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, 300 insertions, 0 deletions

diff --git a/drivers/fpga/Kconfig b/drivers/fpga/Kconfig
new file mode 100644
index 000000000..748fc210c
--- /dev/null
+++ b/drivers/fpga/Kconfig
@@ -0,0 +1,300 @@
+# SPDX-License-Identifier: GPL-2.0-only
+#
+# FPGA framework configuration
+#
+
+menuconfig FPGA
+	tristate "FPGA Configuration Framework"
+	help
+	  Say Y here if you want support for configuring FPGAs from the
+	  kernel.  The FPGA framework adds an FPGA manager class and FPGA
+	  manager drivers.
+
+if FPGA
+
+config FPGA_MGR_SOCFPGA
+	tristate "Altera SOCFPGA FPGA Manager"
+	depends on ARCH_INTEL_SOCFPGA || COMPILE_TEST
+	help
+	  FPGA manager driver support for Altera SOCFPGA.
+
+config FPGA_MGR_SOCFPGA_A10
+	tristate "Altera SoCFPGA Arria10"
+	depends on ARCH_INTEL_SOCFPGA || COMPILE_TEST
+	select REGMAP_MMIO
+	help
+	  FPGA manager driver support for Altera Arria10 SoCFPGA.
+
+config ALTERA_PR_IP_CORE
+	tristate "Altera Partial Reconfiguration IP Core"
+	help
+	  Core driver support for Altera Partial Reconfiguration IP component
+
+config ALTERA_PR_IP_CORE_PLAT
+	tristate "Platform support of Altera Partial Reconfiguration IP Core"
+	depends on ALTERA_PR_IP_CORE && OF && HAS_IOMEM
+	help
+	  Platform driver support for Altera Partial Reconfiguration IP
+	  component
+
+config FPGA_MGR_ALTERA_PS_SPI
+	tristate "Altera FPGA Passive Serial over SPI"
+	depends on SPI
+	select BITREVERSE
+	help
+	  FPGA manager driver support for Altera Arria/Cyclone/Stratix
+	  using the passive serial interface over SPI.
+
+config FPGA_MGR_ALTERA_CVP
+	tristate "Altera CvP FPGA Manager"
+	depends on PCI
+	help
+	  FPGA manager driver support for Arria-V, Cyclone-V, Stratix-V,
+	  Arria 10 and Stratix10 Altera FPGAs using the CvP interface over PCIe.
+
+config FPGA_MGR_ZYNQ_FPGA
+	tristate "Xilinx Zynq FPGA"
+	depends on ARCH_ZYNQ || COMPILE_TEST
+	help
+	  FPGA manager driver support for Xilinx Zynq FPGAs.
+
+config FPGA_MGR_STRATIX10_SOC
+	tristate "Intel Stratix10 SoC FPGA Manager"
+	depends on (ARCH_INTEL_SOCFPGA && INTEL_STRATIX10_SERVICE)
+	help
+	  FPGA manager driver support for the Intel Stratix10 SoC.
+
+config FPGA_MGR_XILINX_CORE
+	tristate
+
+config FPGA_MGR_XILINX_SELECTMAP
+	tristate "Xilinx Configuration over SelectMAP"
+	depends on HAS_IOMEM
+	select FPGA_MGR_XILINX_CORE
+	help
+	  FPGA manager driver support for Xilinx FPGA configuration
+	  over SelectMAP interface.
+
+config FPGA_MGR_XILINX_SPI
+	tristate "Xilinx Configuration over Slave Serial (SPI)"
+	depends on SPI
+	select FPGA_MGR_XILINX_CORE
+	help
+	  FPGA manager driver support for Xilinx FPGA configuration
+	  over slave serial interface.
+
+config FPGA_MGR_ICE40_SPI
+	tristate "Lattice iCE40 SPI"
+	depends on OF && SPI
+	help
+	  FPGA manager driver support for Lattice iCE40 FPGAs over SPI.
+
+config FPGA_MGR_MACHXO2_SPI
+	tristate "Lattice MachXO2 SPI"
+	depends on SPI
+	help
+	  FPGA manager driver support for Lattice MachXO2 configuration
+	  over slave SPI interface.
+
+config FPGA_MGR_TS73XX
+	tristate "Technologic Systems TS-73xx SBC FPGA Manager"
+	depends on ARCH_EP93XX && MACH_TS72XX
+	help
+	  FPGA manager driver support for the Altera Cyclone II FPGA
+	  present on the TS-73xx SBC boards.
+
+config FPGA_BRIDGE
+	tristate "FPGA Bridge Framework"
+	help
+	  Say Y here if you want to support bridges connected between host
+	  processors and FPGAs or between FPGAs.
+
+config SOCFPGA_FPGA_BRIDGE
+	tristate "Altera SoCFPGA FPGA Bridges"
+	depends on ARCH_INTEL_SOCFPGA && FPGA_BRIDGE
+	help
+	  Say Y to enable drivers for FPGA bridges for Altera SOCFPGA
+	  devices.
+
+config ALTERA_FREEZE_BRIDGE
+	tristate "Altera FPGA Freeze Bridge"
+	depends on FPGA_BRIDGE && HAS_IOMEM
+	help
+	  Say Y to enable drivers for Altera FPGA Freeze bridges.  A
+	  freeze bridge is a bridge that exists in the FPGA fabric to
+	  isolate one region of the FPGA from the busses while that
+	  region is being reprogrammed.
+
+config XILINX_PR_DECOUPLER
+	tristate "Xilinx LogiCORE PR Decoupler"
+	depends on FPGA_BRIDGE
+	depends on HAS_IOMEM
+	help
+	  Say Y to enable drivers for Xilinx LogiCORE PR Decoupler
+	  or Xilinx Dynamic Function eXchange AIX Shutdown Manager.
+	  The PR Decoupler exists in the FPGA fabric to isolate one
+	  region of the FPGA from the busses while that region is
+	  being reprogrammed during partial reconfig.
+	  The Dynamic Function eXchange AXI shutdown manager prevents
+	  AXI traffic from passing through the bridge. The controller
+	  safely handles AXI4MM and AXI4-Lite interfaces on a
+	  Reconfigurable Partition when it is undergoing dynamic
+	  reconfiguration, preventing the system deadlock that can
+	  occur if AXI transactions are interrupted by DFX.
+
+config FPGA_REGION
+	tristate "FPGA Region"
+	depends on FPGA_BRIDGE
+	help
+	  FPGA Region common code.  An FPGA Region controls an FPGA Manager
+	  and the FPGA Bridges associated with either a reconfigurable
+	  region of an FPGA or a whole FPGA.
+
+config OF_FPGA_REGION
+	tristate "FPGA Region Device Tree Overlay Support"
+	depends on OF && FPGA_REGION
+	help
+	  Support for loading FPGA images by applying a Device Tree
+	  overlay.
+
+config FPGA_DFL
+	tristate "FPGA Device Feature List (DFL) support"
+	select FPGA_BRIDGE
+	select FPGA_REGION
+	depends on HAS_IOMEM
+	help
+	  Device Feature List (DFL) defines a feature list structure that
+	  creates a linked list of feature headers within the MMIO space
+	  to provide an extensible way of adding features for FPGA.
+	  Driver can walk through the feature headers to enumerate feature
+	  devices (e.g. FPGA Management Engine, Port and Accelerator
+	  Function Unit) and their private features for target FPGA devices.
+
+	  Select this option to enable common support for Field-Programmable
+	  Gate Array (FPGA) solutions which implement Device Feature List.
+	  It provides enumeration APIs and feature device infrastructure.
+
+config FPGA_DFL_FME
+	tristate "FPGA DFL FME Driver"
+	depends on FPGA_DFL && HWMON && PERF_EVENTS
+	help
+	  The FPGA Management Engine (FME) is a feature device implemented
+	  under Device Feature List (DFL) framework. Select this option to
+	  enable the platform device driver for FME which implements all
+	  FPGA platform level management features. There shall be one FME
+	  per DFL based FPGA device.
+
+config FPGA_DFL_FME_MGR
+	tristate "FPGA DFL FME Manager Driver"
+	depends on FPGA_DFL_FME && HAS_IOMEM
+	help
+	  Say Y to enable FPGA Manager driver for FPGA Management Engine.
+
+config FPGA_DFL_FME_BRIDGE
+	tristate "FPGA DFL FME Bridge Driver"
+	depends on FPGA_DFL_FME && HAS_IOMEM
+	help
+	  Say Y to enable FPGA Bridge driver for FPGA Management Engine.
+
+config FPGA_DFL_FME_REGION
+	tristate "FPGA DFL FME Region Driver"
+	depends on FPGA_DFL_FME && HAS_IOMEM
+	help
+	  Say Y to enable FPGA Region driver for FPGA Management Engine.
+
+config FPGA_DFL_AFU
+	tristate "FPGA DFL AFU Driver"
+	depends on FPGA_DFL
+	help
+	  This is the driver for FPGA Accelerated Function Unit (AFU) which
+	  implements AFU and Port management features. A User AFU connects
+	  to the FPGA infrastructure via a Port. There may be more than one
+	  Port/AFU per DFL based FPGA device.
+
+config FPGA_DFL_NIOS_INTEL_PAC_N3000
+	tristate "FPGA DFL NIOS Driver for Intel PAC N3000"
+	depends on FPGA_DFL
+	select REGMAP
+	help
+	  This is the driver for the N3000 Nios private feature on Intel
+	  PAC (Programmable Acceleration Card) N3000. It communicates
+	  with the embedded Nios processor to configure the retimers on
+	  the card. It also instantiates the SPI master (spi-altera) for
+	  the card's BMC (Board Management Controller).
+
+config FPGA_DFL_PCI
+	tristate "FPGA DFL PCIe Device Driver"
+	depends on PCI && FPGA_DFL
+	help
+	  Select this option to enable PCIe driver for PCIe-based
+	  Field-Programmable Gate Array (FPGA) solutions which implement
+	  the Device Feature List (DFL). This driver provides interfaces
+	  for userspace applications to configure, enumerate, open and access
+	  FPGA accelerators on the FPGA DFL devices, enables system level
+	  management functions such as FPGA partial reconfiguration, power
+	  management and virtualization with DFL framework and DFL feature
+	  device drivers.
+
+	  To compile this as a module, choose M here.
+
+config FPGA_MGR_ZYNQMP_FPGA
+	tristate "Xilinx ZynqMP FPGA"
+	depends on ZYNQMP_FIRMWARE || (!ZYNQMP_FIRMWARE && COMPILE_TEST)
+	help
+	  FPGA manager driver support for Xilinx ZynqMP FPGAs.
+	  This driver uses the processor configuration port(PCAP)
+	  to configure the programmable logic(PL) through PS
+	  on ZynqMP SoC.
+
+config FPGA_MGR_VERSAL_FPGA
+	tristate "Xilinx Versal FPGA"
+	depends on ARCH_ZYNQMP || COMPILE_TEST
+	help
+	  Select this option to enable FPGA manager driver support for
+	  Xilinx Versal SoC. This driver uses the firmware interface to
+	  configure the programmable logic(PL).
+
+	  To compile this as a module, choose M here.
+
+config FPGA_M10_BMC_SEC_UPDATE
+	tristate "Intel MAX10 BMC Secure Update driver"
+	depends on MFD_INTEL_M10_BMC_CORE
+	select FW_LOADER
+	select FW_UPLOAD
+	help
+	  Secure update support for the Intel MAX10 board management
+	  controller.
+
+	  This is a subdriver of the Intel MAX10 board management controller
+	  (BMC) and provides support for secure updates for the BMC image,
+	  the FPGA image, the Root Entry Hashes, etc.
+
+config FPGA_MGR_MICROCHIP_SPI
+	tristate "Microchip Polarfire SPI FPGA manager"
+	depends on SPI
+	help
+	  FPGA manager driver support for Microchip Polarfire FPGAs
+	  programming over slave SPI interface with .dat formatted
+	  bitstream image.
+
+config FPGA_MGR_LATTICE_SYSCONFIG
+	tristate
+
+config FPGA_MGR_LATTICE_SYSCONFIG_SPI
+	tristate "Lattice sysCONFIG SPI FPGA manager"
+	depends on SPI
+	select FPGA_MGR_LATTICE_SYSCONFIG
+	help
+	  FPGA manager driver support for Lattice FPGAs programming over slave
+	  SPI sysCONFIG interface.
+
+config FPGA_MGR_EFINIX_SPI
+	tristate "Efinix FPGA configuration over SPI"
+	depends on SPI
+	help
+	  FPGA manager driver support for Efinix FPGAs configuration over SPI
+	  (passive mode only).
+
+source "drivers/fpga/tests/Kconfig"
+
+endif # FPGA