )
+		     __field( unsigned int, lun )
+		     __field( sector_t,     rq_sector )
+		     __field( unsigned int, wp_offset )
+		     __field( unsigned int, good_bytes )
+	    ),
+
+	    TP_fast_assign(
+		__entry->host_no	= cmnd->device->host->host_no;
+		__entry->channel	= cmnd->device->channel;
+		__entry->id		= cmnd->device->id;
+		__entry->lun		= cmnd->device->lun;
+		__entry->rq_sector	= rq_sector;
+		__entry->wp_offset	= wp_offset;
+		__entry->good_bytes	= good_bytes;
+	    ),
+
+	    TP_printk("host_no=%u, channel=%u id=%u lun=%u rq_sector=%llu" \
+		      " wp_offset=%u good_bytes=%u",
+		      __entry->host_no, __entry->channel, __entry->id,
+		      __entry->lun, __entry->rq_sector, __entry->wp_offset,
+		      __entry->good_bytes)
+);
+#endif /* _SD_TRACE_H */
+
+/* This part must be outside protection */
+#undef TRACE_INCLUDE_PATH
+#define TRACE_INCLUDE_PATH ../../drivers/scsi
+#include <trace/define_trace.h>