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, 93 insertions, 0 deletions

diff --git a/drivers/edac/Makefile b/drivers/edac/Makefile
new file mode 100644
index 000000000..a37534300
--- /dev/null
+++ b/drivers/edac/Makefile
@@ -0,0 +1,93 @@
+#
+# Makefile for the Linux kernel EDAC drivers.
+#
+# Copyright 02 Jul 2003, Linux Networx (http://lnxi.com)
+# This file may be distributed under the terms of the
+# GNU General Public License.
+#
+
+obj-$(CONFIG_EDAC)			:= edac_core.o
+
+edac_core-y	:= edac_mc.o edac_device.o edac_mc_sysfs.o
+edac_core-y	+= edac_module.o edac_device_sysfs.o wq.o
+
+edac_core-$(CONFIG_EDAC_DEBUG)		+= debugfs.o
+edac_core-$(CONFIG_EDAC_SCRUB)		+= scrub.o
+edac_core-$(CONFIG_EDAC_ECS)		+= ecs.o
+edac_core-$(CONFIG_EDAC_MEM_REPAIR)	+= mem_repair.o
+
+ifdef CONFIG_PCI
+edac_core-y	+= edac_pci.o edac_pci_sysfs.o
+endif
+
+obj-$(CONFIG_EDAC_GHES)			+= ghes_edac.o
+
+edac_mce_amd-y				:= mce_amd.o
+obj-$(CONFIG_EDAC_DECODE_MCE)		+= edac_mce_amd.o
+
+obj-$(CONFIG_EDAC_AL_MC)		+= al_mc_edac.o
+obj-$(CONFIG_EDAC_AMD76X)		+= amd76x_edac.o
+obj-$(CONFIG_EDAC_CPC925)		+= cpc925_edac.o
+obj-$(CONFIG_EDAC_I5000)		+= i5000_edac.o
+obj-$(CONFIG_EDAC_I5100)		+= i5100_edac.o
+obj-$(CONFIG_EDAC_I5400)		+= i5400_edac.o
+obj-$(CONFIG_EDAC_I7300)		+= i7300_edac.o
+obj-$(CONFIG_EDAC_I7CORE)		+= i7core_edac.o
+obj-$(CONFIG_EDAC_SBRIDGE)		+= sb_edac.o
+obj-$(CONFIG_EDAC_PND2)			+= pnd2_edac.o
+obj-$(CONFIG_EDAC_IGEN6)			+= igen6_edac.o
+obj-$(CONFIG_EDAC_E7XXX)		+= e7xxx_edac.o
+obj-$(CONFIG_EDAC_E752X)		+= e752x_edac.o
+obj-$(CONFIG_EDAC_I82875P)		+= i82875p_edac.o
+obj-$(CONFIG_EDAC_I82975X)		+= i82975x_edac.o
+obj-$(CONFIG_EDAC_I3000)		+= i3000_edac.o
+obj-$(CONFIG_EDAC_I3200)		+= i3200_edac.o
+obj-$(CONFIG_EDAC_IE31200)		+= ie31200_edac.o
+obj-$(CONFIG_EDAC_X38)			+= x38_edac.o
+obj-$(CONFIG_EDAC_I82860)		+= i82860_edac.o
+obj-$(CONFIG_EDAC_AMD64)		+= amd64_edac.o
+
+obj-$(CONFIG_EDAC_PASEMI)		+= pasemi_edac.o
+
+mpc85xx_edac_mod-y			:= fsl_ddr_edac.o mpc85xx_edac.o
+obj-$(CONFIG_EDAC_MPC85XX)		+= mpc85xx_edac_mod.o
+
+layerscape_edac_mod-y			:= fsl_ddr_edac.o layerscape_edac.o
+obj-$(CONFIG_EDAC_LAYERSCAPE)		+= layerscape_edac_mod.o
+
+skx_edac_common-y			:= skx_common.o
+
+skx_edac-y				:= skx_base.o
+obj-$(CONFIG_EDAC_SKX)			+= skx_edac.o skx_edac_common.o
+
+i10nm_edac-y				:= i10nm_base.o
+obj-$(CONFIG_EDAC_I10NM)		+= i10nm_edac.o skx_edac_common.o
+
+imh_edac-y				:= imh_base.o
+obj-$(CONFIG_EDAC_IMH)			+= imh_edac.o skx_edac_common.o
+
+obj-$(CONFIG_EDAC_HIGHBANK_MC)		+= highbank_mc_edac.o
+obj-$(CONFIG_EDAC_HIGHBANK_L2)		+= highbank_l2_edac.o
+
+obj-$(CONFIG_EDAC_OCTEON_PC)		+= octeon_edac-pc.o
+obj-$(CONFIG_EDAC_OCTEON_L2C)		+= octeon_edac-l2c.o
+obj-$(CONFIG_EDAC_OCTEON_LMC)		+= octeon_edac-lmc.o
+obj-$(CONFIG_EDAC_OCTEON_PCI)		+= octeon_edac-pci.o
+obj-$(CONFIG_EDAC_THUNDERX)		+= thunderx_edac.o
+
+obj-$(CONFIG_EDAC_ALTERA)		+= altera_edac.o
+obj-$(CONFIG_EDAC_SIFIVE)		+= sifive_edac.o
+obj-$(CONFIG_EDAC_ARMADA_XP)		+= armada_xp_edac.o
+obj-$(CONFIG_EDAC_SYNOPSYS)		+= synopsys_edac.o
+obj-$(CONFIG_EDAC_XGENE)		+= xgene_edac.o
+obj-$(CONFIG_EDAC_TI)			+= ti_edac.o
+obj-$(CONFIG_EDAC_QCOM)			+= qcom_edac.o
+obj-$(CONFIG_EDAC_ASPEED)		+= aspeed_edac.o
+obj-$(CONFIG_EDAC_BLUEFIELD)		+= bluefield_edac.o
+obj-$(CONFIG_EDAC_DMC520)		+= dmc520_edac.o
+obj-$(CONFIG_EDAC_NPCM)			+= npcm_edac.o
+obj-$(CONFIG_EDAC_ZYNQMP)		+= zynqmp_edac.o
+obj-$(CONFIG_EDAC_VERSAL)		+= versal_edac.o
+obj-$(CONFIG_EDAC_LOONGSON)		+= loongson_edac.o
+obj-$(CONFIG_EDAC_VERSALNET)		+= versalnet_edac.o
+obj-$(CONFIG_EDAC_CORTEX_A72)		+= a72_edac.o