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, 36 insertions, 0 deletions

diff --git a/arch/x86/Makefile_32.cpu b/arch/x86/Makefile_32.cpu
new file mode 100644
index 000000000..c5aa169b5
--- /dev/null
+++ b/arch/x86/Makefile_32.cpu
@@ -0,0 +1,36 @@
+# SPDX-License-Identifier: GPL-2.0
+# CPU tuning section - shared with UML.
+# Must change only cflags-y (or [yn]), not CFLAGS! That makes a difference for UML.
+
+tune		= $(call cc-option,-mtune=$(1),$(2))
+
+ifdef CONFIG_CC_IS_CLANG
+align		:= -falign-functions=0 $(call cc-option,-falign-jumps=0) $(call cc-option,-falign-loops=0)
+else
+align		:= -falign-functions=0 -falign-jumps=0 -falign-loops=0
+endif
+
+cflags-$(CONFIG_M586TSC)	+= -march=i586
+cflags-$(CONFIG_M586MMX)	+= -march=pentium-mmx
+cflags-$(CONFIG_M686)		+= -march=i686
+cflags-$(CONFIG_MPENTIUMII)	+= -march=i686 $(call tune,pentium2)
+cflags-$(CONFIG_MPENTIUMIII)	+= -march=i686 $(call tune,pentium3)
+cflags-$(CONFIG_MPENTIUMM)	+= -march=i686 $(call tune,pentium3)
+cflags-$(CONFIG_MPENTIUM4)	+= -march=i686 $(call tune,pentium4)
+cflags-$(CONFIG_MK6)		+= -march=k6
+# Please note, that patches that add -march=athlon-xp and friends are pointless.
+# They make zero difference whatsosever to performance at this time.
+cflags-$(CONFIG_MK7)		+= -march=athlon
+cflags-$(CONFIG_MCRUSOE)	+= -march=i686 $(align)
+cflags-$(CONFIG_MEFFICEON)	+= -march=i686 $(call tune,pentium3) $(align)
+cflags-$(CONFIG_MCYRIXIII)	+= $(call cc-option,-march=c3,-march=i486) $(align)
+cflags-$(CONFIG_MVIAC3_2)	+= $(call cc-option,-march=c3-2,-march=i686)
+cflags-$(CONFIG_MVIAC7)		+= -march=i686
+cflags-$(CONFIG_MATOM)		+= -march=atom
+
+# Geode GX1 support
+cflags-$(CONFIG_MGEODEGX1)	+= -march=pentium-mmx
+cflags-$(CONFIG_MGEODE_LX)	+= $(call cc-option,-march=geode,-march=pentium-mmx)
+# add at the end to overwrite eventual tuning options from earlier
+# cpu entries
+cflags-$(CONFIG_X86_GENERIC) 	+= $(call tune,generic,$(call tune,i686))