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, 134 insertions, 0 deletions

diff --git a/Documentation/ABI/testing/sysfs-class-bdi b/Documentation/ABI/testing/sysfs-class-bdi
new file mode 100644
index 000000000..0d2abd88a
--- /dev/null
+++ b/Documentation/ABI/testing/sysfs-class-bdi
@@ -0,0 +1,134 @@
+What:		/sys/class/bdi/<bdi>/
+Date:		January 2008
+Contact:	Peter Zijlstra <a.p.zijlstra@chello.nl>
+Description:
+
+Provide a place in sysfs for the backing_dev_info object.  This allows
+setting and retrieving various BDI specific variables.
+
+The <bdi> identifier can be either of the following:
+
+MAJOR:MINOR
+
+	Device number for block devices, or value of st_dev on
+	non-block filesystems which provide their own BDI, such as NFS
+	and FUSE.
+
+MAJOR:MINOR-fuseblk
+
+	Value of st_dev on fuseblk filesystems.
+
+default
+
+	The default backing dev, used for non-block device backed
+	filesystems which do not provide their own BDI.
+
+What:		/sys/class/bdi/<bdi>/read_ahead_kb
+Date:		January 2008
+Contact:	Peter Zijlstra <a.p.zijlstra@chello.nl>
+Description:
+	Size of the read-ahead window in kilobytes
+
+	(read-write)
+What:		/sys/class/bdi/<bdi>/min_ratio
+Date:		January 2008
+Contact:	Peter Zijlstra <a.p.zijlstra@chello.nl>
+Description:
+	Under normal circumstances each device is given a part of the
+	total write-back cache that relates to its current average
+	writeout speed in relation to the other devices.
+
+	The 'min_ratio' parameter allows assigning a minimum
+	percentage of the write-back cache to a particular device.
+	For example, this is useful for providing a minimum QoS.
+
+	(read-write)
+
+What:		/sys/class/bdi/<bdi>/min_ratio_fine
+Date:		November 2022
+Contact:	Stefan Roesch <shr@devkernel.io>
+Description:
+	Under normal circumstances each device is given a part of the
+	total write-back cache that relates to its current average
+	writeout speed in relation to the other devices.
+
+	The 'min_ratio_fine' parameter allows assigning a minimum reserve
+	of the write-back cache to a particular device. The value is
+	expressed as part of 1 million. For example, this is useful for
+	providing a minimum QoS.
+
+	(read-write)
+
+What:		/sys/class/bdi/<bdi>/max_ratio
+Date:		January 2008
+Contact:	Peter Zijlstra <a.p.zijlstra@chello.nl>
+Description:
+	Allows limiting a particular device to use not more than the
+	given percentage of the write-back cache.  This is useful in
+	situations where we want to avoid one device taking all or
+	most of the write-back cache.  For example in case of an NFS
+	mount that is prone to get stuck, or a FUSE mount which cannot
+	be trusted to play fair.
+
+	(read-write)
+
+What:		/sys/class/bdi/<bdi>/max_ratio_fine
+Date:		November 2022
+Contact:	Stefan Roesch <shr@devkernel.io>
+Description:
+	Allows limiting a particular device to use not more than the
+	given value of the write-back cache.  The value is given as part
+	of 1 million. This is useful in situations where we want to avoid
+	one device taking all or most of the write-back cache.  For example
+	in case of an NFS mount that is prone to get stuck, or a FUSE mount
+	which cannot be trusted to play fair.
+
+	(read-write)
+
+What:		/sys/class/bdi/<bdi>/min_bytes
+Date:		October 2022
+Contact:	Stefan Roesch <shr@devkernel.io>
+Description:
+	Under normal circumstances each device is given a part of the
+	total write-back cache that relates to its current average
+	writeout speed in relation to the other devices.
+
+	The 'min_bytes' parameter allows assigning a minimum
+	percentage of the write-back cache to a particular device
+	expressed in bytes.
+	For example, this is useful for providing a minimum QoS.
+
+	(read-write)
+
+What:		/sys/class/bdi/<bdi>/max_bytes
+Date:		October 2022
+Contact:	Stefan Roesch <shr@devkernel.io>
+Description:
+	Allows limiting a particular device to use not more than the
+	given 'max_bytes' of the write-back cache.  This is useful in
+	situations where we want to avoid one device taking all or
+	most of the write-back cache.  For example in case of an NFS
+	mount that is prone to get stuck, a FUSE mount which cannot be
+	trusted to play fair, or a nbd device.
+
+	(read-write)
+
+What:		/sys/class/bdi/<bdi>/strict_limit
+Date:		October 2022
+Contact:	Stefan Roesch <shr@devkernel.io>
+Description:
+	Forces per-BDI checks for the share of given device in the write-back
+	cache even before the global background dirty limit is reached. This
+	is useful in situations where the global limit is much higher than
+	affordable for given relatively slow (or untrusted) device. Turning
+	strictlimit on has no visible effect if max_ratio is equal to 100%.
+
+	(read-write)
+What:		/sys/class/bdi/<bdi>/stable_pages_required
+Date:		January 2008
+Contact:	Peter Zijlstra <a.p.zijlstra@chello.nl>
+Description:
+	If set, the backing device requires that all pages comprising a write
+	request must not be changed until writeout is complete.
+
+	(read-only)