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, 217 insertions, 0 deletions

diff --git a/Documentation/ABI/testing/sysfs-class-net-statistics b/Documentation/ABI/testing/sysfs-class-net-statistics
new file mode 100644
index 000000000..53e508c69
--- /dev/null
+++ b/Documentation/ABI/testing/sysfs-class-net-statistics
@@ -0,0 +1,217 @@
+What:		/sys/class/net/<iface>/statistics/collisions
+Date:		April 2005
+KernelVersion:	2.6.12
+Contact:	netdev@vger.kernel.org
+Description:
+		Indicates the number of collisions seen by this network device.
+		This value might not be relevant with all MAC layers.
+
+What:		/sys/class/net/<iface>/statistics/multicast
+Date:		April 2005
+KernelVersion:	2.6.12
+Contact:	netdev@vger.kernel.org
+Description:
+		Indicates the number of multicast packets received by this
+		network device.
+
+What:		/sys/class/net/<iface>/statistics/rx_bytes
+Date:		April 2005
+KernelVersion:	2.6.12
+Contact:	netdev@vger.kernel.org
+Description:
+		Indicates the number of bytes received by this network device.
+		See the network driver for the exact meaning of when this
+		value is incremented.
+
+What:		/sys/class/net/<iface>/statistics/rx_compressed
+Date:		April 2005
+KernelVersion:	2.6.12
+Contact:	netdev@vger.kernel.org
+Description:
+		Indicates the number of compressed packets received by this
+		network device. This value might only be relevant for interfaces
+		that support packet compression (e.g: PPP).
+
+What:		/sys/class/net/<iface>/statistics/rx_crc_errors
+Date:		April 2005
+KernelVersion:	2.6.12
+Contact:	netdev@vger.kernel.org
+Description:
+		Indicates the number of packets received with a CRC (FCS) error
+		by this network device. Note that the specific meaning might
+		depend on the MAC layer used by the interface.
+
+What:		/sys/class/net/<iface>/statistics/rx_dropped
+Date:		April 2005
+KernelVersion:	2.6.12
+Contact:	netdev@vger.kernel.org
+Description:
+		Indicates the number of packets received by the network device
+		but dropped, that are not forwarded to the upper layers for
+		packet processing. See the network driver for the exact
+		meaning of this value.
+
+What:		/sys/class/net/<iface>/statistics/rx_errors
+Date:		April 2005
+KernelVersion:	2.6.12
+Contact:	netdev@vger.kernel.org
+Description:
+		Indicates the number of receive errors on this network device.
+		See the network driver for the exact meaning of this value.
+
+What:		/sys/class/net/<iface>/statistics/rx_fifo_errors
+Date:		April 2005
+KernelVersion:	2.6.12
+Contact:	netdev@vger.kernel.org
+Description:
+		Indicates the number of receive FIFO errors seen by this
+		network device. See the network driver for the exact
+		meaning of this value.
+
+What:		/sys/class/net/<iface>/statistics/rx_frame_errors
+Date:		April 2005
+KernelVersion:	2.6.12
+Contact:	netdev@vger.kernel.org
+Description:
+		Indicates the number of received frames with error, such as
+		alignment errors. Note that the specific meaning depends on
+		on the MAC layer protocol used. See the network driver for
+		the exact meaning of this value.
+
+What:		/sys/class/net/<iface>/statistics/rx_length_errors
+Date:		April 2005
+KernelVersion:	2.6.12
+Contact:	netdev@vger.kernel.org
+Description:
+		Indicates the number of received error packet with a length
+		error, oversized or undersized. See the network driver for the
+		exact meaning of this value.
+
+What:		/sys/class/net/<iface>/statistics/rx_missed_errors
+Date:		April 2005
+KernelVersion:	2.6.12
+Contact:	netdev@vger.kernel.org
+Description:
+		Indicates the number of received packets that have been missed
+		due to lack of capacity in the receive side. See the network
+		driver for the exact meaning of this value.
+
+What:		/sys/class/net/<iface>/statistics/rx_nohandler
+Date:		February 2016
+KernelVersion:	4.6
+Contact:	netdev@vger.kernel.org
+Description:
+		Indicates the number of received packets that were dropped on
+		an inactive device by the network core.
+
+What:		/sys/class/net/<iface>/statistics/rx_over_errors
+Date:		April 2005
+KernelVersion:	2.6.12
+Contact:	netdev@vger.kernel.org
+Description:
+		Indicates the number of received packets that are oversized
+		compared to what the network device is configured to accept
+		(e.g: larger than MTU). See the network driver for the exact
+		meaning of this value.
+
+What:		/sys/class/net/<iface>/statistics/rx_packets
+Date:		April 2005
+KernelVersion:	2.6.12
+Contact:	netdev@vger.kernel.org
+Description:
+		Indicates the total number of good packets received by this
+		network device.
+
+What:		/sys/class/net/<iface>/statistics/tx_aborted_errors
+Date:		April 2005
+KernelVersion:	2.6.12
+Contact:	netdev@vger.kernel.org
+Description:
+		Indicates the number of packets that have been aborted
+		during transmission by a network device (e.g: because of
+		a medium collision). See the network driver for the exact
+		meaning of this value.
+
+What:		/sys/class/net/<iface>/statistics/tx_bytes
+Date:		April 2005
+KernelVersion:	2.6.12
+Contact:	netdev@vger.kernel.org
+Description:
+		Indicates the number of bytes transmitted by a network
+		device. See the network driver for the exact meaning of this
+		value, in particular whether this accounts for all successfully
+		transmitted packets or all packets that have been queued for
+		transmission.
+
+What:		/sys/class/net/<iface>/statistics/tx_carrier_errors
+Date:		April 2005
+KernelVersion:	2.6.12
+Contact:	netdev@vger.kernel.org
+Description:
+		Indicates the number of packets that could not be transmitted
+		because of carrier errors (e.g: physical link down). See the
+		network driver for the exact meaning of this value.
+
+What:		/sys/class/net/<iface>/statistics/tx_compressed
+Date:		April 2005
+KernelVersion:	2.6.12
+Contact:	netdev@vger.kernel.org
+Description:
+		Indicates the number of transmitted compressed packets. Note
+		this might only be relevant for devices that support
+		compression (e.g: PPP).
+
+What:		/sys/class/net/<iface>/statistics/tx_dropped
+Date:		April 2005
+KernelVersion:	2.6.12
+Contact:	netdev@vger.kernel.org
+Description:
+		Indicates the number of packets dropped during transmission.
+		See the driver for the exact reasons as to why the packets were
+		dropped.
+
+What:		/sys/class/net/<iface>/statistics/tx_errors
+Date:		April 2005
+KernelVersion:	2.6.12
+Contact:	netdev@vger.kernel.org
+Description:
+		Indicates the number of packets in error during transmission by
+		a network device. See the driver for the exact reasons as to
+		why the packets were dropped.
+
+What:		/sys/class/net/<iface>/statistics/tx_fifo_errors
+Date:		April 2005
+KernelVersion:	2.6.12
+Contact:	netdev@vger.kernel.org
+Description:
+		Indicates the number of packets having caused a transmit
+		FIFO error. See the driver for the exact reasons as to why the
+		packets were dropped.
+
+What:		/sys/class/net/<iface>/statistics/tx_heartbeat_errors
+Date:		April 2005
+KernelVersion:	2.6.12
+Contact:	netdev@vger.kernel.org
+Description:
+		Indicates the number of packets transmitted that have been
+		reported as heartbeat errors. See the driver for the exact
+		reasons as to why the packets were dropped.
+
+What:		/sys/class/net/<iface>/statistics/tx_packets
+Date:		April 2005
+KernelVersion:	2.6.12
+Contact:	netdev@vger.kernel.org
+Description:
+		Indicates the number of packets transmitted by a network
+		device. See the driver for whether this reports the number of all
+		attempted or successful transmissions.
+
+What:		/sys/class/net/<iface>/statistics/tx_window_errors
+Date:		April 2005
+KernelVersion:	2.6.12
+Contact:	netdev@vger.kernel.org
+Description:
+		Indicates the number of packets not successfully transmitted
+		due to a window collision. The specific meaning depends on the
+		MAC layer used.  On Ethernet this is usually used to report
+		late collisions errors.