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, 117 insertions, 0 deletions

diff --git a/drivers/thermal/intel/Kconfig b/drivers/thermal/intel/Kconfig
new file mode 100644
index 000000000..347c59bc8
--- /dev/null
+++ b/drivers/thermal/intel/Kconfig
@@ -0,0 +1,117 @@
+# SPDX-License-Identifier: GPL-2.0-only
+config INTEL_POWERCLAMP
+	tristate "Intel PowerClamp idle injection driver"
+	depends on X86
+	depends on CPU_SUP_INTEL
+	depends on CPU_IDLE
+	select POWERCAP
+	select IDLE_INJECT
+	help
+	  Enable this to enable Intel PowerClamp idle injection driver. This
+	  enforce idle time which results in more package C-state residency. The
+	  user interface is exposed via generic thermal framework.
+
+config X86_THERMAL_VECTOR
+	def_bool y
+	depends on X86 && CPU_SUP_INTEL && X86_LOCAL_APIC
+
+config INTEL_TCC
+	bool
+	depends on X86
+
+config X86_PKG_TEMP_THERMAL
+	tristate "X86 package temperature thermal driver"
+	depends on X86_THERMAL_VECTOR && NET
+	select THERMAL_NETLINK
+	select INTEL_TCC
+	default m
+	help
+	  Enable this to register CPU digital sensor for package temperature as
+	  thermal zone. Each package will have its own thermal zone. There are
+	  two trip points which can be set by user to get notifications via thermal
+	  notification methods.
+
+config INTEL_SOC_DTS_IOSF_CORE
+	tristate
+	depends on X86 && PCI
+	select IOSF_MBI
+	select INTEL_TCC
+	help
+	  This is becoming a common feature for Intel SoCs to expose the additional
+	  digital temperature sensors (DTSs) using side band interface (IOSF). This
+	  implements the common set of helper functions to register, get temperature
+	  and get/set thresholds on DTSs.
+
+config INTEL_SOC_DTS_THERMAL
+	tristate "Intel SoCs DTS thermal driver"
+	depends on X86_64 && PCI && ACPI && NET
+	select INT340X_THERMAL
+	select INTEL_SOC_DTS_IOSF_CORE
+	help
+	  Enable this to register Intel SoCs (e.g. Bay Trail) platform digital
+	  temperature sensor (DTS). These SoCs have two additional DTSs in
+	  addition to DTSs on CPU cores. Each DTS will be registered as a
+	  thermal zone. There are two trip points. One of the trip point can
+	  be set by user mode programs to get notifications via Linux thermal
+	  notification methods.The other trip is a critical trip point, which
+	  was set by the driver based on the TJ MAX temperature.
+
+config INTEL_QUARK_DTS_THERMAL
+	tristate "Intel Quark DTS thermal driver"
+	depends on X86_INTEL_QUARK
+	help
+	  Enable this to register Intel Quark SoC (e.g. X1000) platform digital
+	  temperature sensor (DTS). For X1000 SoC, it has one on-die DTS.
+	  The DTS will be registered as a thermal zone. There are two trip points:
+	  hot & critical. The critical trip point default value is set by
+	  underlying BIOS/Firmware.
+
+menu "ACPI INT340X thermal drivers"
+source "drivers/thermal/intel/int340x_thermal/Kconfig"
+endmenu
+
+config INTEL_BXT_PMIC_THERMAL
+	tristate "Intel Broxton PMIC thermal driver"
+	depends on X86 && INTEL_SOC_PMIC_BXTWC
+	select REGMAP
+	help
+	  Select this driver for Intel Broxton PMIC with ADC channels monitoring
+	  system temperature measurements and alerts.
+	  This driver is used for monitoring the ADC channels of PMIC and handles
+	  the alert trip point interrupts and notifies the thermal framework with
+	  the trip point and temperature details of the zone.
+
+config INTEL_PCH_THERMAL
+	tristate "Intel PCH Thermal Reporting Driver"
+	depends on X86 && PCI
+	select ACPI_THERMAL_LIB if ACPI
+	help
+	  Enable this to support thermal reporting on certain intel PCHs.
+	  Thermal reporting device will provide temperature reading,
+	  programmable trip points and other information.
+
+config INTEL_TCC_COOLING
+	tristate "Intel TCC offset cooling Driver"
+	depends on X86
+	select INTEL_TCC
+	help
+	  Enable this to support system cooling by adjusting the effective TCC
+	  activation temperature via the TCC Offset register, which is widely
+	  supported on modern Intel platforms.
+	  Note that, on different platforms, the behavior might be different
+	  on how fast the setting takes effect, and how much the CPU frequency
+	  is reduced.
+
+config INTEL_HFI_THERMAL
+	bool "Intel Hardware Feedback Interface"
+	depends on NET
+	depends on CPU_SUP_INTEL
+	depends on X86_THERMAL_VECTOR
+	select THERMAL_NETLINK
+	help
+	  Select this option to enable the Hardware Feedback Interface. If
+	  selected, hardware provides guidance to the operating system on
+	  the performance and energy efficiency capabilities of each CPU.
+	  These capabilities may change as a result of changes in the operating
+	  conditions of the system such power and thermal limits. If selected,
+	  the kernel relays updates in CPUs' capabilities to userspace.